Advances in Food Research Volume 30

Download Advances in Food Research Volume 30

Post on 27-Dec-2016




33 download



    VOLUME 30

  • This Page Intentionally Left Blank


    VOLUME 30

    Edited by

    C. 0. CHICHESTER Univerdy of Rhode Island

    Kingston, Rhode island

    E. M. MRAK Universiv of California

    Davis, California

    Editorial Board


    B. S. SCHWEIGERT Universiv of California

    Davis, California



    ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

    Orlando San Diego New York Austin London Montreal Sydney Tokyo Toronto


    ACADEMIC PRESS, INC Orlando. Florida 32887

    United Kingdom Edition published by ACADEMIC PRESS INC. ( L O N D O N ) LTD. 24-28 Oval Road. London NWI 7DX

    LIBRARY OF C o N c R t s s CATALOG C A R D NUMBER 48-7808

    ISBN 0-12-016430-2

    PRlNTCD IN 111F LINITtO S l A l k b 0 1 AMFRICA

    86 87 88 89 Y X 7 6 5 4 1 ? I



    Sulfites In Foods: Uses, Analytical Methods, Residues, Fate, Exposure Assessment, Metabolism, Toxicity, and


    Steve L. Taylor, Nancy A. Higley, and Robert K. Bush

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 4

    111. Safety of Sulfites in Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , . 32 IV. Possible Substitutes and Their Limitations . . . . . 61 V. Future Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

    References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

    11. Uses of and Exposure to Sulfites in Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Maillard Reactions: Nonenzymatic Browning in Food Systems with Special Reference to the Development of Flavor

    James P. Danehy

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Chemistry of Browning in Model Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    111. Role of Browning in Specific Food Systems . . . . . . . . . . . . . . . . . IV. Browning, Nutrition, and Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Trends in Continuing Research . . . . . . . . . . . . . .

    Postharvest Changes in Fruit Cell Wall

    Melford A. John and Prakash M. Dey

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Components of Primary Cell Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    111. Structure of Primary Cell Wall IV. Fruit Development . . . . . . . . . . V. Concluding Remarks . . . . . . . . . . . . .

    References . . . . . . . .

    71 84 91

    120 123 124

    139 140 149 168 178 180


  • vi C 0 N T E N T S

    Soy Sauce Biochemistry

    Tarnotsu Yokotsuka

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    111. Recent Resear logical Advances in Shoyu Manufacture . . . . . . . . . . . IV. Color of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Flavor Evaluation of Koikuchi Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    VI. Volatile Flavor Ingredients of Koikuchi Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    VIII. Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    V11. Safety Problem of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    New Protein Foods: A Study of a Treatise

    Harold L. Wilcke, C. E. Bodwell, Daniel T. Hopkins, and Aaron M . Altschul

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. The Energy-Protein Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    111. Food Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Conventional Sources of Protein Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Reflections on Foods from Animal S

    VI. New Protein Foods Based on Plant Sources ................................ VII. Properties of Plant Protein Products . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    I96 204 209 24 I 257 261 287 30 I 313

    332 332 334 335 352 354 360 378 38 I

    INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

  • WILLIAM VERE CRUESS 1886-1 968


    The field of Food Science and Technology is a relatively new one and it is well that its few pioneers not be forgotten. To this end, Dr. Sam Goldblith described in Volume 27 of Advances in Food Research the life and accomplishments of Dr. Samuel C. Prescott, one of the fathers of modem food science and tech- nology, whose life spanned an era from the first use of the term microbe to beyond the discovery of DNA in 1953. The life span of another great pioneer in the field, Dr. William V. Cruess, covered the same period, from 1886 to 1968.

    While Prescott was working in the East, at the Massachusetts Institute of Technology, on problems related to sanitation and food preservation, Cruess entire career was spent on the West Coast, most of it at the University of California.

    Early in his career as a chemist, Cruess worked primarily on improving Cal-



    ifornia wines. During the time that prohibition made teaching and research in winemaking illegal, he turned his talents to another area, embarking on a pro- gram of intensive research and development in the field of food preservation that yielded revolutionary theoretical and practical results.

    Thanks to Cruess and his co-workers, the sun drying of fruits, for example, was replaced in California by mechanical dehydration, providing better product of uniform quality. He was also a pioneer in developing new products such as canned fruit cocktail and nectars from surplus fruits. Then, too, he is remem- bered as one of the early great teachers of food science and technology.


    Cruess was born on August 9, 1886, in a farming area called Indian Valley, near the town of San Miguel in the Central Coastal area of California. The soil yielded reluctantly and it was not easy to make a living in Indian Valley. In his memoirs Cruess mentions that the family subsisted mostly on red beans, salt pork, homemade bread, and once in a great while a little quail or dove. Fruit and vegetables were scarce.

    Very dry years on the farm were common. In 1888 the rainfall for the crop year was only 2 inches. Cattle died, wells went dry, and the principal food that year was boiled whole wheat brought in from elsewhere.

    Cruess attended a one-room grammar school with 20 students, about three miles from his home. One teacher taught all classes from kindergarten through the eighth grade. Normally, Cruess walked to school and back, but on rainy days he was allowed to ride horseback or use the family buggy.

    Although Cruess was eager to join his five fellow grammar school graduates in the Paso Robles High School, he remained out of school for 15 months, working as a farmhand and cutting firewood to earn enough money to pay for his room, board, books, and clothing.

    Boys in Indian Valley put in long days in those times, often sleeping in the haystacks so they could get to work early the next day. There was no danger of rain because it was dry country, and they were so tired after haying, harvesting, and hauling sacks of wheat or barley to town they had no trouble getting to sleep. They were up at 6 A.M. and worked until about five in the afternoon. During the harvesting season, they often started on the combine harvester about 5

    Cruess entered Paso Robles High School in the fall of 1902. During the first year he lived with a family in town, working part-time for the landlord to help pay for part of his $15-a-month bill for room and board. In high school he took courses in algebra, geometry, Latin, Spanish, history, chemistry, physics, and English.



    In his memoirs Cruess makes some interesting comments about automobiles in those days. A brand new phenomenon, automobile operation was forbidden in his area during daylight because they frightened horses drawing wagons or buggies into running away. Horses had the right-of-way.


    All during high school Cruess dreamed of going on to the University of California. A serious difficulty, however, was that he had no money. His father offered to sell one of their best horses and borrow necessary additional funds, but it seemed more sensible to Cruess to keep the horse and work on the farm rather then go into debt for his college expenses. He delayed going to college for a year in order to earn enough money to pursue a higher education.

    During this period, he worked as a harvest hand in summer, and in the fall he moved to the city of Oakland, where he worked in a car barn. He earned enough in 15 months to cover school expenses and room and board for the first year in college.

    Late in the summer of 1907, he went to the University of California at Berkeley and enrolled in chemistry. He had intended to enroll in mining en- gineering, but the dean of the College of Chemistry, Professor Edmund ONeill, who had been a classmate and close friend of his father in grammar school, persuaded him to major in chemistry because of the demand for chemists. Cruess followed ONeills advice and never regretted it.

    At the beginning of the second semester Dean ONeill offered Cruess a part- time job in the chemistry department as assistant to a lecturer. Cruess accepted with pleasure, for the offer included rent-free use of two rooms on the top floor of the chemistry building. As Cruess put it, he was well fixed for living quarters. At night the campus watchman often dropped in to chat with him and other students andeven to play a game of cards with them. The watchman was a Civil War veteran and Cruess learned a great deal from him about that war. As Cruess put it, This was an interesting association in the Chemistry Department.

    Thanks to his job at the University during the school term and to field work in the summers, Cruess was able to finance his education and even to graduate with a few dollars in his pocket.

    While in college he joined the La Junta Club, a social house club that later became a chapter of the national Sigma Phi fraternity. A fellow member was Earl Warren, who became Governor of California and subsequently Chief Justice of the U.S. Supreme Court.

    During his senior year, 1910- 191 1, Cruess held a part-time job with Professor M. E. Jaffa, who gave several courses in nutrition and was head of the Food and Drug Laboratory of the State Board of Health, then located on the Berkeley


    campus of the University of California. Assisting Jaffa in analyzing feed stuffs such as alfalfa hay and cottonseed meal taught the young chemist a great deal about proteins, fat, sugar, crude fiber, and so on in animal feeds.

    Twenty years later in 1931, while working full time for the University at Berkeley, he earned his Ph.D. in biochemistry from Stanford University. His thesis was concerned with the chemistry of the bitter principle in olives.


    Cruess first appointment after graduation was in the Division of Viticulture and Enology in the College of Agriculture at Berkeley. He had taken courses under Professors Frederick Bioletti and Hans Holm in zymology, or winemak- ing. The supposedly temporary appointment as a substitute for Professor Holm who was on a 1-year leave of absence became permanent when Holm resigned before the end of the year to take a position with a university in New England.

    Cruess job included two lectures and two laboratory periods a week in zymology. Instruction covered making culture media, sterilizing Petri dishes, isolating pure cultures of yeast and wine bacteria, yeast spore formation, the fermentation of grape must for wine, and acetic acid fermentation. Cruess wrote that with Professor Biolettis advice and assistance he managed to get by. Zymology was a long way from chemistry, although his knowledge of chemistry was extremely valuable.

    Cruess was Assistant in Zymology from 1911 to 1914, Assistant Professor from 1914 to 1918, Associate Professor from 1918 to 1929, and Professor of Food Technology from 1934 to 1955, when he assumed the Emeritus title.

    Shortly after his appointment in 191 1 Cruess was asked by Professor Bioletti to do some work on controlling fermentation in a small California winery about 30 miles from Berkeley. Spending several days a week that year at the winery, he learned the rudiments and operations in commercial wine making, and es- pecially, the use of pure yeast and control of fermentation with SO,. It was an excellent experience for him.

    The next year he conducted further research at the Swett winery near Mar- tinez, California. The owner was a son of John Swett, founder of Californias school systemand a friend of John Muir, the great naturalist. This afforded Cruess an opportunity to meet Muir, who- told him of the marvels of the high Sierra mountains of California. Muirs descriptions of them made such an im- pression on Cruess that in later years he spent a great deal of time walking the trails, fishing the streams and lakes, skiing on the slopes, and climbing the peaks of the Sierras. He loved to rough it in the mountains and his lovely wife was always with him; but although she loved the beauty of the Sierras, roughing it in the wilderness did at times reach the limit of endurance.


    The experiments at the Swett winery were concerned with the clarification of fresh grape juice, the use of pure yeast cultures in winemaking, and the recovery of residual wine from pressed red grape pomace by use of a diffusion battery system.

    Cruess experience with wine also included studies made in the laboratory at Berkeley with various yeast strains, SO,, clarification, and other winemaking problems. He made a collection of yeasts and some of them survived in laborato- ry cultures all through the years of prohibition.

    It was during his trips to the wineries in the Napa Valley area that Cruess met a lovely young school teacher, Marie Gleason, whom he married in 1917. Marie Cruess became a wonderful partner for the professor. She took a great interest in his students and they deeply appreciated it. She was an accomplished artist and her work was often exhibited. One of her paintings of Cruess now hangs in the lobby of Cruess Hall on the Davis Campus of the University of California.


    The 19 I8 Constitutional amendment prohibiting alcoholic beverages put an end to winemaking until its repeal early in Franklin Roosevelts administration.

    Meanwhile, there was a need for instruction and research on the preparation and preservation of unfermented fruit products. Consequently, several years before actual repeal of the 18th Amendment, Cruess initiated a lecture course entitled Zymology 116, which was concerned with the canning, sun drying, dehydration, and production of juices and other unfermented products from fruit; and the class was well attended.

    Research in dehydration resulted in the forced-draft, counter-current-tunnel dehydrator which Cruess collaborated on with A. W. Christie, P. F. Nichols, and E. M. Mrak.

    Cruess early emphasis in departmental research, therefore, was in the preser- vation and utilization area. Californias agriculture at that time was to a consider- able extent tree-fruits oriented, principally toward fresh markets. Only perfect fruits qualified for this type of marketing, leaving behind large quantities of so- called culls. Converting those to acceptable consumer products by known processing techniques or by devising new methods was a personal challenge he accepted and pursued throughout his career. Problems encountered in this area could only be solved by the acquisition of new knowledge, so new staff members were appointed to provide the appropriate research specialty. His early lead- ership in faculty growth prevailed throughout the entire history of the department.

    This early work was the beginning of food technology in California, though the subject was called fruit products in those days and some members of the


    faculty of the College of Agriculture felt that it was not a respectable area worthy of teaching and research. At the time, it took a good deal of courage on the part of Cruess to continue in the new field that is known today as food science and technology. In those days, it was alright to teach and conduct research on fertilizers and even manure, but not on the food we eat. How times and attitudes have changed, for food science is indeed a respectable field today! It was the courage and perseverance of Cruess and a few others that helped to bring about this change.


    Shortly after the repeal of prohibition, the Department of Fruit Products was established in Berkeley. The name of the department was soon changed to Food Technology. Cruess was chosen to head and build a good department and this he did.

    When the Food Products Department was first established, the curriculum was essentially practical, lacking both breadth and a firm theoretical foundation. In strengthening the department, Professor Cruess built a staff of young men well- trained in the basic sciences, some of whom also had the practical outlook of the food plant operator. Under his guidance, the curriculum gradually acquired greater breadth and depth; with greater emphasis being placed on the basic disciplines-chemistry, biochemistry, mathematics, and engineering-the groundwork was laid for what I term today the full-spectrum program in food science and technology. In brief, Professor Cruess shared with Dean Prescott of MIT the honor of having placed the entire field of food technology, as we know it today, on a firm basis-truly, a great achievement.

    Aside from his notable professional accomplishments, Professor Cruess was distinguished by a great dedication to his students. He was a demanding teacher who expected much of his students; but in turn he gave generously to them of his time and interest. He guided them, both professionally and personally, and in some cases even provided financial assistance to enable them to complete their studies. He and his charming wife often entertained his students in their home, inviting as many as 50 or 100 for barbecues, even dances. A few of his students were even married in his home. The sincere mutual respect and affection be- tween Professor Cruess and his students offers, perhaps, both a lesson and hope to those who are now concerned about faculty-student relations in the high- speed, impersonal environment of the multiversity.

    After 15 years of prohibition, there were few experienced wine makers in this country but there were plenty of home wine makers, bathtub gin artists, so to speak, and people who thought they knew how to make wine commercially but


    did not. They made more vinegar and other undrinkable liquids than potable wine. As a result, Cruess and members of his department, especially Maynard Joslyn and George Marsh, spent much of their time during the first few years after repeal, instructing California wine makers in the basic principles and prac- tices of making sound wines.

    With Joslyn and Marsh, Cruess also did a great deal of work on the freezing of California fruits and vegetables. They collaborated in early investigations on the preservation of perishable fruits and vegetables by freezing both for home and commercial use.

    Among his other distinctions, Cruess was the first person in the United States to work on problems relating to the processing of olives. He spent a great deal of time on the bacteriological and chemical aspects of olive processing. Later, Dr. Reese Vaughn joined him in conducting research on olive processing.

    To learn more about the wine, olive, and food industries in Europe, Cruess took several sabbatical leaves to make observations in Spain, France, Italy, Denmark, Sweden, Norway, England, Ireland, Canada, and even Egypt. His Egyptian visit included two months of lectures.

    Later he made trips to Hawaii to obtain a firsthand view of problems relating to the preservation of fruits and nuts in that area. These included, in particular, the production and treatment of macadamia nuts.


    Through the years Cruess published a great deal, including more than 600 scientific and applied papers and books. His most important book, published in 1923, was Commercial Fruit and Vegetable Products. It was a first and a monument in the area of commercial food practices and was translated into several other languages. Other books were Principles of Wine Making, Methods of Wine Analyses, Laboratory Manual for Fruit and Vegetable Products, Home and Farm Food Preservation, and Technology of Wine Making. The last revision of Commercial Fruit and Vegetable Products was published in 1958 and a new book, Technology of Wine Making, was published in 1960.

    He also did much to help establish the Fruit Products Journal. He published much in this journal, thus initiating the new area for publication of research in food science. In fact, he was considered the guardian of the journal and the field it covered.

    Cruess was very active in the development of the Institute of Food Tech- nologists. The first organizational meetings were held at MIT under the guidance of Dean Prescott. When it was decided to make a national organization, Cruess was heavily involved. He eventually became the national president and had


    much to do with the organization of the northern and southern California sections of the Institute.


    Over the years Cruess received many honors. These included the Nicholas Appert and Babcock Awards from the Institute of Food Technologists; Chevalier et officer du Merite Agricole of France; election to the American Academy of Microbiology, New York Academy of Science; and Academia Italian della Vita e del Vino of Italy.

    He was awarded the LL.D degree by the University of California, Davis, in 1960. The citation read: Alumnus of the University in the Class of 191 1. A member of her faculty for more than forty years, a biochemist in the Experiment Station, for some years department chairman, and now Professor Emeritus, of Food Science and Technology. Holder of the Babcock-Hart Award, of the first Nicholas Appert Medal, an Award given for your outstanding contributions to food technology, and of citations from several branches of the armed services for your work during World War 11. A highly productive research scientist, you have admirably combined the advancement of science with service to California agriculture.

    In 1952 the new Food Technology building at Davis was named Cruess Hall. He also received the Service Award of the 49ers of the Canning Industry, and a recognition from the California Farm Bureau Federation, the Food and Container Institute of the U.S. Armed Forces, the Raisin Industry, the Dried Fruit Associa- tion of California, and the Fig Institute. In spite of these many honors, Cruess was a modest person with great humility, as his acceptance address of the Appert Award makes clear:

    The speaker feels honored far beyond his just due in having been selected to receive the Nicholas Appert Award of the Institute of Food Technologists for 1942; the first of the series of yearly awards estab- lished by the Chicago Section of I.F.T. There are many in our organi- zation who are much more worthy of this recognition. Also, it should be stated that an investigators reputation often depends not so much on his own accomplishment as on those of his immediate associates. He may, as the titular head of a laboratory, symbolize its achievements and receive the honors that should be shared with his co-workers. The present case is no exception.

    It is difficult to express adequately in a few words the extent of Professor Cruess service to the University of California, to the food industry, to those who


    had the privilege to know him personally, particularly students and fellow work- ers, and to the larger community of mankind. His career was marked by distinc- tion in all phases; and his modesty, selflessness, and dedication, as well as his professional accomplishments and talents, won him the highest regard of his students, colleagues, and friends.

    He was also a leader in developing interest and leaders in the field of food technology of promising young scientists, who have done much to advance the field. Some of these are: C. 0. Chichester, A. W. Christi, M. A. Joslyn, G. L. Marsh, E. M. Mrak, P. F. Nichols, H. J. Phaff, J. Irish, andmany, manyothers.

    He was indeed a great man.

    CRUESS, WILLIAM VERE 1886-1968

    B.S.-University of California, 191 1 Ph.D.-Stanford University, 1931 Assistant in Zymology, 191 1-1912 Assistant Professor of Zymology, 1913- 1920 Assistant Professor of Fruit Products, 1920- 1921 Associate Professor of Fruit Products, 192 1- 1934 Chemist in the Experiment Station, 1925-1945 Biochemist in the Experiment Station, 1945-1954 Professor of Fruit Technology, 1934- 1945 Professor of Food Technology, 1945- 1954 Emeritus, 1954


  • This Page Intentionally Left Blank




    *Food Research Institute, University of Wisconsin, Madison, Wisconsin 53706 $William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin 53705 #Department of Medicine, University of Wisconsin, Madison, Wisconsin 53706

    1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Uses of and Exposure to Sulfites in Foods

    B. Natural Occurrence of Sulfites in Foods . . . . . . . . . . . . . . . . . . . . . . . . C. History of Use of Sulfiting Agents as Food Ingredients D. Current Food Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Methods for Measurement of Sulfite Residue Levels . . . . . . . . . . . . . . F. Chemistry of Sulfites and Fate in Foods . . . G. Treatment Levels versus Residual Levels . . . . . . . . . . . . . . . . . . . . . . . H. Exposure Assessments Safety of Sulfites in Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Metabolism of Sulfites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Human Challenge Trials C. Animal and Cellular Toxi D. Hypersensitivity to Ingested Sulfites . . . . . . . . . . . . . . . . . . . . . . . . . . . Possible Substitutes and Their Limitations A. Control of Enzymatic Browning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Control of Nonenzymatic Browning . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Use as Antioxidants or Reducing Agents D. Use as an Antimicrobial Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Use as a Bleaching Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    A. Description . . . . . . .



    V. Future Research Needs . . . .

    2 4 4 6 7 8

    17 21 30 30 32 32 38 39 47 61 61 62 62 62 63 63 64

    ?Resent address: International Flavors & Fragrances, Research & Development, Union Beach, New Jersey 07735.


    Copyright 6 1986 by Academic Press, Inc. All rights of reproduction in any form rewrved.



    Sulfiting agents have a long history of use as food ingredients. The term sulfiting agents refers to sulfur dioxide (SO,) and several forms of inorganic sulfite that liberate SO, under the conditions of use. In addition, naturally occur- ring sulfites are present in many foods. The yeast cultures used in the fermenta- tion of wines and beers naturally produce a portion of the sulfites found in these products.

    Sulfiting agents are added to foods for many important technical purposes, including the control of enzymatic and nonenzymatic browning, antimicrobial action, antioxidant and reducing agent uses, bleaching agent uses, and a variety of processing aid uses. In many products, the sulfites serve more than one purpose. Alternatives to the sulfiting agents are not easy to identify. Possible alternatives usually provide a narrower range of benefits, are often less effective, and are almost always more expensive.

    Sulfiting agents are currently used in a wide variety of food products. Data on the treatment levels for the sulfites and residual sulfites are not available for some food products, and wide variations in treatment modes and levels for particular products are known to occur in the industry.

    The analysis of sulfite residues in foods is confused by the rapid reaction between sulfiting agents and various food components. Sulfites react readily with reducing sugars, carbonyls, and proteins to yield a variety of organic combined sulfites. Analytical procedures are available for free SO, (SO, and the vari- ous inorganic sulfite salts) and total SO, (free SO, plus some of the combined forms of sulfite). Processing, storage, and preparation act to lower the available residual levels of sulfites in foods. The actual levels of free and total SO, in a particular food product are dictated by the extent of absorption of the sulfites during treatment, the nature of the processing treatment following sulfite addi- tion, and the conditions of storage. The actual levels of free and total SO, remaining in foods at the point of consumption have received less attention. The fate of sulfites added to specific foods is a largely unexplored area of study. The rapid reaction of sulfites with food components would be expected to leave little free SO, in the product at the point of consumption (Green, 1976; Joslyn and Braverman, 1954; Schroeter, 1966).

    Recently, the safety of the continued use of sulfites in foods has been ques- tioned on the basis of their alleged role in the initiation of asthmatic reactions in certain sensitive individuals. Numerous cases of sulfite-induced asthma have been reported in the medical literature since 1977 (Baker etal . , 1981; Buckley et al., 1985; Bush et al., 1986; Stevenson and Simon, 1981b; Towns and Mellis, 1984; Twarog and Leung, 1982) and additional anecdotal reports have been made to the Food and Drug Administration. These cases of sulfite-induced


    asthma were confirmed by positive challenges with capsules or solutions contain- ing inorganic sulfites. Only a small subgroup of the asthmatic population has sensitivity to sulfites in capsules.

    Important questions remain regarding the possibility that sulfited foods might initiate asthmatic reactions in these sensitive individuals. Although some of the described patients report asthmatic reactions to sulfited foods and many of the anecdotal cases involve suspicions of reactions to sulfited foods, only a few controlled challenges with sulfited foods have been performed with sensitive asthmatics (Howland and Simon, 1985; Seyal et af., 1984).

    We speculate that the degree of hazard posed to sulfite-sensitive asthmatics by sulfited foods may be considerably diminished by the reactions of the sulfites with food components. The majority of the challenges described in the medical literature thus far have involved inorganic sulfites in capsules or in acidic solu- tions. There can be little doubt that some asthmatics are sensitive to free in- organic sulfites, although this may be related to the conversion of inorganic sulfite salts to SO, at acidic pHs. However, as mentioned, most sulfited foods contain little free inorganic sulfite; sulfited lettuce is an exception (Taylor er af., 1985). These free inorganic sulfite residues would probably induce reactions in a manner similar to sulfites in capsules. The combination of sulfites with food components would drastically lower the free SO, content of most foods, thereby limiting exposure to these free forms of sulfites. Further research is needed to determine the effects of ingestion of combined forms of sulfites on the sulfite- sensitive asthmatics.

    Concerns have also arisen in recent years regarding the possibility that many consumers may be exceeding the Acceptable Daily Intake (ADI) for sulfites, although the recent report of the Ad Hoc Review Group on the Reexamination of the GRAS (Generally Recognized as Safe) Status of Sulfiting Agents indicates that these concerns are probably unwarranted (Life Sciences Research Office, 1985). They estimate that total intake of sulfites as SO, is about 10 mg/cap- ita/day, which is well below the AD1 of 42 mg for a 60-kg person (Life Sciences Research Office, 1985). Again, the same questions arise about the relative con- tributions of free and combined sulfites to the total sulfite intake. The intake of combined sulfite likely exceeds the intake of free sulfite by many fold. Because of their stabilities, the combined forms of sulfites would likely pose a lower hazard to consumers than free sulfites. More research will be required to firmly establish the relative toxicity of the free and combined sulfites.

    In this article, the current uses of sulfites in foods will be examined. The critical issue of exposure assessment will be explored in a review of the fate of sulfites in foods, residual levels, and analytical methodology. The questions about the safety of the use of sulfites in foods will be tackled by reviewing the available information on the metabolism of free and combined sulfites, the


    toxicity of free and combined sulfites, and the hypersensitivity reactions among certain asthmatics. In the final section of this review, the remaining unresolved issues will be highlighted with a discussion of future research needs. In our opinion, further research is necessary before decisions can be made on the future regulatory status of sulfites, although tremendous pressure is being exerted on the Food and Drug Administration to make a decision on the status of sulfites. More information is needed on the responses of sulfite-sensitive asthmatics to sulfited foods, the comparative reactivities of asthmatics to free and combined sulfites, the comparative toxicities of free and combined sulfites, the fate of sulfites in a variety of foods, and the extent of consumer exposure to free and combined sulfites.



    Sulfur dioxide and several forms of inorganic sulfites that liberate sulfur dioxide under the conditions of use are food additives known collectively as sulfiting agents. Sulfur dioxide (SO,), potassium bisulfite (KHSO,), potassium metabisulfite (K,S,O,), sodium bisulfite (NaHSO,), sodium metabisulfite (Na,S,O,), and sodium sulfite (Na,SO,) are listed in the Code of Federal Reg- ulations (CFR) as GRAS provided that they are not used in meats or other foods recognized as a source of thiamine. However, the GRAS status of these sulfiting agents is currently being reexamined, and changes may be made (Life Sciences Research Office, 1985). In addition, other sections of the CFR specifically allow the use of sulfiting agents in a variety of food-related processes. A list of the CFR sections and the processes covered by each section is provided in Table I. Note that all of the GRAS sulfiting agents are presently allowed for use for certain of these purposes. Potassium sulfite (K,SO,) and sulfurous acid (H,SO,), which are not GRAS substances, are specifically allowed for use only in the processing of caramel. Sulfiting agents are also permitted for use in wine and beer, although the Bureau of Alcohol, Tobacco, and Firearms (BATF) has proposed that the use of sulfites in wine and beer be curtailed to some extent (Anonymous, 1984). Presently, the levels of use of the sulfiting agents in most foods are not strictly limited by regulation. Exceptions are glucose syrup, dextrose monohydrate, and wine where the maximum allowable residual levels of SO, are specified, and food starch bleaching where the treatment level of SO, is controlled to a max- imum of 0.05%. It should be noted that the levels of sulfites used in some products such as wines are self-limiting because of organoleptic considerations.

    The theoretical yield of sulfur dioxide varies for the different forms of the




    CFR section Subject Sulfiting agents allowed

    2lCFR 182.3616 GRAS status KHSO3 2lCFR 182.3657 GRAS status K2S205 2lCFR 182.3739 GRAS status NaHS03

    2lCFR 182.3798 GRAS status Na2S03 2lCFR 182.3862 GRAS status so2 2lCFR 73.85(2) Caramel H2SO3, Na2SO3. K2SO3 2lCFR 168.111 Dextrose monohydrate SO2 (20 ppm maximum residual) 2lCFR 168.120 Glucose syrup SO2 (40 ppm maximum residual) 2lCFR 172.892 Food starch bleaching agents SO2 (0.05% maximum) 2 ICFR 173.3 1O(c) Boiler water additives Na2S205, Na2S03 2lCFR 177.1200(c) Cellophane NaHS03, Na2S03 2lCFR 177.1400 Water-soluble, hydroxyethyl All GRAS sulfiting agents

    2lCFR 182.3766 GRAS status Na2S205

    cellulose film

    sulfiting agents, as outlined in Table 11. Consequently, different treatment levels are required with the various sulfiting agents to yield equivalent doses of the active agent, SO,. For comparative purposes, it is helpful to calculate treatment levels on the basis of percentage of theoretical yield of SO,. However, it must be realized that these theoretical yields would almost never be achieved in food applications because the sulfiting agents react rapidly with food components, can be volatilized into the atmosphere, or can oxidize to sulfate. As will be empha- sized later, these reactions are dependent on a number of variables, including pH, temperature, and storage time.


    ~~~~~~ ~

    Theoretical yield Approximate solubility Chemical Formula of so2 (%) @/lo0 ml H20)

    Sulfur dioxide so2 100.00 11 at 20C Sodium sulfite, anhydrous Na2S03 50.82 28 at 40C Sodium sulfite, heptahydrate Na2S03 . 7H20 25.41 24 at 25C Sodium bisulfite NaHS03 61.56 300 at 20C Sodium metabisulfite Na2S205 67.39 54 at 20C Potassium metabisulfite K2S205 57.60 25 at 0C Potassium bisulfite KHSO3 53.32 -

    ,Z From Green (1976).


    The Food Chemicals Codex supplies specifications for the food grades of four of the sulfiting agents. In general, food grade sulfiting agents must be at least 90% pure to meet these standards. Some problems arise in the definition of sodium bisulfite, since there is some doubt about the existence of sodium bisulfite in the solid state. It may exist entirely as sodium metabisulfite or as a mixture of bisulfite and metabisulfite (Green, 1976). For that reason, the Food Chemicals Codex defines the purity of sodium bisulfite on the basis of SO, equivalents.


    In addition to their use as food additives, it must be remembered that the sulfites can also occur naturally in foods. Foods contain a variety of sulfur- containing compounds, including the sulfur amino acids, sulfates, sulfites, and sulfides. These sulfur-containing compounds are interconvertible in some food systems that possess the appropriate enzymes.

    The natural occurrence of sulfites in foods has been most thoroughly studied in alcoholic beverages such as wine and beer (Eschenbruch, 1974). The ability of yeasts to produce sulfite has been known since the end of the last century. Sulfite arises from sulfate via a multienzyme pathway. The sulfite can be converted into methionine and cysteine, but sulfite always exists in the fermentation medium. Sulfite can also be converted into H,S and other sulfides, which are organolep- tically undesirable in wines and beers. Most strains of Saccharomyces cerevisiae generate between 10 and 30 ppm SO,, although some strains producing in excess of 100 ppm SO, have been identified (Eschenbruch, 1974).

    Sulfite serves several functions in wine, including antimicrobial functions, prevention of browning, and binding of acetaldehyde (Eschenbruch, 1974). However, sulfite can be detected organoleptically if the concentration becomes too high; the threshold is thought to be about 50 ppm as free sulfite. Because of its adverse effects on the organoleptic quality of the wine and the potential for abuse of sulfites in making wine from inferior grapes, many countries have imposed strict limitations on the amount of residual SO, allowed in wine. In the Federal Republic of Germany, for example, the limits are 50 ppm of free sulfite and 300 ppm for total sulfite in wines of the Qualitatswein class. In the United States, the upper limit is 350 ppm as total residual SO,, although BATF is proposing an upper limit of from 125 to 175 pprn (Anonymous, 1984).

    Consequently, the formation of sulfite by yeasts must be critically controlled. The choice of yeast strains is important since they can vary by an order of magnitude in their capacities for sulfite formation (Dott et al., 1976; Eschen- bruch, 1974; Rankine and Pocock, 1969). The high sulfite formation by certain yeast strains can be attributed to several metabolic differences related to sulfite


    metabolism (Dott et al., 1977; Eschenbruch and Bonish, 1976; Heinzel and Truper, 1976). However, both low- and high-sulfite-forming strains are equiv- alent in their abilities to reduce sulfite to the obnoxious sulfides (Dott and Truper, 1976). The extent of sulfite formation is also related to other factors, including the amount of sulfite-binding compounds produced in the fermentation (Rankine and Pocock, 1969; Weeks, 1969). Acetaldehyde, pyruvic acid, and a- ketoglutaric acid bind SO, and serve to control the quantities of free SO, in the fermentation medium (Burroughs and Sparks, 1973a-c; Rankine and Pocock, 1969; Weeks, 1969). Only the free SO, has antimicrobial properties.

    The situation is much the same in beer except that lower levels of SO, are produced during beer fermentation (Hysert and Morrison, 1976). The SO, is derived mainly from sulfate and also serves as a precursor for sulfides.

    Wine and beer cannot be made without formation of sulfites. In beer, residual total SO, levels ranged from 0.2 to 11 ppm in the absence of added SO, in one study (Hysert and Morrison, 1976), although higher natural levels of formation might be expected to occur. Much of the residual sulfite in beer is in the combined state (Chapon et al . , 1982). In wine, natural SO, formation can account for 15-125 ppm of residual SO, in the finished product. In other food products fermented by yeasts, we would expect that SO, formation from sulfate would occur naturally, although we are not aware of studies confirming this suspicion.


    The sulfiting agents have enjoyed a long history of effective use as food and drug ingredients. Ancient Greeks wrote about the use of SO, for the fumigation of houses. The Romans and Egyptians are supposed to have used SO, for the sanitation of wine vessels (Roberts and McWeeny, 1972). Its use as a food preservative dates to at least 1664 when cider was added to flasks while they still contained SO, (Evelyn, 1664). The inorganic sulfites appeared as food additives at a much later date. The first years of use of the various inorganic sulfiting agents in the United States are as follows: sodium bisulfite, 1921; sodium sulfite, 1930; potassium and sodium metabisulfite, 1939 (Subcommittee on Review of the GRAS List, 1972). The sulfiting agents were first used in wine and beer. Among nonalcoholic products, the sulfiting agents were first used on dried fruits and vegetables in all likelihood. However, their use in foods spread rapidly as a consequence of the absence of toxic hazards and their widespread functional effectiveness. In the decade between 1960 and 1970,. a 30-70% increase in the amounts of several sulfiting agents used annually in the United States was ob- served (Subcommittee on Review of the GRAS List, 1972), a testament to the



    (UNITED STATES, 1976)

    Amount produced Sulfiting agent (Ib)

    Sodium bisulfite 4,900,000 Sodium metabisulfite 92,000 Potassium metabisulfite 220,000 Sulfur dioxide 2,200,000 Sodium sulfite 15,000

    growing utilization of these additives. Table 111 contains data on the production of sulfur dioxide and the inorganic sulfiting agents for use in foods. Foods represent one of the larger uses for these chemicals. SO, and sodium bisulfite are produced and used in far larger quantities than most of the other sulfiting agents.


    1 . Overview of Current Applications

    Sulfiting agents are used in foods for many important purposes: the inhibition of nonenzymatic browning, the inhibition of various enzymatic reactions includ- ing enzymatic browning, inhibition and control of microorganisms, an antioxi- dant and reducing agent including dough conditioning, a bleaching agent, a processing aid, and several secondary uses including pH control agent and sta- bilizing agent. Each of the general categories will be discussed with a brief explanation of the mechanism of action of the sulfiting agents in effecting these changes. As might be expected for any group of substances that possesses so many useful properties, an enormous number of specific applications have been found for sulfiting agents in foods. Later in this section, we will make an attempt to identify these applications and the treatment levels associated with each ap- plication. Other uses for the sulfiting agents have been devised, and these uses will be described, although we are not certain that they are being used in the food industry. Finally, the sulfiting agents provide other benefits in foods beyond those which are readily identified with the sulfiting agents. These benefits will be identified and discussed.

    Several previous reviews on the applications for sulfiting agents in foods have appeared (Green, 1976; Joslyn and Braverman, 1954; Roberts and McWeeny, 1972; Schroeter, 1966). Based on our current knowledge, these reviews are inadequate and should be considered to be out of date (Joslyn and Braverman,


    1954; Schroeter, 1966), oriented toward the British food industry (Green, 1976; Roberts and McWeeny, 1972), or incomplete. Many additional applications of the sulfiting agents have now been identified, although the previous reviews do provide much valuable information on some of the major uses of the sulfiting agents in foods.

    2 . Inhibition of Nonenzymatic Browning

    Nonenzymatic browning is a term used to describe a family of diverse reac- tions that commonly involve the formation of carbonyl intermediates and brown, polymeric pigments. Examples include the reactions between amino acids and reducing sugars and carmelization of sugars. The chemistry of the reactions involved is complex and not completely understood. An excellent review of the chemistry of nonenzymatic browning and the effects of the sulfites on these reactions was prepared by McWeeny et al. (1974). The sulfites can be used to control nonenzymatic browning because of their ability to react with the carbonyl intermediates. A variety of carbonyl intermediates can be formed during the nonenzymatic browning process, including reducing sugars, simple carbonyls, dicarbonyls, and a,P-unsaturated carbonyls. The sulfites can react with all of these intermediates and thus block formation of the brown pigments. Wedzicha et al. (1984) developed a kinetic model for the inhibition of nonenzymatic browning by sulfites. Reaction of sulfites with the carbonyls generated by non- enzymatic browning accounts for most of the loss of sulfites in dehydrated vegetables (Wedzicha et a l . , 1984). Some of the sulfite-carbonyl reaction prod- ucts are more stable than others. The sugar hydroxysulfonates formed between reducing sugars and sulfites are the least stable, although they are quite stable at acid pHs. The sulfonated carbonyls formed on reaction of the sulfites with the a, @-unsaturated carbonyls are extremely stable and the reaction is generally considered to be irreversible. The differences in stability of the sulfite addition products can influence the effectiveness of the sulfites in certain food applica- tions. For example, sulfite can almost totally inhibit nonenzymatic browning of glucose-glycine solutions because of the irreversible reaction of the sulfites with the a$-unsaturated carbonyl intermediates of this reaction (Wedzicha and Mc- Weeny, 1974a). The glucose-glycine system typifies the situation that occurs in dehydrated potatoes. A kinetic model for the glucose-glycine reaction and its inhibition by sulfites has been developed by Wedzicha (1984). On the other hand, sulfites can only retard the formation of browning pigments in the ascorbic acid-glycine reaction because the principal intermediates of this reaction are dicarbonyl compounds which react reversibly with the sulfites (Wedzicha and McWeeny, 1974a). The ascorbic acid-glycine system is typical of the situation existing in fruit juices and drinks.

    Sulfites find wide use as inhibitors of nonenzymatic browning. They have


    been used for this purpose to control discoloration of wines, dried fruits, dehy- drated vegetables, dehydrated potatoes, coconut, pectin, some varieties of vin- egar, and white grape juice. Sulfites are also used to control the commercial carmelization process. In warmer climates, sulfites can be used to control non- enzymatic browning in fruit juices and drinks (Joslyn and Braverman, 1954). Sulfites are also used to control juice color formation in the production of beet sugar (McGinnis, 1982).

    3. Inhibition of Various Enzymatic Reactions

    SO, and sulfites can act as inhibitors of numerous enzymatic reactions, includ- ing polyphenoloxidase, ascorbate oxidase, lipoxygenase, peroxidase, and thia- mine- dependent enzymes. The actions of the sulfiting agents on oxidizing enzyme systems have been reviewed by Haisman ( 1974).

    Inhibition of polyphenoloxidase is useful in the control of enzymatic brown- ing. Polyphenoloxidase catalyzes the oxidation of mono- and ortho-diphenols to quinones. The quinones can cyclize, undergo further oxidation, and condense to form brown pigments. The mechanism of action of the sulfites in preventing enzymatic browning is not known, but very likely involves several different types of actions. Sulfites may directly inhibit the enzyme; potassium meta- bisulfite has recently been shown to inhibit strawberry polyphenoloxidase at 10 mM concentrations (Wesche-Ebeling and Montgomery, 1983). Sulfites may also interact with the intermediates in the enzymatic browning reaction and prevent their participation in the reactions leading to formation of the brown pigments. For example, sulfites may combine with the quinones and prevent their participa- tion in the further oxidation, cyclization, and condensation reactions. Evidence for the formation of quinone-sulfite complexes has been reviewed (Haisman, 1974). Alternatively, the sulfites may simply act as reducing agents promoting the reaction of the quinones back to the original phenols. The level of sulfites necessary to prevent enzymatic browning depends on the nature of the available substrate. When only monophenols such as tyrosine are present, fairly low levels of sulfite are effective. Potatoes are an example of this situation. When diphenols are present, much higher concentrations of sulfites are necessary. An example of this situation would be guacamole. The sulfites do not irreversibly inhibit the enzymatic browning reaction so the required concentrations are also dependent on the length of time that the reaction must be inhibited.

    Inhibition of enzymatic browning is the primary reason for using sulfites in salad bar items, including cut fruits, lettuce, and guacamole. Sulfites have also been used to prevent enzymatic browning in prepeeled potatoes, sliced potatoes, cut apples and other fruits supplied to the baking industry (Ponting et al . , 1971), fresh mushrooms (Komanowsky et al., 1970), and table grapes (Nelson, 1983).


    A similar reaction occurs in shrimp where enzymatic tyrosine oxidation leads to black spot formation. The reaction is catalyzed by tyrosinase, a type of polyphenoloxidase. Black spot formation in shrimp can be controlled by the addition of sulfites (Fieger, 1951).

    Sulfites can also prevent the oxidation of ascorbate by ascorbate oxidase and other enzymes. Ascorbate levels decrease very quickly following the maceration of plant tissues due to the action of ascorbate oxidase. Sulfite addition preserves ascorbate and can be used in potato, pumpkin, cauliflower, tomato, and green and red pepper products (Haisman, 1974).

    Sulfites can also inhibit lipoxygenase, an enzyme known to cause formation of off-flavors during postharvest storage of vegetables such as peas (Haisman, 1974). Treatment with sulfites will prevent formation of these off-flavors, an added benefit to their use in dehydrated peas and other vegetables.

    Anaerobic bacterial fermentation can be inhibited in grape juice by sulfites. This inhibition is essential to the production of wines. The mechanism of this inhibition is not entirely understood, but it is partially due to the destruction of thiamine, which serves as an essential cofactor for several of the fermentative enzymes (Haisman, 1974). These enzymes are thus inhibited.

    4. Inhibition and Control of Microorganisms

    The sulfites play crucial roles in the inhibition of bacteria in several food pro- cesses. In winemaking, the sulfites are employed to prevent undesirable bacterial fermentation of the grape or fruit juice. Sulfites are also essential in the corn steeping process used to facilitate removal of the corn starch; the sulfites prevent bacterial growth in the steep liquor (Schroeter, 1966). The application of sulfites to table grapes is critical to prevent bacterial and mold growth (Nelson, 1983; Nelson and Ahmedullah, 1973, 1976). Although not a common practice in the United States, sulfites have been widely used to prevent mold damage in fruits prior to jam production (Roberts and McWeeny, 1972). Sulfites have also found use in the prevention of postharvest deterioration of fruits used for the production of juices (Moms et a f . , 1979).

    The use of sulfites as antimicrobial agents has been reviewed by Roberts and McWeeny (1972), Joslyn and Braverman (1954), and Ingram (1959). The sul- fites are selective antimicrobial agents with more inhibitory effect on acetic acid bacteria, lactic acid bacteria, and various molds than on yeasts (Joslyn and Braverman, 1954). This selectivity enhances their value in the control of undesir- able fermentation in winemaking. The mechanism of the antimicrobial action of the sulfites is not well understood. However, several factors are known to control the antimicrobial efficacy of the sulfites. One of the more important factors is pH which controls the form of sulfite present in the food. Apparently, H,SO, is the


    active form of the sulfites in terms of their antimicrobial actions (Carr et al., 1976; Ingram, 1959), so lower pHs enhance the antimicrobial effect. The com- bination of sulfites with food components also affects their antimicrobial activity (Ingram, 1959; Joslyn and Braverman, 1954). The sulfite adducts have no anti- microbial activity. Consequently, more sulfite is required to preserve a glucose syrup than a sucrose syrup, since sulfites will combine with glucose but not sucrose (Ingram, 1959). Considerable sulfite must be added to wine because of the binding of the sulfites to fermentation products such as acetaldehyde. The volatilization of SO, from acidic products also affects the level retained for antimicrobial action.

    Sulfites can have some detrimental effects as a result of their antimicrobial actions. In red wines, high levels of SO, inhibit the desirable malolactic fermen- tation, which serves to reduce the acidity of wines produced in cool regions (Liu and Gallander, 1983).

    Although we know of no practical use of this antimicrobial activity, sulfites also inactivate certain types of enteroviruses including poliovirus type I, cox- sackievirus type A9, and echovirus type 7 (Salo and Cliver, 1978).

    5. Antioxidant and Reducing Agent Uses

    The antioxidative effects of the sulfites are partially responsible for their preserving effect on ascorbate and their inhibition of nonenzymatic and en- zymatic browning. The ability of the sulfiting agents to promote the reduction of the oxidized quinones to reduced phenols is one of the mechanisms available for the inhibition of these processes by the sulfites. Sulfites also prevent the oxida- tion of essential oils and carotenoids, which would generate off-flavors (Baloch et al., 1977; Roberts and McWeeny, 1972). A major function of SO, in beer is the inhibition of oxidative changes that are considered undesirable to flavor development (Roberts and McWeeny, 1972; Schroeter, 1966).

    Sulfites are widely-used as dough conditioners in the baking industry for biscuits, crackers, cookies, and frozen pizza doughs and pie crusts. In these products, sulfites act by breaking the disulfide bonds in the gluten fraction of the dough (Wade, 1972). The sulfites also promote disintegration of the protein matrix during the corn steeping process, which facilitates rapid hydration, soft- ening of the kernel, and extraction of the starch (Schroeter, 1966). The sulfites may exert this action via their ability to reduce disulfide bonds, although we know of no direct proof for this possibility. SO, has also been used to improve the extraction of pectins from various sources through its ability to depolymerize the pectins (Roberts and McWeeny, 1972).


    6. Bleaching Agent Uses

    The major application of the bleaching properties of the sulfites is the bleach- ing of cherries for the production of maraschino cherries and g lad fruit products (Josyln and Braverman, 1954; Weigand, 1946). The sulfiting agents are also reported to bleach pectins (Roberts and McWeeny, 1972). The uniformity and translucency of color of orange, lemon, grapefruit, and citron peel are improved by storage in a sulfite brine (Cruess and Glickson, 1932). Sulfur dioxide can also be used as a bleaching agent for food starches (Table I). The bleaching of table grapes during sulfite fumigation is considered detrimental to quality (Nelson, 1983).

    7. Use as a Processing Aid

    Many of the applications of sulfites fall into the category of processing aids. This particular category of use for sulfiting agents is difficult to define and variations probably exist in its definition within the food industry. Obviously, sulfite residues can originate from the use of sulfited products in the formulation of the end product. Examples would include the use of beet sugar or corn syrup in a variety of products and the use of maraschino cherries in fruit cocktail. Typically, SO, residue levels from such uses would be rather low. Further investigation of the use of sulfiting agents as processing aids will be necessary to obtain a better picture of the extent of such uses and their contribution to con- sumer exposure.

    8. Secondary Uses

    This category of uses of the sulfiting agents is diverse because these additives have many desirable secondary benefits beyond the primary reasons for their use. Examples would include their facilitation of corn starch extraction (Schroeter, 1966), a secondary benefit to the primary purpose of preventing microbial growth in the corn steep liquor. Another example would be the control of excess alkalinity and the improvement in boiling properties of beet sugar juice, a sec- ondary benefit to the primary purpose of control of color formation in the juice (McGinnis, 1982). Many additional examples could be selected.

    9. Specific Applications and Treatment Levels

    The above discussion clearly shows that sulfiting agents are used in the food industry for a variety of products and for many different reasons. The Federation


    of American Societies for Experimental Biology (FASEB) panel attempted to identify these applications and the residue levels resulting from each use (Life Sciences Research Office, 1985). This information may not be representative of the entire food industry, since variations exist in the use of sulfiting agents, the type of sulfiting agents employed, the treatment levels, and the means of apply- ing the sulfiting agents to the foods, all of which would affect residual levels. Sulfite uses have been identified in baked goods and baking mixes, alcoholic and nonalcoholic beverages, coffee and tea, condiments and relishes, dairy product analogs, prepared fish and shellfish products, fresh fish and shellfish, fresh fruits and fruit juices, fresh vegetables, gelatins, grain products, gravies and sauces, jams and jellies, nuts and nut products, processed fruits and fruit juices, pro- cessed vegetables and vegetable juices, snack foods, soups and soup mixes, sugar, and sweet sauces, toppings, and syrups. These categories correspond to those listed in the CFR (21CFR 170.3). In Table IV, specific uses of the sulfiting agents within each of these categories are identified, and reported residual levels and exposure estimates are given (Life Sciences Research Office, 1985); most of the information in Table IV was obtained from the FASEB compilations (Life Sciences Research Office, 1985). Some of the information in Table IV needs to be verified for accuracy, but it is probably the best and most complete survey of sulfite use ever conducted. Also, the uses and residual levels may not be repre- sentative of the entire industry. A major deficiency has been the lack of analyses of sulfited foods at the point of consumption. Storage, processing, and prepara- tion can affect residual sulfite levels in the product prepared for consumption. Further research will be needed to determine, with greater accuracy, actual consumer exposure to sulfites.

    10. Other Uses for Sulfiting Agents in Foods

    Certain other uses have been developed for the sulfiting agents in foods. We are not certain that these processes are actually being used in the food industry, but they are feasible. Two examples will be cited, although many more appear in the literature. A procedure for improved color retention in canned garbanzo beans that involves a presoak in NaHSO, has been developed (Daoud et al., 1977; Luh et al., 1978). The pink discoloration noted with certain varieties of canned pears can be prevented by use of SO, (Chandler and Clegg, 1970).

    11. Additional Benefits of Sulfiting Agents in Foods

    Sulfites provide additional benefits in foods beyond those already discussed. To our knowledge, they are not used in foods for these purposes, so these benefits might best be classified as fortuitous or potential uses. The carcinogenic



    Estimated level in product

    as consumedb Food intake Sulfite intakec Category Subcategory (ppm SO2) (g/capita/day) (mg/capita/day)

    Baked goods and baking mixes

    Beverages, nonalcoholic

    Beverages, alcoholic

    Coffee and tea


    Dairy analogs Fish and shellfish

    Fresh fruit

    Dried fruitd Fruit juices

    Cookies Cake with dried carrots Crackers Sheeting doughs Pie dough Pizza ciust Pie crust Tortilla shells Total Cola and pepper Lemon-lime Orange Root beer Ginger ale Grape Juice-containing carbonated

    Beer Wine Instant tea Liquid concentrated tea Tea leaves Olives Pickles/relishes Salad dressing mix (dry) Vinegar


    Malt vinegar Wine vinegar

    Filled milk Dried cod Shrimp Fruit salad Grapes Total Apple concentrate, imported Cheny-beny Grape, red or purple Grape, white, white sparkling,

    pink sparkling, or red spark- ling

    5 10 5


    TABLE 1V (Conrinued)

    Estimated level in product

    as consumedb Food intake Sulfite intakec Category Subcategory (ppm SO2) (gkapitalday) (mg/capita/day)

    Frozen fruit Canned fruit Gelatin

    Grain products

    Gravies, sauces

    Jams and jellies

    Nut products Protein isolates, soy Snack foods

    Soft candy

    Sugars and frostings Sweet sauces

    and syrups


    Fresh vegetables

    Canned vegetables

    Lemon, nonfrozen Lime, nonfrozen Unspecified Maraschino cherries Total Flavored mix Unflavored mix Corn starch and modified corn

    starch Hominy Spinach pasta Meat, tomato, milk, buttery,

    Domestic Imported Pectin Coconut

    and specialty

    Apple bits Crackers, filled Corn-based snacks Pretzels Caramel Dry mix Sugars, average Corn syrup High fructose corn syrup Fruit topping Maple syrup Molasses Total Mushrooms Salad bar lettuce Cole slaw Guacamole Potatoes, whole, peeled Total Potatoes Pickled cocktail onions Pickled peppers Sauerkraut (in glass) Sauerkraut (in cans)

    800 160

    5 50

    5 4 5


    20 5


    5 13.6

    > 10-50 5 2-5 -

    0.5 0.5 - -

    0.5 7

    30 3 60 20

    125 16 13

    950 < 10-250 - -

    5 10 30 30 30 12

    0.12 0.04 0.5 0.3 6.6 - -


    0.1 0.1 3

    5.7 0.06

    0.5 0.25

    0.29 0.88



    - -

    9.3 99 26 50 3.4 3 0.27

    39.1 0.2 -

    - - -

    65 0.27 -

    - 0.12 1.1

    0.096 0.016 0.025 0.015 0.033 - -


    0.002 0.001 0.225

    0.029 0.001

    0.003 0.017

    0.0001 0.0005



    - -

    0.005 0.70 0.78 0.150 0.203 0.06 0.033 0.63 -


    - - -

    0.325 - -

    - - -


    TABLE IV (Continued)

    Estimated level in product

    as consumedb Food intake Sulfite intakec Category Subcategory (ppm SO*) (glcapitalday) (mg/capita/day)

    Dried vegetables Potatoes I5 3.2 0.24 Frozen vegetables Potatoes 20 1.7 0.034 Vegetable juices Total 5 10 0.05

    - - Sauerkraut juice 100

    a Compiled largely from the report entitled The Reexamination of the GRAS Status of Sulfiting Agents, prepared by the Life Sciences Research Office (Federation of American Societies for Experimental Biology), January, 1985.

    Reported as ppm of SO, equivalents using total SO2 data in most cases. Reported as SO2 equivalents. Includes dried apples, apricots, peaches, golden raisins, cherries, figs, pears, tropical fruit (pineapple,

    mango, papaya), dried fruit mixes, cake with dried fruit, and dried fruit fillings. Exceptions are dark raisins and prunes which are not sulfited.

    mycotoxins, aflatoxins B, and G,, can be degraded by bisulfite, although rather high levels are required (Doyle and Marth, 1978a,b). The complete destruction of aflatoxin B, in corn requires up to 10% bisulfite treatment for 72 hr, but chickens fed this corn were unaffected (Hagler et al . , 1982). The bisulfite addi- tion product with aflatoxin B , has been identified (Hagler et al., 1983). Naturally occurring aflatoxin M, in milk can be destroyed by sulfite addition also (Ap- plebaum and Marth, 1982). Burroughs (1977) showed that the mycotoxin, pa- tulin, could be destroyed by SO,, but concentrations of 2000 ppm were neces- sary to remove 15 ppm patulin from apple juice and cider. In brewing, the presence of SO, in production malts seems to control the formation of N-nitro- sodimethylamine (NDMA) in the malt (Lukes et al., 1980). Malt systems con- taining 30 mg/kg residual SO, always had less than 10 pg/kg NDMA. At lower SO, levels, progressively higher concentrations of NDMA were found. The mutagenicity of coffee can be suppressed by sulfites in the concentration range of about 200 ppm (Suwa er al., 1982).


    1 . Free versus Total SO,

    Numerous methods for SO, and sulfite residue analysis have been developed over the years, and it is beyond the scope of this review to provide a critical analysis of all of these methods. However, comments must be made on the relative merits of measuring total SO, versus free SO,.


    Sulfites exist in foods as sulfurous acid, inorganic sulfites, and a variety of forms of combined sulfites. Complex equilibria dependent on a number of fac- tors control the amount of sulfite in each of these states, as will be discussed in the next section. Methods for the measurement of free SO, are aimed at detection of undissociated sulfurous acid, bisulfite ions, and sulfite ions. Methods for the measurement of total SO, are aimed at detection of these substances plus some of the combined forms of the sulfites. Generally, the methods for analysis of total SO, can be subdivided into two groups: those in which the combined SO, is liberated by distillation from acid, and those in which the combined SO, is liberated by treatment of an extract with excess alkali. SO, will not be liberated from all forms of combined sulfite by either of these treatments; some combined sulfites are quite stable. The levels of combined sulfites are not included in some methods of analysis as a distinct determination. They are often calculated from the differences between total SO, and free SO,, and thus are underestimates representing only the dissociable forms of combined sulfites. Many methods exist for the measurement of free, total, and combined SO,. The most widely used methods were developed many years ago and were critically reviewed by Josl yn and Braverman ( 1954).

    2 . Analysis of Free SO,

    Methods for the measurement of free SO, are largely variations on the original iodometric titration method developed by Ripper (1892). The titration is done following acidification which is relied upon to reduce the rate of dissociation of combined sulfites. Several problems can occur with the measurement of free SO, by these procedures: Iodine-reducing substances other than SO, may be present, recombination of SO, with carbonyls may occur, oxidation to sulfate may occur, some weakly bound forms of SO, may be dissociated, such as the anthocyanin complexes in wine, and colored substances may interfere (Bruno et al., 1979; Joslyn and Braverman, 1954; Ponting and Johnson, 1945; Vahl and Converse, 1980). Oxidation to sulfate (Joslyn and Braverman, 1954) and the presence of other iodine-reducing substances (Ponting and Johnson, 1945) can be controlled. However, the method is also subject to considerable error and poor precision, at least for the analysis of SO, in wine (Vahl and Converse, 1980). We have found this method to be unsuitable for the determination of free SO, in many foods. Fujita et al. (1979) noted that extreme care must be taken to prevent dissociation of combined sulfites in these methods. Improved methods, including aeration- oxidation procedures (Buechsenstein and Ough, 1978; Fujita et al., 1979; Ogawa et al., 1979; Mitsuhashi et al . , 1979; Rankine and Pocock, 1970), polar- ographic methods (Bruno et al., 1979), flame photometric detection by gas chromatography (Hamano et al . , 1979; Mitsuhashi ef al., 1979), ion chro-


    matography (Anderson et al., 1983), and enzymatic assays with sulfite oxidase (Beutler, 1984), have been developed, but not widely tested with a variety of foods.

    3. Analysis of Total SO,

    The most commonly used method for analysis of total SO, residues is the Monier-Williams method (Monier-Williams, 1927), an acidic distillation pro- cedure. The Monier-Williams method for many years has been considered to be the most reliable method for SO, analysis in foods (Joslyn and Braverman, 1954). However, the method is extremely time-consuming and laborious. That has led to considerable effort in the search for suitable alternatives. Although the method is reproducible and apparently free from most types of interference, it is not the perfect method. First, it does not truly measure total SO,, since some forms of combined sulfites are not dissociated during the acidic distillation procedure (Jennings et al., 1978; Wedzicha and Bindra, 1980). Also, the Monier-Williams procedure can be subject to interference by other sulfur com- pounds in foods (Mitsuhashi et al., 1979; Wedzicha and Bindra, 1980). Several improved methods for the analysis of total SO, have been developed, including alkali treatment coupled with idometric titration (Ponting and Johnson, 1945), the Rankine aeration-oxidation method (Buechsenstein and Ough, 1978; Fujita et al., 1979; Mitsuhashi et al., 1979; Ogawa et al., 1979; Rankine and Pocock, 1970), polarographic methods (Bruno et al., 1979), colorimetric procedures (Jennings et al . , 1978; Nury et al., 1959), combustion in oxygen with measure- ment of the sulfur in the gaseous products and the residue (Wedzicha et al . , 1984), sulfite oxidase assays (Beutler, 1984), and flame photometric detection after gas chromatography (Hamano et al . , 1979; Mitsuhashi et al., 1979). These methods have generally been shown to yield results similar to the Monier- Williams method. Several other methods, including a microdiffusion method (Adachi et al., 1979), a direct colorimetric method (Adachi et al., 1979), and gas-sensing probes (Jennings et al., 1978), have been found to have inaccuracies or limited uses. Some of these alternative methods allow simultaneous determin- ation of free, combined, and total SO, (Bruno et al . , 1979; Buechsenstein and Ough, 1978; Fujita et al . , 1979; Mitsuhashi et al., 1979; Ogawa et al., 1979; Ponting and Johnson, 1945; Rankine and Pocock, 1970). However, none of them is suitable for the measurement of all forms of combined sulfite.

    4. A Critical Appraisal of Sulfite Analysis

    For unknown reasons that probably date back many years, the measurements of free and total sulfite residues in foods are referred to as free and total SO,


    analyses. The reason probably relates to the release of SO, under the conditions of the assay. However, SO, is not the form that actually exists in foods. The free SO, methods are actually detecting residues of free inorganic sulfite salts. It would be preferable to refer to them as assays of free sulfite (or free sulfite as SO,) rather than free SO,. The total SO, methods are detecting the free sulfite residues as well as some of the combined forms of sulfite. The combustion method described by Wedzicha et al. (1984) may detect most of the combined forms of sulfite. These methods should be referred to as assays of total sulfite (or total sulfite as SO,) rather than total SO,, although the use of the adjective total may be misleading, since not all forms of combined sulfites can be detected with these assays.

    Considerable data exist in the literature on residual SO, levels in foods (some of that data can be found in Table IV). However, for several reasons, we have some reservations about these data. Many methods exist for the measurement of residual sulfite levels in foods, and correlations between the various methods have not been established for most foods. Therefore, it is difficult to evaluate the validity of some of the published residue data. Some of the methods used to generate these residue data have subsequently been shown to give erroneous results. Second, very little residue data are available for sulfited foods obtained from the marketplace. Much of the available data were obtained from products sampled immediately after processing. Therefore, the effects of storage and any differences with standard, present-day commercial practices have not been taken into account. Much of the available residue data are from fairly old studies, and treatment conditions have probably changed in the intervening years. As men- tioned previously, the effects of preparation on residual sulfite levels have essen- tially been ignored in previous work.

    However, the most important issue surrounds the question of what should actually be measured. Some of the previous data are for free SO, and some are for total SO,. As detailed in Section 111, we speculate that the free inorganic sulfites are much more likely to be implicated in the toxic and asthmatic reactions than are the organic sulfite adducts. For consumer exposure assessments, one would ideally wish to measure exposure to the forms of sulfite that cause toxic or hypersensitivity reactions, most likely the free inorganic sulfite residues or the free sulfite residues plus those combined forms of sulfite that would be expected to release SO, in vivo during the digestive or metabolic processes (see Section 111). Most of the exposure assessments have used total SO, residues. Yet, the conditions of the Monier-Williams distillation procedure are much more severe than any conditions that would be encountered in vivo. Therefore, the Monier- Williams procedure may detect some combined forms of sulfite that would not be expected to release SO, in vivo. More information is needed on the in vivo stability of various combined forms of sulfite in comparison to their recovery in


    the Monier-Williams or other procedures for total sulfite analysis. While the Monier-Williams procedure may detect excessive amounts of SO, for accurate exposure assessments, the free sulfite analyses may not be suitable because they do not detect the combined forms of sulfite that release SO, in vivo.

    The confusion is magnified by the numerous variations and modifications of the procedures for the determination of free and total sulfite. The alternative and/or modified procedures in many cases yield different results for the same sample. Standardized Association of Official Analytical Chemists (AOAC) pro- cedures exist for both free and total SO, analysis (Horwitz, 1975), but many analysts use modifications of these methods. Therefore, it is difficult to compare the data of one analyst with those of another. This uncertainty further confuses the situation for those performing exposure assessments.


    I . EfSect of pH on Su&tes

    Sulfur dioxide dissolves readily in water-producing sulfurous acid, H,SO,. On treatment with alkali, sulfurous acid yields sulfites, bisulfites, and meta- bisulfites. These inorganic forms of sulfites are in equilibrium with one another in aqueous solutions and the concentration of each species is dependent on pH. At high pHs, SO$- is the predominant species, while at very low pHs, H,SO, predominates. At intermediate pHs, HSO, predominates, reaching a maximum concentration at pH 4.0. SO, can be evolved from H,SO,, but only at acid pHs. Table V shows the maximum percentage of SO, that could be liberated at various pHs. Note that no SO, can be evolved from a solution until the pH drops to pH 4.0 or below. A more complete discussion of the distribution of the various forms of sulfite as a function of pH can be found elsewhere (Green, 1976; Joslyn and Braverman, 1954).


    PH Free SO2 (%)

    I 86 2 37 2.5 16 3 6 4 0.5

    From Green (1976).


    2. Fate of Sulfites in Foods

    The sulfites react readily with a variety of food constituents, including al- dehydes, ketones, reducing sugars, unsaturated organic compounds, browning intermediates, proteins, and anthocyanins, to name but a few. The extent of reaction is dependent on the pH, temperature, concentration of sulfite, and the reactive components of the food matrix. An equilibrium always exists between the combined and free forms of the sulfites, although some of the reactions are virtually irreversible, while others are more readily reversible (Wedzicha et al., 1984). The reactions remove free sulfites from the food which often diminishes their effectiveness in the food product. The dissociable, combined forms of sulfite can serve as a reservoir for free sulfite, but the irreversible reactions remove sulfite permanently from the pool of free SO,. As we noted earlier, most of the desirable actions of the sulfites are dependent on the free forms; the combined sulfites are usually ineffective. Therefore, treatment levels for specific foods have historically been chosen to provide an active, residual level of free SO, throughout the typical shelf life of the product.

    3. Reactions of Sulfites with Carbonyls

    Sulfites are known to have a particular affinity for reactions with aldehydes and ketones. The primary products of these reactions are hydroxysulfonates (Joslyn and Braverman, 1954). Burroughs and Sparks (1973a-c) have made a very thorough study of the reactions between carbonyls and sulfites. The reaction rates between carbonyls and sulfites are fast, and the equilibrium constants overwhelmingly favor the hydroxysulfonates in the range of pH 1-8 (Burroughs and Sparks, 1973a). At higher pHs, more dissociation occurs (Burroughs and Sparks, 1973a). Temperature changes in the range of 0-60C have little effect on the stability of acetaldehyde hydroxysulfonate, but pyruvic acid hydroxysulfo- nate shows progressively greater dissociation at higher temperatures (Adachi et al., 1979). Burroughs and Sparks (1973b,c) have succeeded in developing a model for the sulfite-binding properties of wines and ciders on the basis of the concentrations of various types of carbonyl compounds. Acetaldehyde is the primary sulfite-binding substance in these beverages, since it is a primary prod- uct of the fermentation process (Burroughs and Sparks, 1973b). The ability of sulfites to inhibit the nonenzymatic and enzymatic browning reactions is largely due to their reactions with the carbonyl intermediates produced in these reactions (Haisman, 1974; McWeeny et al . , 1974). Wedzicha et al. (1984) demonstrated that the reaction of sulfites with carbonyls generated from nonenzymatic brown- ing was the major reaction of sulfites in dehydrated vegetables.

    The hydroxysulfonates of carbonyl compounds are rather stable reaction prod- ucts. The most stable carbonyl hydroxysulfonates would be those formed with


    the a,P-unsaturated carbonyl intermediates of the browning reaction (McWeeny et al., 1974). However, acetaldehyde hydroxysulfonate should be considered an extremely stable reaction product (Adachi et al., 1979; Burroughs and Sparks, 1973a). Although some of the other carbonyl hydroxysulfonates are somewhat less stable, such as the pyruvic acid product (Adachi et a l . , 1979; Burroughs and Sparks, 1973a), this entire class of reaction products is stable. They would not be expected to dissociate to any great extent in the stomach, and dissociation in the small intestine would not even be particularly rapid.

    4 . Reactions of Sulfites with Reducing Sugars

    The sulfites are also capable of reactions with reducing sugars, such as glucose. Sugar hydroxysulfonates are not formed as readily as carbonyl hy- droxysulfonates; a considerable molar excess of reducing sugar is usually re- quired (Adachi et al . , 1979). The glucose hydroxysulfonates are also much less stable than their carbonyl counterparts (Green, 1976; Joslyn and Braverman, 1954). At the neutral pH of 7.0, the reaction between sulfites and glucose would be very rapid, but dissociation would predominate, leaving the bulk of the sulfite in the free form (Green, 1976). At pH 4 to 5 , the reaction rate would still be fairly fast and dissociation would be much slower, thus favoring the combined form (Green, 1976). At pH 2, the reaction rate would be slow but the product, once formed, would be stable (Green, 1976). The sugar hydroxysulfonates, though not as stable as the carbonyl hydroxysulfonates, would still be stable in the stomach, although dissociation in the small intestine would be predicted.

    5. Reactions of Sulfites with Proteins and Amino Acids

    The disulfide bonds of cystine, peptides, and proteins can be cleaved by sulfite, resulting in the formation of a thiol (R-SH) and an S-sulfonate (R-SSO,) (Means and Feeney, 1971). The reaction goes essentially to completion with free cystine at physiological pH, but the disulfide bonds of many proteins are unreac- tive presumably because of steric hindrance or an unfavorable electronic environ- ment (Gunnison, 1981). Denatured proteins are more susceptible to this sul- fitolysis reaction.

    Sulfites also react with cysteinyl residues of proteins and peptides (Green, 1976; Schroeter, 1966). The reaction product with cysteine is P-carboxyethylthi- osulfonate (Schroeter, 1966). The formation of amine bisulfites has also been reported from the reaction of sulfites with tertiary amines and Schiff's bases arising from the browning reaction (Joslyn and Braverman, 1954). These reac- tions with amines and thiols would predominate in food systems such as flour doughs where gluten is a major constituent (Thewlis and Wade, 1974).

    Methionine can be oxidized to methionine sulfoxide by sulfites via a free


    radical mechanism (Yang, 1970, 1984). Tryptophan can be destroyed by a sulfite-induced free radial reaction (Yang, 1973, 1984). The relative stabilities of the sulfite adducts with proteins and amino acids are unknown.

    6. Reactions of Sulfites with Vitamins

    Sulfites can apparently react with a number of vitamins, including thiamine, ascorbic acid, vitamin B,,, and vitamin K.

    The sulfitolysis of thiamine has been known for many years and is considered deleterious, since it destroys the nutritive value of this vitamin (Schroeter, 1966). For this reason, sulfites cannot legally be added to foods, such as meat, that are considered to be prime sources of thiamine in the United States. Such additions are allowed in other countries. The sulfitolysis of thiamine is an irreversible nucleophilic attack of sulfite on the quaternary nitrogen of the thiazole ring; the products of the reaction are pyrimidine sulfonic acid and 4-methyl-5P-hydroxy- ethylthiazole. The kinetics of this reaction are first order with respect to both reactants, and the half-life of 10 p M thiamine in the presence of 1 mM sulfite is about 13 hr (Gunnison, 1981).

    Sulfites can also react with the nonenzymatic browning products of ascorbic acid. Wedzicha and McWeeny (1974a,b) investigated the reaction products of ascorbic acid, amino acids, and sulfite; a large portion of the total sulfite residues obtained from such reactions could not be recovered efficiently by the acidic distillation procedure of Monier-Williams (1927). They succeeded in identifying 3-deoxy-4-sulfopentosulose and 3-deoxy-4-sulfohexosulose as two extremely stable sulfite adducts in the reaction mixtures and sulfited, dehydrated cabbage (McWeeny, 1979; McWeeny et al., 1974; Wedzicha and McWeeny, 1974b). Up to 80% of the SO, in dehydrated cabbage may be in the form of such stable adducts (McWeeny, 1979). The sulfite adducts are extremly stable in vivo (Walker et al., 1983a).

    Cobalamins react readily with sulfites. Most of the studies involve hy- droxycobalamin, a naturally occumng form of vitamin B ,2 (cyanocobalamin). A complex is readily formed between hydroxycobalamin and sulfite (Kaczka et al . , 1950). The reaction has a very favorable equilibrium constant (Firth et al., 1969). Cobalamins can actually catalyze the oxidation of sulfite to sulfate (Kaczka et al., 1950; Smith et al., 1952).

    Sulfite adds reversibly to the 2,3 double bond of menadione (vitamin K3), a water-soluble synthetic form of vitamin K (Shih and Petering, 1973). The sulfo- nate adduct is probably dissociable in the body, since the adduct can serve as a source of vitamin K when fed to animals (Nir et al., 1978). The reaction rate for the naturally occumng fat-soluble forms of vitamin K (vitamins K, and K,) with sulfite is unknown.


    Sulfite also forms reversible adducts with the pyridine and flavin nucleotides. Sulfite adds to the 3,4 double bond of nicotinamide adenine dinucleotide (NAD) (Shih and Petering, 1973). The sulfite-NAD adduct is not particularly stable, but its stability is enhanced if the NAD is associated with an enzyme (Gunnison, 1981). With the flavin nucleotides, flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), the addition reaction occurs at the N-5 atom of the isoalloxazine ring (Muller and Massey, 1969). The flavin nucleotide-sulfite adducts are even less stable than the NAD-sulfite adduct, but again the stability is enhanced if they are bound to protein (Muller and Massey, 1969). The reac- tions of sulfite with enzyme-bound NAD, FAD, and FMN result in inhibition of the enzyme (Gunnison, 1981).

    Bisulfite has been shown to form nucleophilic adducts with folate and di- hydrofolate (Vonderschmitt et al . , 1967). The reaction can occur at pH 6.5, but, because of an unfavorable equilibrium, it requires a considerable excess of bisulfite to obtain any appreciable quantity of adduct. The adduct is also unstable in the presence of oxygen.

    Sulfites can also destroy p-carotene, the precursor of vitamin A (Yang, 1984; Wedzicha and Lamikanra, 1983). The mechanism apparently involves free radi- cals. Sulfite-induced lipid peroxidation can initiate this reaction.

    7. Reactions of Sulfites with Nucleic Acids and Nucleotides

    A variety of reactions have been described between sulfite and nucleic acids or nucleotides (Gunnison, 198 1; Hayatsu, 1976). Certain of these reactions are thought to be responsible for the mutagenicity of sulfites which is observed in certain systems (see Section III,C,4).

    Sulfites can add reversibly to the 5,6 double bonds of uracil and cytosine and their derivatives (Hayatsu, 1976; Shapiro et al . , 1970b, 1973). With uracil, the reaction is most rapid at pH 7, is readily reversible at pHs above and below pH 7, and is dependent on relatively high sulfite concentrations (Hayatsu et a l . , 1970; Pitman and Jain, 1979; Shapiro et a l . , 1976). Uridine can be regenerated from the sulfonate adduct by removal of free sulfite from the reaction mixture (Rork and Pitman, 1974). The cytidine adducts are most stable at acid pH, and their formation is also dependent on high concentrations of sulfite (Shapiro et a l . , 1974). The cytidine adducts are unstable at physiological pH (Shapiro et a l . , 1974). RNA and single-stranded DNA can be deaminated via this mechanism (Shapiro et al . , 1970a, 1973).

    The 5,6-dihydrocytosine-6-sulfonate can be deaminated under some condi- tions to yield the uracil adduct (Shapiro et al . , 1974). The deamination is cata- lyzed by sulfite and occurs optimally at pH 5 (Shapiro e f al . , 1974). At low sulfite concentrations and physiological pH, the rate of this conversion is ex-


    tremely slow (Slae and Shapiro, 1978). However, this conversion is thought to be responsible for some of the observed mutagenic effects of sulfites (see Section III,C ,4).

    Sulfite can also catalyze the transamination of cytosine and its derivatives with primary and secondary amines to produce N4-substituted cytosines (Shapiro and Gazit, 1977). 5,6-Dihydrocytosine-6-sulfonate is an intermediate in this reaction also. The reaction is slow and requires high concentrations of sulfite, but can occur at physiological pH. Polycytidylic acid can also be involved in such reactions.

    Gunnison (198 1) suggests that sulfite-induced mutations may also be the result of deamination of 5-methylcytosine to thymine.

    In certain reaction mixtures, aerobic oxidation of sulfite to free radicals can occur (Hayatsu, 1976). These free radicals can lead to a variety of reactions with nucleic acids and their derivatives, including cleavage of the glycosidic linkages of uridine and cytidine (Kitamura and Hayatsu, 1974), fission of the chains of polyuridylic acid and polycytidylic acid (Kitamura and Hayatsu, 1974), breakage of the phosphodiester bonds in DNA in phage T7 (Hayatsu and Miller, 1972), and addition to the double bonds of 4-thiouracil and 6-isopentenyladenosine, minor base constituents of yeast transfer RNAs (Stacey and Harris, 1963).

    The reactions of sulfites with nucleic acids can modify the activities of these biomolecules in in vitro systems. These modifications are detailed in the review by Gunnison (1981). Shapiro and Braverman (1972) demonstrated that sulfite adduct formation in poly(U) interferes with its ability to form helical complexes with poly(A) and the ability of poly(U) to code for phenylalanine incorporation into protein. Similar modification of messenger RNA from coliphage MS2 and ribosomal RNA from Escherichia coli led to decreased incorporation of amino acids into proteins (Braverman et al., 1975). Uracil-sulfonate adducts can also interfere with the DNA polymerase reaction (Kai et al., 1974). Sulfite-induced free radical reactions can inhibit the transforming activity of DNA from Bacillus subtilis (Inoue et al., 1972), can inactivate bacteriophage A (Kudo et al., 1978), and can cross-link the maturation and coat proteins with nucleic acids in RNA bacteriophage MS2 (Turchinsky et al., 1974).

    8. Reactions of Sulfites with Pigments

    Anthocyanins and other phenols in wines can also react with sulfites (Bur- roughs, 1975; Somers and Evans, 1977). Anthocyanin complexes with sulfites are quite unstable, dissociating even under acidic conditions (Burroughs, 1975). Therefore, the titrimetric methods for free SO, would be expected to give er- roneously high results when applied to red wines (Burroughs, 1975). The antho- cyanin-sulfite complexes would be expected to dissociate readily on exposure to stomach acid.


    9. Reactions of Sulfites with Fatty Acids

    Sulfites in the concentration range of 0.5-10 mM can induce the oxidation of unsaturated fatty acids in corn oil and free linolenic acid (Kaplan et al., 1975; Lamikanra, 1982), presumably through a free radical mechanism. Sulfites also enhanced lipid peroxidation in a rat liver homogenate (Inouye et al., 1980). Southerland et al. (1982) showed that bisulfite could react directly with the double bonds of mono- and polyunsaturated fatty acids. This reaction could be inhibited by vitamin E and butylated hydroxytoluene.

    10. Fate of Sulfites in Specific Foods

    The proportion of added sulfite existing in the combined form would be variable from food to food. In most foods, the combined forms of sulfite would be expected to predominate. An exception would be sulfited lettuce, where virtually all of the sulfite is present in the free form (Taylor et al . , 1985). A comparison of free, combined, and total SO, residue levels in several foods as determined by the aeration-oxidation method is interesting (Mitsuhashi et al., 1979). The percentage of the total SO, existing as free SO, is 2.3% for white wine, 22.3% for concentrated orange juice, 14.8% for molasses, 34.4% for corn starch, and 32.3% for frozen peeled shrimp. The remainder of the sulfite is in the combined form.

    The fate of sulfites in foods is thus an extremely complex situation. The combination with organic constituents, the equilibrium between the various in- organic forms, the volatilization of SO,, and the oxidation to sulfate can all be important reactions. The relative importance of these reactions is dependent on the particular food involved. Although many studies have reported residual SO, levels in foods, a few studies with radiolabeled sulfites stand out as the only systematic investigations of the fate of sulfites in foods.

    Thewlis and Wade (1974) studied the fate of radiolabeled sulfites in hard sweet biscuit doughs. The results indicated that 93% of the sulfite sulfur added to the dough remained in the biscuit. About 30% of the sulfite was converted to sulfate and another 73% appeared as combined sulfites, probably S-sulfopro- teins. Less than 0.2% remained, as inorganic sulfite (free SO,) in the finished, cooked biscuits.

    Gilbert and McWeeny (1976) examined the fate of radiolabeled sulfites in dehydrated vegetables (cabbage, carrots, and potatoes). .4lthough there were some methodological problems with this study, between 35 and 45% of the sulfite was free or dissociable sulfite in the dehydrated potatoes, while a small percentage was converted to sulfate and the remainder was stable sulfite adducts. The effects of storage, preparation, and cooking on the residual level in the dehydrated products were not examined.


    Further evaluation of the fate of [35S]sulfite in dehydrated cabbage and carrots was conducted by Wedzicha et al. (1984). In cabbage, 28.9% of the added sulfite remained as sulfite, and 8.7% was oxidized to sulfate. C-Sulfonates accounted for 34.3%, while 29.0% was attributed to a combination of disulfites hydroxysulfonates, and S-sulfonates. In carrots, only 3.0% of the added sulfite remained as sulfite, while 10% was oxidized to sulfate. The higher level of remaining sulfite in cabbage was attributed to the higher pH of blanching (pH 9- 10) for this product by comparison to carrots (pH 5). With carrots, C-sulfonates accounted for 5.0% of the added sulfite, while 81.0% was attributed to the combination of disulfites, hydroxysulfonates, and S-sulfonates. In dehydrated cabbage, the number of combined sulfite products was large; eight radioactive peaks were obtained by ion-exchange chromatography.

    McWeeny et al. (1980) also studied the fate of sulfites during production of strawberry jam from sulfited strawberries. The strawberries absorbed 78% of the radiolabeled sulfite from the sulfite liquor, but 98.5% of this sulfite was either reacted chemically or lost during jam production: 50% was boiled off (jam is acidic, so volatilization of SO, is likely), 38% was converted to combined sulfite adducts during storage of the strawberries, 11 % was converted to combined sulfite adducts during jam making, and only 1.5% remained as inorganic sulfite in the finished jam. In a study of the fate of sulfites in dried fruit, McBean (1967) showed that

    much of the SO, was lost on drying and that of the remaining sulfite, 80-90%, was in the combined form. Further losses of sulfite occurred during storage. In shrimp, most of the residual sulfite is located on the shell (Weingartner et

    al., 1977). That sulfite is mostly removed with the melting ice water, and further reductions are accomplished with rinsing in water or hypochlorite solutions (Weingartner et al . , 1977). With 1.25% bisulfite dips, only 11.7% of the sulfite residue on the shell remained after 15 days of storage on ice at 2"C, while no residual sulfite was noted in the muscle following 10 days of storage (Weingartner et al . , 1977).

    11. Critical Factors in Determination of the Fate of Sulfites in Foods

    The first critical step in determination of the fate of sulfites in foods is the absorption of the sulfites from dip solutions or SO, from the atmosphere into the product. With dried fruits, the concentration of absorbed SO, is a function of the sulfite concentration in the dip solution, the immersion time, and the pH (Staf- ford et al., 1972). SO, absorption is higher at pH 2.5 than at pH 4.5 (Stafford et al., 1972). With table grapes, sulfitation is accomplished by release of SO, from in-package generators during transport and storage (Nelson, 1983; Nelson and Ahmedullah, 1973, 1976). The released SO, in the atmosphere of the container disappears within 3 weeks (Nelson and Ahmedullah, 1976). The level of sulfites


    in the grapes is of course proportional to the amount of SO, exposure and varies from 17 to 40 ppm, depending on the temperature treatment and package type (Nelson and Ahmedullah, 1973).

    The second important factor in determining the fate of sulfites in foods is the nature of the processing treatments. As can be seen from the studies already cited (Gilbert and McWeeny, 1976; McWeeny et al., 1980; Thewlis and Wade, 1974; Wedzicha et al., 1984), sulfite levels can be altered in a number of ways: (1) The sulfites can be physically lost as SO, if the pH of the product drops below pH 4.0, especially if the product is heated; (2) much of the sulfites in nonacid products can be converted into combined sulfite adducts, many of which remain to be characterized; (3) some of this combined sulfite will be in the form of extremely stable products, which cannot be recovered by conventional methods, so it will be lost as far as analysis is concerned; and (4) oxidation of sulfite to sulfate can occur in some foods and may be particularly significant in wines and flour doughs, perhaps because it is catalyzed enzymatically in these foods. Leaching of sulfite brine solutions is also an important step in lowering sulfite residuals in many products. An example would be maraschino cherries.

    The third step in determining the fate of sulfites in foods is the effect of storage on residual sulfite levels. Storage almost always diminishes the amount of in- organic sulfite or free SO, in the product. In bottled red wine, free SO, is lost on storage; the loss is correlated with the oxygen content of the wine and presum- ably represents conversion to sulfate (Jacobs, 1976). In dehydrated potatoes, 46- 68% losses in residual sulfite were noted on 24 weeks of storage at 75F, depending on the type of storage container (Lisberg and Chen, 1973). In dried apricots, loss of SO, is also dependent on the type of package, with greater losses from air-permeable packages than air-impermeable packages (Davis et al., 1973). In oxygen-permeable packages, some of the loss was attributed to sulfate formation (Davis et al., 1973). Between 50 and 80% of the total SO, residue is lost from dried apricots in 48 weeks of storage at 25C (Davis et al., 1973). In dried apples, the loss of residual SO, was shown to be a function of storage temperature (Sayavedra and Montgomery, 1983). At 1C, virtually no SO, was lost in 400 days of storage, while at 38C, 80% of the residual SO, was lost within 400 days (Sayavedra and Montgomery, 1983).

    The final step in determining the fate of sulfites in foods is the effect of preparation on residual sulfite levels. This final step has received little attention until the recent concern over sulfites and virtually nothing has been published on the subject. The cooking of sulfited Thai noodles diminishes the total SO, level by about 70% (Kingkate et al. , 1981).

    Apparently, processing, storage, and preparation act largely to lower the lev- els of residual sulfite in foods. The actual amounts of free and total sulfite available at the point of consumption have received little study, but it is probable that the lowest free sulfite concentrations would exist at that point. If, as we


    propose in Sections 111 and IV, the inorganic sulfites are more toxic than com- bined sulfites and are responsible for the asthmatic reactions, then this situation is very beneficial. Much more study of the effects of storage and preparation on residual sulfite levels will be necessary, along with confirmation of our hypoth- esis that combined sulfite adducts are not triggers of the asthmatic response.


    Treatment levels bear virtually no resemblance to residual levels for the sul- fites. As we have noted, numerous factors and reactions can affect residual levels of the sulfites. The fate of sulfites in foods would be different for each food and each set of treatment conditions. Subsequent storage and preparation conditions would also affect residual levels.

    Much is heard about the losses in SO, that occur during processing, storage, and preparation. As we have described in the previous section, true losses of sulfite as SO, can occur in foods with pHs below pH 4. The pH-dependent losses of SO, can be influenced by time, temperature, humidity, light, and other factors. True losses of SO, can also occur through physical separation processes such as leaching. Losses of sulfite by oxidation to sulfate will occur in many foods, but seem to be most important in fermented foods where conversion back to sulfite is always possible. However, in the majority of cases, most of the lost SO, is really not lost from the product at all. It is merely converted into stable combined sulfites that will not release the SO, under the usual assay conditions. These losses may be important in terms of converting the sulfites into forms that are less likely to initiate an asthmatic or toxic response. Thus, the substantial losses in inorganic sulfites that occur between treatment and consumption may be more important in terms of lowering the hazard of sulfites in foods.

    In the future, emphasis should be placed on assessment of consumption levels rather than treatment levels. Further knowledge of the role of different forms of sulfites will be necessary before we will know exactly what we should be mea- suring, however (see Section 11,E,4).


    I . Consumer Exposure to Su@tes from Foods

    The residue data available for assessing consumer exposure to sulfites in foods are woefully inadequate. We have already detailed our misgivings about the methods of measurement, the reliance on total instead of free and combined SO,, and the lack of information on the effects of storage and preparation on residue levels.

    The Codex Alimentarius Commission (1975) has established a potential daily


    intake for sulfiting agents as SO, of 2.1 mg/kg or 126 mg for a 60-kg person. This level would represent the upper limit of sulfite intake in the population. The Expert Panel on Food Safety & Nutrition of the Institute of Food Technologists (1975) agreed with this estimate of the upper limit of intake. However, they noted that wide variations in sulfite intake would occur in the population. The panel concluded that most Americans consume no more than 10- 15 mg total SO, per day or about 0.2 mg/kg for a 60-kg man. The Ad Hoc Review Group on the Reexamination of the GRAS Status of Sulfiting Agents (Life Sciences Research Office, 1985) estimated total sulfite intake of 10 mg/capita/day from the con- sumption of food, wine, and beer, with food accounting for perhaps 6 mg of this total. They estimated that the 99th percentile intake for total SO, does not exceed 180 mg/capita/day. While these estimates may be reasonable, we believe that, based on our misgivings about the sulfite residue data, a continuing evaluation of the levels of consumer exposure to sulfites would be desirable. In particular, it would be interesting to have data on consumer exposure to inorganic and com- bined sulfite residues in addition to data on total SO, residues.

    Obviously, some foods will contribute more heavily to consumer sulfite ex- posure than others. Many sulfited foods have rather low residue levels. Dried fruits, dehydrated vegetables, dehydrated potatoes, wine, and sulfited restaurant salads and potatoes will contribute higher levels of sulfites than most other products. If these foods are used routinely in the diet, the 10- to 15-mg estimates could easily be exceeded. In a hypothetical meal with 250 ml of wine containing 150 pprn total SO,, a restaurant salad (100-g serving) containing 600 ppm total SO,, and dried apricots (25-g serving) containing 2500 ppm total SO,, a total SO, intake of 160 mg would be achieved.

    The joint FAO/WHO Expert Committee on Food Additives (1974) established an acceptable daily intake (ADI) for sulfiting agents as SO, of 0.7 mg/kg or 42 mg for a 60-kg man. The Committee used animal toxicity data to arrive at this AD1 (see Section 111,C). The AD1 could be exceeded by one sulfited restaurant salad, several dried apricots, or 250 ml of wine having 175 ppm SO,. The 99th percentile of total SO, intake (180 mg/capita/day) is equivalent to 3 mg/kg, although the average per capita intake of total SO, of 10 mg/day equates to only 0.17 mg/kg (Life Sciences Research Office, 1985).

    2. Consumer Exposure to Sulfites from Other Sources

    The major alternate sources of sulfites and SO, are drugs and the atmosphere. The use of sulfites in pharmaceuticals has been reviewed by Schroeter (1966). Their use dates back to at least 1940 and probably earlier. They are used in drugs primarily as antioxidants. Sulfiting agents have been used in sympathomimetic amines, sulfonamides, antibiotics, steroids, vitamins, dextrose solutions, eye


    medications, antisyphilitic drugs, indigocarmine solutions, anticoagulants, heparin, phenothiazine compounds, bronchodilators, local anesthetics, and mor- phine (Schroeter, 1966). The sulfites are commonly, but not universally, used in such drugs. The levels of use of sulfites in drug formulations have not been strictly controlled. They have historically been used at the minimum concentra- tions necessary to provide adequate antioxidant protection. The concentration of sulfites in the majority of the drug formulations has been in the range of 0.01- 1 .O%, although a few may contain higher levels (Life Sciences Research Office, 1985; Schroeter, 1966). Exposure via drugs can be high but would be sporadic for most individuals. However, some patients would have continual exposure to sulfited drugs over an extended period of time. With drugs, the routes of ex- posure may be respiratory and intravenous as well as oral.

    SO, is evolved into the atmosphere from volcanoes and industrial processes. Estimates are that 146 million tons of SO, are emitted annually into the atmo- sphere from industrial processes and motor vehicles (Stem et al., 1973). The chief sources are coal and oil burning and production of sulfuric acid. Estimates are that another 1.5 million tons of SO, are released into the atmosphere each year from volcanic activity, although considerable variations in volcanic activity occur. National standards require that the average level of SO, in the atmosphere over a 24-hr period not exceed 0.14 ppm on more than one day each year and that the average annual atmospheric concentration not exceed 0.03 ppm. Daytime SO, levels in industrialized areas could be expected to exceed 0.14 ppm on occasion. Workplace environments can also be contaminated with SO,, with several parts per million occurring at times. The threshold limit value (TLV) for SO, in the workplace is 2 ppm averaged over an 8-hr day (American Conference of Government Industrial Hygienists, 1982). Oxidation of SO, in the atmosphere to sulfuric acid occurs and is the focus of concerns regarding acid rain. The rate of oxidation is faster at higher humidity levels and may be as great as OS%/min in fogs. Exposure to atmospheric SO, occurs via the respiratory route rather than the oral route.

    The contributions of drug sulfite residues and atmospheric sulfur oxides to the total intake of sulfite have not generally been included in exposure assessments for food additives. Estimates of daily exposure to SO, can be derived from the 0. I4-ppm, 0.03-ppm, and 2-ppm limits described above. Respectively, such exposures would provide total SO, exposures of 3.8 mg/24 hr, 0.85 mg/24 hr, and 24 mg/8 hr (Life Sciences Research Office, 1985).


    A. METABOLISM OF SULFITES The key to the understanding of sulfite toxicity may lie in elucidation of sulfite

    metabolism. Several researchers have proposed that defects in sulfite metabolism


    among certain segments of the human population may put them at greater risk to the possible toxic effects of sulfite ingestion (Calabrese et al . , 1981; Jacobsen et al., 1984). Certainly defective sulfite metabolism in experimental animals can be correlated with enhanced sulfite toxicity (Cohen et al . , 1973). Considerable information is available on the metabolism of free inorganic sulfites. Unfortu- nately, much less information is known concerning the metabolism of combined sulfites.

    I. Free Sulfites

    Free sulfite is metabolized principally by sulfite oxidase, more precisely known as su1fite:cytochrome-c oxidoreductase or sulfite:O, oxidoreductase (EC The product of this oxidative reaction is sulfate, which can be rapidly excreted in the urine. Sulfite oxidase is present in all animal species that have been examined, although species variations are observed in the levels of activity. The rat possesses the highest level of sulfite oxidase activity, having 3-5 times greater activity than rabbits or rhesus monkeys (Gunnison et al., 1977) and approximately 10-20 times greater activity than humans (Johnson and Ra- jagopalan, 1976a,b). Sulfite oxidase can be found in most mammalian tissues except blood, with the highest activities found in the liver, followed by the kidney (MacLeod et al., 1961; Johnson et al., 1977). The enzyme is localized subcellularly in the intermembranous spaces of the mitochondria (Gunnison, 1981; Kessler and Rajagopalan, 1972).

    Sulfite oxidase has been purified from several sources (Cohen and Fridovich, 1971a; Johnson and Rajagopalan, 1976a; Kessler and Rajagopalan, 1972) and studied extensively. The enzyme has a molecular weight of 1 15,000- 120,000, depending on the species, comprised of subunits of approximately 55,000 MW each. Sulfite oxidase is a molybdoprotein (Cohen et al., 1971; Kessler and Rajagopalan, 1972) that also possesses a heme molecule in addition to the apoenzyme (Cohen and Fridovich, 1971b). The reaction mechanism involves the transfer of electrons from sulfite to the Mo6+ site in the enzyme. The electrons are then transferred to the heme moiety and from there to cytochrome c, a constituent of the mitochondria1 respiratory chain. This leads to the ultimate production of sulfate and one molecule of ATP, with the reduction of 4 0, to H,O.

    Sulfite oxidase appears to be the major metabolic pathway for both endoge- nous and exogenous sulfite (Gunnison, 198 1). Endogenous sulfite arises from the metabolism of the sulfur-containing amino acids, cysteine and methionine. The conversion of sulfite to sulfate via sulfite oxidase is the final step in the catabolism of these amino acids.

    Endogenously produced sulfite cannot normally be detected in blood or other tissues due to its rapid oxidation to sulfate and excretion in the urine (Gunnison et


    al. , 1981b). It has been estimated that humans excrete about 25 mmol(2400 mg) in their urine each day, the majority (up to 24 mmol) of which is generated from endogenous sulfite (Institute of Food Technologists Expert Panel on Food Safety and Nutrition, 1975). Rats generate about 0.5 mmol sulfite/kg/day based on urinary sulfate excretion in animals in sulfur balance (Whiting and Draper, 1980). Apparently, sulfite oxidase was evolved to protect animals from endoge- nously produced sulfite (Gunnison, 1981). This same enzyme serves to metabo- lize exogenous sulfite as well.

    The capacity of sulfite oxidase is very high in mammalian species (Gunnison, 198 1; Institute of Food Technologists Expert Panel on Food Safety and Nutrition, 1975). Based on projections from in virro assays of sulfite oxidase, Cohen et al. (1973) calculated that the enzyme could theoretically oxidize sulfite at a rate of 750 mmol/kg/day (48 g of SO,/kg/day). Using perfused dog livers, Wilkins et al. (1968) demonstrated that sulfite could be oxidized at a rate of 0.8 mmol/kg/hr, which equates to a daily rate of 19 mmol/kg (1 200 mg of SO,/kg/day). Oshino and Chance (1975) showed that perfused rat livers were capable of even faster sulfite oxidation, with a rate of 2.4 mmol/kg/hr or 58 mmol/kg/day (3700 mg of SO,/ kg/day). In experiments with intact animals, Yokoyama et al. (1971) and Bhagat and Lockett (1960) observed that dog and rats, respectively, could metabolize inhaled SO, and ingested bisulfite to sulfate readily, with the majority of the dose appearing in the urine as sulfate within a short time after administration. Gibson and Strong (1973) observed that the majority of an oral dose of sulfite equivalent to 50 mg SO,/kg was excreted in the urine as sulfate within 24 hr. They could not detect urinary sulfite, indicating extremely efficient oxidative metabolism. Sulfite is absorbed very rapidly from the gastrointestinal tract (Bhagat and Lockett, 1960; Gibson and Strong, 1973), so only a small portion of any oral dose is excreted in the feces (Gibson and Strong, 1973). Gunnison and Palmes (1976) and Gunnison et al. ( 1 977) evaluated the rate of metabolism of intravenous doses of sulfite in rabbits, rats, and rhesus monkeys. They observed that large intravenous doses of sulfite could be oxidized to sulfate within minutes. The biological half-lives for plasma sulfite were calculated to be 1-2, 3-4, and 10 min in rats, rabbits, and rhesus monkeys, respectively, after intravenous doses of 0.3-0.6 mmol/kg (Gun- nison et al., 1977). The half-lives of sulfite increased somewhat with increasing doses of sulfite, probably because of feedback inhibition of sulfite oxidase by sulfate (Gunnison et al . , 1977). These studies collectively indicate that animals possess sufficient sulfite oxidase to handle both endogenously produced sulfite and rather substantial amounts of exogenous sulfite in addition.

    Perhaps one precautionary note should be added regarding the capacity of sulfite oxidase. Oshino and Chance (1975) demonstrated with perfused rat livers that the rate of hepatic sulfite oxidation is limited by diffusion. Only 25-40% of the infused sulfite dose was removed from the perfusate by the liver on each


    pass. The rates of hepatic sulfite oxidation were similar for normal livers and livers from rats with diminished sulfite oxidase levels when sulfite was infused slowly. Therefore, the diffusion-limited metabolism of sulfite would indicate that, despite the capacity of hepatic sulfite oxidase, some sulfite will pass through the liver unmetabolized. The significance of this finding is uncertain. Obviously, extrahepatic tissues possess considerable sulfite oxidase (MacLeod et af., 1961; Johnson et af., 1977). The studies with intact animals would suggest that sulfite is metabolized rapidly in spite of this diffusion limitation. However, in sulfite-sensitive asthmatics, this limitation means that some absorbed sulfite would be expected to pass through the liver and reach other tissues such as the lungs.

    Free plasma sulfite originating from either endogenous or exogenous sources is not detectable in normal rats, mice, or rabbits, but can be detected in rhesus monkeys challenged with 2 mmol sulfite/kg/day in the drinking water (Gunnison and Palmes, 1973, 1976). Gunnison et af. (1981a) were able to measure up to 70-800 p M sulfite in the plasma of normal and sulfite oxidase-deficient rats after a gastric challenge with 2-9 mmol sulfitelkg. Free plasma sulfite has also been observed in a child with severe sulfite oxidase deficiency (Shih et af., 1977).

    Several other pathways exist for the metabolism of sulfite in addition to the sulfite oxidase pathway. The minor pathways are metabolism to thiosulfate and formation of S-sulfonate compounds. Thiosulfate is produced from the reaction of sulfite with 3-mercaptopyruvate (3-MP), a reaction catalyzed by 3-mercap- topyruvate sulfurtransferase (EC (Westley, 1980). 3-MP arises from cysteine catabolism, as does sulfite. Thiosulfate is detected only at very low levels in the urine of normal humans or rats (Gunnison, 1981; Gunnison et af., 1981b), but is excreted in large amounts among sulfite oxidase-deficient indi- viduals of both species (Gunnison et af., 1981b; Shih et af., 1977). Thiosulfate is somewhat unstable in urine and is likely excreted at greater rates than those measured (Gunnison et af., 1981b). However, urinary excretion of thiosulfate was unchanged from normal in the siblings of sulfite oxidase-deficient patients (Irreverre et af., 1967), indicating that thiosulfate excretion is probably not affected by heterozygous deficiencies of sulfite oxidase.

    S-Sulfonate compounds are formed nonenzymatically by the reaction of sulfite with the disulfide bonds of proteins, cysteine, and glutathione (Cecil, 1963). Urinary cysteine S-sulfonate cannot be detected in normal humans or rats, but has been detected in individuals with sulfite oxidase deficiency in both species (Gunnison et al., 1981b; Johnson et al. , 1980; Shih ef al., 1977). However, only the S-sulfonates with low molecular weights would be excreted in the urine, so such measurements may not be totally indicative of S-sulfonate formation. Since S-sulfonates can be formed in the extracellular compartments, this pathway may have some significance in the disposition of exogenous sulfite, even with the


    existence of high levels of sulfite oxidase. S-Sulfonates have been found in the plasma of rabbits and rhesus monkeys after exposure to sulfite by ingestion or injection and to SO, by inhalation (Gunnison and Palmes, 1973, 1974, 1978). Plasma S-sulfonates are not detectable in rats unless the gastrointestinal tract is bypassed by parenteral administration of sulfite (Gunnison and Palmes, 1978). This has led to speculation that S-sulfonate formation is greater in animals with comparatively lower sulfite oxidase levels (Gunnison et al., 1977). The high activity of sulfite oxidase in rat small intestine (Johnson et al., 1974) may prevent plasma S-sulfonate formation in that species. Protein S-sulfonate com- pounds are very stable in vivo by comparison to sulfite; their biological half-lives can be 1-3 days (Gunnison and Farruggella, 1979; Gunnison and Palmes, 1973, 1978). The mechanism for clearance of protein S-sulfonates from the body is not known.

    A better evaluation of the comparative importance of the sulfate, thiosulfate, and S-sulfonate pathways for excretion of ingested and endogenously produced sulfite can be obtained by comparing the urinary excretion rates. In normal humans, urinary S-sulfonate cannot be detected (Shih et al., 1977; Johnson et al., 1980). Humans have been observed to excrete thiosulfate at a rate of 32 k 13 pmo1/24 hr (Sorbo and Ohman, 1978). Urinary excretion of sulfate in humans can reach 25,000 kmo1/24 hr (Institute of Food Technologists Expert Panel on Food Safety and Nutrition, 1976), although some of this undoubtedly represents ingestion of sulfate as such.

    Obviously, normal individuals have a tremendous capacity to metabolize sul- fite by several different pathways, with the sulfite oxidase pathway being the most important. However, profound sulfite oxidase deficiency has been reported on several occasions in humans (Duran et al., 1979; Irreverre et al., 1967; Mudd et al., 1967; Ogier et al., 1982; Shih et al., 1977). The deficiency is charac- terized by increased urinary excretion of sulfite, thiosulfate, and cysteine S- sulfonate and decreased excretion of sulfate. The deficiency is congenital and may be due to defects in the apoenzyme or in the metabolism of the essential molybdenum cofactor. The individuals with profound sulfite oxidase deficiency exhibit dislocated ocular lenses and severe neurological abnormalities resulting in mental and physical retardation. In at least one case, death occurred at an early age (Institute of Food Technologists Expert Panel on Food Safety and Nutrition, 1975). The levels of sulfite oxidase in humans can be determined using cultured skin fibroblasts (Shih et al., 1977). Using this technique, Shih et al. (1977) identified the parents of a child with sulfite oxidase deficiency as probable heterozygotes by virtue of their intermediate levels of the enzyme. Recently, Jacobsen et al. (1984) have demonstrated that sulfite-sensitive asthmatics may also have heterozygous levels of sulfite oxidase activity.

    Sulfite oxidase deficiency can be produced in rats by feeding diets high in tungsten (W) relative to molybdenum (Mo) (Johnson et al., 1974). The W


    competes with Mo for the Mo-binding site on sulfite oxidase. The loss of func- tional sulfite oxidase is slow; Gunnison et al. ( I98 1 b) observed a half-life of 4 days for hepatic sulfite oxidase at a W:Mo ratio of 5800. Eventually with pro- longed administration of high W:Mo diets, a steady-state level of sulfite oxidase activity is reached at about 1% of the normal adult level for the W:Mo ratio of 5800 (Gunnison et al . , 198 1 b). The amount of activity at the steady-state level is a function of the W:Mo ratio (Gunnison et al., 1981b). Gunnison (1981) has argued convincingly that the sulfite oxidase-deficient rat could serve as a model for sulfite metabolism and toxicity studies in humans, since the activity of the enzyme can be adjusted via the W:Mo ratio of the diet to approximate various human conditions: normal, heterozygous deficient, and deficient. The normal rat possesses considerably more sulfite oxidase activity than the human (Johnson and Rajagopalan, 1976a,b), making it a poor choice for an animal model. The sulfite oxidase-deficient rat displays increased plasma sulfite levels after an exogenous challenge of sulfite and increased urinary excretion of thiosulfate and S-sulfonate (Gunnison et al., 1981b). In the absence of an exogenous sulfite challenge, elevated plasma sulfite is not observed in these rats until the sulfite oxidase activity drops to I-2% of normal rat levels, a testament to the efficiency of endogenous sulfite oxidation (Gunnison et al., 1981b). Apparently, a very small amount of sulfite oxidase is needed to prevent the escape of endogenous sulfite from the cell.

    2 . Combined Sulfites

    Very little information is available on the metabolism of the various combined forms of sulfite. Intuitively, one would expect that some combined sulfites would be metabolized in a manner similar to inorganic sulfites. An example would be glucose hydroxysulfonate; the pH stability of glucose hydroxysulfonate would seem to indicate hydrolysis to the free sulfite under the neutral pH condi- tions of the small intestine. However, this probable instability of the reducing sugar hydroxysulfonates has not been established in vivo through experimenta- tion.

    Gibson and Strong (1974) demonstrated that sulfited proteins may be metabo- lized to sulfate in a manner similar to inorganic sulfite. Oral administration of 35S-labeled sulfited rat serum protein to rats was followed by excretion of 40- 55% of the label in the urine within 24 hr. Most of the urinary label was sulfate. Most of the remaining label was found in the feces; very little (7-17%) was incorporated into body tissues. Sulfited proteins are apparently metabolized to sulfate rather quickly. However, additional studies are needed to determine if other sulfited proteins are metabolized in the same manner as the sulfited rat semm proteins.

    The metabolism of 3-deoxy-4-sulfohexosulose (DSH), a product of the reac-


    tion between an intermediate in the Maillard browning reaction and sulfite (Wedzicha and McWeeny, 1974a,b), has been recently studied in detail (Walker et al., 1983a). [I4C]DSH was administered orally to mice and rats. In mice, 50% of the dose was found in the feces, 13% in the cage washings, and 29% in the urine. In rats, fecal excretion varied between 58.5 and 73%, with urinary excre- tion between 16.5 and 31%. Carcass levels of label in both rats and mice were less than 0.1% after 72 hr. The results indicate that DSH is poorly absorbed, but once absorbed is rapidly metabolized and/or excreted. However, significant transient levels of radioactivity were observed in liver, kidney, lung, and pan- creas of both species and in rat testes and mouse bladder.

    Much further work is needed to define the metabolism of the combined forms of sulfite that predominate in foods.


    Several human challenge trials have been conducted with sulfites and sulfited beverages. Many of these human challenge trials were conducted in the early part of this century; a good review of these studies is provided by Cluzan ef al. (1965). In the earliest of these studies, Leuch (1 895) challenged 30 volunteers with a small amount of wine (30-50 ml) containing various quantities of SO,. He observed that, with SO, quantities above 45 mg of free SO,, the subjects complained of throat irritation, stomach bum, and headache. Walbaum (1906) also reported gastrointestinal irritation in subjects who received much larger amounts-310-620 mg of sodium sulfite at 4 times per day over a 4-day inter- val. Wiley (1907) challenged 12 subjects over a 20-day period with increasing amounts of SO, in aqueous solution (80-100 mg) or in the form of capsules of sodium sulfite (1 15-760 mg). He reported some subjective complaints such as headache, hearing impairment, and weakness, but also observed renal impair- ment and hypochromic anemia. Rost and Franz (1913) challenged human volun- teers with 1.0 g (17 mg/kg) of sodium sulfite per day and noted no gastroin- testinal complaints, but vomiting occurred when the dosage was increased to 4- 5.8 g/day (70-100 mg/kg). This result was substantiated by Lafontaine and Goblet (1953, who observed gastrointestinal irritation and vomiting in humans when sulfite doses in excess of 3.5 mg/kg were administered. While these experiments were interesting and established that very high doses of sulfite are acutely toxic to humans, one must recognize that methods for sulfite analysis may have been rather crude at the time of the early experiments and that the prolonged trial of Wiley (1907) may have been compromised by the known destruction of thiamine by sulfite. Wiley would not have known that thiamine supplementation was necessary.


    Perhaps the best human challenge trial was conducted by Hotzel er al. (1969). They placed 12 volunteers on a normal diet for 15 days followed by a thiamine- deficient diet for 15 days. Six subjects were then challenged with beverages that provided 400 mg of SO, per day (50 mg as NaHSO, and 350 mg as sodium glucose hydroxysulfonate) for 25 days; the other 6 subjects received the same beverages containing no added SO, for 25 days. Sulfite administration was then discontinued for 10 days, and the subjects were given oral thiamine supplementa- tion for 2 days. Clinical examinations, enzyme activity measurements, hema- tocrit values, and erythrocyte counts were taken throughout the experiment on all volunteers. None of the subjects exhibited any abnormalities or changes that could be attributed to sulfite. This study has been taken to indicate that sub- chronic administration of sulfite to humans is without effect even when the subjects are thiamine deficient (Life Sciences Research Office, 1976).


    1 . Acute Toxicity

    The free inorganic sulfites do not have a high degree of acute toxicity. Most of the LD50 studies have involved intraperitoneal or intravenous routes of admin- istration rather than the more relevant oral route of administration. A tabulation of the LD,,s of the sulfiting agents is provided in Table VI. The general order of acute toxicity is intravenous > intraperitoneal > per 0s. By sonie routes of administration in some species, a dose killing 50% of the animals was not achieved; these doses are reported in Table VI as LD,,,. The LD,, values do not always agree when independent studies are compared such as the LD,,s for intravenous administration of Na,SO, to mice (Table VI). Many factors could explain the discrepancies. In particular, it must be remembered that sulfites are unstable in aqueous solutions, so any storage of the solution before dosing would result in a loss of sulfite and an apparent decrease in toxicity. Cohen et al. (1973) determined that the intraperitoneal LD,, of NaHSO, was 181 mg/kg (1 11 mg SO,/kg) in sulfite oxidase-deficient rats as compared to 473 mg/kg (291 mg SO,/ kg) in normal rats.

    The acute toxicities of the combined forms of sulfite have received little study. Lewis and Tatken (1979) list an LD,,, of 1220 mg/kg as SO, for oral admin- istration of acetaldehyde hydroxysulfonate in the rabbit. Walker et al. (1983b) could not determine an oral LD,, for DSH in rats or mice and conclude that the oral LD,, for DSH exceeds 5 g/kg. These scattered results would tend to indicate that some of the combined sulfites are less toxic than the inorganic sulfites, but further studies are needed on additional compounds and on other routes of



    LD5,, LD,,, SO2 equiv. Species Route" Chemical (mg/kg) (mg/kg) (mglkg) Reference





    Guinea pig




    iP iv iv iv iv

    PO Po


    iP iP iv iv Po Parenteral iP iv Po

    iv iv iv iv iv




    sc iP sc iv

    NaHS03 NaHS03 Na2S0 Na2S03 Na2S03 Na2S03 . 7 H 2 0 SO26

    s02c NaHS03 NaHS03 NaHS03 Na2S03 K2S205

    Na2S205 NaHS03 NaHS03 Na2S03 Na2S03 Na2S03

    NaHS03 Na2S03 Na2S03 Na2S03 Na2S03 Na2S03 NaHS03 Na2S03 Na2S03 . 7 H 2 0

    Na2 s 2 O5

    675 130 130 155 175 277

    1040 2000 650 473 115 I15

    1800 500 300 65 - -


    95 95



    - -

    - 244 - -

    416 80 66 79 89 70

    1040 2000 400 29 1 71 58

    1037 337 I85 40

    1435 I52 33

    129 58 48

    102 305 66 1 102 I50 661 189

    Wilkins et al. (1968) Hoppe and Goble (1951) Lewis and Tatken (1979) Jaulmes (1970) Hoppe and Goble (195 I ) Nofre er al. (1963) Jaulmes ( 1970) Jaulmes ( 1970) Wilkins et al. (1968) Cohen et al. (1973) Hoppe and Goble (1 95 I ) Lewis and Tatken (1979) Lauteaume er al. ( 1 969) Ezrielev (1968) Wilkins et al. (1968) Hoppe and Goble ( I 95 1) Lewis and Tatken ( 1 979) Lewis and Tatken (1979) Lewis and Tatken (1979) Lewis and Tatken (1979) Hoppe and Goble (1951) Lewis and Tatken (1979) Lewis and Tatken (1979) Lewis and Tatken (1979) Lewis and Tatken (1979) Lewis and Tatken (1979) Wilkins et a/. (1968) Lewis and Tatken (1979) Lewis and Tatken (1979)

    ip, Intraperitoneal; iv, intravenous; PO, per 0s; sc, subcutaneous. As a 6.5% aqueous solution. As a 3.5% aqueous solution.

    administration before firm conclusions can be drawn. Walker (1984) notes that much less information is available on the toxicity of the combined sulfites than is known about the reactions leading to their formation.

    2 . Subchronic and Chronic Toxicity

    Numerous subchronic and chronic toxicity studies have been conducted on the free inorganic sulfites. For the purposes of this review, the early studies will be


    ignored because of the distinct possibility that many of the toxic manifestations were the result of thiamine deficiency, since the impact of sulfite on thiamine was not recognized at that time. Some of these studies have been reviewed elsewhere (Cluzan et al., 1965; Ti1 et al., 1972a).

    The more recent studies of subchronic and chronic toxicity of sulfites gener- ally fall into two categories: those in which the sulfite was administered with the drinking water and those in which the sulfite was administered with the diet. Both of these approaches have disadvantages. Sulfites are unstable in drinking water; Lockett and Natoff (1960) observed a 20% decline in sulfite levels within 48 hr. Some investigators have ignored the stability problems, making their studies difficult to interpret. The drinking water approach has been favored by some investigators because it avoids the problem of thiamine destruction that is inherent with the incorporation of sulfites into the diet. Gunnison et al. 11981a) showed that sulfites do not destroy thiamine systemically, although Gunnison (1981) notes that sulfite ingested with drinking water might destroy some thiamine in the stomach. The incorporation of sulfites into the diet is also fraught with difficulties, since the sulfites are extremely reactive with other dietary components. These reactions can substantially decrease the free sulfite content of the diet and makes interpretation of the results difficult.

    Many of the recent studies have focused on attempts to confirm the finding reported by Fitzhugh et al. (1946), who administered NaHSO, to rats in their diets for up to 1 year. The diet was often left in the feeder cups unchanged for up to 1 week, which resulted in losses through reaction of up to 75% of the sulfite (Gunnison, 1981). The diets contained 0.05-2.0% NaHSO, (0.08-13 mmol/kg/day) originally. Fitzhugh et al. (1946) noted toxic manifestations at bisulfite levels above 0.1 % that included growth retardation, clinical polyneuritis, spectacle eyes, bleached incisor teeth, brown uteri, atrophy of various viscera, calcified renal tubular casts, atrophy of bone marrow and bone, myocardial necrosis and fibrosis, and gastric squamous epithelial hyperplasia. These results have been questioned because of the diminishing levels of sulfite in the diets and the probable destruction of thiamine in the diet. Fitzhugh et al. (1946) attempted to correct the thiamine deficiency through supplementation. Polyneuritis was not observed in the supplemented animals, but the other toxic manifestations per- sisted. Gunnison (198 1) has questioned whether the thiamine supplementation was sufficient to entirely correct the deficiency. Based on the severity of the manifestations observed in this experiment by comparison to others (see later), we would echo these sentiments and further note that other dietary factors might have been affected by the storage of diet in the feeder cups for prolonged periods which could lead to other deficiencies. Bhagat and Lockett (1964) noted that diet prepared with metabisulfite and stored at room temperature would quickly become deficient in thiamine. On prolonged storage of 3-4 months at room temperature, the diets would cause problems, such as chronic diarrhea, that could not be


    reversed by thiamine supplementation (Bhagat and Lockett, 1964). This is an indication that other factors in the diet may also be destroyed by sulfite addition and contribute to the toxicological evaluations if diets are not properly prepared and stored.

    The results of Fitzhugh et al. (1 946) have not been corroborated in other chronic toxicity studies. Three of these studies have involved the incorporation of sulfites into the drinking water (Cluzan et al., 1965; Lauteaume ef al., 1965; Lockett and Natoff, 1960). Lockett and Natoff (1960) administered 0, 375, and 750 ppm of SO, as Na,S,O, in the drinking water of rats in a 3-year multigenera- tion study. They observed no effects of sulfite on growth, food intake, fecal output, fertility, weight of the newborn, growth during lactation, or any of the pathological signs noted earlier by Fitzhugh et al. (1946). The study of Lockett and Natoff (1960) was compromised by the losses of SO, in the drinking water (10% in 24 hr) and the fact that many of their animals developed respiratory ailments during the course of the experiment. Cluzan et al. (1965) conducted a multigeneration study in rats over a 20-month period, administering 700 ppm of SO, as K,S,O,. They found no evidence of toxicity as mortality, growth rate, feed and water consumption, organ weights, hematological values, clinical symptoms, and reproductive capacity were equivalent to controls. Cluzan et al. (1965) did not provide any evidence for the stability of sulfites in their experi- ment. Lauteaume et al. (1965) administered sulfites by gastric intubation to rats over a 2-year period at a rate of 3 m1/100 g body weight/day. The rats were divided into three groups that received (1) water and 450 ppm SO,, (2) red wine with 110 pprn SO,, or (3) red wine with 450 ppm SO,. The sulfites did not affect growth rates, reproduction, or the development of macroscopic or microscopic lesions.

    The most thorough evaluations of the chronic toxicity of sulfites were per- formed by Ti1 et al. (1972a,b) using incorporation of Na,S,O, into the diets of rats and pigs. Losses of sulfite through reactions with other dietary components were minimized by frequent diet preparation and frozen storage. The amount of sulfite loss was measured and the data were reported using the corrected values. Thiamine was added to the diets 'in sufficient quantities to overcome any thiamine destruction by sulfite. On a percentage basis, sulfite losses were great- est at low sulfite concentrations, while thiamine losses were highest at high sulfite concentrations. Sulfite losses ranged from 4.5 to 22%, while thiamine losses ranged from 1.7 to 15.4%.

    In the rat study (Ti1 ef al., 1972a), the added levels of Na,S,O, were 0.125, 0.25,0.50, 1 .O, and 2.0%. The study was conducted over a period of 2 years and involved three generations. The 2.0% Na,S,O, diet caused slight growth retar- dation in the F, and F, generations, but had no effect on the F, generation. Part of this effect is explained on the lower birth weights in the F, and F, generation,


    although other reproductive effects were absent. Occult blood was observed in the feces of rats receiving the 1.0% and 2.0% Na,S,O, diets. Kidney weights were slightly increased with the 2% diet in the F, females only, and this change was not accompanied by any functional or histopathological changes in the kid- neys. Histopathological observations were largely normal except for the exis- tence of hyperplasia in the fore and glandular stomachs of the rats receiving 1% and 2% Na,S,O, diets. This hyperplasia was noted in all three generations and was observed to a lesser extent in the forestomachs only of some rats on 0.5% Na2S,0, diets. Beems et al. (1983) have further examined this hyperplastic response and concluded that it involves chief cells, but the mechanism of the response remains unknown. The no-effect level from the rat study was 0.25% Na,S,O, in the diet, which is equivalent to 72 mg SO,/kg/day after conversion and correction for sulfite losses. The Joint FAO/WHO Expert Committee on Food Additives (1974) used the results of this experiment to establish the AD1 of 0.7 mg SO,/kg by simply applying a 100-fold safety factor to the no-effect level obtained by Ti1 er al. (1972a).

    Although this experiment is the most carefully controlled study of the chronic toxicity of sulfites in existence, it has been criticized. Hickey er al. (1976) point out that the levels of sulfite oxidase in humans are much lower than the levels in normal rats, so a study of sulfite toxicity using ,normal rats is not justified. Subchronic toxicity studies with sulfite oxidase-deficient rats clearly demonstrate that such animals are more susceptible to the toxic effects of sulfites (Gunnison er al., 1981b). However, the 100-fold safety factor is intended partly to correct for such differences in detoxification pathways. The study of Ti1 et al. (1972a) should be recognized as a study of the toxicity of total sulfite rather than free sulfite, however. Ti1 er al. (1972a) analyzed for sulfite residues in their diets by the method of Reith and Willems (1968), which detects total sulfite levels. Therefore, some of the Na,S,O, added to the diet may have reacted with dietary components but would be recovered as SO, during the analytical procedure. In all likelihood, Ti1 er al. (1972a) underestimated the degree of free sulfite loss by reaction with dietary components.

    Ti1 et al. (1972b) also conducted a chronic toxicity study in pigs. The tech- niques were identical to those used in the rat study (Ti1 et al . , 1972a). A 48-week feeding period was employed. The results varied somewhat, however. Some growth retardation was noted in diets having 0.83 and 1.72% residual sulfite, although this was due to diminished food intake, as a later paired feeding trial did not demonstrate any differences in growth rates or food conversion. Organ to body weight ratios were increased at the 0.83% and 1.72% levels for liver, kidney, heart, and spleen, although this is ascribed to the lower body weights. In contrast to the rat study, no occult blood was observed in the feces. Histo- pathological examinations were normal except for mild inflammation and hyper-


    plasia in the stomach at the 0.83% and 1.72% levels at both the 15-week and 48- week observation periods.

    Subchronic toxicity studies were also conducted by Ti1 et al. (1972a,b) on both rats and pigs. In rats (Ti1 et al., 1972a), high sulfite levels (0-8%) were fed in the diet for 10-56 days. Diets containing 6% sulfite caused marked growth depression, reduced food intake, and lowered food conversion efficiency. Severe anemia, increased spleen weights, and slightly elevated leukocyte counts were also observed. The hyperplasia of the forestomach was found with 1% sulfite or more, while glandular stomach hyperplasia, hemorrhagic erosions, necrosis, and inflammation were found with 4% sulfite or more. Forestomach ulcers and papillomatous elevations occurred at 6 and 8% sulfite. All of these effects were reversible. In pigs (Ti1 ef al., 1972b), the changes observed after 15 weeks of feeding were similar to those encountered after 48 weeks of feeding. Bhagat and Lockett (1964) observed diminished growth rates in rats fed 0.6% Na,S,O, in the diet over a 5- to 7-week period, but this effect could be reversed by supple- mentation with thiamine. Gunnison et al. (198 la) confirmed the observation of anemia in rats and attributed it to the interaction of sulfite with dietary factors, perhaps vitamin B12. Gunnison et al. (1981a) conducted their experimeots with sulfite oxidase-deficient rats and showed elevated excretion of S-sulfonates can occur after administration of low levels of sulfite (0-3.5 mmol/kg/day). Ob- viously, sulfite oxidase-deficient rats are more susceptible to the toxic effects of oral sulfite, and Gunnison (1981) has suggested their use in sulfite toxicity studies.

    Few experiments have been conducted on the chronic and subchronic tox- icities of combined forms of sulfite. Dietary studies such as those by Ti1 et al. (1972a,b) are probably tests of the toxicity of some mixture of free and combined sulfites. Gibson and Strong (1973) used glucose hydroxysulfonate in some of their metabolism studies. Glucose hydroxysulfonate is likely to be stable to stomach acid, but likely decomposes to free sulfite in the neutral pH of the small intestine. They found no histological abnormalities in the livers and kidney of rats dosed with glucose hydroxysulfonate for 30 days. Walker et al. (1983b) did not observe any adverse effects after oral administration of DSH to rats for 14 days.

    3. Carcinogenicity

    Tumorigenic effects were not encountered in any of the chronic toxicity tests described above. In addition, Tanaka et al. (1979) failed to find any tumors in a carcinogenicity test of K,S20, in mice; 0, 1, and 2% K,S,OS was administered in the drinking water. Gunnison et al. (1981a) noted a small incidence (4/ 149) of mammary adenocarcinoma in sulfite oxidase-deficient rats as compared to O / 143


    in controls after 5 months of feeding of tungsten, but the effect was not statis- tically significant.

    4. Mutagenicity

    The mutagenicity of free inorganic sulfites has been extensively studied. The subject has been reviewed in detail elsewhere (Gunnison, 1981; Shapiro, 1977), and no attempt will be made here to provide such detail. The reactions of sulfite with nucleic acids were covered in Section II,F,7. The mutagenicity of the sulfites is thought to originate from the deamination of cytosine to uracil. The involvement of sulfite-induced deamination of 5-methylcytosine to thymine in the mutagenic process has also been considered, but the cytosine-to-uracil con- versions are thought to be quantitatively more important (Wang and Ehrlich, 1980; Wang ef al., 1980).

    Sulfites are capable of inducing mutations in vitro in several mutagenicity test systems, including E. coli, y phage, T4 phage,yeast, and Vicia faba root mer- istems (Chambers et al., 1973; Dorange and Dupuy, 1972; Hayatsu and Miura, 1970; Mukai et al., 1970; Njagi and Gopalan, 1982; Summers and Drake, 1971). However, these experiments required high concentrations of sulfite and acid pHs in the vicinity of pH 5 . When incubations were performed at neutral pH, no measurable mutagenic response was observed (Mukai et al., 1970). MacRae and Stich (1979) found that sulfite induces dose-related sister chromatid exchange in Chinese hamster ovary cells, but the potency of this induction was relatively weak. Sulfites can also cause chromosome damage when incubated in v i m with oocytes from mice, cows, or sheep (Jagiello et al., 1975). However, they could not induce chromosome aberrations in mouse oocytes cultured in vitro after an intravenous injection of sulfite.

    Despite the evidence for mutagenicity of sulfite in the systems described above, there is no evidence for sulfite-induced mutagenesis in other systems. The Food and Drug Administration contracted for mutagenicity studies in a variety of systems, and the results of these tests are reported in the 1976 GRAS evaluation document (Life Sciences Research Office, 1976). Sodium bisulfite was not mu- tagenic in the host-mediated assay in mice, the dominant lethal assay in rats, the in vivo cytogenetic assay in rats, and human tissue culture cells in vitro (Life Sciences Research Office, 1976). Sodium sulfite and potassium metabisulfite were not mutagenic in vitro in the Ames Salrnonellalmammalian microsome test (Life Sciences Research Office, 1976). Sodium metabisulfite did cause mitotic inhibition and damage to anaphase cells when added to human embryonic lung cells in culture (Life Sciences Research Office, 1976). However, sodium meta- bisulfite was not mutagenic in the host-mediated assay, the dominant lethal assay, or in vivo cytogenetic assays (Life Sciences Research Office, 1976).


    Generoso et al. (1978) showed that sulfite was negative in the dominant lethal assay in mice after intraperitoneal injections. Drosophila ingesting a 0.08 M solution of NaHSO, (5120 ppm SO,) displayed a mutation rate that was not significantly different from controls (Valencia er al., 1972). Renner and Wever (1983) were unable to induce cytogenetic damage as monitored by sister chro- matid exchange, chromosome aberration, and the micronucleus test in sulfite oxidase-deficient mice and Chinese hamsters after intragastric administration of one or two doses of Na,S,O, (330 or 660 mg/kg) in aqueous solutions or fruit juice. Bisulfite in aqueous solutions or in fruit or vegetable juices was not mutagenic to Salmonella typhirnuriurn strains TA 1535, TA 1538, TA 100, or TA 98, but an increase in revertants was obtained with strain his-G46 (Munzer, 1980). In this strain, more revertants were obtained with sulfited fruit or vegeta- ble juices than aqueous solutions of sulfite (Munzer, 1980).

    Some evidence also exists for a comutagenic effect of sulfites (Mallon and Rossman, 198 1). Enhanced ultraviolet mutagenicity was observed in Chinese hamster V79 cells if they were exposed to 10 mM sulfite either during or immediately following irradiation. A twofold increase in mutagenicity was ob- served by comparison to irradiated controls not exposed to sulfite. With E. coli, 100 mM sulfite caused an eightfold increase in mutagenicity. Mallon and Rossman (198 1) obtained evidence implying that sulfite was inhibiting excision repair.

    Sulfite can also be an antimutagen. Sulfite at 200 ppm is able to inhibit the mutagenic effect of coffee in the Salmonellalmammalian microsome system and the induction of prophage A (Suwa et al., 1982). Sulfite also suppressed the mutagenicities of the 1,2-dicarbonyls, diacetyl and glyoxal (Suwa et al., 1982).

    Almost no information is available on the mutagenicity of the various com- bined forms of sulfite. Walker et ul. (1983b) demonstrated that DSH is not mutagenic in the Ames Salmonellulmammalian microsome assay.

    5. Teratogeniciry

    Teratogenicity studies on NaHSO,, Na,S,O,, and K,S,O, have been con- ducted in several species on behalf of the Food and Drug Administration; the results of these evaluations were reviewed in the 1976 GRAS evaluation docu- ment (Life Sciences Research Office, 1976). These sulfites were administered orally to rats and mice on a daily basis on day 6 through day 15 of gestation and similarly in hamsters except on day 6 through day 10 of gestation. The doses (in mg/kg) for mice, rats, and hamsters ranged up to 150, 110, and 120, respec- tively, for NaHSO,; up to 160, 110, and 120, respectively, for Na,S,O,; and up to 125 and 155 in mice and rats, respectively, for K,S,O,. The incidence of teratogenic effects was unchanged from control animals. Maternal and fetal survival were also not affected by these sulfites.


    Dulak et al. (1984) investigated the reproductive toxicology of sulfite in sulfite oxidase-deficient rats. Exposure to sulfites from 3 weeks before mating until day 20 of gestation revealed no reproductive hazards for sulfite. Mating and pregnan- cy rates, gestational weight gain, preimplantation loss, resorbed and dead fetuses, litter size, fetal weights, and malformations were unaffected by sulfite treatment.

    6. Studies in Cell Cultures

    Sulfites have a variety of effects on cultured cells. Sulfites are cytotoxic to mouse fibroblasts, mouse liver cells, HeLa cells, Chorella pyrenoidosa cells, and human lymphocytes in culture (Das and Runeckles, 1974; Thompson and Pace, 1962; Timson, 1973). This cytotoxic effect was observed in the 0.1-20 mM range, although the minimum inhibitory concentrations varied among the different cultures. DNA synthesis can be inhibited in chick embryo fibroblasts by 0.1-1 .O mM sulfite (Chin et al . , 1977). Sulfite can also prevent the adhesion of Chinese hamster cells to the substratum (Kudo et al . , 1980). Kikigawa and Iizuka (1972) showed that 7.5 mM sulfite inhibited the ADP- and collagen- induced aggregation of rabbit platelets.


    1. History of Asthma and Other Adverse Reactions to Sulfites

    Recently, sulfiting agents have been reported to induce asthma when adminis- tered to certain asthmatics (Baker et al . , 1981; Freedman, 1977; Kochen, 1976; Stevenson and Simon, 1981b). The first reports, generated by Kochen (1976) and Freedman (1977), did not immediately attract much attention. However, the simultaneous reports by Allen and Collett (198 1) and Stevenson and Simon (1981a) at the American Academy of Allegy meetings, which linked sulfite ingestion in foods and drugs with asthmatic episodes in several patients, sparked considerable interest and additional research. The evidence linking ingestion of sulfiting agents with exacerbation of asthma in a segment of the asthmatic popu- lation is now compelling, although the role of sulfited foods in the initiation of these reactions has not been clearly established, as will be indicated later. Addi- tionally, sulfiting agents have been implicated in a few rare instances with other types of hypersensitivity reactions, including anaphylactoid reactions, hypoten- sion, and contact sensitivity (Fisher, 1975; Prenner and Stevens, 1976; Rudzki, 1979; Schwartz, 1983), indicating that asthma is not the only adverse reaction to sulfiting agents. However, asthma is very likely to be the most common adverse reaction to the sulfites. In this section, each of the published studies on adverse reactions to ingestion of sulfiting agents will be reviewed, with particular empha-


    sis on its contribution toward evaluating the degree of hazard posed by the use of sulfiting agents in foods. Studies pertaining directly to respiratory exposure to SO, will not be reviewed in detail because SO, is a well-documented hazard to virtually all asthmatics and others when inhaled (Boushey, 1982; Koenig et al., 1980; Linn et al., 1983; Nadel et al., 1965; Sheppard et al., 1980), and tolerance levels have been established for exposure to SO, in the workplace and the ambient air. However, the inhalation route of exposure may have some relevance to the discussion because such exposures might occur from inhaling the air released during the opening of a bag of dried fruit (Werth, 1982) or during ingestion of an acidic beverage (Delohery et al., 1984a).

    2 . The Earliest Reports

    Kochen (1976) reported the case of a child with mild asthma who experienced acute transient episodes of asthma after the consumption of sulfited foods. Con- firmatory sulfite challenges were not conducted. This report was considered to be an isolated, unique, and not fully substantiated case until the later reports began to appear.

    The pioneering study of the induction of asthma by ingested sulfites was published by Freedman (1977). Freedman interviewed 272 asthmatic patients and queried each of these patients about their asthmatic experiences following ingestion of a particular type of orange drink. This type of orange drink, which contains orange juice, sweetener, tartrazine, sodium benzoate, sulfur dioxide, stabilizers, and artificial flavorings, is not available in the United States. Of the 272 patients, 30 (1 1%) reported experiencing asthma soon after ingestion of such orange drinks. Of these 30 patients, 14 volunteered for oral challenges with sulfur dioxide, sodium benzoate, and tartrazine. The challenges were adminis- tered to the patients following an 8-hr period of abstinence from bronchodilators or cromoglycate and a 3-hr period of abstinence from food. Sodium metabisulfite was dissolved in a citric acid-water solution so that the challenge dose was 250 ml containing 100 ppm SO,. This would be equivalent to a dose of 25 mg of SO,. With the addition of citric acid, the pH of the solution was acid, and therefore most of the SO, probably existed as HSO, and H,SO,. Of the 14 patients, 8 showed a decrease in lung function as determined by a drop in their forced expiratory volume in 1 sec (FEV,) as measured by spirometry. Any decrease in FEV, exceeding 12% was considered positive. The group included 5 females and 3 males, and 3 of these patients also developed asthma when chal- lenged with sodium benzoate. On challenge with SO,, the maximal drop in FEV, occurred by 1 1 min (a range of 2-25 min) with measurable decreases often occurring within 1-2 min. The maximal depression in FEV, ranged from 12 to 57%, with an average of 31%. Three patients had decreases in FEV, of less than


    20%. One of these patients, who had a marginal drop in FEV, of 12% on administration of 25 mg of SO,, was challenged with 75 mg of SO, and experi- enced a decrease in FEV, of 37%. Prior administration of sodium cromoglycate protected 4 of 4 patients from the effects of ingested SO,.

    Several features of Freedmans study are subject to criticism and possible misinterpretation. The study is sometimes quoted as being an evaluation of the sensitivity of 272 asthmatics to sulfiting agents. In fact, only 14 patients were actually challenged with sulfites. Freedman used a drop of 12% in FEV, as an indication of a positive response. This is an extremely conservative approach. Most pulmonary specialists would consider a 12% drop as only marginal and would require either a 15 or 20% drop to indicate a positive response. At the 20% level, the number of responders to the 25-mg challenge would drop from 8 to 5. Freedman did not conduct the challenges in either a placebo-controlled or dou- ble-blind manner. Placebo control of such challenges is considered to be the minimal safeguard against biased results and double-blind confirmation of my reactions is preferred (Bush er al., 1986). The pH of the challenge solution may have contributed greatly to the acquired results. SO, will be evolved from an aqueous solution only if the pH is below 4.0. Freedman does not state the pH of his challenge solutions. However, he prepared the solution by dissolving 0.75 g sodium metabisulfite and 0.75 mg citric acid in 1 liter of water and then diluting by a factor of 5. In our hands, such a solution has a pH of 2.94. At this acidic pH, most of the free SO, would be in the HSO, form, with about 10% as H,SO, (Green, 1976; Joslyn and Braveman, 1954). About 6% of the added meta- bisulfite would be evolved as gaseous SO, at this pH. This would be equivalent to 1.5 mg of SO,, a dose sufficient to induce bronchoconstriction in asthmatics if inhaled. Therefore, Freedmans study may simply represent another demonstra- tion of the ability of gaseous SO, to induce asthma.

    Freedman also made some rather intriguing observations which need to be resolved with the subsequent results of Stevenson and Simon (1981b). Freedman observed rather rapid decreases in FEV,, with 6 of the 8 patients reaching maximal loss of lung function within 10 min or less. By contrast, Stevenson and Simon (1981b) measured FEV, at 30-min intervals and observed a slower re- sponse of 15-30 min. The difference may be due to the fact that Freedman used a beverage vehicle, while Stevenson er af. used capsules. The beverage vehicle allowed exposure of the sublingual and buccal mucosa in addition to the gastroin- testinal tract. The rapidity of the response suggested to Freedman that the route of absorption of the sulfite was by inhalation of SO, vaporizing from the solution or absorption of the sulfite through the sublingual and/or buccal mucosa. Based on the pH of his challenge solutions, the most likely possibility is that SO, was vaporized from these solutions and inhaled by the sensitive patients. Variable inhalation of SO, from acidic solutions has now been demonstrated to be the


    mechanism of reaction to these solutions (Delohery et a l . , 1984a,b). Another intriguing aspect of Freedmans work was the blockage of the response by prior administration of sodium cromoglycate. Since cromoglycate acts by stabilizing the mast cell membrane, thereby preventing release of histamine and other medi- ators of the allergic response, the inhibitory action of cromoglycate would possi- bly suggest that mediator release plays a role in the mechanism of the response to ingested sulfiting agents. However, cromoglycate is known to have other ac- tions, including phosphodlesterase inhibition, reduction of mucosal hyperac- tivity, and inhibition of neurological reflexes, so blockage of histamine release may not be the only explanation for the actions of cromoglycate.

    3. 1981 American Academy of Allergy and Immunology Reports

    The earliest reports by Allen and Collett (1981) and Stevenson and Simon (1981a) were brief abstracts of presentations made at the 1981 American Acade- my of Allergy and Immunology meeting. Allen and Collett (1981) reported 2 patients with sensitivity to sodium metabisulfite. One of these patients had had asthmatic reactions elicited by sulfites in foods, while the other patient had experienced asthma following administration of drugs containing sulfites. The sensitivity to sulfites was confirmed by double-blind challenges with capsules containing 500 mg of sodium metabisulfite. The 500-mg challenge dose is rather high by comparison to the amounts used by Freedman (1977), Stevenson and Simon (1981a,b), and the levels presently being used by the Australian group (Baker and Allen, 1982; Delohery et a l . , 1984a,b). The patients described here were also sensitive to tartrazine, aspirin, and sodium benzoate. Stevenson and Simon (1981a) identified 4 asthmatic patients with sulfite sensitivity. They also employed capsule challenges, but used potassium metabisulfite. The threshold doses for decreases in lung function ranged from 10 to 50 mg. These patients were not found to be sensitive to sodium benzoate, aspirin, tartrazine, or mono- sodium glutamate.

    4 . Further Reports from Australia

    A later report by Allens group describes in detail the cases of 2 sulfite- sensitive patients (Baker et a l . , 1981). It is not clear if these patients are identical to the ones described in the earlier abstract. The first case was a 67-year-old female who had experienced asthma after ingesting a crabmeat salad prepared with vinegar dressing. A subsequent challenge of this patient with capsules of sodium metabisulfite confirmed the existence of an asthmatic reaction related to the consumption of sulfites. The second case was a 23-year-old female whose asthmatic symptoms worsened on ingestion of wine. A subsequent challenge


    with a capsule containing 500 mg of sodium metabisulfite confirmed the exis- tence of the asthmatic reaction to sulfites. The challenges were done double blind, with lactose as the negative control.

    A third report from the Australian group, also in abstract form, details their experiences with metabisulfite challenges through early 1982 (Baker and Allen, 1982). By this time, they had identified 8 patients with asthmatic sensitivity to oral challenge with metabisulfite (presumably the sodium salt). Of the 8 patients, 3 were also sensitive to aspirin and other food additives, including tartrazine and benzoate. The challenge protocol had been modified to include administration of graded doses starting at 10 mg and progressing through 300 mg, with lung function evaluations at 0.5-hr intervals. Curiously, 4 of the 8 sensitive patients did not react to a 300-mg capsule challenge, but did react to a 25-mg challenge of metabisulfite dissolved in 50 ml of 0.5% citric acid. This mode of administration is quite similar to that of Freedman (1977). The reactions to acidic sulfite solu- tions occurred within 1-5 min, while positive capsule challenges showed a 20- to 30-min lag period. They conclude that the response to acidic sulfite solutions is due to inhalation of vaporized SO,.

    More recent reports from the Australian group (Delohery et af., 1984a,b) delve more deeply into the comparative responses to capsule versus beverage challenges. Acidic solutions of metabisulfite were able to provoke asthma in 60% of all asthmatics, a much higher percentage than found with capsule chal- lenges (Delohery et al., 1984a). A comparison of sulfite reactors with asthmatics not reactive to sulfites revealed that both groups were equally sensitive to inhaled SO,, but that the sulfite reactors were the only group responsive to ingestion of sulfited acidic beverages. The sulfite reactors responded to a mouthwash with a sulfite solution, but not to sulfite solution administered directly into the stomach via a nasogastric tube. It must be assumed that this group of reactive asthmatics does not have any capsule reactors because they would be predicted to respond to any direct gastric challenge. Delohery et al. (1984a.b) conclude that the bev- erage reactors are inhaling SO, as they swallow, while nonreactors can swallow without inhalation. Allen and Delohery (1985) revealed that these asthmatics do not respond to sulfited acidic beverages if they take a deep breath and hold it before using a sulfite mouthwash. The existence of such a high percentage of beverage reactors is somewhat surprising, although all asthmatics respond to inhaled SO,. However, the practical significance of this type of sulfite sensitivity is uncertain. Delohery et af. (1984a,b) used challenges of 50 mg of metabisulfite in a citric acid solution. It is unlikely that asthmatics would routinely encounter such levels of free sulfite in most beverages. Wine might easily contain 50 mg of total sulfite per serving, but the majority of this sulfite would be in the form of combined sulfites. Still, this type of sensitivity may explain the common com- plaints of asthmatics about adverse reactions to the ingestion of wines.


    Allen and Delohery ( 1985) also investigated the mechanism involved in reac- tions to sulfite in capsules. After ingestion of 25- to 50-mg capsules of meta- bisulfite, 4-50 ppm of SO, could be detected in the stomach via a nasogastric tube. They speculate that SO, is evolved from metabisulfite by the action of stomach acid and that the SO, can be inhaled following eructation. Unfortunate- ly, they did not measure SO, concentrations in the nasopharynx after capsule ingestion.

    5 . Further Reports from the La Jolla Group

    Stevenson and Simon (1981b) also published a more detailed account of their initial findings. Descriptions of 5 sulfite-sensitive patients are provided in this report. Four of the patients were identical to the ones described in their earlier abstract. The challenges were performed with capsules of potassium meta- bisulfite. Graded doses starting at l mg were employed, with the doses increas- ing to 5, 10, 25, and 50 mg of K,S,O, until an asthmatic response was noted. The 5 patients had asthmatic reactions beginning at 15-30 min after administra- tion of the threshold dose. The threshold dose was 10 mg for 2 of the patients, 25 mg for another 2 patients, and 50 mg for the fifth patient. Since several doses were administered at 30-min intervals, it is possible, though unlikely, that the patients were reacting to an accumulated dose rather than the last dose adminis- tered. Falls in FEV, ranged from 23 to 49%. The challenges were placebo controlled, reproducible, and blinded to some. extent. This experimental design was imperative, since all of these patients were severe asthmatics who required steroids for control. Such asthmatics would be predicted to be unstable, so repeat challenges and blinded challenges were necessary. These patients were not sen- sitive to aspirin, tartrazine, or monosodium glutamate.

    Stevenson and Simon (1981b) attempted unsuccessfully to define the mecha- nism of action of potassium metabisulfite in these patients. Evidence for an IgE- mediated reaction could not be found. In fact, no evidence could be found that mediator release is involved in the reaction. Cutaneous testing with 0.02 mg of K,S,O, given intradermally was negative in the 4 tested patients. Incubation of peripheral basophils with K,S,O, in concentrations up to 0.01 M failed to induce histamine release. These tests would be positive in reactions involving mediator release whether IgE-mediated or not. Despite the lack of evidence for an IgE- mediated reaction among the patients studied by Stevenson and Simon (1981b), systemic sensitivity beyond altered lung function was noted in all of their pa- tients. The systemic symptoms were flushing, weakness, and hypotension. These symptoms can be involved in IgE-mediated reactions or other reactions involving mediator release. Stevenson and Simon (1981b) hypothesize that po- tassium metabisulfite acts via stimulation of the cholinergic reflex arc. This


    stimulation would account for some of the observed symptoms, including bron- choconstriction. However, it is difficult to explain hypotension on this basis. The therapeutic effectiveness of atropine is also consistent with this mechanism. Some evidence suggests that inhaled SO, activates irritant receptors in the bron- chial tubes and that these receptors may activate the cholinergic reflex arc (Boushey, 1982; Nadel et al., 1965). However, other theories of the actions of inhaled SO, also exist (Boushey, 1982). Further proof will be needed before cholinergic stimulation will be accepted as the mode of action of ingested sulfi tes.

    In a 1981 abstract from the American Academy of Allergy and Immunology meeting, Simon et al. (1982) presented the first indication of the prevalence of sensitivity to ingested sulfites among asthmatics. A total of 61 asthmatics chosen randomly were challenged with potassium metabisulfite capsules containing 10, 25, 50, 100, and 200 mg K,S,O, at 30-min intervals. A positive reaction was defined as a fall in FEV, of at least 25%. Challenges were placebo controlled and single blind, with repetition of any positive response in a second challenge. Of the 61 patients, 5 (8.2%) reacted to K,S,O,. The reactions were milder than those encountered in their earlier studies (Stevenson and Simon, 1981a,b), and the threshold doses tended to be higher. This study would suggest that the prevalence of sulfite sensitivity among asthmatics is rather high. However, we question whether the population of asthmatics used in this survey was truly random. Many of the asthmatics used in this survey had severe asthma, and the study group was probably not a true cross section of the entire asthmatic popula- tion.

    The La Jolla group presented three abstracts at the 1984 American Academy of Allergy and Immunology meeting (Goldfarb and Simon, 1984; Jacobsen et al . , 1984; Simon et al., 1984). Goldfarb and Simon (1984) evaluated the com- parative sensitivities of sulfite-sensitive asthmatics (SSA) as a function of the route of exposure. Six SSA were used in this study; all 6 SSA had reacted to capsule challenges with 10-50 mg of sulfite, with a fall in FEV, of >25%. The minimum provoking dose for a beverage challenge was approximately one-half that of the capsule challenge. Inhalation of nebulized sulfite solutions provoked reactions at one-tenth to one-one hundredth of the capsule challenge dose. None of these SSA reacted to subcutaneous administration of sulfites at doses up to 10 times higher than their provoking capsule dose. Obviously, inhalant exposures are the most hazardous to SSA. Inhalant exposures could be encountered through the use of bronchodilator solutions preserved with sulfites (Koepke et al., 1983). Usually, the bronchodilating effect of the active ingredient would overwhelm the bronchoconstricting effect of sulfite, although a few patients seem to suffer paradoxical bronchoconstriction when treated with sulfited bronchodilators (Koepke et al. , 1984a; Simon, 1985).


    Jacobsen et af. (1984) focused their efforts on elucidation of the mechanism of sulfite sensitivity in asthmatics. Using skin biopsies from SSA, Jacobsen et al. ( 1984) cultured skin fibroblasts and demonstrated that these cells had diminished levels of sulfite oxidase by comparison to cells cultured from normal individuals and asthmatics without sulfite sensitivity. The depressed levels of sulfite oxidase may indicate that these individuals are heterozygous for a deficiency of the enzyme. The diminished sulfite oxidase levels could compromise the detoxifica- tion of sulfite in these individuals (see Section IILA), but further studies will be needed to define the full implications of this finding.

    Jacobsen et af. (1984) were also able to show that cyanocobalamin (vitamin B,,) can protect SSA from the effects of ingested sulfite. Up to 50 mg of vitamin B,, orally was necessary to block the reaction to ingestion of a 50-mg capsule of K,S,O,. The vitamin B,, action was catalytic, as demonstrated by the observa- tion that 5 mg would protect against 50 mg of K,S,O,. The B,, effect is probably associated with the known ability of the cobalamins to catalyze the oxidation of sulfite to sulfate (see Section II,F,6). The effective dose of vitamin B,, is far in excess of the recommended dietary allowance for this vitamin. It is even in excess of the levels of vitamin B,, used to treat pernicious anemia. However, such pharmacological doses of vitamin B,, may provide a convenient means of prophylaxis for SSA. R. A. Simon (personal communication) is now counseling his SSA patients to take 5 mg of vitamin B,, before eating a restau- rant meal.

    Simon et al. (1984) evaluated the effectiveness of a variety of possible block- ing agents on sulfite-induced asthma among SSA. Vitamin B,,, atropine, cro- molyn, and doxepin were all effective blocking agents. These agents were effec- tive irrespective of the route of administration of the sulfite. The effectiveness of all four agents is rather surprising, since they have different modes of action. Atropine is an anticholinergic agent, cromolyn is a mast cell membrane stabilizer and calcium channel blocker, doxepin is a broad spectrum antihistamine, and vitamin B,, catalyzes sulfite oxidation. Rather high doses of these blocking agents were necessary, and it is possible that at such high doses these agents could have additional effects beyond those just mentioned. This experiment does not provide many clues to the mechanism of action of sulfites in provoking asthma in these subjects. The equivalent effectiveness of these agents toward ingested versus inhaled sulfite is also rather surprising, since the mechanisms of the two routes of exposure are almost certain to be different.

    At the 1985 American Academy of Allergy and Immunology meeting, this group presented two additional abstracts on sulfite sensitivity (Howland and Simon, 1985; Simon, 1985). One of these reports involved the description of two cases of paradoxical bronchoconstriction in sulfite-sensitive asthmatics after ad- ministration of a sulfited bronchodilator (Simon, 1985). Howland and Simon


    (1985) challenged 5 sulfite-sensitive asthmatics with 3 oz of sulfited lettuce containing 80-90 mg of bisulfite (calculated by assessing the amount of sulfite solution not recovered after drainage). All 5 patients experienced pronounced decreases in lung function; the mean FEV, decrease was 44%, with a range of 31-64%. Untreated lettuce had no effect. This challenge study demonstrates conclusively that sulfited lettuce can elicit asthmatic reactions. Whether other sulfited foods will elicit these reactions is not known. Sulfited lettuce contains appreciable quantities of free SO, (Taylor et al., 1985), which may enhance the likelihood that lettuce will initiate asthmatic reactions by comparison to other sulfited foods.

    6. Other Reports of Sulfite-Induced Asthma

    Several additional studies on the prevalence of sulfite sensitivity were con- ducted after Simon et al. (1982) reported that 8.2% of all asthmatics might be affected. Bush et al. (1985) showed that sensitivity to encapsulated sulfites was an appreciable risk only for those patients who require steroids for the control of their symptoms. Among 83 steroid-dependent or severe asthmatics, the preva- lence of sulfite sensitivity was 8.4%. Among 120 mild or nonsteroid-dependent asthmatics, the prevalence of sulfite sensitivity was only 0.8%. Mild asthmatics make up about 80% of all asthmatics, so the overall prevalence for the total population of asthmatics is estimated to be 1.8% from this study. Buckley et al. (1985) selected 134 patients from a total clinic population of 1073 asthmatic subjects; 50/134 or 37% reacted to oral challenges with capsules of K,S,O,. This suggests a minimal prevalence of 4.6% (50/1073). However, as with the population examined by Simon et al. (1982), there is no indication that the patient population evaluated by Buckley et al. (1985) is representative of the overall asthmatic population. Towns and Mellis (1984) performed oral sulfite challenges with both capsules and citric acid solutions of Na,S,O,. None of the children developed asthma after challenge with capsules, but 19 of 29 (66%) experienced a significant decrease in FEV, after challenge with an acidic sulfite solution. This confirms previous suggestions that many more asthmatics are sensitive to acidic solutions of sulfite by comparison to encapsulated sulfites (Delohery et al., 1984a,b).

    Other case reports of asthmatic sensitivity to sulfites have also appeared (Sprenger et al . , 1985; Altman et al., 1985; Schwartz and Chester, 1984; Koepke et al . , 1984; Yang et al., 1985; Werth, 1982; Twarog and Leung, 1982). One was a patient with a history of asthma that worsened with ingestion of certain foods, particularly dried apricots and Catawba grape juice (Werth, 1982). The patient also experienced flushing during these episodes. Both of these foods are sulfited. Occasional asthmatic attacks were experienced following ingestion


    of wine, beer, cheese, blueberries, apples, and strawbenies. Of these foods, only wine, beer, and possibly freshly cut fruits would be expected to contain residual SO,. Symptoms of asthma were produced in the patient by sniffing a freshly opened bag of dried apricots. Oral challenge with capsules of potassium meta- bisulfite at doses up to 50 mg were negative. Inhalation of nebulized K,S,05 in water induced a rapid decline in FEV, . Apparently, this patient is another exam- ple of an individual who responds to inhaled SO, but not to ingested sulfites. He constitutes further proof for our suggestion that two groups of sulfite-sensitive asthmatics exist.

    Another case was reported by Twarog and h u n g (1982). This patient had perennial asthma and had experienced several adverse reactions to drugs that contained sodium bisulfite or sodium metabisulfite. Oral challenge of this patient with sodium metabisulfite in water revealed that a 5-mg dose caused a 52% drop in FEV,. The reaction to a 5-mg dose makes this patient the most sensitive described so far. Flushing was also noted. In addition, this patient may be unique, since evidence of mediator release in response to the sulfites was ob- tained in her case. Skin testing with sodium bisulfite at 0.1 mg/ml resulted in a definite wheal and flare reaction. Sodium bisulfite at concentrations of lop3- 10- M also caused release of histamine from this patients leukocytes. For both skin testing and leukocyte histamine release, control tests on other individuals were negative. These findings do not constitute proof for the existence of an IgE- mediated or type I reaction, however, because no evidence for the existence of a specific antibody was obtained. However, this patient seems to be unique, since Stevenson and Simon (198 lb) found no evidence of mediator release in 4 of their sulfite-sensitive patients. This patient probably represents a small subgroup of sulfite-sensitive patients. Apparently, the majority do not react via mediator release, but obviously some patients may mount such responses. This patient was challenged with sodium metabisulfite in water, a slightly acidic solution. It is difficult to determine if her response was due to inhalation of SO, or ingestion of sulfites. The ingestion route would seem most probable, since a 10-min lapse occurred between administration of the dose and the fall in FEV,. Also, SO, would not be evolved from a water solution, which would have a pH of greater than 4.0.

    Altman et al. (1985) and Sprenger et al. (1985) provide some additional evidence for the possibility of mediator release in the pathogenesis of sulfite- induced asthma. Sprenger et al. (1985) describe a single patient with sensitivity to both inhaled SO, and aqueous solutions of K,S,O, (the pH was not specified). In this patient, an increase in the level of neutrophil chemotactic activity (NCA) in the serum was observed 2 hr after the maximal decline in FEV, . Altman et al. (1985) identified 3 additional patients with similar patterns of sensitivity along with increased serum NCA. NCA can be released from mast cells with appropri-


    ate antigen challenges. However, these findings are somewhat confusing. The increase in NCA in serum did not correspond in time to the decreased lung function. Also, the pH of the sulfite solutions is not provided, so it is impossible to know if these patients fall in the small group with sensitivities to encapsulated sulfites or the large group with sensitivities to ingestion of acidic sulfited beverages.

    The data from Sprenger et al. (1985) suggest that patients with sensitivities to ingested sulfites would also display inhaled sulfite sensitivity. Koepke et al. (1984b) performed inhalation challenges on 3 sulfite-sensitive (by capsule chal- lenge) and 10 nonsulfite-sensitive asthmatics. All 3 sulfite-sensitive asthmatics and 4 of the 10 others had declines in FEV, of 20% or greater. The remaining 6 asthmatics had diminished lung function also, but it had not reached a 20% decrease at the administered levels of sulfite. Again, these data suggest that all patients with reactions to ingested sulfites will respond to inhaled sulfites. Schwartz and Chester (1984) obtained some conflicting information. Six asth- matics who developed airway obstruction after ingesting solutions of K,S,O, were subjected to inhalation challenge. Only 3 of the 6 patients responded to both ingestion and inhalation challenges with sulfite. These data suggest that a positive oral sulfite challenge is usually but not invariably accompanied by a positive aerosol challenge.

    Yang et al. (1985) identified 3 sulfite-sensitive asthmatics using oral chal- lenges with K,S,O, capsules. Two of the patients had positive intradermal skin tests to 1 mg/ml solutions of K,S,O,. Passive transfer was also demonstrated with unheated serum from one of these patients. They conclude that IgE mecha- nisms may play a role in a subset of sulfite-sensitive asthmatics.

    Several reviews on asthmatic reactions to sulfites have appeared (Bush et al., 1986; Schwartz, 1984; Simon, 1984; Stevenson and Simon, 1984; Twarog, 1983).

    7. Other Adverse Reactions to Sulfites

    Asthma has not been the only adverse reaction associated with ingestion of sulfites, although it appears to be the most common. Prenner and Stevens (1976) reported a patient who experienced urticaria and pruritis, swelling of the tongue, difficulty in swallowing, and tightness in the chest after ingestion of a sulfited restaurant salad. The patient had a positive scratch test to 0.2 mg of sodium bisulfite. An oral challenge with 10 mg of sodium bisulfite produced itching, nausea, flushing, cough, tightness in the throat, and erythema. Passive transfer testing was also positive. The passive transfer test indicates the presence of a serum factor involved in this patients response to sulfites. However, even in this case, this cannot be construed as definite evidence of an IgE-mediated reaction,


    although it is suggestive of such a reaction. Prenner and Stevens (1976) men- tioned in their report that several food handlers had described instances of contact sensitivity from handling sulfite solutions. Fisher (1975) had previously reported a case of eczema in a food handler, which had been attributed to bisulfite exposure. Rudzki (1979) recently identified sulfites as contact allergens as well. Several other cases of urticaria and angioedema attributed to sulfites have been described (Allen et al., 1984; Habernicht et al . , 1983; Huang and Fraser, 1984). Habernicht et al. (1983) described two women who reported urticaria and an- gioedema after ingestion of sulfited foods. One of these patients developed urticaria and burning of the scalp within 15 min following challenge with 25 mg K,S,O, in a capsule. Allen et al. (1984) note that urticaria can be induced by sulfite challenges, but that larger doses are usually required than those used in challenges of asthmatic subjects. Huang and Fraser (1984) suggested that sub- cutaneous administration of sulfites could provoke urticaria, angioedema, and laryngeal edema in sensitive individuals. Subcutaneous injection of 1.8 ml of lidocaine, which contains 0.5 mg of NaHSO,, produced palmar pruritis in a patient. No controlled challenge was administered. Yang et al. (1985) described a single patient with urticaria and angioedema after oral challenge with K,S,O, capsules.

    Very recently, another type of adverse reaction to sulfites has been described (Schwartz, 1983). Two patients were identified with anaphylactic-like reactions possibly associated with restaurant meals. The first patient had experienced an episode of clammy skin, weakness, headache, chest tightness, tachycardia, and a feeling of dissociation from his body commencing 10 min after eating a restau- rant salad. The second patient had developed dizziness, nausea, palpitations, hives, dysphagia, chest tightness, and dyspnea after a restaurant meal of shellfish and salad. Both patients were administered single-blind, placebo-controlled chal- lenges with metabisulfite (Na or K salt not specified). With increasing doses in the range of 10-50 mg, progressively worsening hypotension was observed in both patients. Abdominal distress, nausea, dizziness, and weakness were also noted. Not all of the symptoms from the restaurant episodes were seen in the challenges, but this may have been due to the exposure to lower doses of sulfites in the challenges. Neither patient experienced asthma and neither had positive skin tests, so these reactions were not IgE mediated. These cases are the first reports of hypotension without asthma following challenge with sulfites. Steven- son and Simon (1981b) noted hypotension in some of their patients who also experienced asthma on challenge with sulfites. The frequency of the hypotensive response to sulfites is unknown.

    Sulfiting agents may also cause problems when administered as a component of a drug formulation. The most common manifestation is asthma, as described previously. The use of bisulfite in epidural anesthetics has recently been associ-


    ated with paralysis of the lower extremities (Wang et al., 1984). This rare reaction occurs when the anesthetic is accidentally injected into the subarachnoid space. The paralytic condition was duplicated in rabbits by injecting 1.2-2.4 mg of sodium bisulfite into the lumbar subarachnoid space.

    Flaherty et al. (1985) described an unusual case of sulfite sensitivity in a patient with underlying liver disease (sclerosing cholangitis) and ulcerative col- itis. This patients liver condition was observed to worsen after ingestion of home-preserved juices and restaurant salads. These episodes were often accom- panied by palmer and plantar erythema with pruritis. The liver function tests in this patient improved on a sulfite-free diet. An increase in serum levels of liver enzymes was noted after challenge with 500 mg of metabisulfite. This increase could be blocked by prior administration of 3 mg of vitamin B,*.

    Sulfites have been evaluated for their possible role in other conditions as well. Sonin and Patterson (1985) failed to trigger episodes of idiopathic anaphylaxis in 12 patients using oral challenges with Na,S,O, in lemonade. Similarly, Meggs et al. (1985) could find no role for sulfites in the elicitation of idiopathic anaphylaxis in challenges of 25 patients with capsules of NaHSO,. However, plasma his- tamine levels were elevated twofold in 23 of the 25 patients following bisulfite challenge. Eight patients with systemic mastocytosis were subjected to similar challenges, and no evidence was found to implicate sulfites in this condition (Meggs et al., 1985). Like the patients with idiopathic anaphylaxis, plasma histamine levels were elevated twofold in 7 of the 8 patients with systemic mastocytosis after bisulfite challenge.

    8. Sensitivity to Suljited Foods

    Many of the reported sulfite-sensitive asthmatics provide a history of asth- matic reactions to foods that are suspected to contain sulfite residues. Their sensitivities to free inorganic sulfites have been documented through capsule and/or beverage challenges. However, there are only two reports of controlled challenges to a sulfited food or beverage unless one wants to count the challenges performed with sulfites in citric acid solutions or lemonade. We do not believe that the challenges with citric acid solutions are representative of the situation that exists with most foods, since most sulfited foods have pHs above 4.0 and would not spontaneously liberate SO,. The first study of the sensitivity of asth- matics to sulfited foods or beverages was conducted by Seyal et al. (1984) with wine. The subjects were asked to drink 4 oz of white wine containing 140 mg/liter of SO,. Only 1/25 asthmatics and 0/25 controls developed asthma following the challenge. Unfortunately, Seyal et al. (1984) did not prescreen their asthmatic population for sulfite senstivity, so the number of sulfite-sensitive asthmatics in their group is unknown and probably small. The study would have


    been strengthened considerably if it had been conducted on a group of sulfite- sensitive asthmatics. Also, Seyal et al. (1984) considered a drop in FEV, of 12% or greater as a positive response. As noted previously, most pulmonary spe- cialists would require a drop of at least 15-20% to signal a positive response. Therefore, the single responder in this study may be questionable. The other challenge study of sulfited food was conducted by Howland and Simon (1985) with sulfited lettuce; the results were described earlier.

    As discussed earlier, SO, and the inorganic sulfites react rapidly with food components. The rate, completeness, and products formed by these reactions are dependent on the pH, temperature, sulfite concentrations, type and concentration of various food components, and other factors. The primary products in many foods are the hydroxysulfonates of aldehydes, ketones, and reducing sugars. In vitro experiments have shown that these sulfite addition compounds are rather stable in dilute acid at room temperature (Adachi et al., 1979; Burroughs and Sparks, 1973; Green, 1976; Joslyn and Braverman, 1954). Therefore, they would not be expected to liberate SO, in the stomach under its acidic conditions. Some release of sulfites from sugar hydroxysulfonates might be predicted to occur in the neutral pH conditions of the small intestine. However, other hy- droxysulfonates would be stable even under these conditions. Very recent work indicates that one hydroxysulfonate is not metabolized at all in rats or mice after feeding in the diet (Walker et af., 1983a). Sulfite addition compounds can also be formed with amino acids and proteins (Green, 1976; Schroeter, 1966). The stability of these adducts in gastric acid is not known, but they are probably more stable than many of the hydroxysulfonates. The question then centers on the role of the sulfite addition products in the induction of the asthmatic response. The answer to that question is not known.

    Some added SO, and inorganic sulfites remain in the food product in the uncombined state. This free SO, would probably react much like the sulfites ingested in capsules. Only if the food was below pH 4.0 would the gas, SO,, be evolved from the food in the oral cavity. In other foods, free SO, would exist primarily as HSO, and SO:-. Gaseous SO,, if it exists in the food, would likely pose a hazard to asthmatics who might inhale it during consumption of the food (Delohery et al., 1984a,b; Towns and Mellis, 1984). Other forms of free SO, would pose a possible hazard only to those individuals with sensitivites to sulfites in capsules. The combination of sulfites with food components would drastically lower the free SO, content of most sulfited foods (lettuce is an exception), thereby limiting exposure to these free forms of the sulfiting agents. Even with the free sulfites, the food matrix may diminish the degree of sen- sitivity by slowing rates of absorption and access to the sites of action.

    The degree of hazard posed to sulfite-sensitive asthmatics by sulfited foods can only be established with actual food challenges of sensitive patients. Since


    foods vary in the nature of their combined sulfites and in the amount of residual free sulfites, such challenges will need to be performed with a variety of sulfited foods. We expect, on the basis of chemical considerations, that most sulfited foods will be much less hazardous than equivalent amounts of sulfites in cap- sules. Again, sulfited lettuce may be an exception, since it contains a high proportion of free SO, (Taylor et al., 1985).


    If the current GRAS review leads to some limitation on the continued use of sulfites, it will be necessary to consider alternatives. This section is designed to present some possible alternatives and their limitations. A complete substitute for the sulfiting agents which would possess all of the desirable properties of this group of food additives will be virtually impossible to find. Replacements for each of the individual benefits provided by the sulfiting agents might be identi- fied. However, in many foods, sulfiting agents are used for more than one purpose, e.g., the use in white wines for both its antimicrobial and antibrowning properties. The potential substitutes are also less effective and more costly in most cases. Many of the suggestions presented in this section were obtained from the review by Roberts and McWeeny (1972).


    Enzymatic browning will be inhibited by any process that destroys or inacti- vates the enzyme. Blanching would obviously work, but is impractical for use on fresh fruits and vegetables. Acid denaturation of the enzyme is also feasible, e.g., with application of lemon juice or vinegar. Since polyphenoloxidase is dependent on cupric ions for activity, the removal of these metallic ions may inhibit the process. EDTA would serve this purpose. Citric acid and tartaric acid also work in this manner. The activity of the enzyme can be slowed by lowering the pH through the addition of acids or fermentation. Lowering the water activity of the food can also diminish this reaction, but again dehydration is often not practical. Removal of oxygen works quite well, since oxygen is a required substrate for polyphenoloxidase. Reducing agents can be effective in converting the quinones back to the diphenols. Ascorbate and cysteine have been used in this manner. On cut surfaces, the sulfites have the advantage of being able to penetrate quickly into the cellular matrix, a property not shared by their sub- stitutes. Hence, the substitutes are less effective.

    Several alternatives using combinations of the above materials for control of enzymatic browning have been developed. One procedure involves ascorbate


    and CaCl,. Another alternative is a combination of phosphate, citric acid, dex- trose, aluminum sulfate, and sodium erythrobate. Both of these alternative meth- ods work primarily on the basis of acidification with reducing activity. Ponting er al. (1971) pioneered use of ascorbate and calcium in the preservation of sliced apples. Montgomery (1983) recently noted that enzymatic browning of pear juice concentrate can be prevented by cysteine. Cysteine does not work well on cut surfaces because of its lack of cellular penetration.


    Nonenzymatic browning can be controlled by (1) elimination of the active compounds, (2) lowering pH, (3) separation of the active species, or (4) behydra- tion to low water activities. The removal of sugars can be be effected by fermen- tation, glucose oxidase, or leaching, as is done with potatoes. Theoretically, cysteine could be effective in competing for reaction with reducing sugars, but this has never been evaluated as a practical alternative. The browning reaction can also be slowed by the addition of acids such as lactic, tartaric, citric, acetic, or ascorbic acids, or aluminum sulfate. Physical separation of the active species can occasionally work if, e.g., the sugar is in the sauce and the amino acid is in the entree. Dehydration to less than 4% water will also inhibit nonenzymatic browning, but is not economically feasible. The replacement of sulfites in the control of nonenzymatic browning will be difficult because none of the above treatments is as effective or universally applicable.


    Ascorbic acid can replace sulfites as antioxidants in beer, but naturally occur- ring levels of SO, in beer may make replacement unnecessary. Sulfites have the added advantage of controlling nitrosamine formation in the malt (Lukes et al., 1980). As a reducing agent, it may be possible to replace sulfites with cysteine or other mercaptans, although these substitutes have undesirable organoleptic prop- erties and undesirable color and texture.


    For wines and corn steep liquors, alternative agents will be difficult to find. Other antimicrobial agents will inhibit undesirable fermentations, but all have drawbacks in terms of expense, stability, specificity, or objectionable off-fla- vors. Agents such as lactic acid or sorbic acid may be useful in lowering the necessary levels of SO,. In table grapes, the gaseous nature of SO, is indispens- able and a substitute will be difficult to identify.



    For the bleaching of cherries in the production of maraschino cherries, SO, has no equal. Sodium chlorite is useful for secondary bleaching (Beavers and Payne, 1968), but is far inferior for primary bleaching.


    Despite their long history of use as food additives, much remains to be learned about sulfites which would be helpful to the present concerns about their safety. Better information is needed on the issues of consumer exposure assessment and the toxicity and hypersensitivity reactions to sulfited foods.

    For the purpose of improving consumer exposure assessments, better analyt- ical methods and more analytical data are needed. The methods should empha- size determination of both free and combined sulfites. In particular, better meth- ods are needed for the determination of combined or total sulfites. The combined sulfites appear to be less toxic than the free sulfites, so analytical data for both free and combined (or total) sulfite are needed. The analytical data should em- phasize samples taken from typical points of consumption so that the losses on storage and preparation can be taken into account. As part of this effort, further investigations into the fate of sulfites in specific foods are needed. Emphasis should be placed on the identification of combined forms of sulfite so that their toxicity might be evaluated.

    From the viewpoint of toxicity assessment, further work is especially needed on the assessment of the toxicity of the combined sulfites. Since the bulk of sulfite ingestion is in the form of combined sulfites, the general lack of such information makes hazard evaluation virtually impossible. The toxicological studies should probably be focused on chronic and subchronic toxicity and should emphasize the oral route of administration. Further toxicological com- parisons are needed in sulfite oxidase-deficient versus normal animals.

    On the hypersensitivity issue, a variety of unknowns remain. The major issue will be the determination of the responsiveness of sulfite-sensitive asthmatics to sulfited foods in controlled challenge trials. Only through the use of such chal- lenges will the tolerance of these asthmatics for sulfited foods become available. In all likelihood, many sulfited foods will contain such low residual levels that they will not elicit asthma in these patients. The issue of the incidence of sulfite sensitivity in the asthmatic population remains to be answered as well. The incidence among mild asthmatics is unknown, although it appears as though the severe asthmatics are most likely to be sulfite reactors. The existence of more than one type of sulfite-sensitive asthmatic, acidic beverage reactors versus


    capsule reactors, seems likely from present data, but more studies are needed to establish the mechanisms (e.g., site of exposure) responsible for the existence of these two groups. The mechanism of action of encapsulated sulfites in inducing asthma in some asthmatics remains a mystery. Effective treatment may depend on the elucidation of this mechanism. Lastly, the existence of other types of hypersensitivity responses to sulfites has been established, but more studies are needed to establish the prevalence of such reactions.


    Adachi, T., Nonogi, H., Fuke, T., Ikuzawa, M., Fujita, K., Izumi, T., Hamano, T., Mitsuhashi, Y., Matsuki, Y., Suzuki, H., Toyoda, M., Ito, Y., and Iwaida, M. 1979. On the combination of sulphite with food ingredients (aldehydes, ketones, and sugars). Z. Lebensm. Unters. Forsch.

    Allen, D. H., and Collett, P. 1981. Life-threatening asthmatic reactions to the food and drug

    Allen, D., and Delohery, J. 1985. Metabisulfite induced asthma. J . Allergy Clin. Immunol. 75, 145. Allen, D. H., Van Nunen, S., Loblay, R., Clarke, L., and Swain, A. 1984. Adverse reactions to

    foods. Med. J. Ausr. 141, S37442. Altman, L. C., Sprenger, J. D., Marshall, S., Johnson, L. E., Koenig, J., and Pierson, W. E. 1985.

    Neutrophil chemotactic activity in sulfite sensitive patients. J . Allergy Clin. Immunol. 75, 144. American Conference of Governmental Industrial Hygients. 1982. Threshold limit values for chem-

    ical substances in work air. Am. Conf. Govt. I d . Hyg., Cincinnati. Ohio. Anderson, C., Warner, C. R., Daniels, D. H., and Padgett, K. L. 1983. Analysis of sulfites in foods

    by ion chromatography. AOAC Annu. Meet., 97th. Washington, D . C . , Ocr. 6. Anonymous 1984. BATF proposes lowering sulfite content of wine. Food Chem. News Oct. 1, 20-

    22. Applebaum, R. S., and Marth, E. H. 1982. Use of sulphite or bentonite to eliminate aflatoxin MI

    from naturally contaminated raw whole milk. Z . Lebensm. Unters. Forsch. 174, 303-305. Baker, G. J . , and Allen, D. H. 1982. The spectrum of metabisulfite induced asthmatic reactions;

    their diagnosis and management. Ausr. N. Z. Med. J . 12, 213. Baker, G. J., Collett, P., and Allen, D. H. 1981. Bronchospasm induced by metabisulfite-containing

    foods and drugs. Med. J. Aust. 2, 614-616. Baloch, A. K . , Buckle, K. A., and Edwards, R. A. 1977. Stability of @-carotene in model systems

    containing sulphite. J. Food Technol. 12, 309-316. Beavers, D. V., and Payne, C. H. 1968. Upgrades brined cherries. Food Eng. 40, 84-85. Beems, R. B., Spit, B. J., K&ter, H. B. W. M., and Feron, V. J. 1983. Nature and histogenesis of

    Beutler, H. 0. 1984. A new enzymatic method for determination of sulphite in food. Food Chem.

    Bhagat, B., and Luckett, M. F. 1960. The absorption and elimination of metabisulphite and thiosul-

    Bhagat, B., and Lockett, M. F. 1964. The effect of sulphite in solid diets on the growth of rats. Food

    Boushey, H. A. 1982. Bronchial hyperreactivity to sulfur dioxide: Physiologic and political implica-

    168, 200-205.

    preservative, sodium metabisulfite. J . Allergy Clin. Immunol. 67, 7 I .

    sulfite-induced gastric lesions in rats. Exp. Mol. Parhol. 36, 3 16-325.

    15, 157-164.

    fate by rats. J . Pharm. Pharmacol. 12, 690-694.

    Cosmet. Toxicol. 2, 1-13.

    tions. J. Allergy Clin. Immunol. 69, 335-338.


    Braverman, B., Shapiro, R., and Szer, W. 1975. Modification of E . coli ribosomes and coliphage MS2 RNA by bisulfite: Effects on ribosomal binding and protein synthesis. Nucleic Acids Res.

    Bruno, P., Caselli, M., Di Fano, A., and Traini, A. 1979. Fast and simple polarographic method for determination of free and total sulphur dioxide in wines and other common beverages. Analyst

    Buckley, C. E., Saltzman, H. A., and Sieker, H. 0. 1985. The prevalence and degree of sensitivity

    Buechsenstein, J. W., and Ough, C. S . 1978. SO2 determination by aeration-oxidation: A com-

    Burroughs, L. F. 1975. Determining free sulfur dioxide in red wine. Am. J. Enol. Vitic. 26, 25-29. Burroughs, L. F. 1977. Stability of patulin to sulfur dioxide and to yeast fermentation. J. Assoc. Of.

    Anal. Chem. 60, 100-103. Burruughs, L. F., and Sparks, A. H. 1973a. Sulphite-binding power of wines and ciders. I. Equi-

    librium constants for the dissociation of carbonyl bisulphite compounds. J . Sci. Food Agric. 24,

    Burroughs, L. F., and Sparks, A. H. 1973b. Sulphite-binding power of wines and ciders. 11. Theoretical considerations and calculation of sulphite-binding equilibria. J. Sci. Food Agric. 24,

    Burroughs, L. F., and Sparks, A. H. 1973c. Sulphite-binding power of wines and ciders. 111. Determination of carbonyl compounds in a wine and calculation of its sulphite-binding power. J. Sci. Food Agric. 24, 207-217.

    Bush, R. K., Taylor, S . L., and Busse, W. W. 1986. A critical evaluation of clinical trials in reactions to sulfites. J. Allergy Clin. Immunol. (in press).

    Calabrese, E., Sacco, C., Moore, G., and DiNardi, S. 1981. Sulfite oxidase deficiency: A high risk factor in SO,, sulfite, and bisulfite toxicity? Med. Hypotheses 7 , 133-145.

    Cam, J. G., Davies, P. A., and Sparks, A. H. 1976. The toxicity of sulphur dioxide toward certain lactic acid bacteria from fermented apple juice. J. Appl. Bucreriol. 40, 201-212.

    Cecil, R. 1963. Intramolecular bonds in proteins. I. The role of sulfur in proteins. In The Proteins (H. Neurath, ed.), Vol. 1, 2nd Ed., p. 379. Academic Press, New York.

    Chambers, R. W., Aguyagi, S. Y., Furukawa, Y., Zawadska, H., and Bhanet, 0. S . 1973. Inactiva- tion of valine acceptor activity by a cystosine into uracil missense change in the anticodon of yeast valine transfer RNA. J . Biol. Chem. 248, 5549-5551.

    Chandler, B. V., and Clegg, K. M. 1970. Pink discoloration in canned pears. 111. Inhibition by chemical additives. J . Sci. Food Agric. 21, 323-328.

    Chapon, L., Chapon, S., and Djenga, N. 1982. Free sulfite in beers-kinetic study. J. Am. Soc. Brew. Chem. 40, 31-39.

    Chin, S . , Bissell, M. J., and Bassham, J. A. 1977. The consequences of bisulfite exposure in primary chick embryo fibroblast in culture. Bull. Environ. Conrum. Toxicol. 18, 749-757.

    Cluzan, R., Causeret, J., and Hugot, D. 1965. Potassium metabisulfite: Long-term study of toxicity in the rat. Ann. Biol. Anim. Bioch. Biophys. 5 , 267-281.

    Codex Alimentarius Commission. March 1975. Potential daily intake of food additives. Joint FAO/WHO Food Standards Programme, Codex Committee on Food Additives, 10th session, p. 5 . Report CXlFA 75/5. Food and Agriculture Organization, Rome.

    Cohen, H. J., and Fridovich, 1. 1971a. Hepatic sulfite oxidase. Purification and properties. J. Biol. Chem. 246, 359-366.

    Cohen, H. J., and Fridovich, 1. 1971b. Hepatic sulfite oxidase. The nature and function of the heme prosthetic groups. J. Biol. Chem. 246, 367-373.

    Cohen, H. J. , Fridovich, I., and Rajagopalan, K. V.1971. Hepatic sulfite oxidase: A functional role for molybdenum. J. Biol. Chem. 246, 374-382.

    2, 501-507.

    104, 1083-1087.

    to ingested sulfites. J. Allergy Clin. Immunol. 7 5 , 144.

    parison with Ripper. Am. J . Enol. Vitic. 29, 161-164.

    187- 198.



    Cohen, H. J., Drew, R. T., Johnson, J . L., and Rajagopalan, K. V. 1973. Molecular basis of the biological function of molybdenum. The relationship between sulfite oxidase and the acute toxicity of bisulfite and SOz. Proc. Natl. Acad. Sci. U.S.A. 70, 3655-3659.

    Cruess, W. V., and Glickson, D. 1932. Observations on brining and candying of citron. Fruir Prod.

    Daoud, H. N., Luh, B. S., and Miller, M. W. 1977. Effect of blanching, EDTA and NaHS03 on color and vitamin B6 retention in canned garbanzo beans. J. Food Sci. 42, 375-378.

    Das, G., and Runeckles, V. C. 1974. Effect of bisulfite on metabolic development in synchronous Chorella pyrenoidosa. Environ. Res. 7, 353-362.

    Davis, E. G., McBean, D. M., Rooney, M. L., and Gipps, P. G. 1973. Mechanisms of sulphur dioxide loss from dried fruits in flexible films. J. Food Technol. 8 , 391-405.

    Delohery, J . , Castle, W.. Simmul, R., and Allen, D. 1984a. Metabisulfite and SO2 reactivity in asthmatics. J . Allergy Clin. Immunol. 73, 136.

    Delohery, I., Simmul, R., Castle, W. D., and Allen, D. 1984b. The relationship of inhaled sulfur dioxide reactivity to ingested metabisulfite sensitivity in patients with asthma. Am. Rev. Respir.

    Dorange, J.-L., and Dupuy, P. 1972. Mise en evidence dun action mutaghe de sulfite de sodium sur la levure. C. R . Acad. Sci. Ser. D 274, 2798-2800.

    Don, W., and Truper, H. G. 1976. Sulfite formation by wine yeasts. 111. Properties of sulfite reductase. Arch. Microbiol. 108, 99-104.

    Don, W., Heinzel, M., and Truper, H. G . 1976. Sulfite formation by wine yeasts. I. Relationships between growth, fermentation, and sulfite formation. Arch. Microbiol. 107, 289-292.

    Don, W., Heinzel, M., and Truper, H. G. 1977. Sulfite formation by wine yeasts. IV. Active uptake of sulfate by low and high sulfite producing wine yeasts. Arch. Microbiol. 112, 283- 285.

    Doyle, M. P., and Marth, E. H. 1978a. Bisulfite degrades aflatoxin: Effect of temperature and concentration of bisulfite. J. Food Pror. 41, 774-780.

    Doyle, M. P., and Marth, E. H. 1978b. Bisulfite degrades aflatoxin: Effect of citric acid and methanol and possible mechanisms of degradation. J . Food Pror. 41, 891-896.

    Dulak, L., Chaing, G., and Gunnison, A. F. 1984. A sulphite oxidase-deficient rat model: Re- productive toxicology of sulphite in the female. Food Chem. Toxicol. 22, 599-607.

    Duran, K., Korteland, J., Beemer, F. A., Heiden, C. v. d., de Bree, P. K., Brink, M., and Wadman, S. K. 1979. Variability of sulfituria: Combined deficiency of sulfite oxidase and xanthine oxidase. In Models for the Study of Inborn Errors of Metabolism (F. A. Hommes, ed.), pp. 103-107. Elsevier, Amsterdam.

    Eschenbruch, R . 1974. Sulfite and sulfide formation during winemaking-a review. Am. J. Enol. Viric. 25, 157-161.

    Eschenbruch, R., and Bonish, P. 1976. Production of sulphite and sulphide by low- and high- sulphite forming wine yeasts. Arch. Microbiol. 107, 299-302.

    Evelyn, J. 1664. Pomona, or an appendix concerning fruit trees. In relation to cider and several ways of ordering it. Supplement, Aphorisms Concerning Cider, p. 24. Martin & Allestry, London.

    Ezrielev, G. I. 1968. Contribution to the toxicology of sodium metabisulphite. Pharmacol. Toxicol. 31, 120.

    Fieger, E. A. 1951. Cause and prevention of black spot on shrimp. Ice Refrig. 120, 49-50. Firth, R. A., Hill, H. A. 0.. Pratt, J. M., Thorp, R. G., and Williams, R. J. P. 1969. The chemistry

    of vitamin BIZ. Part XI. Some further formation constants. J. Chem. SOC. A, 381-386. Fisher, A. A. 1975. Contact dermatitis due to food additives. Curis 16, 691. Fitzhugh, 0. G., Knudsen, L. F., and Nelson, A. A. 1946. The chronic toxicity of sulfites. J .

    J . 12, 17-18, 25.

    Dis. 130, 1027-1032.

    Pharmacol. Exp. Ther. 86, 37-48.


    Flaherty, M., Stormont, J. M., and Condemi, J. J. 1985. Metabisulfite (MBS)-associated hepatotox- icity and protection by cobalamins (B12). J . Allergy Clin. Immunol. 75, 198.

    Freedman, B. J . 1977. Asthma induced by sulphur dioxide, benzoate and tartrazine contained in orange drinks. Clin. Allergy 7, 407-4 15.

    Fujita, K., Ikuzawa, M., Izumi, T., Hamano, T., Mitsuhashi, Y., Matsuki, Y., Adachi, T., Nonogi, H., Fuke, T., Susuki, H., Toyoda, M., Ito, Y., and Iwaida, M. 1979. Establishment of a modified Rankine method for the separate determination of free and combined sulphites in foods. 2. Lebensm. Unters. Forsch. 168, 206-2 11.

    Generoso, W. M., Huff, S. W., and Cain, K. T. 1978. Tests on induction of chromosome aberra- tions in mouse germ cells with sodium bisulfite. Murar. Res. 56, 363-365.

    Gibson, W. B . , and Strong, F. M. 1973. Metabolism and elimination of sulphite by rats, mice and monkeys. Food Costnet. Toxicol. 11, 185-198.

    Gibson, W. B . , and Strong, F. M. 1974. Accumulation of ingested sulphite- and sulphate-sulphur and utilization of sulphited proteins by rats. Food Comer. Toxicol. 12, 625-640.

    Gilbert, J . , and McWeeny, D. J. 1976. The reactions of sulphur dioxide in dehydrated vegetables. J . Sci. Food Agric. 27, 1145- 1146.

    Goldfarb, G., and Simon, R. A. 1984. Provocation of sulfite sensitive asthma. J . Allergy Clin. Immunol. 73, 135.

    Green, L. F. 1976. Sulphur dioxide and food preservation-a review. Food Chem. 1, 103-124. Gunnison, A. F. 1981. Sulphite toxicity: A critical review of in virro and in vivo data. Food Cosmer.

    Gunnison, A. F., and Farruggella, T. J. 1979. Preferential S-sulfonate formation in lung and aorta. Chem. -Biol. Inreracr. 25, 27 1-277.

    Gunnison, A. F., and Palmes, E. D. 1973. Persistence of plasma S-sulfonates following exposure of rabbits to sulfite and sulfur dioxide. Toxicol. Appl. Pharmacol. 24, 266-278.

    Gunnison, A. F., and Palmes, E. D. 1974. S-Sulfonates in human plasma following inhalation of sulfur dioxide. Am. Ind. Hyg. Assoc. J . 35, 288-291.

    Gunnison, A. F., and Palmes, E. D. 1976. A model for the metabolism of sulfite in mammals. Toxicol. Appl. Pharmacol. 38, 1 11-126.

    Gunnison, A. F., and Palmes, E. D. 1978. Species variability in plasmas-sulfonate levels during and following sulfite administration. Chem.-Biol. Interact. 21, 315-329.

    Gunnison, A. F., Bresnahan, C. A., and Palmes, E. D. 1977. Comparative sulfite metabolism in the rat. rabbit, and rhesus monkey. Toxicol. Appl. Pharmacol. 42, 99-109.

    Gunnison, A. F., Dulak, L., Chiang, G., Zaccadi, J., and Farruggella, T. J. 1981a. A sulfite oxidase-deficient rat model: Subchronic toxicology. Food Cosmer. Toxicol. 19, 221-232.

    Gunnison, A. F., Farruggella. T. J., Chiang, G., Dulak, L., Zaccardi, J . , and Birkner, J. 1981b. A sulfite-oxidase-deficient rat model: Metabolic characterization. Food Cosmer. Toxicol. 19,209- 220.

    Habernicht, H. A., Preuss, L., and Lovell, R . G. 1983. Sensitivity to ingested metabisulfites: Cause of brofichospasm and urticaria. Immunol. Allergy Prac. 5 , 243-245.

    Hagler, W. M., Jr., Hutchins, J. E., and Hamilton, P. B. 1982. Destruction of aflatoxin in corn with sodium bisulfite. J. Food Pror. 45, 1287-1291.

    Hagler, W. M., Jr., Hutchins, J. E., and Hamilton, P. B. 1983. Destruction of aflatoxin B, with sodium bisulfite: Isolation of the major product aflatoxin BIS. J. Food Prof. 46, 295-300.

    Haisman, D. R. 1974. The effect of sulphur dioxide on oxidizing enzyme systems in plant tissues. J . Sci. Food Agric. 25, 803-810.

    Hamano, T., Mitsuhashi, Y., Matsuki, Y., Ikuzawa, M., Fujita, K., Izumi, T., Adachi, T., Nonogi, H., Fuke, T., Suzuki, H., Toyoda, M., Ito, Y., and Iwaida, M. 1979. Application of gas chromatography for the separate determination of free and combined sulphites in foods. Z. Lebensm. Unrers. Forsch. 168, 195-199.

    Toxicol. 19, 667-682.


    Hayatsu, H. 1976. Bisulfite modification of nucleic acids and their constituents. frog. Nucleic Acid Res. Mol. Biol. 16, 75-124.

    Hayatsu, H., and Miller, R. C., Jr. 1972. The cleavage of DNA by the oxygen-dependent reaction of bisulfite. Biochem. Biophys. Res. Commun. 46, 120-124.

    Hayatsu, H., and Miura, A. 1970. The mutagenic action of sodium bisulfite. Biochem. Biophys. Res. Commun. 39, 156-160.

    Hayatsu, H., Wataya, Y., Kai, K., and lida, S. 1970. Reaction of sodium bisulfite with uracil, cytosine, and their derivatives. Biochemistry 9, 2858-2865.

    Heinzel, M., and Truper, G. 1976. Sulfite formation by wine yeasts. 11. Properties of ATP-sul- furylase. Arch. Microbiol. 107, 293-297.

    Hickey, R. J., Clelland, R. C., Bowers, E. J., and Boyce, D. E. 1976. Health effects of atmospheric sulfur dioxide and dietary sulfites. Arch. Environ. Health 31, 108-1 12.

    Hoppe, J. O., and Goble, F. C. 1951. The intravenous toxicity of sodium bisulfite. J. Pharmacol. Exp. Ther. 104, 101-106.

    Honvitz, W. (ed.) 1975. Official Methods of Analysis, 12th Ed., pp. 336-338. Association of Official Analytical Chemists, Washington, D.C.

    Hotzel, D., Muskat, E., Bitsch, I . , Aign, W., Althoff, J.-D., and Cremer, H. D. 1969. Thiamin- mangel and Unbedenklichkeit von Sulfit fiir den Menschen. Inr. 2. Viruminforsch. 39, 372- 383.

    Howland, W. C., and Simon, R. A. 1985. Restaurant-provoked asthma: Sulfite sensitivity? J . Allergy Clin. Immunol. 75, 145.

    Huang, A. S., and Fraser, W. M. 1984. Are sulfite additives really safe? N. Engl. J. Med. 311,542. Hysert, D. W., and Momson, N. M. 1976. Sulfate metabolism during fermentation. Am. Soc. Brew.

    Chem. J . 34, 25-3 I . Ingram, M. 1959. Technical aspects of the commercial use of antimicrobial chemicals as food

    preservatives. Chem. Ind. 552-557. Inoue, M., Hayatsu, H., and Tanooka, H. 1972. Concentration effect of bisulfite on the inactivation

    of transforming activity of DNA. Chem.-Biol. Interact. 5 , 85-95. Inouye, B., Morita, K., Ishida, T., and Ogata, M. 1980. Cooperative effect of sulfite and vanadium

    compounds on lipid peroxidation. Toxicol. Appl. Pharmucol. 53, 101-107. Institute of Food Technologists Expert Panel on Food Safety & Nutrition. 1975. Sulfites as food

    additives. Food Technol. 29, 117-120. Irreverre, F., Mudd, S. H., Heizer, W. D., and Laster, L. 1967. Sulfite oxidase deficiency: Studies

    of a patient with mental retardation, dislocated ocular lenses, and abnormal urinary excretion of S-sulfo-L-cysteine, sulfite, and thiosulfate. Biochem. Med. 1, 187-216.

    Jacobs, D. D. 1976. Effect of dissolved oxygen on free sulfur dioxide in red wines. Am. J. Enol. Vitic. 27, 42-43.

    Jacobsen, D. W., Simon, R. A,, and Singh, M. 1984. Sulfite oxidase deficiency and cobalarnin protection in sulfite-sensitive asthmatics (SSA). J . Allergy Clin. Immunol. 73, 135.

    Jagiello, G . M., Lin, J. S., and Ducayen, M. B. 1975. SOz and its metabolites: Effects on mam- malian egg chromosomes. Environ. Res. 9, 84-93.

    Jaulrnes, P. 1970. Etat actuel des techniques pour &placement de Ianhydride sulfureux. Bull. Off. Int. Vigne Vin 478, 1333.

    Jennings, N., Bunton, N. G., Crosby, N. T., and Alliston, T. G. 1978. A comparison of three methods for the determination of sulphur dioxide in food and drink. J. Assoc. Publ. Anal. 16,

    Johnson, J. L., and Rajagopalan, K. V. 1976a. Purification and properties of sulfite oxidase from

    Johnson, J. L., and Rajagopalan, K. V. 1976b. Human sulfite oxidase deficiency. Characterization


    human liver. J . Clin. Invesr. 58, 543-550.

    of the molecular defect in a multicomponent system. J . Clin. Invesr. 58, 551-556.


    Johnson, J. L., Rajagopolan, K. V., and Cohen, H. J. 1974. Molecular basis of the biological function of molybdenum. Effect of tungsten on xanthine oxidase and sulfite oxidase in the rat. J. Biol. Chem. 249, 859-866.

    Johnson, J. L . , Jones, H. P., and Rajagopalan, K. V. 1977. In vitro reconstitution of demolyb- dosulfite oxidase by a molybdenum cofactor from rat liver and other sources. J. Biol. Chem. 252,4994-5003.

    Johnson, J. L., Waud, W. R., Rajagopalan, K. V., Duran, M., Beemer, F. A, , and Wadman, S. K. 1980. Inborn errors of molybdenum metabolism: Combined deficiencies of sulfite oxidase and xanthine dehydrogenase in a patient lacking the molybdenum cofactor. Proc. Natl. Acad. Sci.

    Joint FAO/WHO Expert Committee on Food Additives. 1974. Toxicological evaluation of certain food additives with a review of general principles and of specifications. 17th Report, Food and Agriculture Organization, Rome.

    Joslyn, M. A., and Braverman, J. B. S. 1954. The chemistry and technology of the pretreatment and preservation of fruit and vegetable products with sulfur dioxide and sulfites. Adv. Food Res. 5,

    Kaczka, E. A., Wolf, D. E., Kuehl, F. A., and Folkers, K. 1950. Vitamin BIZ: Reactions of cyano- cobalamin and related compounds. Science 112, 354-355.

    Kai, K., TSUNO, T., and Hayatsu, H. 1974. The effect of bisulfite modification on the template activity of DNA for DNA polymerase I. Nucleic Acids Res. 1, 889-899.

    Kaplan, D., McJilton, C., and Luchtel, D. 1975. Bisulfite induced lipid oxidation. Arch. Environ. Health 30, 507-509.

    Kessler, D. L., and Rajagopalan, K. V. 1972. Purification and properties of sulfite oxidase from chicken liver. J. Biol. Chem. 247, 6566-6573.

    Kikigawa, K., and Iizuka, K. 1972. Inhibition of platelet aggregation by bisulfite-sulfite. J. Pharm. Sci. 61, 1904-1907.

    Kingkate, A , , Jaengdawang, C., Chakrangkoon, P., Halilamian, C., and Toyoda, M. 1981. Re- sidual sulfur dioxide in some Thai noodles. J . Food Pror. 44, 334-336.

    Kitamura, N., and Hayatsu, H. 1974. Cleavage of the glycosidic linkage of pyrimidine ribonucleosides by the bisulfite-oxygen system. Nucleic Acids Res. 1, 75-86.

    Kochen, J. 1976. Sulfur dioxide, a respiratory tract irritant, even if ingested. Pediatrics 52, 145- 146.

    Koenig, J. Q., Pierson, W. E., and Frank, R. 1980. Acute effects of inhaled SOz plus NaCl droplet aerosol on pulmonary function in asthmatic adolescents. Environ. Res. 22, 145-153.

    Koepke, J. W., Selner, J. C., and Dunhill, A. L. 1983. SO2 derived from bronchodilator solutions. J . Allergy Clin. Immunol. 71, 147.

    Koepke, J. W., Christopher, K. L., Chai, H., and Selner, J. C. 1984a. Dose-dependent bron- chospasm from sulfites in isoetharine. J . Am. Med. Assoc. 251, 2982-2983.

    Koepke, J. W., Selner, J. C., Christopher, K., and Glover, G. 1984b. Inhaled metabisulfite sen- sitivity. J. Allergy Clin. Immunol. 73, 135.

    Komanowsky, M . , Talley, F. B., and Eskew, R. K. 1970. Air drying of cultivated mushrooms. Food Technol. 24, 80-84.

    Kudo, I . , Miura, A., and Hayatsu, H. 1978. Inactivation of bacteriophage X during the aerobic oxidation of bisulfite. Environ. Res. 16, 205-215.

    Kudo, I., Hayatsu, H., Yokoiyama, A., and Kuroda, Y. 1980. Effect of bisulfite on cell-to- substratum adhesion of Chinese hamster cells in culture. J . Pharmacobiodyn. 3, 74-78.

    Lafontaine, A., and Goblet, J. 1955. La toxicit6 des sulfites. Arch. Belg. Med. SOC. Hyg. Med. Trav. Med. Leg. 13, 281-287.

    Lamikanra, 0. 1982. Sulfite-induced oxidation and browning of linolenic acid. J. Food Sci. 47,

    V.S.A. 77, 3715-3719.


    2025-2027, 2032.


    Lauteaume, M.-Th., Ramel, P., Jaulmes, P., and Manin, D. 1969. Dttermination et comparaison des DL 50 du mktabisulfite de potassium de lethanol et de leur combinaison (hydroxydthane- sulfonate de potassium) par voie orale sur le rat de souche Wistar. Ann. FalsiJ Expert. Chim.

    Leuch, 1895. Korresp. bl. Schweiz. Aertze. 609. Cited in Cluzan, R., Causeret, J . , and Hugot, D. 1965. Potassium metabisulfite: Long-term study of toxicity in the rat. Ann. Biol. Anim. Bio- chern. Biophys. 5 , 267-281.

    Lewis, R. J., and Tatken, R. L. 1979. Registry of Toxic Effects of Chemical Substances-1979. U.S. Dept. Health and Human Services, U.S. Govt. Printing Office, Washington, D.C.

    Life Sciences Research Office. 1976. Evaluation of the health aspects of sulfiting agents as food ingredients. Prepared for the Food and Drug Administration under FDA contract No. 223-75-2004, Bethesda, Maryland. 25 pp.

    Life Sciences Research Office. 1985. The reexamination of the GRAS status of sulfiting agents. Prepared for the Food and Drug Administration under FDA contract No. 223-83-2020, Beth- esda, Maryland. 96 pp.

    Linn, W. S., Shamoo, D. A., Spier, C. E., Valencia, L. M., Anzar, U. T. , Venet, T. G., and Hackney, J . D. 1983. Respiratory effects of 0.75 ppm sulfur dioxide in exercising asthmatics: Influence of upper-respiratory defenses. Environ. Res. 30, 340-348.

    Lisberg, G., and Chen, T . 4 . 1973. Storage stability of dehydrated potato granules packaged in cans and cartons. J. Food Sci. 38, 363-364.

    Liu, J.-W. R., and Gallander, J. F. 1983. Effect of pH and sulfur dioxide on the rate of malolactic fermentation in red table wines. Am. J . Enol. Viric. 34, 44-46.

    Lockett, M. F., and Natoff, I. L. 1960. A study of the toxicity of sulfite. J. Phurm. Phurmacol. 12, 488-496.

    Luh, B. S., Karbassi, M., and Schweigert, B. S. 1978. Thiamine, riboflavin, niacin and color retention in canned small white and garbanzo beans as affected by sulfite treatment. J . Food Sci. 43,431-434.

    Lukes, B., OBrien, T. J., and Scanlan, R. A. 1980. Residual sulfur dioxide in finished malt: Colorimetric determination and relation to N-nitrosodimethylamine. Am. SOC. Brew. Chem. J .

    McBean, D. M. 1967. Levels of free and combined sulfur dioxide in fruits during sulfuring and

    McGinnis, R. A. 1982. Beet-Sugar Technology, p. 265. Beet Sugar Development Foundation,

    MacLeod, R. M., Farkas, N., Fridovich, I., and Handler, P. 1961. Purification and properties of

    MacRae, W. D., and Stich, H. F. 1979. Induction of sister chromatid exchanges in Chinese hamster

    McWeeny, D. J . 1979. The chemical behavior of food additives. Proc. Nurr. SOC. 38, 129-133. McWeeny, D. J., Knowles, M. E., and Hearne, J. F. 1974. The chemistry of non-enzymic browning

    in foods and its control by sulphites. J. Sci. Fd. Agric. 25, 735-746. McWeeny, D. J., Shepard, M. J., and Bates, M. L. 1980. Physical loss and chemical reactions of

    SO2 in strawberry jam production. J. Food Technol. 15, 613-617. Mallon, R. G. , and Rossman, T. B. 1981. Bisulfite (sulfur dioxide) is a comutagen in E . coli and in

    Chinese hamster cells. Mutar. Res. 88, 125-133. Means, G. E., and Feeney, R. E. 1971. Chemical Modification of Proteins, p. 152. Holden-Day,

    San Francisco. Meggs, W. J. , Atkins, F. M., Wright, R. H., Fishman, M., Kaliner, M. A, , and Metcalfe, D. D.

    1985. Sulfite challenges in patients with systemic mastocytosis (SM) or unexplained anaphylax- is (UEA). J. Allergy Clin. Immunol. 75, 144.

    62, 231-241.


    38, 146-148.

    drying. Food Technol. 21, 1402-1406.

    Fort Collins, Colorado.

    hepatic sulfite oxidase. J . Biol. Chem. 236, 1841-1846.

    cells by the reducing agents bisulfite and ascorbic acid. Toxicology 13, 167-174.


    Mitsuhashi, Y., Hamano, T., Hasegawa, A., Tanaka, K., Matsuki, Y., Adachi, T., Obara, K., Nonogi, H., Fuke, T., Sudo, M., Ikuzawa, M., Fujita, K., Izumi, T., Ogawa, S., Toyoda, M., Ito, Y., and Iwaida, M. 1979. Comparative determination of free and combined sulphites in foods by the modified Rankine method and flame photometric detection gas chromatography. Z. Lebensm. Vnters. Forsch. 168, 299-304.

    Monier-Williams, G. W. 1927. Determination of Sulphur Dioxide in Foods. Public Health and Medical Subjects, Rept. No. 43, pp. 1-56. Great Britain Ministry of Health, London.

    Montgomery, M. W, 1983. Cysteine as an inhibitor of browning in pear juice concentrate. J. Food Sci. 48, 951-952.

    Moms, J. R., Cawthon, D. L., and Fleming, J. W. 1979. Effects of temperature and SOz addition on quality and postharvest behavior of mechanically-harvested juice grapes in Arkansas. J. Am. SOC. Hortic. Sci. 104, 166-169.

    Mudd, S. H., Irreverre, F., and Laster, L. 1967. Sulfite oxidase deficiency in man: Demonstration of the enzymatic defect. Science 156, 1599-1601.

    Mukai, F., Hawryluk, I., and Shapiro, R. 1970. The mutagenic specificity of sodium bisulfite. Biochem. Biophys. Res. Commun. 39, 983-988.

    Miiller, F., and Massey, V. 1969. Flavin-sulfite complexes and their structures. J. Biol. Chem. 244,

    Miinzer, R. 1980. Untersuchungen zur mutagenen Wirkung von Bisulfit. Lebensmittel-Wiss. Tech- nol. 13, 219-220.

    Nadel, J. A., Salem, H., Tamplin, B., and Tokiwa, Y. 1965. Mechanism of bronchoconstriction during inhalation of sulfur dioxide. J. Appl. Physiol. 20, 164-167.

    Nelson, K. E. 1983. Effects of in-package sulfur dioxide generators, package liners, and temperature on decay and desiccation of table grapes. Am. J. Enol. Vitic. 34, 10-16.

    Nelson, K. E., and Ahmedullah, M. 1973. Effect of temperature change on the release rate of sulfur dioxide from two-stage sodium bisulfite generators. Am. J . Enol. Viric. 24, 75-80.

    Nelson, K. E., and Ahmedullah, M. 1976. Packaging and decay-control systems for storage and transit of table grapes for export. Am. J. Enof. Vitic. 27, 74-79.

    Nir, I., Kafri, I., and Cohen, R. 1978. Effect of pelleting on the vitamin K activity of various vitamin K3 preparations in chicks. Poultry Sci. 57, 206-209.

    Njagi, G. D. E., and Gapalan, H. N. B. 1982. Cytogenetic effects of the food preservatives- sodium benzoate and sodium sulphite on Viciu fubu root meristems. Murut. Res. 102, 213-219.

    Nofre, C., Dufour, H., and Cier, A. 1963. Toxicit6 g6nCrale compare6 des anions mintraux chez la souris. C. R. Acud. Sci. 257, 791-794.

    Nury, N. S., Taylor, D. H., and Brekke, J. E. 1959. Modified direct colorimetric method for determination of sulfur dioxide in dried fruit. Agric. Food Chem. 7, 351-353.

    Ogawa, S., Suzuki, H., Toyoda, M., Ito, Y., Iwaida, M., Nonogi, H., Fuke, T. , Obara, K., Adachi, T., Fujita, K., Ikuzawa, M., Izumi, T., Hamano, T., Mitsuhashi, Y., and Matsuki, Y. 1979. Colorimetric microdetermination of sulphites in foods by use of the modified Rankine apparatus. Z. Lebensm. Unters. Forsch. 168, 293-298.

    Ogier, H., Saudubray, J. M., Charpentier, C., Munnich, A., Perignon, J. L., Kesseler, J., and Frezal, A. 1982. Double deficit en sulfite et xanthine oxydase, cause dencephalopathie due a une anomalie hereditaire du metabolisme du molybdene. Ann. Med. Interne 133, 594-596.

    Oshino, N., and Chance, B. 1975. The properties of sulfite oxidation in perfused rat liver: Interaction of sulfite oxidase with the mitochondria1 respiratory chain. Arch. Biochem. Biophys. 170,514- 528.

    Pitman, I. H., and Jain, N. B. 1979. Covalent addition of bisulfite ion to N-alkylated uracils and thiouracils. Aust. J. Chem. 32, 545-552.

    Ponting, J. D., and Johnson, G. 1945. Determination of sulfur dioxide in fruits. Ind . Eng. Chem. 17,




    Ponting, J. D., Jackson, R., and Watters, G. 1971. Refrigerated apple slices: Effects of pH, sulfites and calcium on texture. J. Food Sci. 36, 349-350.

    Prenner, B. M., and Stevens, J. J. 1976. Anaphylaxis after ingestion of sodium bisulfite. Ann. Allergy 37, 180-182.

    Rankine, B. C., and Pocock, K. F. 1969. Influence of yeast strain on binding of sulfur dioxide in wines, and on its formation during fermentation. J. Sci. Food Agric. 20, 104-109.

    Rankine, B. C., and Pocock, K. F. 1970. Alkalimetric determination of sulphur dioxide in wine. Ausr. Wine Brew. Spirir Rev. 29, 40-44.

    Reith, J. F., and Willems, J. J. L. 1968. Ober die Bestimmung der schwefligen Saure in Lebensmit- teln. Z. Lebensm. Unters. Forsch. 108, 270-280.

    Renner, H. W., and Wever, J. 1983. Attempts to induce cytogenetic effects with sulphite in sulphite oxidase-deficient Chinese hamsters and mice. Food Chem. Toxicol. 21, 123-127.

    Ripper, M. 1892. Die schweflige Saure im Weine und deren Bestimmung. J. Prakt. Chem. 46,428- 473.

    Roberts, A. C., and McWeeny, D. J. 1972. The uses of sulphur dioxide in the food industry-a review. J. Food Technol. 7, 221-238.

    Rork, G. S., and Pitman, I. H. 1974. Elimination of bisulfite ion from a series of uracil-bisulfite adducts. Evidence for a two-step mechanism. J . Am. Chem. SOC. 96, 4654-4663.

    Rost, E., and Franz, F. 1913. Vergleichende Untersuchung der phamakologischen Wirkiingen der organisch gehundenen schwefligen Sauren und des neutralen schwefligsauren Natriums. Arb. Kaiserl. Gesundh. 43, 187-303.

    Rudzki, E. 1979. Two not yet described contact allergens: Sodium hyposulfite and tertiary butyl alcohol. Przegl. Dermarol. 66, 375-377.

    Salo, R. J. , and Cliver, D. 0. 1978. Inactivation of enteroviruses by ascorbic acid and sodium bisulfite. Appl. Environ. Microbiol. 36, 68-75.

    Sayavedra, L. A., and Montgomery, M. W. 1983. Effect of moisture, temperature and sulfur dioxide on color of dried applies. Annu. IFT Meet., 43rd, Abstr. No. 301.

    Schroeter, L. C. 1966. Sulfur Dioxide-Applications in Foods, Beverages, and Pharmaceuticals. Pergamon, Oxford.

    Schwartz, H. J. 1983. Sensitivity to ingested metabisulfite: Variations in clinical presentation. J . Allergy Clin. Immunol. 71, 487-489.

    Schwartz, H. J. 1984. Observations on the uses and effects of sulfiting agents in foods and drugs. Immunol. Allergy Prac. 6, 29-34.

    Schwartz, H. J., and Chester, E. H. 1984. Bronchospastic responses to aerosolized metabisulfite in asthmatic subjects: Potential mechanisms and clinical implications. J. Allergy Clin. Immunol. 74,511-513.

    Seyal, M. A., Nagy, S. M., Fletcher M. P., Ough, C. S., and Gershwin, M. E. 1984. Response of asthmatic and normal volunteers to the sulfiting agents of wine. J. Allergy Clin. Immunol. 73, 135.

    Shapiro, R. 1977. Genetic effects of bisulfite (sulfur dioxide). Murar. Res. 39, 149-176. Shapiro, R., and Braverman, B. 1972. Modification of polyuridylic acid by bisulfite: Effect on

    double helix formation and coding properties. Biochem. Biophys. Res. Commun. 47,544-550. Shapiro, R., and Gazit, A. 1977. Crosslinking of nucleic acids and proteins by bisulfite. Adv. Exp.

    Med. Biol. 86A, 633-640. Shapiro, R., Cohen, B. I., and Servis, R. E. 1970a. Specific deamination of RNA by sodium

    bisulphite. Nature (London) 227, 1047-1048. Shapiro, R., Servis, R. E., and Welcher, M. 1970b. Reactions of uracil and cytosine derivatives with

    sodium bisulfite. A specific deamination method. J. Am. Chem. SOC. 92, 422-424. Shapiro, R., Braverman, B., Louis, J. B., and Servis, R. E. 1973. Nucleic acid reactivity and


    conformation. 11. Reaction of cytosine and uracil with sodium bisulfite. J . Eiol. Chem. 248,

    Shapiro, R., DiFate, V., and Welcher, M. 1974. Deamination of cytosine derivatives by bisulfite. Mechanism of the reaction. J . Am. Chem. SOC. 96, 906-912.

    Shapiro, R., Welcher, M., Nelson, V., and DiFate, V. 1976. Reaction of uracil and thymine derivatives with sodium bisulfite. Studies on the mechanism and reduction of the adduct. Biochim. Biophys. Acta 425, 115-124.

    Sheppard, D., Wong, W. S., Uehara, C. F., Nadel, J. A., and Boushey, H. A. 1980. Lower threshold and greater bronchomotor responsiveness of asthmatic subjects to sulfur dioxide. Am. Rev. Respir. Dis. 122, 873-878.

    Shih, N. T., and Petering, D. H. 1973. Model reactions for the study of the interaction of sulfur dioxide with mammalian organisms. Biochem. Biophys. Res. Commun. 55, 1319- 1325.

    Shih, V. E., Abroms, I. F., Johnson, J. L., Carney, M., Mandell, R., Robb, R., Cloherty, J. P., and Rajagopalan, K. V. 1977. Sulfite oxidase deficiency. Biochemical and clinical investigations of hereditary metabolic disorder in sulfur metabolism. N . Engl. J. Med. 297, 1022-1028.

    Simon, R. A. 1984. Adverse reactions to drug additives. J. Allergy Clin. Immunol. 74, 623-630. Simon, R. A. 1985. Reactivity to inhaled BronkosolR in sulfite sensitive asthmatics (SSA). J .

    Allergy Clin. Immunol. 75, 145. Simon, R. A,, Green, L., and Stevenson, D. D. 1982. The incidence of ingested metabisulfite

    sensitivity in an asthmatic population. J . Allergy Clin. Immunol. 69, 118. Simon, R., Goldfarb, G., and Jacobsen, D. W. 1984. Blocking studies in sulfite sensitive asth-

    matics. J. Allergy Clin. Imrnunol. 73, 136. Slae, S., and Shapiro, R. 1978. Deamination of cytidine by bisulphite: Mechanism at neutral pH. J .

    Org. Chem. 43, 4197-4201. Smith, E. L., Ball, S. , and Ireland, D. M. 1952. B12 vitamins (cobalamins). 2. Neutral, basic, and

    acidic cobalamins. Eiochem. J . 52, 395-400. Somers, T. C., and Evans, M. E. 1977. Spectral evaluation of young red wines: Anthocyanin

    equilibria, total phenolics, free and molecular S02. chemical age. J. Sci. Food Agric. 28,

    Sonin, L . , and Patterson, R. 1985. Metabisulfite challenge in patients with idiopathic anaphylaxis. J.

    Sorbo, B.. and Ohman, S. 1978. Determination of thiosulfate in urine. Scand. J. Clin. Lab. Invesr.

    Southerland, W. M., Akogyeram, C. 0.. Toghrol, F., Sloan, L., and Schemer, R. 1982. Interaction of bisulfite with unsaturated fatty acids. J. Toxicol. Environ. Healrh 10, 479-491.

    Sprenger, J. D., Altman, L. C., Koenig, J., and Pierson, W. E. 1985. Simultaneous sulfite and sulfur dioxide (SO2) sensitivity. J. Allergy Clin. Immunol. 75, 144.

    Stacey, F. W., and Harris, J. F., Jr. 1963. Formation of carbon-hetero atom bonds by free radical chain additions to carbon-carbon multiple bonds. In Organic Reactions (A. C. Cope, R. Adams, A. H. Blatt, V. Boekelheide, T. L. Cairns, D. Y. Curtin, and C. Niemann, eds.), Vol. 13, pp. 196-200. Wiley, New York.

    Stafford, A. E., Bolin, H. R., and Mackey, B. E. 1972. Absorption of aqueous bisulfite by apricots. J . Food Sci. 37, 941-943.

    Stem, A. C., Wohlers, H. C., Boubel, R. W., and Lowry, W. P. 1973. Fundamentals of Air Pollution, pp. 30-31. Academic Press, New York.

    Stevenson, D. D., and Simon, R. A. 1981a. Metabisulfite (K2S2OS) sensitivity in asthmatics. J. Allergy Clin. Immunol. 67, 3.

    Stevenson, D. D., and Simon, R. A. 1981b. Sensitivity to ingested metabisulfites in asthmatic subjects. J. Allergy Clin. Immunol. 68, 26-32.



    Allergy Clin. Immunol. 75, 67-69.

    38, 521-527.


    Stevenson, D. D., and Simon, R. A. 1984. Sulfites and asthma. J . Allergy Clin. Immunol. 74,469- 472.

    Subcommittee on Review of the GRAS List (Phase 11). 1972. A comprehensive survey of industry on the use of food chemicals generally recognized as safe (GRAS). Committee on Food Protection, Division of Biology & Agriculture, National Research Council. National Academy of Sciences, Washington, D.C.

    Summers, G. A,, and Drake, J. W. 1971. Bisulfite mutagens in bacteriophage T4. Genetics 68,603- 607.

    Suwa, Y. , Nagao, M., Kosugi, A, , and Sugimura, T. 1982. Sulfite suppresses the mutagenic effect of coffee. Mutaf. Res. 102, 383-391.

    Tanaka, T., Fujii, M., Mori, H., and Hirono, I. 1979. Carcinogenicity test of potassium meta- bisulfite in mice. Ecotoxicol. Environ. Safety 3, 45 1-453.

    Taylor, S. L., Martin, L. B., and Nordlee, J. A. 1985. Detection of sulfite residues in restaurant salads. 1. Allergy Clin. Immunol. 75, 198.

    Thewlis, B. H., and Wade, P. 1974. An investigation into the fate of sulphites added to hard sweet biscuit doughs. J . Sci. Food Agric. 25, 99-105.

    Thompson, J . R., and Pace, D. M. 1962. The effects of sulphur dioxide upon established cell lines cultivated in vitro. Can. J . Biochem. Physiol. 40, 207-217.

    Til, H. P., Feron, V. J., and DeGroot, A. P. 1972a. The toxicity of sulphite. I. Long-term feeding and multigeneration studies in rats. Food Cosmer. Toxicol. 10, 291-310.

    Til, H. P., Feron, V. J . , DeGroot, A. P., and Van der Waal, P. 1972b. The toxicity of sulphite. 11. Short- and long-term feeding studies in pigs. Food Comer. Toxicol. 10, 463-473.

    Timson, J . 1973. Action of sodium sulphite on the mitosis of human lymphocytes. Chromosomes

    Towns, S. J., and Mellis, C. M. 1984. Role of acetylsalicylic acid and sodium metabisulfite in chronic childhood asthma. Pediatrics 73, 631-637.

    Turchinsky, M. F., Kusova, K. S., and Budowsky, E. 1. 1974. Conversion of non-covalent interac- tions in nucleoproteins into covalent bonds: Bisulfite-induced formation of polynucleotide- protein crosslinks in MS2 bacteriophage virions. FEBS Left. 38, 304-308.

    Twarog, F. J. 1983. Metabisulfite sensitivity in asthma: A review. N . Engl. Region Allergy Proc. 4, 100-103.

    Twarog, F. J . , and Leung, D. Y. M. 1982. Anaphylaxis to a component of isoetharine (sodium bisulfite). J. Am. Med. Assoc. 248, 2030-2031.

    Vahl, J. M., and Converse, J. F. 1980. Ripper procedure for determining sulfur dioxide in wine: A collaborative study. J . Assoc. 08. Anal. Chem. 63, 194-199.

    Valencia, R., Abrahamson, S., and Wagoner, P. 1972. Testing for sodium bisulfite-induced germ line mutations in Drosophila melanogasrer. In Annual Report, Food Research Institute- 1972, Part I. Project Reports, pp. 108-1 11. University of Wisconsin, Madison.

    Vonderschmitt, D. J., Vitols, K. S., Huennekens, F. M., and Scrimgeour, K. G. 1967. Addition of bisulfite to folate and dihydrofolate. Arch. Biochem. Biophys. 122, 488-493.

    Wade, P. 1972. Action of sodium metabisulphite on the properties of hard sweet biscuit dough. J . Sci. Food Agric. 23, 333-336.

    Walbaum, H. 1906. Arch. Hyg. 57, 87. Cited in Cluzan, R., Causeret, J., and Hugot, D. 1965. Potassium metabisulfite: Long-term study of toxicity in the rat. Ann. Biol. Anim. Biochem. Biophys. 5, 267-28 I .

    Walker, R. 1984. Biological and toxicological consequences of reactions between sulphur(1V) oxo- anions and food components. Food Chem. 15, 127-138.

    Walker, R., Mendoza-Garcia, M. A,, Quattrucci, E., and Zerilli, M. 1983a. Metabolism of 3-

    Today 4, 21 1-214.


    deoxy-4-sulphohexosulose, a reaction product of sulphite in foods, by rat and mouse. Food Chem. Toxicol. 21, 291-297.

    Walker, R., Mendoza-Garcia, M. A., loannides, C., and Quattrucci, E. 1983b. Acute toxicity of 3- deoxy-4-sulphohexosulose in rats and mice, and in vitro mutagenicity in the Ames test. Food Chem. Toxicol. 21, 299-303.

    Wang, B. C., Hillman, D. E., Speilholz, N. I., andTurndorf, H. 1984. Chronic neurological deficits and Nescaine-CE-an effect of the anesthetic, 2-chloroprocaine, or the antioxidant, sodium bisulfite? Anesth. Analg. 63, 445-447.

    Wang, R. Y.-H., and Ehrlich, M. 1980. Comparison of bisulfite modification of 5-methylde- oxycytidine and deoxycytidine residues. Fed. Proc.. Fed. Am. Soc. Exp. Eiol. 39, 1882.

    Wang, R. Y.-H., Gehrke, C. W., and Ehrlich, M. 1980. Comparison of bisulfite modification of 5- methyldeoxycytidine and deoxycytidine residues. Nucleic Acids Res. 8, 4777-4790.

    Wedzicha, B. L. 1984. A kinetic model for the sulphite inhibited Maillard reaction. Food Chem. 14,

    Wedzicha, B. L., and Bindra, N. K. 1980. A rapid method for the determination of sulphur dioxide in dehydrated vegetables. J . Sci. Food Agric. 31, 286-288.

    Wedzicha, B. L., and Lamikanra, 0. 1983. Sulfite mediated destruction of p-carotene: The partial characterisation of reaction products. Food Chem. 10, 275-283.

    Wedzicha, B. L., and McWeeny, D. J. 1974a. Non-enzymic browning reactions of ascorbic acid and their inhibition. The production of 3-deoxy-4-sulphopentosulose in mixtures of ascorbic acid, glycine, and bisulphite ion. J. Sci. Food Agric. 25, 577-587.

    Wedzicha, B. L., and McWeeny, D. J. 1974b. Non-enzymic browning reactions of ascorbic acid and their inhibition. The identification of 3-deoxy-4-sulphopentosulose in dehydrated, sulphited cabbage after storage. J. Sci. Food Agric. 25, 589-592.

    Wedzicha, B. L., Lamikanra, O., Herrera, J. C., and Panahi, S . 1984. Recent developments in the understanding of the chemistry of sulphur(1V) oxospecies in dehydrated vegetables. Food Chem. 15, 141-155.

    Weeks, C. 1969. Production of sulfur dioxide-binding compounds and of sulfur dioxide by two Saccharomyces yeasts. Am. J . Enol. Vitic. 20, 32-39.

    Weigand, E. H. 1946. The brined cherry industry. West. Canner Packer 38, 60-63. Weingartner, K. E., Koburger, J. A., Oblinger, J. L., and Knapp, F. W. 1977. Residual bisulfite in

    Werth, G. R. 1982. Inhaled metabisulfite sensitivity. J. Allergy Clin. Immunol. 70, 143. Wesche-Ebeling, P., and Montgomery, M. W. 1983. Extraction and partial characterization of

    strawberry polyphenoloxidase. Annu. IFT Meet., 43rd. Abstr. No. 295. Westley, J. 1980. Rhodanese and the sulfane pool. I n Enzymatic Basis of Detoxification (W. B.

    Jakoby, ed.), Vol. 11, pp. 245-262. Academic Press, New York. Whiting, S. J. , and Draper, H. H. 1980. The role of sulfate in the calciuria of high protein diets in

    adult rats. J. Nutr. 110, 212-222. Wiley, H. W. 1907. Influence of food preservative and artificial colors on digestion and health.

    Sulphurous acid and sulphites. Bur. Chem. U. S. Dept. Agric. Bull. No. 84 (Part Ill). Cited in Cluzan, R., Causeret, J . , and Hugot, D. 1965. Potassium metabisulfite: Long-term study of toxicity in the rat. Ann. Eiol. Anim. Biochern. Eiophys. 5 , 267-281.

    Wilkins, J. W., Jr., Greene, J. A., Jr., and Weller, J. M. 1968. Toxicity of intraperitoneal bisulfite. Clin. Pharmacol. Ther. 9, 328-332.

    Yang, S. F. 1970. Sulfoxide formation from methionine or its sulfide analogs during aerobic oxidation of sulfite. Biochemistry 9, 5008-5014.


    iced Panaeus shrimp. J. Food Pror. 40, 234-235.


    Yang. S. F. 1973. Destruction of tryptophan during the aerobic oxidation of sulfite ions. Environ.

    Yang, S. F. 1984. Reactions of oxidation intermediates of sulphite species with some cellular

    Yang, W. H., Purchase, E. C. R., and Rivington, R. N. 1985. Positive Prausnitz-Kiistner reaction

    Yokoyama, E., Yoder, R. E., and Frank, N. R. 1971. Distribution of 35S in the blood and its

    Res. 6, 395-402.

    components of plants. Food Chem. 15, 113-124.

    in metabisulfite sensitive subjects. J . Allergy Clin. Immunol. 75, 198.

    excretion in the urine of dogs exposed to %02. Arch. Environ. Health 22, 389-395.





    Department of Chemistry, Universdy of Notre Dame, Notre Dame, Indiana 46556

    I. Introduction . . . . . . . . . . . A. Historical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Maillard a nition of Browning Phenomena C. An Overview o ain Research Stream . . . . . . . . . . . . . . . . . . . . . Chemistry of Browning in Model Systems . . . . . . . . . . . . . . . . . . . . . . . . . . A. Preliminary Considerations B. First Steps in the Sequence of Maillard Reactions . . . . . . . . . . . . . . . . C. Some Informative Model Studies D. Empirical Relations between Reactants and Aromas . . . . . . . . . . . . . . Role of Browning in Specific Food Systems . . . . . . . . . . . . . . . . . . . . . . . . . A. Chocolate and Cocoa B. Bread and Other Bake ........................ C. Meat Flavors: Natural and Artificial . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Other Food Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Important Compounds in Browning Flavors . . . . . . . . . . . . . . . . . . . . .

    A. Limited L o s s of Nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Possible Development of Mutagenicity . . . . . . . . . . . . . . . . . . . . . . . . .

    References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .



    IV. Browning, Nutrition, and Health

    V. Trends in Continuing Research . . . . . . . . . . . . . . . . .

    77 77 79 81 84 84 86 89 91 91 93

    100 103 109 113 120 120 122 123 124



    For the individual, especially a civilized person, the selection of foods is determined largely by flavor. Concern for physical well-being, and the nutri-


    Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.


    tional values involved, come far behind flavor as a factor determining human dietary choice.

    The concept of flavor is intrinsically complex, since it comprises the qualita- tive summation of the three distinctly different sensations involved in the normal process of eating: aroma, taste, and touch.' Actually, the additional senses of sight and sound make a substantial contribution to the anticipation and apprecia- tion of flavor: red or brown (not green) meat, green (not yellow) broccoli, and crunchy (not limp) celery and potato chips.

    Fortunately for our ancestors, the appreciation of flavors, either natural ones or those produced by the empirical achievements of culinary skill, did not have to await the belated arrival of chemistry. But the latter does have an important contribution to make, both in understanding what determines all three compo- nents of flavor and in improving them, individually and collectively.

    Rohan (1972) has pointed out ". . . that, whereas much is known about the flavor of chemicals, there is very little known of the basic principles governing the chemistry of flavor." While little is as yet known regarding the interaction of specific molecules in foods with the equally specific taste buds or olfactory receptors, a great deal of tedious and painstaking work has gradually provided the identification of key compounds primarily responsible for well-known fla- vors, e.g., vanilla, clove, and cinnamon.


    (vanilla) eugenol cinnamaldehyde

    (clove) (cinnamon)

    On the other hand, chemistry has demonstrated that some very important aromas, including those of coffee and cocoa, cheeses, meats, baked products, and others, are not produced as a result of the presence of a unique characterizing

    'Aromas (or odors) are the signals perceived by the olfactory organs, tastes are the signals perceived by the lingual organs, and flavors are the simultaneous perceptions of the other two.


    compound. Rather, the aroma is the result of a reproducible blend of a very large number of components in proper balance, no one of which alone would even suggest the familiar aroma.

    It is useful to make a distinction between native flavors and developed flavors. The first, characteristic of fruits and flowers, is determined by the specific molecules produced by the plants as secondary metabolities. The second is illustrated very strikingly by the familiar and unmistakable flavor of maple. The fresh sap of the sugar maple contains no hint of the familiar flavor, nor is it a matter of dilution, for freshly freeze-dried sap, though it tastes sweet, is colorless and lacking in aroma. The boiling process, however, serves not only to remove water, but to induce the chemical reactions that produce the familiar colored and aromatic products. Very different, but equally valid examples are the transforma- tions of green beans into roasted coffee and cocoa and the change of a comparatively bland carcass into roasted turkey.


    These and many other examples of developed flavors are of ancient origin. The commencement of the scientific studies of this general flavor problem proba- bly was initiated by Louis-Camille Maillard (1912a,b). In a misguided attempt to determine the biological synthesis of proteins, he heated concentrated solutions or semidry mixtures of D-glucose with amino acids and observed a gradual darkening, a frothing, and the development of odors somewhat reminiscent of the baking of bread or the roasting of animal or vegetable products. This work attracted sufficient attention to persuade many others to continue the study of what came to be called the Maillard reaction.

    Eventually it became apparent that the work being done and reported on the browning reaction during this early period was distributed among three major categories, as follows:

    1. Controlled studies of model systems designed to provide an experimental basis for the determination of the actual compounds formed and the elu- cidation of the mechanisms by which these compounds arise;

    2. Attempts to inhibit the browning reaction in those systems in which its occurrence renders certain food products inedible; and

    3. Attempts to exploit browning, either by producing products which can be used as flavoring agents or by providing conditions such that the browning reaction takes place during the processing of foods so as to maximize development of desirable flavors.


    Both the volume of the work published and the recognition of its potential importance to the food industry led to the appearance in 1951 of the first general review of the subject (Danehy and Pigman, 1951). Only 2 years later a second review, limited largely to a consideration of model systems and mechanistic interpretations of their reactions, was published (Hodge, 1953). To date more than a dozen general reviews have been published, as have many more reviews directed to special areas in one or another branch of the food industry. Recently two international symposia have been devoted to the Maillard reaction (Ud- devalla, Sweden, and Las Vegas, Nevada) (Eriksson, 1981; Waller and Feather, 1983).

    It is important to emphasize two important characteristics of the browning reactions: First, they can be harmful as well as valuable, and second, high temperatures are not always necessary for the development of browning reaction products. The classic example exemplifying these characteristics is the dried-egg problem of World War 11. In the early 1940s the U.S. Armys Quartermaster Corps found that dehydrated, but not bone-dry eggs became all but inedible when distributed to their field stations, particularly in the South Pacific. It was, howev- er, clearly demonstrated that a slow browning reaction between glucose and nitrogenous constituents of the broken eggs, at ambient temperatures, was re- sponsible; the problem was definitely solved by treating the eggs with glucose oxidase before dehydration to remove the free glucose and so prevent the brown- ing reaction from taking place.

    In 1953, Hodge concluded his review with the statement that . . . the control of browning reactions to produce only wanted flavors and odors is an intriguing possibility. Control of browning to do mans will is the ultimate goal of brown- ing research, but progress toward this goal can be made only as the reaction mechanisms are better understood (p. 941). Not the least part of that under- standing has been achieved by Hodge and his colleagues at the U.S. Department of Agricultures (USDA) Northern Utilization Laboratory in Peoria, Illinois.

    As early as 1947, Barnes and Kaufman published a succinct summary of what had developed in the area of browning reactions since Maillard had published his papers in 1912.

    While the thrust of their summary is negative (browning is responsible for deteriorative changes in food products), Barnes and Kaufman did recognize that the Maillard reaction may also be the contributing factor in the development of many of our characteristic food flavors. Although no evidence was as yet avail- able, there was reason to suspect that the distinctive flavor differences in such items as breakfastfoods, the crust of baked bread, and roasted coffee may be attributed to chemical combinations brought about during the heat treatment operations.



    The many investigations that have been reported during the past 35 years in journal articles, books, and patents can be divided for the most part between three categories:

    1. Detailed, organic chemical studies of the firsr stages of reaction between reducing sugars and various nitrogenous compounds, including amino acids;

    2. Pyrolytic studies of mixtures of reducing sugars and various nitrogenous compounds, including amino acids; and

    3. Identification of the volatile products formed during certain major types of food processing, and the correlation of these compounds, both with flavors and with the hypothetical reactions by which these products and flavors may be formed.

    The first group, organic chemical studies of the first stages of reaction, estab- lished as a first step the formation of N-glycosides by the reaction of reducing sugars with basic nitrogen compounds under very mild conditions and the spon- taneous conversion of the N-glycosides into isomeric forms (the Amadori rear- rangement). These studies were reviewed in great detail by Reynolds (1963) and earlier by Hodge (1953).

    A consensus has long since been reached as to the nature of the main se- quences of chemical reactions involved in the browning of foods, known collec- tively as the Maillard reactions. These chemical sequences are summarized in Figs. 1 and 2.

    During the past decade Hayashi and Namiki have used electron spin resonance (ESR) spectroscopy to study browning reactions in model systems consisting of reducing sugars and alkylamines heated to 98C in aqueous or alcoholic solutions (3 M in each reactant) (see Waller and Feather, 1983, pp. 21-46, for a summary of this work which cites all earlier references). They demonstrated the formation of free radical products at a very early stage and, from the analysis of the spectra, proposed that the radical products are N,N-disubstituted pyrazine cation radicals, assumed to be formed by bimolecular condensation of a two-carbon enaminol. This assumption was supported by the isolation and identification of glyox- alcyclohexylimine. To the extent that this scheme has any relevance in food systems or even in model systems with amino acids (some data on amino acids are included), it should probably be thought of as a concomitant pathway rather than as a revision of the established one.

    The present review is concerned specifically with the role of Maillard reac- tions in developing flavors during the cooking of certain foods. It seems best,





    0-91 ucose =

    R, N / H

    y 2


    c=o I

    HO - C - H I

    H - C - O H I

    H - C - O H I



    1-deoxy-1-N-a1 kyl ami no- - D-fructose- -

    R H

    \ N /

    H / R

    \ JN: C

    H - C - OH - F 1 I H-C-OH b

    H O - C - H I I H O - C - H

    I H - C - O H

    I H -C-OH H - C - O H

    I H2COH

    I H2COH

    N-alkylarnino-D-glucoside a -

    R R /

    N H\, \

    C H H I


    HO - C - H I

    H - C - O H (-H+ I

    H - C -OH I


    + / H

    \ A H y ; H -C- OH

    I H O - C - H

    I H -C- OH

    I H - C - O H

    I H2COH

    FIG. 1. An overview of reactions involved in the nonenzymatic browning of foods. (A) Aldose- amine condensation followed by Amadori rearrangement. (B) Reaction products derived from the aminoglycosides. (C) Oxidative degradation of a-amino acids by reductones: Strecker degradation.


    B 4 H R

    \ N /

    c = o

    HO-C-H I

    I H-C-OH H-$-OH

    1 $-H,O ( 1 , 4 - e l i m i n a t i o n ) 'I H E L \c4 \ R

    I H C \ 4 \

    I C OH


    i +H20

    H O

    I H C

    C OH 1


    \/ +RNH2

    \ p \

    , H O \ c4

    I c = o

    I -H20 1 H \ r"'


    H-C-OH I

    1 ,4 -e l i m i n a t i o n ! 1 H H \ /

    C I1 C

    O=C OH / \

    H H-$-OH I

    Methy l a - d i ca r b o n y l compound

    C I

    c = o I


    F u r t h e r r e a c t i o n s , w i t h and w i t h o u t amines

    3- deoxyhexosone

    FIG. 1B.


    HO R ,R"

    > )c=c ' + o=c HO, R +H20

    R ' ' 'H ,.;".C: , R ' N=C, *' H

    C R

    H' \ NH2 R I' 'C '

    FIG. IC.

    however, to include a brief summary of the studies of model systems, since the results of these studies have provided a basis for interpreting what goes on during the far more complex processes of roasting and baking natural foods.



    The gradual recognition that complex, presumably nonenzymatic transforma- tions which take place in certain foods, responsible for deterioration in some cases and for the development of traditionally valued flavors in others, have a common chemical basis led to the carrying out of experiments in relatively simple model systems.

    It is often difficult to make meaningful comparisons of the results of different


    FIG. 2. ( 1953).

    Hodges view of the pathways by which browning products are formed. From Hodge

    studies, even when the reactants are the same, because of the wide differences in reaction conditions, particularly with regard to concentrations of reactants, tem- peratures, and pH values. In attempting to make such comparisons, it is well to keep in mind three major sets of conditions and the directive influences they have in transforming food products:

    1. Low moisture-high temperatures for relatively short periods of time. These are the conditions which develop flavors and colors that otherwise would not appear at all, e.g., in the roasting of coffee beans, cocoa beans, nuts, grains, and meats. Presumably the earliest stages of the browning sequence (Fig. 1A and


    B) take place rapidly to furnish the intermediates which undergo the final trans- formations (Figs. 1C and 2).

    2. Low moisture-moderate temperatures for relatively long periods of time. These are the conditions that produce deterioration, i.e., off-flavors and un- wanted colors on storage of foods where it is desired to retain the properties as originally packaged. The classic case, referred to previously and studied by Kline and Stewart (1948), is the deterioration of dried eggs.

    3. High moisture (i.e., solutions) over a wide range of temperatures (-20- 110C) and times (hours to weeks). This set of conditions is the least relevant to food conditions, except for beverages, where browning is not usually important, either for good or evil. Yet these are the conditions that have been employed for most of the simple model studies, since they are suitable for the kinds of mea- surements that were made, i.e., development of color, decrease in concentration of reactants, and appearance of products.

    These studies have made it possible to determine the relative activities of specific carbonyl compounds and trivalent nitrogen compounds in various com- binations, usually by measurement of absorption in the visible range and by determination of free amino groups. These studies have also demonstrated un- equivocally the molecular structure of the first intermediates and the conditions under which they form. Further experiments with these isolated intermediates do furnish justification for the reaction sequences shown in Figs. 1B and C and 2.

    The model systems which have been studied largely parallel the food systems, though in simplified form, i.e., mixtures of amino acids and sugars, either in aqueous solutions or in the semisolid state.

    The bulk of the work has been done with glucose, as might have been ex- pected, since it is cheap and readily available and, more importantly, because it is the most widespread and abundant reducing sugar in foodstuffs. Xylose is the runner-up, and fructose is a weak third. It is common knowledge that reduction products (glycerol, sorbitol, etc.) and oxidation products (gluconic acid, tartaric acid, etc.) do not contribute to browning. Indeed, it has been claimed that the latter inhibit browning, probably nonspecifically, by lowering the pH value of the system. However, Nafisi and Markakis (1983) have shown that aspartic and glutamic acids quantitatively inhibit browning in buffered aqueous solutions where the only difference is the presence or absence of the amino acid anions.


    The reaction initiating the sequence between a carbonyl group and a trivalent nitrogen atom is the most thoroughly investigated and best understood of all the reactions. As early as 1963 Reynolds published a review with 140 references,


    limited largely to the studies of reactions of aldoses with amines, the determina- tion of the structures and properties of the first product of reaction (a glycosyla- mine), and the rearrangement of the latter to a more stable ketoseamine.

    Typically, an aldose reacts spontaneously and reversibly with an amine to form an aldosylamine (an N-alkyl glycoside). Much of the earlier work was done with aromatic amines in order to facilitate isolation of the products, but a wide variety of primary and secondary aliphatic amines have also been used. In particular, the esters of amino acids readily yield crystalline glucosylamines. In neutral or alkaline solution these glycosylamines exhibit no reducing properties, as would be expected, since they are the nitrogen analogs of the 0-glycosides. Like the latter, the glycosylamines are readily hydrolyzed by acids to the parent aldose and amine.

    The stability of the glycosylamines is limited. Even in the dry or nearly dry state at 25C, they rearrange spontaneously to 1-amino- 1 -deoxy-2-ketoses, the so-called Amadori rearrangement (Gottschalk, 1952; Hodge, 1955). The impor- tance of these Amadori compounds for Maillard-type browning was demon- strated by Hodge and Rist (1953). They found that D-glucosylpiperidine was slowly transformed during storage at room temperature into a dark, tarry sub- stance from which 1 Wpiperidino- 1 -deoxy-D-fructose (the Amadori compound) could be isolated in yields of up to 50%. But 2-O-methyl-~-glucosylpiperidine, in which the hydroxyl group on C-2 is blocked, remained colorless and stable after storage for 2 years at 25C. Thus, preventing the Amadori rearrangement of an N-substituted glycosylamine inhibited subsequent browning.

    When the carbonyl compound is a ketose (fructose, for example) rather than an aldose, the initially formed fructosylamine (an N-alkyl fructoside) rearranges in an exactly analogous manner to form a 2-alkylamino-2-deoxy-D-glucose (the Heyns rearrangement) (see Fig. 3) (Heyns and Noack, 1962). These Heyns compounds also are precursors of the browning phenomena.

    Subsequent to the formation of the Amadori products (and presumably of the Heyns products), alternative pathways become available for the next stage in the browning sequence, depending upon whether the enolization of the 2-keto com- pound (for example) involves the C-1 atom or the C-3 atom (Fig. 1B).

    The reductones* shown in Fig. 1 B have been isolated and identified (Hodge et

    *Reductone is the trivial name for 3-hydroxy-2-ketopropanal. By extension, reductones are vicinal dicarbonyl compounds capable of some degree of enolization.


    CH20H CH20H

    I + RNH2 - c=o I . # / R HO-C-N. -H,O L \ I I H 1


    C=N: I / R

    HO-C-H I

    H-C-OH t

    HO-C-H I


    HO-C-H I

    H-:-OH I I I I

    H \ /OH CH20H R

    H-C-N C-N ~ -Ht C=N, 1 H I H


    I +/R I H

    I ../R II ../

    I I I HO-C-H HO-C-H HO-C-H

    H-C-OH H-?-OH H-C-OH


    FIG. 3. The Heyns rearrangement: from a ketose and an amine to a 2-aklylamino-2-deoxyaldose.

    al., 1963). Moreover, their further thermal decomposition, both alone and in the presence of amino compounds, produces compounds identified in food systems. It must be remembered that in food systems the thermal degradation of carbohy- drates alone (Fagerson, 1969) takes place simultaneously with Maillard reactions and provides an independent source of conjugated compounds (reductones, furan derivatives, pyran derivatives, cyclopentene derivatives, and unique com- pounds), some of which are further intermediates and others of which are termi- nal products which contribute to aroma, flavor, and color as well (caramels).

    Insofar as Maillard reactions are concerned, the best understood of the reac- tions subsequent to the thermal fission of the Amadori and Heyns compounds are the Strecker reactions (Fig. lC), which produce aldehydes and new nitrogen- containing compounds. The former are major contributors to aromas and the latter are intermediates for producing additional flavorants.

    Further information and discussion on the compounds contributing to the aroma, flavor, and color of browned foods will be presented in Section 111.



    The study carried out by Rooney ef al. (1967) on model systems is outstand- ing; well conceived, comprehensive, and lucidly written, it is worthwhile sum- marizing it in some detail. Two model systems were used: (1) 0.2 M (equimolar) aqueous solutions of a sugar and an amino acid, pH 5.5,95"C, 12 hr; and (2) 100 g wheat starch mixed with 65 ml of a solution 0.02 M in both sugar and amino acid, pH 5.5, rolled and baked at 425F (218C) for 30 min. The carbonyl compounds produced were separated and determined quantitatively by paper chromatographic methods. The results from both systems were mutually comple- mentary. The aldehyde produced is controlled mainly by the amino acid, while the amount of aldehyde is determined mostly by the type of sugar.

    Alanine, isoleucine, leucine, methionine, phenylalanine, and valine produced predominantly those aldehydes that would have been expected from Strecker degradations. In addition, smaller quantities of acetone, formaldehyde, and other carbonyl compounds were also found. Lysine, arginine, histidine, and tryp- tophan caused rapid and intense browning, but did not produce significant quan- tities of specific carbonyl compounds. Glutamic acid and proline caused rela- tively little browning and a small production of carbonyl compounds. Both with regard to color formation and production of carbonyl compounds, xylose was most reactive, maltose was least reactive, and glucose was intermediate. Iso- leucine, leucine, and lysine produced pleasing aromas, while methionine and phenylalanine gave unpleasant aromas.

    Somewhat earlier, Rothe (1960) and Rothe and Voight (1963) conducted a rather similar, though less comprehensive investigation. Generally, both teams are in agreement, though there are some discrepancies in intensities of browning. Rooney et al. found that lysine, arginine, histidine, and tryptophan caused in- tense browning; Rothe and Voight agreed that lysine caused intense browning, but reported that arginine and histidine browned weakly, and tryptophan, scarce- ly at all. Using xylose only, Rothe and Voight recognized an inverse relation between browning and quantity of aldehyde produced and suggested that this might be accounted for if the aldehydes formed via Strecker degradation were consumed in subsequent pigment formation before they could be swept out.

    At a very early stage in the serious study of the Maillard reaction (1950- 1952), Lea and his colleagues made a remarkably comprehensive and quan- titative study of the reaction between glucose and the polar functional groups of casein (Lea and Hannan, 1950a-c; Lea ef al., 1951). Since they focused sharply on the problems of Maillard chemistry, established several points which are still valid today, and influenced strongly other workers in the field, it is worthwhile summarizing their work. Sodium caseinate (69 g) and glucose (31 g) were


    dissolved in water, adjusted to pH 6.3, lyophilized, and stored at 37C and 70% relative humidity. After 5 days, about two-thirds of the free amino groups had reacted and the product was still colorless and water-soluble. The powerful reducing properties of the product ". . . and the further observation that glucose cannot be regenerated from it by treatment with acid, indicates that the prod- uct . . . is not a simple N-glycoside, although such a substance may well be first formed and immediately undergo isomerization by some intramolecular change such as the Amadori rearrangement" (Lea and Hannan, 1950a, p. 528). After 30 days when the product had become brown and poorly soluble over a wide pH range, about 90% of the lysine, 70% of the arginine, 30% of the histidine, 50% of the methionine, and 30% of the tyrosine had reacted. Acid hydrolysis liberated all of the combined methionine, most of the tyrosine, and 70% of the lysine, but none of the arginine or histidine. There was no demonstrable loss of tryptophan or of total acidic or amide groups.

    When the remaining free glucose was removed by dialysis from the casein- glucose system which had been stored 5 days at 37C and 70% relative humidity and the sample relyophilized and stored again at 37C and 70% relative humidi- ty, the complex browned rapidly, at a rate which indicated that ". . . decom- position of carbohydrate attached to the protein amino groups could account for most of the darkening of a casein-glucose mixture at 37"" (loc. cit.). Fully acetylated casein, stored with glucose in the same manner, browned only very slowly at 37"C, but at 60C the acetylated glucose underwent changes in color, solubility, and amino acid content 20 times faster than the casein-glucose system.

    Since a free hydroxyl group on the C-2 atom of an aldose is essential for an Amadori rearrangement, one would expect that a 2-deoxyaldose would not initi- ate or support Maillard browning. Lea and Rhodes (1952) found that whereas galactose reacted with the free amino groups of casein at a rate very similar to that previously observed with glucose, 2-deoxygalactose reacted with the amino groups considerably more slowly. The development of a brown discoloration, however, was very much more rapid with the modified than with the normal sugar.

    Here, then, are two experimental anomalies which have never been explained: (1) Acetylated casein-glucose at 60C browns 20 times faster than casein- glucose; and (2) casein- and 2-deoxygalactose browns much more rapidly than does casein-galactose.

    These three studies, then, early on gave a sound experimental basis for infer- ring the sequence of chemical reactions responsible for the colors and the flavors, both desirable and undesirable, which are formed during browning in relatively low-moisture (-2-40%) food systems.



    In what was perhaps the earliest report to describe a deliberate attempt to produce aromas useful in foods via the Maillard reaction (Kiely et al . , 1960), 7 sugars were heated individually with 20 amino acids in the presence of 15-50% water, at pH of 4.0,5.0,6.0, and 8.0, at 50, 100, and 150C until a golden color had been reached. While details are not given, it was stated that Although a very careful comparison was made of the eight sugars in the reactions, significant differences in the production of bread aromatics between the sugars was not apparent; there were some differences, however, in the rates of reaction.

    In view of the fact that flavor, and to a lesser extent, color, is what Maillard browning is all about, it is remarkable how few model studies give even a hint as to the aroma or flavor of the products obtained. There are at least two reasons for this situation. First, many chemists have been content to study the chemistry per se without regard to practical aspects. Second, while it is relatively easy to present data on color, at least in terms of absorbance at specific wavelengths, it is not so easy to describe aromas. Even when specific molecules of known odor are identified, this in itself gives no true picture of the overall aroma of the complex product.

    We shall have a good deal more to say about flavors in conjunction with specific food systems, and we shall present some correlation of flavors and aromas with classes of organic compounds and specific members thereof.

    In Table I are summarized those observations reported on aroma in studies of the reaction of one specific amino acid with one specific sugar system.


    Many bland or even downright unpleasant-tasting substances are transformed into some of the most desirable flavors and popular foods by roasting. Thus, those foods representing such different tastes and aromas as chocolate, bread, roast beef, coffee, and toasted nuts have in common the fact that they are products of the Maillard browning reaction. The enormous variety in flavor is due almost entirely to the large number of permutations from the interactions of a relatively few primary reactants and to the importance of balance between the components finally present. The reproducibility obtained in these seemingly chaotic and certainly random systems is as remarkable as the sensitive discrimi- nation of the mammalian olfactory-gustatory system.




    Amino acid Aroma or compounds produced


    a- Aminobutyric


    Aspartic acid Cysteine

    Cystine Glutamic acid Glycine







    Acetaldehyde(Rooney et al.. 1967; Rothe, 1960; Rothe and Voight, 19631 Roasted barley (Rothe and Voight, 1963) Caramel (Wiseblatt and Zoumut, 1963) Propionaldehyde (Rothe and Voight, 1963) Walnuts (Rothe and Voight, 1963) Breadlike (Kiely er al.. 1960) Very weak (Lea and Hannan, 1950b) Very weak (Wiseblatt and Zoumut, 1963) Meaty (Kiely et al.. 1960) Thiol, H2S (Wiseblatt and Zoumut, 1963) Overboiled egg (Arroyo and Lillard, 1970) Cooked meat (Led1 and Severin, 1973, 1974) Burned horn (Rothe and Voight, 1963) Meaty (Kiely et al.. 1960) Chicken broth (Wiseblatt and Zoumut, 1963) Caramel (Kiely et al.. 1960) Baked potato (Wiseblatt and Zoumut, 1963) 2.5-Dimethylpyrazine, trimethylpyrazine (Dawes and Edwards, 1966) Breadlike (Kiely et al., 1960) Very weak (Wiseblatt and Zoumut, 1963) Potato (Kiely et al., 1960) Weak (Wiseblatt and Zoumut, 1963) Cookie- or mushroom-like (Dawes and Edwards, 1966) Fruity (Kiely et al. , 1960) 2-Methylbutanal (Rooney et al.. 1967; Rothe, 1960, Rothe and Voight,

    Crust (Wiseblatt and Zoumut, 1963) Furfural (Rothe and Voight, 1963) Pleasant (Rooney et al.. 1967) Breadlike (Kiely et al.. 1960) 3-Methylbutanal (Rooney et al.. 1967; Rothe, 1960; Rothe and Voight,

    Cheesy (Wiseblatt and Zoumut, 1963) Baked potato (Wiseblatt and Zoumut, 1963) Pleasant (Rooney et al., 1967) Furfural (Rothe and Voight, 1963) Roasted barley (Rothe and Voight, 1963) Dark corn syrup (Wiseblatt and Zoumut, 1963) Pleasant (Rooney et al., 1967) Methional (Rooney et al . , 1967; Rothe, 1960; Rothe and Voight, 1963) Baked potato (Wiseblatt and Zoumut, 1963) Furfural (Rothe and Voight, 1963) Unpleasant (Rooney et al., 1967)





    TABLE I (Conrinued)

    Amino acid Aroma or compounds produced


    Serine Threonine Valine

    Boiled potato (Arroyo and Lillard, 1970) Objectionable (Arroyo and Lillard, 1970) Cabbage (Rothe and Voight, 1963; Lindsay and Lau, 1972) Floral, rose (Kiely et 01.. 1960; Rothe and Voight, 1963) Phenylacetaldehyde (Rooney er al . , 1967; Rothe, 1960; Rothe and Voight,

    Strong hyacinth (Wiseblatt and Zoumut, 1963) 2.5-Dimethylpyrazine (Dawes and Edwards, 1966) Unpleasant (Rooney er al . , 1967) Cornlike (Kiely et al., 1960) Crackers, toast (Wiseblatt and Zoumut, 1963) Cracker odor (Hunter er al., 1969) Strongly browned flour (Rothe and Voight, 1963) Weak breadlike (Wiseblatt and Zoumut, 1963) Very weak (Wiseblatt and Zoumut, 1963) Fruity (Kiely et al., 1960) 2-Methylpropanal (Rooney er al.. 1967; Rothe, 1960; Rothe and Voight,

    Yeasty, protein hydrolyzate (Wiseblatt and Zoumut, 1963) Furfural (Rothe and Voight, 1963) Roasted barley (Rothe and Voight, 1963)




    The carbonyl compound used in each case was ( I ) eight different sugars (no significant influence on flavors produced by different amino acids) (Kiely el al., 1960); (2) dihydroxyacetone (Wiseblatt and Zoumut, 1963; Hunter et al., 1969); (3) fructose (Dawes and Edwards, 1966); (4) glucose (Rothe, 1960; Kobayashi and Fujimaki, 1965; Arroyo and Lillard, 1970; Lindsay and Lau, 1972); (5) xylose (Rooney et al . , 1967; Rothe, 1960; Rothe and Voight, 1963; Led1 and Severin, 1973, 1974); and (6) maltose (Rooney et al., 1967).


    One of the world's most popular flavors, chocolate and C O C O ~ , ~ is determined by a physical-chemical composition which starts with the seeds of the plant Theobroma cacao and continues with an empirical process discovered and per- fected by the Aztecs, or by an earlier society from whom the Aztecs received it.

    Two entirely separate stages are essential for the development of this flavor: the fermentation of the beans (seeds) in their mucilaginous pulp enclosure when the pod is opened, and the roasting of the dried, fermented beans. It has long

    Thocolate is the unctuous, extremely bitter low-melting substance (called chocolate liquor in the industry) obtained by the crushing and milling of roasted cacao beans. Cocoa is the free-flowing powder obtained by the partial defatting of chocolate liquor.


    been known that neither aroma nor aroma precursors are present in unfermented cacao beans which, when roasted, develop an odor reminiscent of broad beans (Rohan, 1963). Substantially the only sugar present in unfermented beans is sucrose, but a mixture of fructose and glucose accounts for most of the sugar in fermented beans (Rohan, 1964; Reineccius er al., 1972a). Also, in going from unfermented to fermented beans, the concentration of free amino acids increases between 3- and 10-fold (Rohan, 1964; Rohan and Stewart, 1965).

    The justification for the preceding summary statement is found in the reports of a well-conceived and carefully executed research program carried out by Rohan. Starting with the variables involved in fermentation techniques used on West African farms, he did the following (Rohan, 1958a,b; Holden, 1959):

    1. Determined the amino acid and sugar contents of unfermented and fer-

    2. Prepared aqueous methanolic extracts of both kinds of beans; 3. Showed that roasting of the dehydrated extract of the second (but not the

    first) kind of bean produced the characteristic cocoa aroma; and 4. Determined the changes in sugar and amino acid contents of the dehydrated

    extract brought about by fermenting and roasting (Rohan, 1963, 1965, 1964, 1967; Rohan and Stewart, 1965, 1966a,b, 1967a-c).

    mented beans;



    Amino acid (g)/IOO g dry substance of extract

    Ratio of fermented Amino acid Fermented Unfermented to unfermented

    Leucine Lysine Phenylalanine Threonine Valine Arginine Glycine Alanine Isoleucine Proline Serine Tyrosine Glutamic acid Histidine

    4.75 0.56 3.36 0.84 2.60 0.35 0.35 3.61 1.68 1.97 1.99 1.21 1.71 0.04

    0.45 0.08 0.56 0.14 0.57 0.08 0.09 I .04 0.56 0.72 0.88 0.57 1.02 0.08

    10 7 6 6 5 4 4 3.5 3 3 2 2 I .5 0.5

    a Based on data of Rohan (1964).


    Table I1 presents a compilation of Rohans data for the increase in free amino acids in the extract brought about by fermentation. Subsequently, Rohan and Stewart (1966a,b) presented data graphically for the gradual destruction of total amino acids and sugars during a roast at 182- 183C of dehydrated extracts of fermented beans. Within 30 min, almost half of the amino acids had disappeared, and only 10% of the reducing sugars were left.

    Mohr et af. (1971) started with an aqueous methanolic extract of defatted, ground Ghana cacao beans prepared just as Rohan (1964) had prepared his extract. But Mohr passed the deep brown extract through a polyamide column to adsorb polyphenolic and quinonoid substances before lyophilizing the almost colorless extract. Mohr extended Rohans study by determining both free and peptide-bound amino acids, both before and after roasting. Mohr heated his reaction mixtures only to 121C, which was reached in 8 min, since he found that at that temperature a thin layer of the dried extract began to brown rapidly and to give off a typical cocoa aroma. Mohrs data are presented in Table 111. The amino acids are arranged in the descending order of free amino acids present before roasting. Cumulative data for changes in amino acids and carbohydrates are presented in Table IV.

    Several conclusions can be drawn from these data. First, without exception, free amino acids are much more sensitive to destruction in this system than are the peptide-bound amino acids. This conclusion might have been inferred from Rohans observation (1964) that only the extract from fermented beans gave rise to cocoa aroma upon roasting, but the fermentation produces reducing sugars from sucrose as well as amino acids from polypeptides, so that conceivably the reducing sugars might have produced aroma at the expense of peptides. Mohrs data show that this unlikely hypothesis is not tenable.

    Second, differences in the stability of amino acids under these conditions are not all that great: from 25% for isoleucine to 68.5% for lysine, over a relatively short period of time. In this system the reducing sugars must be the limiting factor, since the glucose and fructose are completely destroyed or removed.

    Third, neither cystine nor cysteine is reported to be present, and the only other sulfur-containing amino acid, methionine, is present at a much lower concentra- tion than any other amino acid. Clearly, as we shall see later, cocoa would probably have a considerably different flavor if cysteine or cystine were present in the fermented beans.

    Although Rohan (1964) had suggested that the operative reaction in the devel- opment of chocolate aroma might be a Strecker degradation of the amino acid fraction, in none of his reports does he give any data on the composition of cocoa volatiles. Bailey et al. (1962) followed the Strecker lead. Using gas chro- matography and mass spectral analysis, they determined that the volatiles from a typical sample of roasted, ground Bahia cocoa contained, in the mole ratio


    UPON ROASTING AT 121C" ~ ~

    Before roasting After roasting Loss upon roasting

    Free Peptide-bound Free Peptide-bound Free Peptide-bound Amino amino acid amino acid amino acid amino acid amino acid amino acid

    acid (mmollkg) (mmollkg) F b (mmollkg) (mmol/kg) Fb (%) (%)

    Leu Ala Phe Glu Ser Val TY r Thr N H 3

    pro Ileu LYS ASP Arg GlY His Met

    102.1 84.3 73.6 63.5 63.3 50.3 41.9 38.0 37.0 33.7 30.0 26.6 24.9 23.5 14.4 7.5 1.9

    108.1 144.2 83.9

    302.4 113.0 158.5 41.5 91.2

    150.0 81.3 84.0 63.0

    302.7 55.5

    139.9 16.7 7.1

    1 . 1 51.5 I .7 53. I I .4 15.2 4.7 26.9 1.7 34.7 3.1 32.3 0.9 16.6 2.4 15.1 4.0 15.5 2.4 21.6 2.8 22.5 2.4 8.0

    12.2 16.0 2.3 13.6 9.7 7.9 2.2 3.0 3.7 0.9

    90.9 130.7 73.1

    301.4 115.8 143.0 43.0 94.5

    150.0 97.9 68.0 36.9

    243.5 49.1

    129.6 10.9 7.3

    I .8 49.7 2.4 37.0 I .9 65.7

    11.2 57.7 3.3 45.2 4.4 35.8 2.6 60.5 6.2 60.3 9.7 58.0 4.5 35.9 3.0 25.0 4.6 68.5

    15.2 35.8 3.6 42.2

    16.4 45.2 3.6 60.0 8.1 52.6

    16.4 9.4

    13.0 0 2 9.8 0 0 0 0

    19.2 41.5 19.6 11.6 7.4

    34.8 0

    a Based on data of Mohr et al. (1971). F, Peptide-bound amino acid/free amino acid.




    Component Before roasting After roasting Decrease

    (mmol/kg) (mrnol/kg) (a)

    Total free amino acids 717 364 49.2 Total peptide-bound amino acids 1868 1789 4.2 Glucose 167 0 100 Fructose 556 14 97.5 sucrose 32 30 6.0 Citric acid 378 367 3 .O

    Dry substance of this extract amounted to -5% of the weight of the shelled cacao beans

    Based on the data of Mohr et al. (1971). (fermented and air dried).

    shown in parentheses, isovaleraldehyde (42.0), isobutyraldehyde ( 15.4). pro- pionaldehyde (13.0), methanol (9. I ) , acetaldehyde (7.0), methyl acetate (6.3), butyraldehyde (3.0), diacetyl (2.8), and at least eight other assorted compounds in lesser amounts, none of them containing nitrogen. The first, second, and fifth most prominent compounds identified could be related to leucine, valine, and alanine as their precursors.

    Obviously, however, while a synthetic mixture corresponding to the above would be fragrant, it would certainly not suggest the aroma of cocoa, based on the work reported by many others. What then is the chemical basis for the aroma and flavor characteristic of cocoa and chocolate products? During the period of 1964-1976, more than a dozen reports addressed themselves to this problem. They ranged from the herculean labors of the Firmenich group, which carried out a classical fractionation of 750 kg of Arriba (Venezuelan) cocoa, which con- firmed the presence of 43 compounds previously reported by others and identi- fied 29 compounds not previously reported (Dietrich et al., 1964), to the power- fully instrumented investigations (high-resolution gas-liquid chromatography, infrared spectroscopy, and mass spectrometry) which, to date, have claimed the identification of more than 300 constituents of cocoa volatiles. The methods of extraction have commonly employed aqueous ethanol (from 70 to 92C v/v) (Dietrich et a f . , 1964), steam distillation (Darsley and Quesnel, 1972), co- distillation with 1,2-propanediol (Flament and Stoll, 1967) or with ethanol (van der Wal et al., 1968, 1971), and supercritical carbon dioxide (Vitzthum et al., 1975). In some cases the extracts were first fractionated into neutral, acidic, and basic fractions (Dietrich et al., 1964; Rizzi, 1967; van F'raag et al. , 1968). In other cases an acidic extraction was employed in order to give only a basic fraction (Stoll et al., 1967a; Reineccius et al . , 1972).


    The cumulative result of all this effort is the reasonably sure establishment that at least 350 organic molecules are present in cocoa volatiles in at least detectable amounts and that a goodly, though indeterminable number of them are final products of the Maillard reactions. Few, if any, of these molecules would be odorless. But how they combine in intensity and specificity to produce the instantly recognizable aroma and flavor of chocolate is still unknown.

    In 1964, Dietrich et al. suggested that their failure to reconstitute the aroma of chocolate from the 72 components known to them at that time could be attributed to the fact that other components had escaped them. Seven years later van der Wal et al. (1971), with semiquantitative data on 181 compounds, made an attempt to duplicate the aroma concentrate using the gas chromatogram as a guide to estimate the proportions and amounts of the constituents involved. Although this synthetic mixture was reminiscent of cocoa, it lacked the pro- nounced aroma of the extract and was easily distinguishable from it. They concluded from this that probably important aroma components still await detection.

    In view of the fact that holding large numbers of organic compounds of diverse functionality in a homogenous system at room temperature, much less at about 100C, is conducive to chemical reaction, we should consider at least two other reasons for the lack of success in attempts to reconstitute the aroma of cocoa. First, some of the compounds actually contributing to the aroma of cocoa may have decomposed and are no longer present in the extract analyzed. Second, some of the compounds identified in the extract may be artifacts, not actually present in the cocoa, but synthesized during the extraction and working-up process.

    Under the circumstances, we shall not list all of the approximately 350 com- pounds which have been claimed to be present in cocoa volatiles. They are listed in overlapping tables in the following papers: Dietrich et al. (1964); Marion et al. (1967); Flament et al. (1967); van der Wal et al. (1968); Vitzthum et al. (1975); Stoll et al. (1967b); Reineccius ef al. (1972); Rizzi (1967). In Table V are listed the classes of compounds, minimum number for each class, identified in the analysis of cocoa volatiles.

    In almost all cases the specification of a compound is purely qualitative, with no information whatsoever as to what fraction of the cocoa is accounted for by that compound. An important and notable exception is the report of Flament et al. (1967), who recorded a large number of substances as "%" of fraction A, which in itself is a 10.0-g concentrate from 204 kg of ground, roasted cacao beans (chocolate liquor). In Table VI are listed those compounds making up the largest part, but not necessarily the most important part, of fraction A, and the percentage of the original chocolate liquor for which they account. These data support the commonly held opinion that carbonyl compounds and pyrazines are




    Hydrocarbons Aliphatic (8) Terpene (6) Aromatic (17)

    Aliphatic (12) T e v m ( 5 ) Aromatic (6)


    Phenols (5) Disulfides (5) Furans ( 15) Pyrazines (34) Oxazoles (4)

    Alcohols Aldehydes Aliphatic (4) Aliphatic (10) Terpene (9) Terpene (2) Aromatic ( 5 ) Aromatic (2)

    Aliphatic (15) Aliphatic (30) Aromatic (14) Aromatic (7)

    Carboxylic acids Esters

    Esters and acetals (12) Sulfides (4) Trisulfides (2) Other sulfur compounds (6) Other 0-heterocycles (6) Pyrroles (9) Nitriles (4) Pyridines (9)

    Number indicates the minimum number of compounds identified.

    major contributors to the aroma of cocoa and show that they are effective in the range of parts per 10 million.

    Flavor is aroma plus taste, and it is important to remember that cocoa itself is quite bitter. Three diketopiperazines, cyclo(-Asn-Pro-), cycle(-Ala-Gly), and cycld-Asn-me-), have been isolated from roasted, but not unroasted, cacao


    Percentage Percentage Compound of A, of a chocolate liquorb

    Trimethylpyrazim Tetrameth y lpyrazine 2.5- + 2,bDimethylpyrazine 2,5-Dimethyl-3-ethylpyrazi~ Acetophemme 2-Methylbutanal 3-Methylbutanal 2-Phenylethyl acetate 2,5-Dimethyl-3-isoamylpyrazi~

    Ethyl caprate 3-Hydroxy-2-butanoW

    21.0 20.3 7.3 5.6 3.7 3.5 2.5 2. I I .4 I .4 1.1

    10.5 (10-5)

    3.65 (10 -5 ) 2.8 (10 -5 ) 1.85 (10-5) 1.75 (10-5) 1.25 (10-5) 1.05 (10-5) 7.0 (10-9 7.0 (10-9 5.5 (10-9


    0 Based on the data of R a n t et 01. (1967). b Percentage of Al X (10/204.oOO) = Percentage of A1 (0.5)(10-5).


    beans. Carefully planned and carried out experiments indicate that the bitter taste of cocoa is due to the simultaneous presence of 30-50 ppm of a diketopiperazine and 100 ppm of theobromine, the characteristic xanthine of cacao beans. Di- ketopiperazines containing the phenylalanyl residue resemble the bitter principle of cocoa most closely (Pickenhagen et al., 1975).

    Mohr et al. (1976) have isolated several peptides from fermented cacao beans, and they have found that when these peptides, along with the amino acids prominent in fermented cacao beans, are pyrolyzed with fructose, the resulting aroma is much closer to that of roasted cacao beans than when peptides or amino acids alone are pyrolyzed with fructose.

    Many of the major classes of compounds and even specific compounds found in cocoa volatiles are also found in the volatiles of other browned food products. We shall deal with these flavor compounds later.


    Browning plays an essential role in the development of flavor in bread, one of mans most important foods. As early as 1910 the occurrence of maltol and isomaltol in bread as natural flavorants was reported. Not until more than 40 years later did research on the flavor of bread enter a sustained phase, which, lasted about 15 years (1953-1969). Some of the important model studies were carried out by investigators primarily interested in bread baking and allied prob- lems (Rooney et al . , 1967; Rothe, 1960; Kiely et al . , 1960; Wiseblatt and Zoumut, 1963; Johnson and Miller, 1961).

    Baker et al. (1953) demonstrated that both fermentation and the formation of a brown crust are essential for satisfactory flavor. Bertram (1953) addressed him- self to an immediate practical problem: Why did the flour from a certain strain of low-protein Dutch wheat, upon baking, give a crust with a gray color? He showed that the addition of either dried egg white or wheat gluten to the flour gave crusts with normal brown color. The results prompted him to carry out some model experiments in which mixtures of wheat starch, different sugars, and dried egg white or amino acids were heated, using bicarbonate rather than yeast as a leavening agent. The results are the first unequivocal demonstration of the importance of the Maillard reaction in crust color.

    Aldehyde formation received a great deal of attention in these studies of bread aroma (Rothe, 1960; Rothe and Thomas, 1959, 1963; Wiseblatt and Kohn, 1960; Wiseblatt, 1960a), since they are prominent and easy to identify semiquan- titatively by trapping them as 2,4-dinitrophenylhydrazones, followed by chro- matography. There was an early consensus that while they are formed in the crust by Strecker degradation of amino acids they are withdrawn into the crumb upon cooling and storage of the bread. Thomas and Rothe (1957) emphasized the


    importance of furfural, which is not formed by a Strecker degradation, and showed that addition of 0.7% of xylose to the flour increases the total aldehyde content of the bread volatiles 4-fold and the furfural content 10-fold; arabinose, sorbose, fructose, and glucose were less effective.

    But it was soon recognized that aldehydes, other carbonyls, and ethanol are not the whole story in terms of flavor. Wiseblatt and Kohn (1960) found that neither the actual distillate containing these compounds nor several synthetic blends of them have proved of any value in enhancing the palatability of a bland chemically leavened bread. Gradually the authenticated list of compounds found in preferments, dough, oven vapors, and bread (Wiseblatt, 1960b, 1961; Wick et al., 1964; Johnson et al., 1966) has grown to include more than 70, including alcohols, aldehydes, ketones, carboxylic acids, esters, and a very few mis- cellaneous compounds: methyl mercaptan, hydrogen sulfide, maltol, and iso- maltol. Is the small number of miscellaneous compounds realistic or have signifi- cant compounds been missed by inadequate analytical methods? On the basis of the evidence available, the aroma constituents of bread appear to differ qualita- tively from those of cocoa, roasted nuts, and cooked meat most strikingly in the complete absence of nitrogenous constituents, particularly pyrazines.

    However, Mulders et al. (1972a,b; Mulders and Dhont, 1972; Mulders, 1973a) made a determined gas chromatographic study of the constituents in the vapor above fresh white bread and their odor values. They found 52 compounds, of which 42 had not been reported previously. Several of them were pyrazines, lactones, and derivatives of furan or pyrrole. While great differences existed in the quantities of components between individual loaves, although the baking protocol had been rigorously standardized, the odor was quite similar for all. An

    - aqueous synthetic mixture, prepared in such a way that the chromatogram of its vapor was identical to the average chromatogram of bread vapors, had an odor which scarcely resembled that of bread; it was rather doughlike. Therefore, the components detected in a normal vapor sample cannot account for the charac- teristic odor of fresh white bread. The odor of the synthetic mixture changed from doughlike to breadlike upon addition of a particular gas chromatographic fraction of a white bread extract.

    The work of Hodge and Moser (1961) confirms the contribution of maltol and isomaltol to bread aroma 50 years after the demonstration of their presence. It has been shown repeatedly that L-proline is particularly important as a precursor of bread aroma constituents (Wick et al . , 1964; Morimoto and Johnson, 1966).

    There have been a number of investigations of the effect on bread aroma of adding sugars or amino acids to a dough before baking (Thomas and Rothe, 1957; Linko et al., 1962, 1963). The most recent report (Salem et al . , 1967) is the most comprehensive and summarizes the situation very well. Systematically, 0.02 mol of an amino acid and 0.02 mol of either glucose or xylose were added


    to a dough containing 700 g flour, processed and baked according to a straight dough procedure. Reflectance data on the top and bottom crusts showed that addition of amino acids increased the intensity of the crust color in all cases. Methionine and arginine, with added glucose, produced the darkest color. Pro- line had little effect on the crust color. Generally, xylose-amino acid mixtures gave darker crusts than glucose-amino acid mixtures, probably because xylose is nonfermentable and more reactive in browning.

    Analytical data indicate that the composition of carbonyl compounds in crust and crumb vary with the amino acid added, with glucose as the sugar added, with the effect much more pronounced in the crust than in the crumb. The contents of furfural and 5-hydroxymethylfurfural (HMF) were less than the control in all cases. This is not surprising, since these compounds serve as intermediates and undergo further condensation with free amino groups, which are more abundant with the deliberate addition of amino acids. Alanine and valine increased the yield of acetone as well as the expected Strecker aldehydes. Addition of leucine and isoleucine doubled the presence of the Strecker aldehydes. Lysine increased the concentration of all carbonyl compounds three- to fourfold. Histidine in- creased both acetone and aldehyde three- to fourfold and isobutyraldehyde and isovaleraldehyde twofold. Interestingly enough, proline, which everyone agrees has a definite positive effect on the aroma, had a modest increasing effect only on acetone; for other carbonyl compounds the content was either the same or lower than in the control.

    Parallel experiments in which xylose rather than glucose was the sugar gave generally comparable results. But xylose definitely gave lower levels of al- dehydes than did glucose. The most plausible explanation of this anomaly is that xylose is so reactive that most of the carbonyl compounds formed in the early stages of baking were volatilized or reacted quickly with free amino groups to give melanoidins, consistent with the dark color of the crusts.

    Although aldehydes are produced during fermentation of the dough, they are volatilized during the later stages of fermentation and the early stages of baking. Addition of isovaleraldehyde to the dough did not increase the isovaleraldehyde content of the crust or crumb. But addition of leucine to the dough produced a two- to threefold increase in the isovaleraldehyde content of the crust. While the sugar added (xylose or glucose) had no effect on the aroma, the addition of the following amino acids did produce these significant effects: leucine and iso- leucine, cheeselike; phenylalanine, floral; methionine, obnoxious; other amino acids, subtle indescribable aromas. From these results it appears possible to alter bread aroma and flavor by the addition of amino acids to bread formulas. The possibilities for enhancing the aroma of toasted bread by amino acid addition appear promising.



    Meat is the muscular tissue of common domestic animals, which are consid- ered and used as food for human consumption. Meat is at least a three-phase system consisting of (1) a hydrophilic, but water-insoluble fibrous protein net- work, (2) a hydrophobic fat deposit held together by membranes, and (3) an aqueous solution containing many soluble low-molecular-weight compounds.

    Raw meat has very little aroma at room temperature, although it is usually possible to distinguish beef, pork, lamb, and chicken by sniffing. The taste of raw meat, which is not at all palatable to most human beings, can be described as somewhat salty and metallic. Raw meat must be cooked in some fashion in order to develop any organoleptically acceptable odor and flavor. Clearly, then, the unheated tissues must contain precursors which undergo thermally induced chemical reactions; these reactions produce both volatile compounds with desir- able aromas and nonvolatile compounds which influence the taste; the combina- tion of these two categories determines the flavor. As we proceed, striking analogies between the development of flavors in the cooking of meat and in the two food systems we have already discussed (cocoa and baked cereal products) will become apparent.

    I . Precursors of Flavor in Meat

    In a series of investigations (Hornstein and Crowe, 1960; Wood, 1961; Macy et al., 1964a,b; Wasserman and Gray, 1965; Landmann and Batzer, 1966; Zaika et al., 1968; Wasserman, 1972; Wasserman and Spinelli, 1970; Jarboe and Malbrouk, 1974), it was definitely established that the extraction of minced lean meat (beef, pork, and lamb) with cold water gave solutions that contained salt, lactic acid, glycoproteins, inosinic acid, taurine, glutamine, asparagine, glucose, and some amino acids. Aging, particularly in the case of beef, is important in the development of the various flavor precursors (Wasserman, 1972). Thus, glyco- gen undergoes glycolysis to lactic acid almost completely within 24 hr after slaughter. Partial autolysis of the proteins and nucleic acids gives an assortment of peptides and amino acids from the former, and a mixture of inosinic acid and its fragments (inosine, hypoxanthine, ribose-5-phosphate, and ribose) from the latter.

    When the filtered aqueous extract of ground beef is heated, a sequence of aromas is developed, beginning with a faint, bloodlike aroma in the barely warm solution, passing through a phase in the boiling solution which gives off the aroma of boiled beef, and terminating in the hot, dried, brown residue (100- 150C) with an aroma resembling broiled steak. When this original filtered


    aqueous extract was subjected to dialysis against water in cellulosic sausage casings and the low-molecular-weight fraction that passed through the membrane was lyophilized to a white powder and then pyrolyzed, it underwent the typical Maillard browning to produce a strong aroma of broiled meat. This aroma was substantially the same, whether the original lean meat was beef or pork. This may result from the fact that the amino acid contents of beef, pork, and lamb are semiquantitatively rather similar.

    It appears, then, that there is a general meaty aroma common to beef, pork, and lamb (and probably poultry), attributable to the pyrolysis of the mixture of low-molecular-weight nitrogenous and carbonyl compounds extracted from the lean meat by cold water. But the aromas of roast beef, roast pork, roast lamb, and roast chicken are unmistakably different. The chemical compositions of the muscular fat deposits of these animals differ appreciably, and it is to these lipid components that we must look to account for the specific flavor differences. Heating the carefully separated fat alone does not give a meaty aroma at all, much less an animal-specific one. It is the subsequent reactions of pyrolysis products of nonlipid and lipid components that give the characteristic aromas and flavors of roasted meats (Wasserman and Spinelli, 1972).

    Since it is precisely at the surface of roasting meat that water concentrations are lowest and temperatures are highest, it is at the meat surface that the flavor- and color-generating activities during roasting are most prominent. This situation is analogous to the formation of crust and aroma in bread and other baked cereal products. The same facts also account for the significant difference between the flavor of roasted and boiled meats.

    2. Compounds Associated with the Flavors and Aromas of Cooked Meat

    Paralleling the studies of the volatile products of roasted cacao beans and of baked cereal products and using the same techniques, a great deal of effort has gone into the determination of the compounds present in the volatile fractions of cooked meat. Most of these have been concerned only with beef, either roasted or boiled, but chicken has also received appreciable attention (Wilson and Katz, 1972). Several lists of compounds isolated from the volatiles of cooked beef have been published (Hen and Chang, 1970; MacLeod and Coppock, 1976; Chang and Peterson, 1977), both cumulative and newly isolated ones. The totals for chicken (as of 1972) and for beef (as of 1977) are more than 200 each. It must be emphasized again that these are qualitative identifications, not quantitative accountings.

    These cumulative tables for cooked meat volatiles are very difficult to dis- tinguish from those published somewhat earlier for cocoa volatiles. Indeed, the larger cumulative tables (Wilson and Katz, 1972; MacLeod and Coppock, 1976)


    resemble somewhat abridged versions of the Aldrich and Eastman Kodak cata- logs of organic chemicals. Meaningful comparisons are hindered by two quite different facts. First, there is usually no hint as to the fraction of the meat or even the fraction of the volatiles that is comprised by a given compound. Second, we are probably getting a great deal of noise with the signals; i.e., there must be many compounds which. even though they have odors, would not be missed if they were absent. We simply cannot believe that more than 200 compounds are required to produce any one of the distinctive roasted food aromas.

    But the human nose has no difficulty in distinguishing chocolate from roast beef, and the flavor chemist is trying to catch up with this degree of discrimina- tion.

    Chang and Peterson (1977) have suggested that lactones, acyclic sulfur com- pounds, nonaromatic heterocyclic compounds containing ring S, N, or 0 atoms, and aromatic heterocyclic compounds containing ring S, N, or 0 atoms may be important contributors to meat flavor, even though none of them alone tastes anything like cooked meat. Wilson er al. (1973) identified 46 sulfur-containing compounds from the volatiles of lean beef pressure-cooked with water at 163 and 182C. Most of these compounds were thiophene or thiazole derivatives, but acyclic thiols, methyl sulfide, and four disulfides were also present. In addition, as we shall see later, sulfur compounds (especially cysteine) play a key role in manufacturing artificial meat flavors. Tonsbeek er al. (1968, 1969) have isolated 4-hydroxy-5-methyl-2,3-dihydrofuran-3-one and 4-hydroxy-2,5-dimethyl-2,3- dihydrofuran-3-one, particularly pungent compounds, from cooked beef. Pyrazines are a particularly important class of flavor compounds, but it was not until 1971 that their presence in beef volatiles was reported (Flament and Ohloff, 1971). and by 1973 a total of 33 had been identified (Mussinan er al., 1973). Recently, Flament er al. (1977) identified several pyrrolo[l ,2-a]pyrazines. Sev- eral other papers have contributed additional data on compounds isolated from volatiles of cooked beef (Watanabe and Sato, 1972; Brinkman er al., 1972; Schutte and Koenders, 1972; Shibamoto, 1980a; Hartman er al., 1983; Galt and MacLeod, 1984).

    Despite the large amount of qualitative if not quantitative data on the chemical composition of the volatiles from cooked meat, no one has yet claimed anything like a duplication of a meat aroma by the combination of the pure chemicals identified in meat aromas. Once more, the parallel with cocoa and baked prod- ucts is striking. Just one example points up the elusive relationship between chemical compounds and food flavors. Hydrogen sulfide has the odor which does characterize rotten eggs, yet it appears to be a necessary component of meat aromas. Its odor threshold is 10 ppb, but its concentration in freshly cooked chicken is 20 to 100 times greater. It is generally agreed that the aroma of a food is the sensed perception of an extremely complex interaction of many compo-


    nents, but one reads between the lines the disappointment of some who report new compounds and note that they do not have a meatlike aroma. Chang and Peterson (1977) suggest the justifiable fear that some components may have been decomposed or missed, but their hope is less justifiable that a unique component may still be found which alone or in combination will have a characteristic beef aroma.

    3. Artificial Meat Flavors

    In the light of what has just been presented regarding the chemical origin of the natural flavors of cooked meat, it is not surprising that the heating of semidry mixtures of hydrolyzed vegetable proteins (HVPs) with reducing sugars gives rise to an aroma and flavor somewhat similar to that of cooked meat. HVPs are produced on an industrial scale by hydrolyzing soy protein, wheat gluten, or corn gluten in hot aqueous hydrochloric acid, neutralizing the excess acid with sodium hydroxide, and evaporating to dryness, which yields a mixture of amino acids and sodium chloride (HVPs).

    The shelf item packets of soup mixes and gravy mixes found in all groceries are practical examples of Maillard technology as testified by the food ingredient disclosures on their labels: HVPs, usually cysteine or cystine, glucose, some- times arabinose, inosinates or guanylates, and less important adjuncts.

    The key involvement of organic sulfur compounds in development of meatlike flavors was announced simultaneously in 1960 by several investigators. In what was the earliest paper to describe deliberate attempts to produce aromas useful in foods via Maillard reactions, Kiely et al. (1960) noted that both cysteine and cystine gave meaty odors when heated with reducing sugars. May et al. received several equivalent patents (May and Akroyd, 1959a,b; May and Morton, 1960; May, 1960; Morton et al., 1960) in which they claimed that heating cysteine or cystine with furan or substituted furans, pentoses, or glyceraldehyde gave a meatlike flavor.

    Hsieh et al. (1980a,b) have experimented with the development of a synthetic meat flavor mixture by using surface response methodology.

    It is precisely to the production of meatlike flavors that the great majority of patents based on the Maillard reaction have been directed. Most of the latter indicate cysteine or cystine as the essential sulfur-containing compound. Other patents claim alternative sources for sulfur, e.g., derivatives of mercaptoacetal- dehyde (Broderick and Linteris, 1960), mercaptoalkyl amines (Ohwa, 1972), S- acetylmercaptosuccinic acid (Mosher, 1973), 2,2- bis(thieny1)tetrasulfide (Katz er al., 1972), a sulfide (Heyland and Cerise, 1979), and hydrogen sulfide (heated with aqueous xylose without any amino acid (Gunther, 1972).

    Several patents (Bidmead et al . , 1968; Giacino, 1968) claim the contribution


    to meatlike flavors made by thiamine when it is present in the standard pyrolytic mixture. Arnold et al. (1969) have reported on the volatile flavor compounds produced by the thermal degradation of thiamine alone. It is generally agreed that the presence of methionine, the other sulfur-containing amino acid in the flavor- developing mixture, produces negative and/or undesirable results.

    Two patents claim that the addition of a ypyrone to meat itself before cooking prevents the development of warmed-over flavors (Sato and Hegarty, 1974, 1976).

    Since the great majority of the patents dealing with the application of Maillard technology to the production of artificial flavors are concerned specifically with meatlike flavors, it is appropriate here to comment on the significance of patents covering reaction flavors, as they are known in the trade.

    During the past 30 years several hundred patents have been granted worldwide for processes and products based on nonenzymatic browning technology. But Chemical Abstracts has not abstracted many more than 100 of them, since they abstract only the first issued of several equivalent patents and list the later ones in a patent concordance.

    There appear to be about 45 standard patents, i.e., patents which specify mixing one or more amino acids with one or more carbonyl compounds and heating, with some or all of the operating conditions given, i.e., temperature, time, water content, pH, and sometimes additives. Although slight changes in initial composition and reaction conditions produce appreciable changes in the flavor and aroma of the reaction products, not one of these patents gives a rigid, controlled specification. The wide ranges of operating conditions and the numer- ous alternatives offered produce such a complete overlap between these patents that not even an expert chemist and a wily lawyer could distinguish one from the other.

    These patents surely have very little value, either from the standpoint of the patent holders or from the standpoint of those who might hope to leam by studying them. All the standard patents say substantially the same thing, and they contain little, if anything, that was not fully disclosed in earlier model studies. It is very doubtful if any one of them could be upheld in court in view of the prior published art. Nor is it likely that a holder of one of these patents could sue successfully for infringement, for two quite different reasons. First, the extreme complexity of the composition of the reaction products would make it impossible to determine by examination how they were made. Second, in view of the redundancy of the patents, it would be overwhelmingly difficult to determine whose patent was being infringed.

    In order to get reproducible results, it is essential to exercise the most precise control at every stage of the process. The basis for this is not provided by any of the patents that characteristically give broad ranges. Maillard reaction products


    arc being manufactured commercially today by detailed proprietary processes that are not described by any patent.

    The reactions of sugars with amines and ammonia to form glycosylamines and Schiff s bases (Ellis and Honeyman, 1959) and to form nitrogen-containing heterocyclic compounds (Grimmet, 1965; Kort, 1970) have been known for more than a century. Recently, Shibamoto and Bernhard (1976, 1977a) wrote that a systematic investigation of reaction parameters for controlling the com- position of pyrazine products and maximizing the yields in the sugar-amine model system could be a key element to understanding the mechanism of pyrazine formation and consequently the characteristics of smoky or roasted flavors of foods. They did reinvestigate the glucose-ammonia-water system. Holding glucose at molar concentration, they systematically varied the con- centration of ammonia from 0.1 to 15 M, the temperature from -5 to 160C (mostly 100"C), and the reaction time from 15 min to 30 days (mostly 2 hr). At lOO"C, increasing the concentration of ammonia increased pyrazine formation up to 8 M NH,, beyond which the pyrazine level remained approximately constant (at - I % , based on glucose). The distribution pattern of the pyrazines was independent of reaction conditions. The principal products, in decreasing order, were 2-methyl-, 2,6-dimethyl-, 2,5-dimethyl-, unsubstituted, 2,3-dimethyl-, and trimethylpyrazine.

    Based on this systematic study of glucose, Shibamoto and Bernhard (1977b) investigated the heating at 100C for 2 hr of molar solutions of mannose, galac- tose, rhamnose, fructose, 2-deoxyglucose, xylose, arabinose, glyceraldehyde, dihydroxyacetone, sorbitol, and glycerol in 8 M aqueous ammonia. Mannose, galactose, and fructose gave -1% total pyrazines (based on sugar), as had glucose. Both pentoses gave slightly higher yields (- 1.2%), but rhamnose gave a surprising 12.5%. 2-Deoxyglucose gave only a 0.5% yield, probably by reason of the blocking of the Amadori rearrangement early in the reaction sequence. Glyceraldehyde gave 0.6% and dihydroxyacetone a 1.2% yield. Sorbitol and glycerol, as might have been expected, gave no pyrazines. It was explicitly stated that no imidazoles or piperazines were detected.

    Recently, Shibamoto et al. (1979) have given a sobering lesson on the impor- tance of experimental details in determining the outcome of an investigation. In the earlier studies, the cooled solutions had been extracted with four 50-ml portions of H,CCI,. In this study, however, continuous extraction of a heated glucose-ammonia-water system with H,CCI, for 16 hr gave an extract from which 54 compounds were isolated and determined quantitatively. In addition to the pyrazines already reported, pyrroles and imidazoles, including 2-meth- ylimidazole, which makes up 72.5% of the total area of the chromatographic peaks, were found.


    Several studies on model systems have been focused directly on the production of meatlike flavors. Two of these have reported the 41 sulfur-containing com- pounds and 27 non-sulfur-containing compounds identified when cysteine and xylose are heated together (Ledl and Severin, 1974; Ledl et al., 1973). Another patent reports not only on heating cysteine with xylose, but on heating of a cysteine-xylose-HVP system as well and lists the 24 sulfur-containing com- pounds identified in the reaction mixture (Mussinan and Katz, 1973).

    Shibamoto and Russell ( 1976) heated an aqueous glucose-ammonia-hydro- gen sulfide solution at 100C for 2 hr. Of the 34 major components identified, 2- methylthiophene accounted for 24.9% of the area of the chromatographic peaks; ethyl sulfide, thiophene, furfural, and 2-acetylfuran each accounted for 10-1 1%; methyl sulfide and 2,5dimethylthiophene, -7% each. The reaction mixture as a whole was deemed by sensory panel evaluation to have a cooked beef odor.

    Once more, although the distributions are expressed quantitatively, there is no information on the yields of those interesting compounds based on glucose, ammonia, and hydrogen sulfide. However, we noted earlier that in their studies of the systems, carbohydrates-ammonia, Shibamoto and Russell found that the amounts of total pyrazines produced, based on the sugars, were in the range of 1-2%.

    Wilsons review (1975) of thermally produced imitation meat flavors, though more than 10 years old, is still well worthwhile consulting. Shibamoto (1980a) lists 161 heterocyclic compounds alone which have been found in cooked meats.


    1. CofSee

    The great importance of coffee has prompted a large amount of research and development involving all aspects of coffee aroma and flavor, including the determination of the aroma constituents and the formulation of concentrated coffee flavors. When Stoll ef af. (1967b; Goldman ef al., 1967) published the results of their monumental study in 1967, they identified 240 volatile constitu- ents, 174 of them for the first time, but they cited almost 150 previously pub- lished reports.

    The roasting of the green inedible coffee beans to produce a fragrant brown product invites comparison with cocoa, but the differences are more prominent than the similarities. The cacao beans are imbedded in a mucilaginous pulp and undergo a spontaneous fermentation as soon as the pod is open. The coffee bean is really a berry, and fermentation plays no significant role in its processing. While chocolate and the cocoa derived from the cacao bean are highly nutritious


    product^,^ the solubles extracted in a cup of coffee, while very flavorful, are devoid of nutritive value. Both cacao beans and green coffee beans contain relatively small amounts of reducing sugars, but appreciable amounts of sucrose which have quite different fates in the two different beans. During the fermenta- tion that follows the opening of the cacao pods (see Section III,A), the rise in temperature kills the seeds so that the invertase within the seeds transforms the sucrose almost completely into glucose and fructose. Feldman et al. (1 969) noted that during the roasting of coffee, the sucrose is quickly pyrolyzed, falling from 4.6% (dry basis) for Colombian green beans or 5.5% for Santos to 0.2-0.3% for a medium roast and less than 0.1% for a dark roast. Feldman also states that the major water-soluble polysaccharide of green coffee beans is arabinogalactan, and that arabinose practically disappears during the roasting of the water-soluble fraction. It may well be that the arabinose residues of this polymer are responsi- ble for Maillard reactions that generate aroma constituents during the roasting of coffee, for evidence from other systems indicates clearly that sucrose is non- contributory (cf. the cocoa and baking systems as well as the model studies).

    Despite the probability that Maillard reactions do play a role in coffee flavor, they have attracted remarkably little attention from the investigators of coffee flavor (Pokorny et al., 1974, 1975). For this reason we must necessarily devote little space to coffee in this report. But much effort is doubtless continuing in the field of coffee research. Efforts are being made to improve extracts and to improve coffee flavors, and new compounds are still being identified in the volatile fraction. Kung (1974) has isolated 3-hydroxy-3-pentene-2-one, the sta- ble, enol form of 2,3-pentanedione, which has a buttery, caramel aroma. It has been known for a long time, of course, that phenolics are important constituents of coffee, e.g., up to 7.5% of chlorogenic acid in green beans, almost half of which survives in roasted coffee and is extracted into the brew.

    0 II

    HO C, H

    HO -f-Jcec < :,o 0 OH OH d HO

    Chlorogenic acid - 4Shelled, dried cacao beans contain about 48% fat, 31% carbohydrate, 10% protein, 3% ash, and

    8% moisture. Both chocolate, which contains all the fat, and cocoa, whose fat content is reduced to 10-18% fat, are totally consumed in food products, while most of the coffee, the grounds, is discarded after extraction.


    There are hundreds of patents having to do with coffee products, but we have found only two in the past 18 years which claim the production of an artificial coffee flavor based on Maillard technology, and both of them completely lack credibility.

    But coffee is big business and is one of the major food crops. The world harvest in 1979-1980 has been reported by the USDAs Foreign Agriculture Service to have been 7.95 X lo6 bags of 60 kg each, equal to 5.3 X 106 tons. While it is no more likely that artificial coffee could make a dent in the coffee market than artificial chocolate has been able to capture even a small fraction of the chocolate market, nevertheless artificial coffee flavors of quality, if they became available, would be useful and marketable items. For this reason the major flavor houses will doubtless continue to cany out substantial research of this kind. To what extent they will look to Maillard technology we have no way of knowing.

    2 . Maple Syrup

    It has long been known that the characteristic and unmistakable flavor of maple sugar and maple syrup is not present in the sap as drawn freshly from the tree, but that it gradually develops during the boiling process (Nelson, 1928; Findlay and Snell, 1935).

    Compared with the other food systems we have been considering, the aroma and flavor complex of maple syrup appears to be a relatively simple one.

    Largely through the work of Underwood et al. (1961a,b; Underwood and Filipic, 1963, 1964, 1965; Filipic and Underwood, 1964), it was shown that the flavor components of maple syrup could be extracted effectively by chloroform and that the three major components are degradation products of lignin: vanillin, syringaldehyde, and dihydroconiferyl alcohol. Subsequently, the same group discovered a number of other odorous compounds in much smaller amounts (Filipic er al., 1965, 1969): 3-methyl-2-cyclopenten-2-01- 1-one, acetoin, syr- ingyl methyl ketone, acetol, vanillyl methyl ketone, sugar pyrolysis products, and several carboxylic acids.

    But the enigma remains, as it does regarding the aromas of all these other food systems of interest. Reconstitution of the compounds extracted from maple syrup does not duplicate the natural product organoleptically , although it has been claimed (Filipic er al., 1965; Pittet et al., 1970) that 3-methyl-2-cyclopenten-2- 01-1-one is the character impact note of maple flavor. This compound was manufactured and sold by the Dow Chemical Company under the trade name Cyclotene. About 12 years ago Dow sold its proprietary interest in this com- pound to Gliddens, now Durkee Industrial Food, SCM Corporation, Cleveland, Ohio. Taste panel evaluation of the ethyl homolog (the same molecule, with an


    ethyl group in place of the methyl group) indicates that it is more intense, sweeter, and softer than Cyclotene (Pittet el al., 1970).

    Do Maillard reactions have anything to do with development of maple flavor? The USDA group, whose results we have been discussing, nowhere mentions this aspect. Their work does not give so much as a nitrogen analysis for maple sugar, much less information on protein or amino acid content. Yet as early as 1948, Kremers, in what was probably the first patent based on Maillard tech- nology, claimed an imitation maple flavor produced by allowing a-amino acids of three to six carbon atoms to react with a reducing sugar or precursor thereof at 100- 170C; serine, threonine, a-methyl-a-aminobutyric acid, a-aminoisobutry- ic acid, and especially a-aminobutyric acid are most effective in producing the best resemblance to maple. Only one other patent on maple flavor has been found which is related to this technology (Naghski and Willits, 1959). It is worth noting as possibly indicative of Maillard technology that when maple syrup is stripped of maple flavor by extraction with chloroform, it is quite feasible to produce more flavor by autoclaving the marc (extracted syrup). Thus, chloroform does not extract the presursor(s) of the flavor (Filipic and Underwood, 1964).

    Finally, because of the generic relation of honey to maple syrup, one of the standard Maillard patents claims that a honey flavoring is produced by the reaction of at least one monosaccharide with phenylalanine or one of its deriva- tives, such as tyrosine, in large amounts of boiling water (Morton and Sharples, 1959).

    The production of pure maple syrup today is well under 5 million gallons annually. Much of it goes into blends, and the production of the purely artificial maple syrups may well be larger than the production of real maple syrup.

    3. Peanuts

    Mason et al. (1966, 1967, 1969; Newel1 et al., 1967) began an investigation of the flavor components of roasted peanuts in 1965, which followed the now familiar pattern of volatilization, gas chromatography, spectroscopy, and chem- ical analyses of aqueous extracts of raw peanuts, to determine possible precur- sors. The volatiles contain substituted pyrazines, pyrroles, and carbonyl com- pounds. The principal nitrogenous precursors are thought to be aspartic acid, asparagine, glutamic acid, glutamine, histidine, and phenylalanine, obtained by the decomposition of two fairly good-sized peptides which have been isolated and subjected to amino acid analysis. The interpretation throughout is the current view of Maillard browning reactions. Mason et al. noted (1969) that the . . . flavor precursors described are amazingly similar to those found in cocoa by Rohan and Stewart. More refined analyses of compounds present in the volatiles have appeared (Ballschmieter, 1972; Brown et al., 1972).


    There are no reports, either in the journal or patent literature, of attempts to reconstitute peanut flavor or to develop artificial flavors for peanuts or other roasted nuts.

    4. Potato Products

    The volatile components of raw, baked, or boiled potatoes (Buttery et al., 1970) are not of Maillard origin, for the internal temperatures reached are not sufficiently high and the water content is too high to allow Maillard reactions to take place. Nevertheless, 31 pyrazines and 3 thiazoles were identified in the volatiles isolated from baked potato (Coleman and Ho, 1980). But the volatiles of potato chips and French fried potatoes certainly arise from Maillard reactions (Fitzpatrick et al., 1965; Buttery and Ling, 1972). Maga and Sizer (1979) found that during the extrusion of potato flakes from 70 to 16O"C, the total pyrazine content of the product increased from 0.005 to 2.50 ppm. Decreasing moisture levels at the highest extrusion temperature increased pyrazine content >59 ppm. On the other hand, dehydrated potatoes are low in fresh potato aroma because of loss of volatiles during the drying process and may develop off-flavors during processing or storage as a consequence of Maillard reactions on a long-term low- temperature basis. Sapers (1975) has reviewed these problems and some of the attempts to solve them.


    Since man began to cook meat and bake bread, he has been familiar with the aromas and flavors produced by Maillard browning reactions. Maltol in about 1912 was probably the first distinct molecular entity to be recognized as a contributor to these characteristic aromas. With the onset of sustained chemical investigations of the Maillard reaction in the 1940s, it became possible to isolate many volatile compounds, to come to some understanding of how they are formed, and even to develop a basis for exercising some measure of control over the processes by which browned foods are prepared. Clearly some kinds of compounds are much more important than others in the determination of aromas and flavors, and most of these important ones have pleasant characteristic odors. Yet it is curious that almost none of them is really characteristic of any specific food product. Many of these compounds contribute indispensably to a given flavor, but individually none of these compounds would ever be mistaken for any one of the natural flavors.

    We are left with the enigma that a relatively large number of aromas and flavors which any normal human could recognize and distinguish immediately


    II C




    S TH










    E A



    S OF B





    D PR








    Literature Structural form

    ula O


    All brow

    ned foods Rooney et a!. , 1967; R

    othe, 1960; R

    othe and Voight,

    1963; Bailey et al., 1962;

    Link0 et al., 1962; Fuji-

    Aliphatic aldehydes



    R-CHC dl


    isobutyraldehyde." 2-m

    ethylb~tyraldehyde,~ isovaleraldehyde," and pbenylacetaldehydeO


    maki et al.. 1968




    atic aldehydes V



    Ph, or a1 k


    Cocoa, nuts

    van Praag et al., 1968; Schreiber et a

    l., 1974 H

    all 1974a,b



    ood and Filipic, 1963




    ood and Fitipic, 1963


  • a 0'



















    - Q 3 8

    a, I

    = 0

    P 2 a


  • TA


    E V

    II (C






    pound Structural form

    ula O

    ccurrence L



    Severin and Seilmeier, 1967;

    Peer et al., 1968; H

    icks ef a

    /. , 1974; Tonsbeek et a



    4-Hydroxy-5-methyl-2,3-di- (continued)


    Me &


    1968, 1969; Hicks and

    Feather, 1975


    ethyl- HO

    Beef, pineapple

    Tonsbeek et a

    l., 1968, 1969; 2,3-dihydro-3-furanone

    Rodin et a

    l., 1965



    Bread, popcorn

    Hodge and M

    oser, 1961; H

    odge and Nelson, 19616,

    1962a,b; Waliradte et a

    f., u;:M

    e 0


    Meat, bread, vegetables

    Jurch and Tatum

    , 197OC

    ; Mills

    er al., 197OC

    ; Shaw et 0


    1971r; Led1 era

    /.. 1976

  • Maltol



    Bread, apples W

    iseblatt, 1961; Patton, 1950; G

    unner et al., 1968


    ,al kyl ,Ac,O


    Pyrroles N

    -Alkyl-2-acylpyrrole ri



    e,Et N


    R I0

    See Footnote Shibam

    oto and Bem

    hard, 1977a,b, 1976; D

    awes and

    Edwards, 1966; R

    izzi, 1967; Flam

    ent and Stoll, 1967; M

    ussinan er al.,

    1973a; Flament et a

    l.. 1977; M

    ason et al.. 1966;

    Velisek et a

    l., 1976; Wang

    and Odell, 1973; M

    aga and Sizer, 1973; Flam

    ent and Stoll, 1967e; Bondarovich er a

    l., 1967; Rizzi, 1968e; K

    oehler et al., 1969; K

    oeh- ler and O

    dell, 1970; Rizzi,

    1972; Takken et al.. 1979

    Severin and Seilmeier. 1967~ Lendvay,







    are constituted in natural systems (i.e., heated, complex natural products) from a relatively small inventory of compounds.

    One substantial achievement of the research on browning is the identification of compounds and classes of compounds which have been added to the repertory of the flavorist. Many of these compounds have been synthesized by classical methods, protected by patents, and compounded into artificial flavors. Hodge et al. (1972) have attempted to systematize this complex field by distinguishing four major kinds of aromas (caramel, nutty or bready, burnt, and variable) and correlating compounds with them on the basis of their organic structure.

    It goes beyond the scope of this report to present or expand this correlation, but in Table VII we do list the main categories of compounds, some specific examples, some mention of their occurrence, and references to their use and preparation, both in the journal and patent literature.


    The presence and the consumption of products of browning reactions in foodstuffs present two possible problems for human health, one nutritional and the other mutagenic.

    Insofar as there might be a problem of either kind, it is necessary to distinguish between foods in which the presence of reaction products is an inevitable (and highly desirable) result of the traditional process employed and foods to which independently produced reaction products have been added as flavorants.

    In the first case it is extremely unlikely that anyone will seriously maintain that the continued use in usual or ordinary amounts of such traditional foods as crusty bread, cooked meat, coffee, roasted nuts, and cocoa presents a danger to human health. Even if some statistically significant evidence could be obtained, the satisfactory record of centuries would make the risk-benefit ratio insignificantly small. In the second case, however, the deliberate addition of independently prepared reaction products as flavorants to foods might be more vulnerable to criticism.


    To what extent is there a nutritional problem? Adrian and Favier ef al. (Adrian and Favier, 1961; Adrian et al., 1961; Adrian and Frangne, 1976; Adrian, 1972, 1973, 1974) have carried out a sustained study of the conditions under which lysine is lost during the preparation of foods and the extent of that loss, and of the effect of Maillard reactions on animal digestibility of dietary proteins. Ludwig (1979a-c; 1980), concerned with the fate of lysine during the production of p- lactoglobulin, has carried out a model study on mixtures of (3-lactoglobulin and


    lactose and has studied the effects of the disulfide groups in the former on the sparing of the lysine. Chichester, Lee, and others (Chichester and Lee, 1980; Kimiagar et al., 1980; Lee and Chichester, 1977; Lee et al., 1977a,b, 1979, 1981; Pintauro et al., 1980, 1983; Tanaka et al., 1977) have carried out an extended study of the nutritional and toxicological effects of long-term feeding of browned products (especially ovalbumin-glucose mixtures stored at 37C and 68% relative humidity) to rats. They have summarized the nutritional conse- quences of Maillard browning as follows: (1) a decrease in the availability of free amino acids, (2) a decrease in the digestibility of proteins, (3) a decrease in biological value of a protein beyond that which can be accounted for by the decreases in digestibility and amino acids, and (4) some undesirable physiologi- cal events when heavily browned foods are fed at moderate levels.

    They concluded that all biometric, clinical biochemical, and histopathological changes resulting from the feeding of Maillard browned proteins in this study can be attributed to nutritional and/or dietary factors.

    Feeney et al. (1975) have reviewed the topic from an emphatically negative viewpoint under the heading, Naturally Occurring Deteriorative Reactions. Recently, it has been shown that roasting casein with glucose over the range of 120-200C for 20 min (Hayase et al. , 1979) causes substantial, but by no means complete decomposition of the amino acid residues. While racemization of ami- no acid residues is emphasized in the title of this article, the extent of racemiza- tion is significant only in the cases of aspartic acid, alanine, and leucine.

    Rhee and Rhee (1981) heated mixtures of oilseed products at 100C with glucose or sucrose and assessed changes in quality by determining the extent of browning, in vitro protein digestibility, available lysine, total amino acids, and computed protein efficiency ratio. While the mixtures containing sucrose changed very little, those containing glucose decreased substantially in protein quality along with the increased intensity of browning.

    Gardner ( 1979) pointed out that nonenzymatic browning has long been recog- nized as a consequence of peroxidizing lipids in the presence of protein, since the oxidation of the double bonds gives aldehydes. For example, the highly unsatu- rated nature of fish lipids results in the browning or rusting of fish. Pokorny et al. (1973, 1977) showed that the effective parameters are the extent of unsatura- tion in the lipids, the degree of lipid peroxidation, and the free amino group concentration in the protein.

    Jokinen et al. (1976) have applied mathematical analysis to a considerable amount of data on the heated system soy protein-glucose-sucrose-potato starch-cellulose in order to determine the effect of temperature (92.5- 142.5C) and a glucose concentration (0-4%) on lysine retention. However, the sophisti- cated treatment of the data does not lead to simple summary conclusions on the sparing of the lysine.

    Thus, it is unlikely that any results on nutritional loss from browning will have


    any influence on human preferences for traditional browned foods. Since fla- vorants are used in relatively small amounts and the flavors are developed only by destruction of amino acids and carbohydrates, it is obvious that nutritional factors have only limited application here if viewed and evaluated scientifically.


    To what extent is there a mutagenic problem? As long ago as 1967 Devik prepared slurries of amino acid and glucose (2: 1) molar ratio) adsorbed on potato starch and held these slurries 20-45 hr at 100C. He then collected the volgtiles by vacuum distillation, examined them by a polarographic method, and found evidence for the presence of N-nitrosamines. However, Heyns and Koch (1970) repeated and extended Deviks work and showed that the supposed N-nitro- samines were actually a mixture of substituted pyrazines. Pintauro et al. (1980; Lee et al . , 1981) have shown no mutagenic response to any of the samples of browned egg albumin-glucose tested.

    An oral report at a meeting of the American Chemical Society by Iwaoka and Meaker (1979) has again suggested the possibility that mutagenic substances can possibly be formed during normal cooking and processing of foods. Recently, Spingarn and Garvie (1979) followed the lead that mutagens are formed during the cooking of meat at temperatures as low as 100C. They used a model system, refluxing six different monosaccharides with 8 M aqueous ammonia for 2 or 6 hr, similar to that described here earlier. The cooled solutions were extracted with dichloromethane, the residue taken up in ethanol, and tested for mutagenicity using Salmonella typhimurium TA 98 (Ames test). In the case of each sugar, strong mutagenic activity with the same unusual strain specificity and the same kinetics of formation as that derived from cooked meat was observed. Mutagen formation followed formation of pyrazines, but it has already been shown that simple alkylpyrazines are nonmutagenic. They hypothesized that the reactions which formed the pyrazines also formed the mutagens. This is where the matter stands today (Barnes et a l . , 1983). The subject deserves and will doubtless receive further study. It is presently impossible to say anything more about a relation between mutagenicity and any specific food system, although Shibamoto (1982) has reveiwed what has been learned about the occurrence of mutagens in model systems.

    Neither the Food and Drug Administration (FDA) nor any other regulatory agency has yet paid any attention to reaction products produced by the Maillard reaction. As we have already pointed out, it is hard to imagine that they would raise objections to the processes by which traditional foods have been prepared for a very long time; but the FDA might well question the sale and distribution of reaction products as flavorants. Presumably these are sold today as artificial flavors without prior submission for approval. One company representative


    expressed his fear of the instability of this situation. On the basis of what is known today, any ban by the FDA would appear to be arbitrary.


    Investigations in the field of Maillard chemistry are now directed to the solu- tion of an ever-widening circle of problems which differ more and more from each other, though each is still attached to the core.

    Organic chemists continue to pyrolyze specific mixtures of nitrogenous com- pounds with carbonyl compounds, hoping to identify intermediates or terminal products whose structures ,are supportive of one of the speculative pathways (Severin and Braeutigam, 1973; Severin and Loidl, 1976; Mulders, 1973a; Shibamoto and Russell, 1977; Sakaguchi and Shibamoto, 1978a; Baltes and Franke, 1978; Otto and Baltes, 1980, 1981; Goodwin, 1983; Lee et al., 1984, Pokorny et ul., 1979; Kawashima et al., 1980; Heyns and Ruediger, 1981; Orsi, 1981). While some of these studies will be substantive contributions to organic chemistry, it is scarcely to be hoped that they will lead to a sharper focus than that presently available for what happens when animal and vegetable products are roasted.

    More hopeful of attaining their ultimate objectives are some of the continuing studies on Amadori compounds (Heyns et al., 1970; Sulser, 1973; Mills, 1979; Anan, 1979; Birch et al.. 1980). In particular, Ciner-Doruk and Eichner (1979; Eichner and Ciner-Doruk, 1979, 198 1) have shown that quantitative determina- tion of Amadori compounds in natural products (e.g., dried tomato paste) can be used for early recognition of quality changes caused by Maillard reactions, since Amadori compounds are formed without an induction period and before any sensory changes appear.

    It appears that there will be a continuing interest in the nutritional significance of the Maillard reaction, i.e., the extent of the loss of essential nutrients, es- pecially lysine, attributable to nonenzymatic browning. Of greater potential sig- nificance, however, is the extent to which browned food products may be a source of mutagens. It could be a strategic weakness of the investigations re- ported and under way that the search for mutagens depends almost entirely on the widely adapted Ames test. It is somewhat disconcerting to read that significant mutagenic activity is produced when starchy foods are prepared by common cooking procedures (Spingarn et al . , 1980a). There remains the much more difficult problem of determining any possible relation between human health and the accumulating data on Ames tests of browned products (Yoshida and Okamo- to, 1980; Spingarn et al., 1980b; Mihara and Shibamoto, 1980; Shibamoto, 1980b; Toda et al., 1981; Shibamoto et al., 1981; Spingarn et al., 1983).

    The suggestion and prediction of Hodge (1953) allows us to close on a positive


    note: . . . the control of browning reactions to produce only wanted flavors and odors is an intriguing possibility.Contro1 of browning to do mans will is the ultimate goal of browning research, but progress toward this goal can be made only as the reaction mechanisms are better understood. Published work along these lines continues (Akiyama er al., 1978; Sakaguchi and Shibamoto, 1978; Baltes, 1979; Pokorny et al., 1979b; Nishimura et al . , 1980; Hsieh er al. , 1980a; Schroedter and Woelm, 1980; Maga, 1981; Pokorny et al., 1981) and may well be more than matched by proprietary R 8z D work which will not be published.


    Adrian, J. 1972. Maillard reaction from the nutritional standpoint. 11. Food behavior. Ind. Aliment. Agric. 89, 1713-1720.

    Adrian, J. 1973. Maillard reaction from a nutritional standpoint. 111. Harmful substances resulting from the Maillard reaction. Ind . Aliment. Agric. 90, 449-455.

    Adrian, I . 1974. Nutritional and physiological consequences of the Maillard reaction. World Rev. Nurr. Diet 19, 71-122.

    Adrian, J . , and Favier, J. C. 1961. Maillard reaction. 1. Behavior of pure lysine. Ann. Nurr. Alimenr. 15, 181-225.

    Adrian, I., and Frangne, R. 1976. Behavior of proteolysates during the Maillard reaction. I. Reac- tivity of a protein as a function of its degree of proteolysis. Ind. Alimenr. Agric. 93, 23-28.

    Adrian, J., Favier, D. C., Carrilho, O., and Jacquot, R. 1961. Stability of lysine heated in the presence of sugars. Arq. Port. Bioquim. 5 , 33-35.

    Akiyama, T., Enomoto, Y., and Shibamoto, T. 1978. A new method of pyrazine synthesis for flavor use. J . Agric. Food Chem. 26, 1176-1179.

    Anan, T. 1979. Isolation and identification of a new Amadori compound from green tea. J. Sci. Food Agric. 30, 906-910.

    Anan, T., Amano, I., and Nakagawa, M. 1981. Relation between changes of chemical composition and taste of green tea. I. Changes in contents of some ingredients during heating of green tea. Nippon Shokuhin Kogyo Gakkaishi 28, 74-18.

    Arnold, R. G., Libbey, L. M., and Lindsay, R. C. 1969. Volatile flavor compounds produced by heat degradation of thiamine. J . Agric. Food Chem. 17, 390-392.

    Arroyo, P. T., and Lillard, D. A. 1970. Identification of carbonyl and sulfur compounds from nonenzymatic browning reactions of glucose and sulfur-containing amino acids. J. Food Sci.

    Bailey, S. D., Mitchell, D. G., Bazinet, M. L., and Weurman, C. 1962. Studies on the volatile

    Baker, J. C., Parker, H. K., and Fortmann, K. L. 1953. The flavor of bread. Cereal Chem. 30,

    Ballschmieter, H. M. B. 1972. Influence of storage and packaging on the volatile compounds of

    BaItes, W. 1979. Roasting aromas. Drsch. Lebensm. Rundsch. 75, 2-7. BaItes, W. 1980. Significance of the Maillard reaction on the development of flavor in foods.

    Baltes, W., and Franke, K. 1978. Model studies on the Maillard reaction. I. Non-volatile reactign

    35, 769-770.

    components of different varieties of cocoa beans. J. Food Sci. 27, 165-170.


    roasted peanuts. Ferre Seifen Anstrichsm. 74, 112-1 19.

    Lebensmitrelchem. Gerichrl. Chem. 34, 39-41.


    products from the reaction of D-glucose with p-chloroaniline. Z . Lebensm. Unrers. Forsch. 167,

    Barnes, H. M., and Kaufman, C. W. 1947. Industrial aspects of the browning reaction. Ind. Eng. Chem. 39, 1167-1 170.

    Barnes, W. E., Spingarn, N. E.. Garvey-Could, C., Vuolo, L. L., Wang, Y. Y., and Weisburger, J. H. 1983. Mutagens in cooked foods: Possible consequences of the Maillard reaction. I n The Maillard Reaction in Foods and Nutrition (G. R. Waller and M. S. Feather, eds.). ACSSymp. Ser. No. 215, 485-506.

    Bertram, G. L. 1953. Studies on crust color. I. The importance of browning reaction in determining the crust color of bread. Cereal Chem. 30, 127-139.

    Bidmead, D. S., Giacino, C., Grossmann, J. D., and Kratz, P. D. 1968. Roasted meat flavor and process for producing the same. U. S. Patent 3,394,016; equivalent to Neth. Appl. 6609520 (Chem. Absrr. 66, 114766).

    Birch, E. J., Lelievre, J., and Richards, E. L. 1980. Thermal analysis of 1-deoxy-1-glycino-o- fructose and 1-P-alanino-I -deOXy-D-fNCtOSe. Curbohydr. Res. 83, 263-272.

    Bondarovich, H. A., Friedel, P., Krampl, V., Renner, I. A., Shephard, F. W., and Gianturco, M. A. 1967. Volatile constituents of coffee. Pyrazines and other compounds. J. Agric. Food Chem.

    Brinkman, H. W., Copier, H., de Leeuw, J. J. M., and Tjan, S. B. 1972. Components contributing to beef flavor. Analysis of the headspace volatiles of beef broth. J. Agric. Food Chem. 20, 177- 181.

    Broderick, J. J., and Linteris, L. L. 1960. Flavoring agents for imparting a meatlike flavor to edible compositions. US. Patent 2,955,041 (Chem. Abstr. 55, 1966~).

    Brown, D. F., Senn, V. J., Stanley, J. B., and Dollear, F. G. 1972. Comparison of carbonyl compounds in raw and roasted runner peanuts. I. Major qualitative and some quantitative differences. J. Agric. Food Chem. 20, 700-706.

    Buttery, R. G., and Ling, L. C. 1972. Characterization of nonbasic steam volatile components of potato chips. J . Agric. Food Chem. 20, 698-700.

    Buttery, R. G., Seifert, R. M., and Ling, L. C. 1970. Characterization of some volatile potato components. J. Agric. Food Chem. 18, 538-539.

    Calame, J. P., Kappeler, H., and Oberhaensli, P. 1973. Flavoring and odorous 2-hydroxy-3-meth- yl-2-cyclopenten-1-one. German Patent 2,313,504 (Chem. Absrr. 80, 3172).

    Chang, S. S. , and Peterson, R. J. 1977. Recent developments in the flavor of meat. J. Food Sci. 42,

    Chichester, C. O., and Lee, T.-C. 1980. Effect of food processing in the formation and destruction of toxic constituents of food. I n Impact of Toxicology on Food Processing (John C. Ayres and John C. Kirschman, eds.), pp. 35-56. AVI, Westport, CT.

    Ciner-Doruk, M., and Eichner, K. 1979. Formation and stability of Amadori compounds in low moisture foods. Z . Lebensm. Unrers. Forsch. 168, 9-20.

    Coleman, E. C., and Ho, C.-T. 1980. Chemistry of baked potato flavor. I . Pyrazines and thiazoles identified in the volatile flavor of baked potato. J. Agric. Food Chem. 28, 66-68.

    Danehy, J. P., and Pigman, W. W. 195 1. Reactions between sugars and nitrogenous compounds and their relationship to certain food problems. Adv. Food Res. 3, 241-290.

    Darsley, R. R., and Quesnel, V. C. 1972. The production of aldehydes from amino acids in roasted cocoa. J. Sci. Food Agric. 23, 215.

    Dawes, I. W., and Edwards, R. A. 1966. Methyl-substituted pyrazines as volatile reaction products of heated aqueous aldose and amino acid mixtures. Chem. Ind. 2203.

    Devik, 0. G. 1967. Formation of N-nitrosamines by the Maillard reaction. Acrn Chem. Scund. 21,


    15, 1093-1097.




    Dietrich, P., Lederer, E., Winter, M., and Stoll, M. I. 1964. Sur Iarome du cacao. Helv. Chim.

    Dworschak, E. 1980. Nonenzyme browning and its effect on protein nutrition. CRC Crir. Rev. Food Sci. Nurr. 13, 1-40.

    Eichner, K., and Ciner-Doruk, M. 1979. Formation and stability of Amadori compounds in low moisture model systems. 2. Lebensm. (Inters. Forsch. 168, 360-367.

    Eichner, K., and Ciner-Doruk, M. 1981. Water activities: Influences on food quality. Formation and decomposition of browning intermediates and visible sugar-amine browning reactions. In Water Activity: Influences on Food Quality, pp. 567-603.

    Einarsson, H., Snygg, B. G., and Eriksson, C. 1983. Inhibition of bacterial growth by Maillard reaction products. J . Agric. Food Chem. 31, 1043-1047.

    Ellis, G. P., and Honeyman, J. 1959. Glycosylamines. Adv. Curbohydr. Chem. 10, 95-168. Eriksson, C. E., ed. 1981. Maillard reactions in food: Chemical, physiological and technological

    Fagerson, I. S. 1969. Thermal degradation of carbohydrates. A review. J . Agric. Food Chem. 17,

    Feeney, R. E., Clary, J. J., and Clark, J. R. 1964. A reaction between glucose and egg-white proteins in incubated eggs. Nurure (London) 201, 192-193.

    Feeney, R. E., Blankenhorn, and Dixon, H. B. F. 1975. Carbonylamine reactions in protein chem- istry. Adv. Prof. Chem. 29, 135-203.

    Feldman, J . R., Ryder, W. S . , and Kung, J. T. 1969. Importance of nonvolatile compounds to the flavor of coffee. J. Agric. Food Chem. 17, 733-739.

    Filipic, V. J., and Underwood, J. C. 1964. Some aromatic compounds in sap composition of maple sap and SYNP. J. Food Sci. 29, 464-468.

    Filipic, V. J., Underwood, J . C., and Willits, C. 0. 1965. Identification of methylcyclopentenolone in maple syrup flavor extract. J . Food Sci. 30, 1008-1015.

    Filipic, V. J. Underwood, J. C., and Dooley, C. J. 1969. Trace components of the flavor fraction of maple SYNP. J. Food Sci. 34, 105-1 10.

    Findlay, G. H., and Snell, J. F. 1935. Some constituents of the sap of the sugar maple (Acer succhurum, Marsh). Can. J . Res. 13B, 269-275.

    Fitzpatrick, T. L., Talley, E. A., and Porter, W. L. 1965. Food color changes. Preliminary studies on the fate of sugars and amino acids in chips made from fresh and stored potatoes. J . Agric. Food Chem. 13, 10-12.

    Flament, 1. 1975. Aroma composition for food and tobacco. Swiss Patent 568,722 (Chem. Abszr. 84, 42237).

    Flament, I., and Ohloff, G. 1971. Sur Iarome de viande de boeuf grillCe. I. Pyrazines. Helv. Chim. Ada. 54, 1911-1913.

    Flament, I., and Stoll, M. 1967. Pyrazines. I. Synthkse de methyl-2-pyrazines alcoylkes en 3, par condensation de Iethylenediamine avec les dioxo-2,3-alcanes. Helv. Chim. Acru 50, 1754- 1758.

    Flament, I., Willhalm, B., and Stoll, M. 1967. Sur Iarome du cacao. 111. Helv. Chim. Acru 50,

    Flament, I., Sonnay, P., and Ohloff, G. 1977. Sur Iarorne de viande de boeuf grillCe. 111. Pyr-

    Fujimaki, M., Kurata, T., and Kato, S. 1968. Reactivities of some carbonyl compounds in Strecker

    Galt, A. M., and MacLeod, G. 1984. Headspace sampling of cooked beef aroma using Texax G. C.

    Gardner, H. W. 1979. Lipid hydroperoxidase reactivity with proteins and amino acids: A review. J .

    Acts 47, 1581-1584.

    aspects. Prog. Food Nurr. Sci. 5, 3 + 501 pp.



    rolo[ 1,2-a]pyrazines, identification et synthese. Helv. Chim. Acru 60, 1872-1883.

    degradation. Agric. Biol. Chem. 32, 46-50.

    J . Agric. Food Chem. 32, 59-64.

    Agric. Food Chem. 27, 220-229.


    Giacino, C. 1968. Poultry flavor composition and process. U S . Patent 3,394,017; equivalent to Neth. Appl. 6603722 (Chem. Abstr. 66, 27875).

    Goldman, I. M., Seibl, J . , Flament, I., Gautschi, F., Winter, M., Willhalm, B., and Stoll, M. 1967. Sur Iarome de cafe. 11. Pyrazines et pyridines. Helv. Chim. Actu 50, 694-705.

    Goodwin, J. C. 1983. Isolation of 3-O-a-~-gluco- and 3-O-~-~-galactopyranosyloxy-2-furyl methyl ketones from nonenzymic browning of maltose and lactose with secondary amino acids. Car- bohydr. Res. 115, 281-287.

    Gottschalk, A. 1952. Some biochemically relevant properties of N-substituted fructosamines derived from amino acids and N-arylglucosamines. Biochem. J . 52, 455-460.

    Greenshields, R. N., and Macgillivray, A. W. 1972. Caramel. I . Browning reactions. Process Biochem. 7 , 11-13, 16.

    Griffin, J. M. 1966. Increasing the flavor strength of anethole, cinnamaldehyde, and methyl salicylate with maltol. U.S. Patent 3,293,045 (Chem. Abstr. 66,45628).

    Grimmett, R. M. 1965. Formation of heterocyclic compounds from carbohydrates and ammonia. Rev. Pure Appl. Chem. 15, 101-108.

    Gunner, S. W., Hand, B., and Sahasrabudhe, M. 1968. Determination of maltol and ethylmaltol in apple juice by gas-liquid chromatography. J. Assoc. 08. Anal. Chem. 51, 959-963.

    Gunther, R. 1972. Flavoring compositions produced by reacting hydrogen sulfide with a pentose. U.S. Patent 3,642,497 (Chem. Absrr. 76, 139285).

    Hall, J. B., and Vock, M. 1972. Food-flavoring compositions containing 3-phenylpentenals. U.S. Patent 3,694,232 (Chem. Abstr. 77, 150681).

    Hall, J. B., and Vock, M. 1974a. 3-Phenylpentenals. U.S. Patent 3,793,376. Hall, J. B., and Vock, M. 1974b. Dialkyl and alkylene acetals of 2- and 3-phenylpentenals. U.S.

    Patent 3,829,504 (Chem. Abstr. 81, 151801). Hartman, G. J.. Jin, Q. Z., Collins, G. J., Lee, K. N., Ho, C.-T., and Chang, S. S. 1983. Nitrogen-

    containing heterocyclic compounds identified in the volatile flavor constituents of roasted beef. J. Agric. Food Chem. 31, 1030-1033.

    Hayase, F., Kato, H., and Fujimaki, M. 1979. Racemization of amino acid residues in casein roasted with glucose and/or methyl linoleate. Agric. Biol. Chem. 43, 2459-2465.

    Hen, K. O., and Chang, S. S. 1970. Meat flavor. Adv. Food Res. 18, 1-83. Heyland, S., and Cerise, L. 1979. Textured protein meat substitute. German Patent 2,837,513

    (Chem. Absrr. 90, 202491). Heyns, K., and Koch, H. 1970. The origin of nitrosamines from the reaction of monosaccharides

    with amino acids (the Maillard reaction). Tetrahedron Lett. 741-744. Heyns, K., and Noack, H. 1962. Die Umsetzung von D-FNCtOSe mit L-Lysin und L-Arginin und

    deren Beziehung zu nichtenzymatsichen Briunungsreaktionen. Chem. Ber. 95, 720-727. Heyns, K., and Ruediger, R. 1981. Reaction of partially alkylated hexuloses with amino acids.

    Liebigs Ann. Chem. 138- 142. Heyns, K., Mueller, G., and Paulsen, H. 1967. Quantitative studies on the reactions of hexoses with

    amino acids. Liebigs Ann. Chem. 703, 202-214. Heyns, K., Heukeshoven, J., and Brose, K. H. 1968. Browning reactions and fragmentation of

    carbohydrates. V. Degradation of fructose amino acids to N-(2-furoylmethyl)amino acids. Inter- mediates in browning reactions. Angew. Chem. Int. Ed. 7 , 628-629.

    Heyns, K., Chiemprasert, T., and Bakes, W. 1970. Non-enzymic browning reactions. VI. Effects of substituents in the o-position of D-glucohexoses on glucoylamine formation and Amadori rear- rangement. Chem. Ber. 103, 2877-2884.

    Hicks, K. B., and Feather, M. S. 1975. Studies on the mechanism of formation of 4-hydroxy-5- methyl-3(2H)-furanone, a component of beef flavor, from Amadori products. J . Agric. Food Chem. 23,957-960.

    Hicks, K. B., Harris, D. W., Feather, M. S., and Leoppky, R. N. 1974. Production of 4-hydroxy-5-


    methyl-3(w)-furanone, a compound of beef flavor, from a I-amino-I-deoxy-D-fructuronic acid. J . Agric. Food Chem. 22, 724-725.

    Hiraoka, T., and Shiozaki, M. 1974. 2-Hydroxy-3-methyl-2-cyclopenten-I-one. Jpn. Kokai 74 30341 (Chem. Abstr. 81, 104827).

    Hodge, J. E. 1953. Dehydrated foods. Chemistry of browning reactions in model systems. Agric. Food Chem. 1, 928-943.

    Hodge, J. E. 1955. The Amadori rearrangement. Adv. Carbohydr. Chem. 10, 169-205. Hodge, J. E., and Moser, H. A. 1961. Flavor of bread and pastry upon addition of maltol, isomaltol,

    and galactosylisomaltol. Cereal Chem. 38, 221-228. Hodge, J. E., and Nelson, E. C. 1961. Preparation and properties of D-galactosylisomaltol and

    isomaltol. Cereal Chem. 38, 207-221. Hodge, J. E., and Nelson, E. C. 1962a. 0-Galactosylisomaltol. U.S. Patent 3,015,654 (Chem.

    Absrr. 56, 13000k). Hodge, J. E., and Nelson, E. C. 1962b. Isomaltol. U. S . Patent 3,054,805 (Chem. Absrr. 58,

    5639b). Hodge, J. E., and Rist, C. E. 1953. The Amadori rearrangement under new conditions and its

    significance for non-enzymatic reactions. J. Am. Chem. SOC. 75, 316-322. Hodge, J. E., Fisher, B. E., and Nelson, E. C. 1963. Dicarbonyls, reductones, and heterocyclics

    produced by reactions of reducing sugars with secondary amine salts. Am. SOC. Brew. Chem.

    Hodge, J. E., Mills, F. D., and Fisher, B. E. 1972. Compounds of browned flavor derived from

    Holden, M. 1959. Processing of raw cocoa. 111. Enzymic aspects of cocoa fermentation. J . Sci. Food

    Homstein, I., and Crowe, P. F. 1960. Flavor studies on beef and pork. J . Agric. Food Chem. 8,

    Hsieh, Y. P. C., Pearson, A. M., and Magee, W. T. 1980a. Development of a synthetic meat flavor mixture by using surface response methodology. J. Food Sci. 45, 1125-1130.

    Hsieh, Y. P. C., Pearson, A. M., Morton, I. D., and Magee. W. T. 1980b. Some changes in the constituents upon heating a model meat flavor system. J . Sci. Food Agric. 31, 943-949.

    Hunter, I. R., Walden, M. K., Scherer, J. R., and Lundin, R. E. 1969. Preparation and properties of 1,4,5,6-tetrahydro-2-acetopyridine, a cracker odor constituent of bread aroma. Cereal Chem.

    Humll, R. F., and Carpenter, K. J. 1977. Maillard reactions in foods. Proc. Int. Symp. Phys. Chem.

    Iwaoka and Meaker. 1979. Researchers say excessive browning of food may produce mutagens.

    Jarboe, I. K., and Malbrouk, A. F. 1974. Free amino acids, sugars, and organic acids in aqueous

    Johnson, J. A., and Miller, B. S . 1961. Browning of baked products. Bakers Dig. (5). 52-59. Johnson, J. A., Rooney, L., and Salem, A. 1966. Chemistry of bread flavor. I n Flavor Chem-

    istry. Adv. Chem. Ser. (56), 153-173. Jokinen, J. E., Reineccius, G. A., and Thompson, D. R. 1976. Losses in available lysine during

    thermal processing of soy protein model systems. J . Food Sci. 41, 816-819. Jurch, G. R., and Tatum, J. H. 1970. Degradation of D-glucose with acetic acid and methylamine.

    Carbohydr. Res. 15, 233-239. Kato, H. 1967. Chemical studies on amino-carbonyl reaction. 111. Formation of substituted pyr-

    role-2-aldehydes by reaction of aldoses with alkylamines. Agric. Biol. Chem. 31, 1086-1090. Kato, H., and Fujimaki, M. 1968. Formation of N-substituted pyrrole-2-aldehydes in the browning

    reaction between D-XYIOS~ and amino compounds. J . Food Sci. 33, 445-449.


    sugar-amine reactions. Cereal Sci. Today 17, 34-40.

    Agric. 10, 691-700.


    46, 189-195.

    Biol. Changes Food Caused Therm. Process 168- 184.

    Food Chem. NewsApril9, 39-40.

    beef extract. J . Agric. Food Chem. 22, 787-791.


    Kato, Y., Watanabe, K., and Sato, Y. 1981. Effect of some metals on the Maillard reaction of

    Katz, I., Evers, W. 1. . and Giacino, C. 1972. Sulfur-containing flavoring compositions. US. Patent

    Katz, I., Wilson, R. A,, Evers, W. J., Vock, M. H., and Verhoeve, G. W. 1973. Heterocyclic

    Kawashima, K., Itoh, H., and Chibata, 1. 1980. Nonenzymatic browning reactions of dihydrox-

    Kiely, P. J., Nowlin, A. C., and Moriarty, I . H. 1960. Bread aromatics from browning systems.

    Kimiagar, M., Lee, T.-C., and Chichester, C. 0. 1980. Long-term feeding effects of browned egg

    Kline, R. W., and Stewart, G. F. 1948. Glucose-protein reaction in dried egg albumen. Ind. Eng.

    Kobayashi. N., and Fujimaki, M. 1965. On the formation of N-acetonylpyrrole on roasting hy-

    Koehler, P. E., and Odell, G. V. 1970. Factors affecting the formation of pyrazine compounds in

    Koehler, P. E., Mason, M. E., and Newell, J. A. 1969. Formation of pyrazine compound in sugar-

    Kort, M. J. 1970. Reactions of free sugars with aqueous ammonia. Adv. Carbohydr. Chem. Bio-

    Kremers, R. E. 1948. Imitation maple flavor. U.S. Patent 2,446,478 (Chem. Absrr. 42, 8006t3. Kung, J. T. 1974. A new caramel compound from coffee. J . Agric. Food Chem. 22, 494-496. Landmann, W. A., and Batzer, 0. F. 1966. Influence of processing procedures on the chemistry of

    meat flavors. J. Agric. Food Chem. 14, 210-214. Lea, C. H., and Hannan, R. S. 1950a. Studies of the reactions between proteins and reducing sugars

    in the dry state. 11. Further observations of the formation of the casein-glucose complex. Biochim. Biophys. Acra 4, 5 18-53 I .

    Lea, C. H., and Hannan, R. S. 1950b. Studies of the reactions between proteins and reducing sugars in the dry state. 111. Nature of the protein groups reacting. Biochim. Biophys. Acra 5 , 433- 456.

    Lea, C. H., and Hannan, R. S. 1950~. Biochemical and nutritional significance of the reaction between proteins and reducing sugars. Nafure (London) 165, 438.

    Lea, C. H., and Rhodes, D. N. 1952. Reactions between proteins and reducing sugars in the dry state. V. Reactions of D-galactose, 2-deoxy-~-galactose, D-glucosamine, and N-acetyl-D-gluco- samine with casein. Biochim. Biophys. Acra 9, 56-60.

    Lea, C. H., Hannan, R. S., and Rhodes, D. N. 1951. Studies of the reactions between proteins and reducing sugars in the dry state. IV. Decomposition of the amino-sugar complex and the reaction of acetylated casein with glucose. Biochim. Biophys. Acra 7 , 366-377.

    Ledl, F., and Severin, Th. 1973. Thermal decomposition of cysteine and xylose in tributyrin. Chem.. Mikrobiol. Technol. Lebensm. 2, 155-160.

    Ledl, F., and Severin, Th. 1974. Sulfur-containing compounds from cysteine and xylose. Z . Lebensm. Unters. Forsch. 154, 29-31.

    Ledl, F., Schnell, W., and Severin, Th. 1976. Studien zur Maillard-Reaktion. XI. Die Bestimmung von 2,3-dihydro-3,5-dihydroxy-6-methyl-4H-pyran-4-one in Nahrungsmitteln. Z . Lebensm. Unters. Forsch. 160, 367-370.

    Lee, T.-C., and Chichester, C. 0. 1977. The nutritional effects of the Maillard browning of foods. I n Non-Conventional Proteins and Foods. NSF-RA-770278. National Science Foundation, Washington, D.C.

    ovalbumin. J . Agric. Food Chem. 29, 540-543.

    3,706,577 (Chem. Absfr. 78, 83062).

    pyrimidine flavoring compositions. U.S. Patent 3,726,692 (Chem. Absrr. 79, 17226).

    yacetone with amino acids or their esters. Agric. Biol. Chem. 44, 1595-1599.

    Cereal Sci. Today 5 , 273-274.

    albumin to rats. 1. Agric. Food Chem. 28, 150-155.

    Chem. 40, 919-922.

    droxyproline with carbonyl compounds. Agric. Biol. Chem. 29, 1059-1060.

    sugar-amine reactions. J. Agric. Food Chem. 18, 895-898.

    amino acid model systems. 1. Agric. Food Chem. 17, 393-396.

    chem. 25, 311-349.


    Lee, C. M., Chichester, C. O., and Lee, T.-C. 1977a. Effect of Maillard reaction products on disaccharidase activities in the rat. J. Agric. Food Chem. 25, 775-778.

    Lee, C. M., Lee, T.-C., and Chichester, C. 0. 1977b. The effect of Maillard reaction products on the absorption of tryptophan. Comp. Biochem. Physiol. A 56, 473-476.

    Lee, C. M., Lee, T.-C., and Chichester, C. 0. 1979. Kinetics of the production of biologically active Maillard browned products in apricots and glucose-L-tryptophan. J. Agric. Food Chem.

    Lee, T.-C., Kimiagar, M., Pintauro, S . J. , and Chichester, C. 0. 1981. Physiological and safety aspects of Maillard browning of foods. Prog. Food Nurr. Sci 5 , 243-256.

    Lee, C. M., Sherr, B., and Koh, Y.-N. 1984. Evaluation of kinetic parameters for a glucose-lysine Maillard reaction. J. Agric. Food Chem. 32, 379-382.

    Lendvay, A. T. 1970. Improving the flavor of baked goods. U.S. Patent 3,499,765 (Chem. Absrr. 72, 110018).

    Lindsay, R. C., and Lau, V. K. 1972. Some reaction products from nonenzymatic browning of glucose and methionine. J . Food Sci. 37, 787-788.

    Lingnert, H., and Waller, G. R. 1983. Stability of antioxidants formed from histidine and glucose by the Maillard reaction. J. Agric. Food Chem. 31, 27-30.

    Linko, Y., and Johnson, J. A. 1963. Changes in amino acids and formation of carbonyl compounds during baking. J. Agric. Food Chem. 11, 150-2.

    Linko, Y., Johnson, J. A,, and Miller, B. S. 1962. The origin and fate of certain c a h n y l com- pounds in white bread. Cereal Chem. 39, 468-476.

    Ludwig, E. 1979a. Studies on the Maillard reaction between P-lactoglobulin and lactose. I. Lysine damage during the production of P-lactoglobulin and during the freeze-drying of a protein- sugar mixture. Nuhrung 23, 425-429.

    Ludwig, E. 1979b. Studies on the Maillard reaction between P-lactoglobulin and lactose. 11. Effects of the thiol and disulfide groups on the properties of the macromolecular reaction products. Nuhrung 23, 531-536.

    Ludwig, E. 1979c. Studies on the Maillard reaction between P-lactoglobulin and lactose. 111. The effects of intermolecular disulfide bridges on the blocking of lysine. Nahrung 23, 707-7 14.

    Ludwig, E. 1980. Studies on the Maillard reaction between P-lactoglobulin and lactose. IV. Spectra and amino acid composition of macromolecular reaction products. Nahrung 24, 399-404.

    Mackod, G., and Coppock, B. M. 1976. Volatile flavor components of beef boiled conventionally and by microwave radiation. J. Agric. Food Chem. 24, 835-843.

    MacLeod, G., and Seyyedain-Ardebili, M. 1981. Natural and simulated meat flavors (with particular reference to beef). CRC Crir. Rev. Food Sci. Nutr. 14, 309-437.

    Macy, R. L., Naumann, H. D., and Bailey, M. E. 1964a. Water-soluble flavor and odor precursors of meat. 1. Qualitative study of certain amino acids, carbohydrates, non-amino acid nitrogen compounds, and phosphoric acid esters of beef, pork, and lamb. J. Food Sci. 29, 136-141.

    Macy, R. L., Naumann, H. D., and Bailey, M. E. 1964b. Water-soluble flavor and odor precursors of meat. 11. Effect of heat on amino nitrogen constituents and carbohydrates in lyophilized diffusates from aqueous extracts of beef, pork, and lamb. J. Food Sci. 29, 142-148.

    Maga, J. A. 1978. Oxazoles and oxazolines in foods. J. Agric. Food Chem. 26, 1049-1050. Maga, J. A. 1981. Chemistry of oxazoles and oxazolines in foods. CRC Crit. Rev. Food Sci. Nurr.

    Maga, J. A., and Sizer, C. E. 1973. Pyrazines in foods. A review. J. Agric. Food Chem. 21,22-30. Maga, J. A., and Sizer, C. E. 1979. Pyrazine formation during the extrusion of potato flakes.

    Maillard, L. C. 1912s. Action des acides amines sur les sucres: Formation des melanoidines par voie

    27, 478-482.

    14, 295-307.

    Lebensm. Wiss. Technol. 12, 15-16.

    methodique. C. R. Acud. Sci. 154, 66-68.


    Maillard, L. C. 1912b. Rtaction gtnkrale des acides amines sur les sucres: Ses constquences biologiques. C. R . Soc. Biol. 72, 599-601.

    Marion, J. P., Miiggler-Chavan, F., Viani, R., Bricout, J., Reymond, D., and Egli, R. H. 1967. Sur la composition de larome de cacao. Helv. Chim. Acfa 50, 1509-1516.

    Mason, M. E., Johnson, B., and Hamming, M. 1966. Flavor components of roasted peanuts. Some low molecular weight pyrazines and a pyrrole. J. Agric. Food Chem. 14, 454-460.

    Mason, M. E., Johnson, B., and Hamming, M. 1967. Volatile components of roasted peanuts. The major monocarbonyls and some noncarbonyl components. J. Agric. Food Chem. 15, 66-73.

    Mason, M. E., Newell, J. A., Johnson, B. R., Koehler, P. E., and Waller, G. R. 1969. Non-volatile flavor components of peanuts. J . Agric. Food Chem. 17, 728-732.

    May, C. G. 1960. Meat-like flavoring substances. U.S. Patent 2,934,435 (Chem. Absrr. 54, 18830). May, C. G., and Akroyd, P. 1959a. Synthetic flavoring agents. German Patent 1,058,824 (Chem.

    May, C. G., and Akroyd, P. 1959b. Flavoring substances and their preparation. U.S. Patent

    May, C. G., and Morton, I. D. j960. Meatlike flavoring substances. U.S. Patent 2,934,436 (Chem.

    Mihara, S., and Shibamoto, T. 1980. Mutagenicity of products obtained from cysteamine-glucose

    Mills, F. D. 1979. Vacuum thermolysis of 1-deOXy-1-SarCOSinO-D-fNCtOse. J. Agric. Food Chem.

    Mills, F. D., Weisleder, D., and Hodge, J. E. 1970. 2,3-Dihydro-3,5-dihydroxy-6-methyl-4H- pyran-4-one, a novel nonenzymatic browning product. Tetrahedron Lett 1243- 1246.

    Mohr, W., Roehrle. M., and Severin, Th. 1971. Die Bildung von Kakaoaroma aus den Vorstufen. Ferre Seifen Ansrrichsm. 73, 515-521.

    Mohr, W., Landschreiber, E., and Severin, Th. 1976. Zur Spezifitat des Kakaoaromas. Ferre Seifen Ansrrichsm. 78, 88-95.

    Morimoto, T., and Johnson, J. A. 1966. Studies on the flavor fraction of bread crust adsorbed by cation-exchange resin. Cereal Chem. 43, 627-643.

    Moroe, T. 1969. Improving the flavor of cocoa and chocolate-flavored materials. U.S. Patent 3,459,556 (Chem. Absrr. 71, 122594).

    Morton, I. D., and Sharples, E. 1959. Honey flavoring. U.S. Patent 2,916,382 (Chem. Absrr. 54, 598 1 d).

    Morton, I. D., Akroyd, P., and May, C. G. 1960. Flavoring substances. British Patent 836,694, equivalent to U.S. Patent 2,934,437 (Chem. Absrr. 54, 17746).

    Mosher, A. J. 1973. Composition having the odor of meat. French Patent 2,178,924 (Chem. Absfr. 81, 24422).

    Mulders, E. J. 1973a. Odor of white bread. IV. Quantitative determination of constituents in the vapor and their odor values. Z. Lebensm. Unters. Forsch. 151, 310-317.

    Mulders, E. J. 1973b. Volatile components from the nonenzymic browning reaction of the cysteine- cystine-ribose system. Z. Lebensm. Unrers.Forsch. 152, 193-201.

    Mulders, E. J., and Dhont, I. H. 1972. Odor of white bread. 111. Identification of volatile carbonyl compounds and fatty acids. Z. Lebensm. Unrers. Forsch. 150, 228-32.

    Mulders, E. J., Maarse, H., and Weurman, C. 1972a. Odor of white bread. I. Analysis of volatile constituents in the vapor and aqueous extracts. Z. Lebensm. Unters. Forsch. 150, 68-74.

    Mulders, E. J., Ten Noever de Brauw, M. C., and Van Straten, S. 1972b. Odor of white bread. 11. Identification of components in pentane-ether extracts. 2. Lebensm. Unrers. Forsch. 150, 306- 310.

    Abstr. 55, 16853).


    Absrr. 54, 18830).

    browning model. J . Agric. Food Chem. 28, 62-6.

    27, 1136-1138.


    Mussinan, C. J., and Katz, I. 1973. Isolation and identification of some sulfur chemicals present in

    Mussinan, C. J., Wilson, R. A., and Katz, I. 1973. Isolation and identification of pyrazines present

    Nafisi, K., and Markakis, P. 1983. Inhibition of sugar-amine browning by aspartic and glutamic

    Naghski, J., and Willits, C. 0. 1959. Maple flavor of desired color and flavor. U.S. Patent

    Nakatani, Y., and Yanatori, Y. 1973. Simple synthesis of 2,3-disubstituted pyrazines. Agric. Biol.

    Nakel, G . M., and Dirks, B. M. 1971a. Food products containing alkyl-substituted pyrazines as

    Nakel, G. M., and Hiler, G. D. 1971b. Organic hisulfide-containing chocolate flavor compositions.

    Nelson, E. K. 1928. Flavor of maple syrup. J. Am. Chem. SOC. 50, 2009-2012. Newell, J. A., Mason, M. E., and Matlock, R. S. 1967. Precursors of typical and atypical roasted

    peanut flavor. J. Agric. Food Chem. 15, 767-772. Nishimura, O., Mihara, S . , and Shibamoto, T. 1980. Compounds produced by the reaction of 2-

    hydroxy-3-methyl-2-cyclopenten-I-one with ammonia and hydrogen sulfide. J . Agric. Food Chem. 28,39-43.

    two model systems approximating cooked meat. J. Agric. Food Chem. 21, 43-45.

    in pressure-cooked beef. J . Agric. Food Chem. 21, 871-872.

    acids. J . Agric. Food Chem. 31, 11 15-1 117.

    2,880,094 (Chem. Abstr. 53, 9516b).

    Chem. 37, 1509-1510.

    flavor agents. U.S. Patent 3,579,353 (Chem. Absrr. 75, 19009).

    U.S. Patent 3,619,210 (Chem. Absrr. 76, 57967).

    Ohwa, T. 1972. Meatlike flavor. Japan. Patent 72 50387 (Chem. Absrr. 79, 145059). Orsi, F. 1981. Derivatographic studies on the Maillard reaction. Nuhrung 25, 519-529. Otto, R., and Bakes, W. 1980. Model experiments for the Maillard reaction. 111. Volatile products

    from the reaction of D-glucose with 4-chloroaniline. Z. Lebensm. Unters. Forsch. 171, 286- 291.

    Otto, R., and Bakes, W. 1981. Model experiments for the Maillard reaction. IV. Production of quinoline from the reaction of hydroxymethylfurfural with p-chloroaniline. 2. Lebensm. Unrers. Forsch. 172, 286-287.

    Patton, S. 1950. The formation of maltol in certain carbohydrate-glycine systems. J . Biol. Chem.

    Pearson, A. M., Harrington, G., West, R. G., and Spooner, M. E. 1962. Browning produced by heating fresh pork. I. The relation of browning intensity to chemical constituents and pH. J. Food Sci. 27, 177-181.

    Peer, H. G., van den Ouweland, G. A. M., and de Groot, C. N. 1968. The reaction of aldopentoses and secondary amine salts, a convenient method of preparing 4-hydroxy-5-methyl-2.3-di- hydrofuran-3-one. Rec. Truv. Chim. 87, 101 1-1016.

    Pickenhagen, W., Dietrich, P., Keil, B., Polonsky, J., Novaille, F., and Lederer, E. 1975. Identifi- cation of the bitter principle of cocoa. Helv. Chim. Acru 58, 1078-1086.

    Pintauro, S. J., Page, G. V., Solberg, M., Lee, T.-C., and Chichester, C. 0. 1980. Absence of mutagenic response from extracts of Maillard browned egg albumin. J . Food Sci. 45, 1442- 1443.

    Pintauro, S. J., Lee, T.-C., and Chichester, C. 0. 1983. Nutritional and toxicological effects of Maillard browned protein ingestion in the rat. I n The Maillard Reaction in Foods and Nutri- tion (G. R. Waller and M. S . Feather, eds.). ACS Symp. Ser. (215). 467-483.

    Pittet, A. O., and Hruza, D. E. 1974. Comparative study of flavor properties of thiazole derivatives. J. Agric. Food Chem. 22, 264-269.

    Pittet, A. O., Rittersbacher, P., and Muralidhara, R. 1970. Flavor properties of compounds related to maltol and isomaltol. J. Agric. Food Chem. 18, 929-933.

    Pokorny, J. 1980. Effect of browning reactions on the formation of flavor substances. Nuhrung 24, 115-127.

    184, 131-134.


    Pokorny, J., Tai, P. T., and Janisek, G. 1973. Non-enzymic browning. 4. Browning reactions of 2- furfuraldehyde with protein. 2. Lebensm. Unters. Forsch. 151, 36-40.

    Pokorny, J., Con, N. H., Bulantova, H., and Janicek, G. 1974. Changes in amino acids and reducing sugars during the roasting of coffee. Nahrung 18, 799-805.

    Pokomy, J., Con, N. H., Smidrkalova, E., and Janicek, G. 1975. Nonenzymic browning. XII. Maillard reactions in green coffee beans on storage. Z . Lebensm. (Inters. Forsch. 158, 87-92.

    Pokorny, J., Luan, N. T., Kondratenko, S. S., and Janicek, G. 1976. Changes in sensory value by interaction of alkanals with amino acids and proteins. Nahrung 20, 267-272.

    Pokomy, J. , Svobodova, H., and Janicek, G. 1977. Reaction of lower alkanals with proteins. Correlation of flavor changes and nonenzymic browning during storage of model systems. Sb. Vys. Sk. Chem.-Technol. Prare, Potraviny E49, 5-21.

    Pokorny, J., Luan, N. T., Bulantova, H., and Davidek, J. 1979a. Reaction of D-arabino-hexosulose with L-lysine. Nahrung 23, 693-700.

    Pokorny, J . , Dvorakova, L., Marcin, A., Bulantova, H., and Davidek, J. 1979b. Sensory profiles of reaction products between amino acids and D-fructose or D-arabino-hexosulose. Nahrung 23,

    Pokorny, J., Bulantova, H., and Davidek, J. 1981. Sensory profiles of browning products from reaction of 2-furaldehyde with L-lysine. Nahrung 25, 531-541.

    Reineccius, G. A., Andersen, D. A., Kavanaugh, T. E., and Keeney, P. G. 1972a. Identification and quantification of the free sugars in cocoa beans. J. Agric. Food Chem. 20, 199-202.

    Reineccius, G. A., Keeney, P. G., and Weissberger, W. 1972b. Factors affecting the concentration of pyrazines in cocoa beans. J. Agric. Food Chem. 20, 202-206.

    Reyes, F. G. R., Poocharoen, B., and Wrolstad, R. E. 1982. Maillard browning reaction of sugar- glycine model systems: Changes in sugar concentration, color, and appearance. J . Food Sci. 47,

    Reynolds, T. M. 1963. Chemistry of nonenzymic browning. I. The reaction between aldoses and

    Rhee, K. S., and Rhee, K. C. 1981. Nutritional evaluation of the protein in oilseed products heated

    Rizzi, G. P. 1967. Occurrence of simple alkylpyrazines in cocoa butter. J. Agric. Food Chem. 15,

    Rizzi, G . P. 1968. Some reactions of methylpyrazines with organolithium reagents. J . Org. Chem.

    Rizzi, G . P. 1971. Aldimines as chocolate-like flavors. US. Patent 3,625,710 (Chem. Abstr. 76, 84684).

    Rizzi, G. P. 1972. A mechanistic study of alkylpyrazine formation in model systems. J . Agric. Food Chem. 20, 1081-1085.

    Rizzi, G. P. 1974. Formation of N-alkyl-2-acylpyrroles and aliphatic aldimines in model nonenzymic browning reactions. J. Agric. Food Chem. 22, 279-282.

    Roberts, D. L. 1968. Pyrazine derivatives for improving flavor and aroma. U. S. Patent 3,402,051 (Chem. Abstr. 70, 27784).

    Rodin, J . O., Himel, C. M., Silverstein, R. M., Leeper, R. W., and Gortner, W. A. 1965. Volatile flavor and aroma constituents of pineapple. I. Isolation and tentative identification of 2,5- dimethyl-4-hydroxy-3(W)-furanone. J. Food Sci. 30, 280-285.

    Rohan, T. A. 1958a. Processing of raw cocoa. I. Small-scale fermentation. J. Sci. Food Agric. 9, 104-111.

    Rohan, T. A. 1958b. Processing of raw cocoa. 11. Uniformity in heap fermentation and development of methods for rapid fermentation of West African Amelonado cocoa. J. Sci. Food Agric. 9,



    amines. Adv. Food Res. 12, 1-52.

    with sugars. J . Food Sci. 46, 164-168.

    549-55 1.

    33, 1333-1337.

    542-55 1. Rohan, T. A. 1963. Precursors of chocolate aroma. J . Sci. Food Agric. 14, 799-805.


    Rohan, T. A. 1964. The percursors of chocolate aroma: A comparative study of fermented and unfermented cocoa beans. J. Food Sci. 29, 456-459.

    Rohan, T. A. 1965. Gas chromatography in the quantitative assessment of the chocolate aroma .potential of unroasted cocoa beans. Food Technol. 19, 1726- 1728.

    Rohan, T. A. 1967. The precursors of chocolate aroma: Application of gas chromatography in following formation during fermentation of cocoa beans. J. Food Sci. 32, 402-404.

    Rohan, T. A. 1972. The chemistry of flavour. In Phytochemical Ecology (J. B. Harborne, ed.), Proc. Phytochem. SOC. Symp., Surrey, I971 pp. 57-69.

    Rohan, T. A., and Stewart, T. 1965. The precursors of chocolate aroma: The distribution of free amino acids in different commercial varieties of cocoa beans. J . Food Sci. 30, 416-419.

    Rohan, T. A., and Stewart, T. 1966a. The precursors of chocolate aroma: Changes in the free amino acids during the roasting of cocoa beans. J. Food Sci. 31, 202-205.

    Rohan, T. A,, and Stewart, T. 1966b. The precursors of chocolate aroma: Changes in the sugars during the roasting of cocoa beans. J. Food Sci. 31, 206-209.

    Rohan, T. A., and Stewart, T. 1967a. The precursors of chocolate aroma: Production of free amino acids during fermentation of cocoa beans. J . Food Sci. 32, 395-398.

    Rohan, T. A,, and Stewart, T. 1967b. The precursors of chocolate aroma: Production of reducing sugars during fermentation of cocoa beans. J. Food Sci. 32, 399-402.

    Rohan, T. A,, and Stewart, T. 1967c. Precursors of chocolate aroma: Studies in the degradation of amino acids during the roasting of Accra cocoa beans. J . Food Sci. 32, 625-629.

    Rooney, L. W., Salem, A , , and Johnson, J. A. 1967. Studies of the carbonyl compounds produced by sugar-amino acid reactions. I. Model systems. Cereal Chem. 44, 539-550.

    Rothe, M. 1960. Volatile aromatic substances in rye bread. Ernahrungsforschung 5 , 131- 142. Rothe, M., and Thomas, B. 1959. The formation, composition, and determination of aromatic

    substances of bread. Nuhrung 3, 1-17. Rothe, M., and Thomas, B. 1963. Aromastoffe des Brotes. Versuch einer Auswertung chemischer

    Geschmacksanalysen mit Hilfe des Schwellenwertes. 2. Lebensm. Untersuch. Forsch. 119,

    Rothe, M., and Voight, I. 1963. Browning and aroma production in the Maillard reaction. Nuhrung

    Rothe, M., Woelm, G., and Bethke, E. 1973. Flavoring white bread. East German Patent 99293 (Chem. Abstr. 81, 90127.

    Sakaguchi, M., and Shibamoto, T. 1978a. Formation of heterocyclic compounds from the reaction of cysteamine and D-glucose, acetaldehyde, or glyoxal. J. Agric. Food Chem. 26, 1179-1 183.

    Sakaguchi, M., and Shibamoto, T. 1978b. Formation of sulfur-containing compounds from the reaction of D-glucose and hydrogen sulfide. J. Agric. Food Chem. 26, 1260-1262.

    Salem, A., Rooney, L. W., and Johnson, J. A. 1967. Studies of the carbonyl compounds produced by sugar-amino acid reactions. 11. In bread systems. Cereal Chem. 44, 576-583.

    Sapers, G. M. 1975. Flavor and stability of dehydrated potato products. J . Agric. Food Chem. 23,

    Sato, K., and Hegarty, G. R. 1974. Inhibition of warmed-over flavor in meats. U.S. Patent

    Sato, K., and Hegarty, G. R. 1976. Inhibition of warmed-over flavor in meats. U.S. Patent

    Schleppnik, A. A. 1969. 4-Pyrones. U.S. Patent 3,474,113. (Chem. Abstr. 72, 3378). Schleppnik, A. A., and Oftedahl, M. L. 1969. 4-Pyrones as flavor enhancers. U.S. Patent 3,455,960

    Schleppnik, A. A., and Oftedahl, M. L. 1970a. 2-Methyl-3-hydroxy-4-pyrone. U.S. Patent


    7, 50-59.

    1027- 1032.

    3,821,444 (Chem. Absrr. 82, 15527).

    3,939,288 (Chem. Abstr. 84, 134271).

    (Chem. Abstr. 72, 21609).

    3,491,122 (Chem. Abstr. 72, 78880).


    Schleppnik, A. A., and Oftedahl, M. L. 1970b. Flavor enhancing 2-methyl-3-hydroxyhydro-4- pyrones. U.S. Patent 3,494,959 (Chem. Absrr. 72, 132521).

    Schroedter, R., and Woelm, G . 1980. Studies on optimal conditions for flavor formation in amino acid-glucose model systems. Nahrung 24, 175-183.

    Schreiber, W. L., Pittet, A. O., and Vock, M. H. 1974. Flavor properties of phenylpentenals. J. Agric. Food Chem. 22, 269-273.

    Schumacher, I. N., Best, F. W., and Green, C. R. 1976. Improving the aroma or taste of a tobacco product or food. Belg. Patent 841,948 (Chem. Absrr. 86, 186179).

    Schutte, L., and Koenders, E. B. 1972. Components contributing to beef flavor. Natural precursors of 1-methylthioethanethiol. J . Agric. Food Chem. 20, 181-184.

    Severin, Th., and Braeutigam, K. H. 1973. Maillard reaction. VII. Formation of p-carboline deriva- tives by reaction of tryptamine with xylose and glucose. Chem. Ber. 106, 2943-2949.

    Severin, Th., and Loidl, A. 1976. Formation of pyridone derivatives from maltose and lactose. XII. Investigations on the Maillard reaction. Z . Lebensm. Unters. Forsch. 161, 119-124.

    Severin, Th., and Seilmeier, W. 1967. Studien zur Maillard-Reaktion. 11. Umwandlung von Pen- tosen unter dem Einfluss von Aminacetaten. Z. Lebensm. Unrersuch. -Forsch. 134, 230-232.

    Shaw, P. E. Tatum, I. H., and Berry, R. E. 1971. 2,3-Dihydro-3,5-dihydroxy-6-methyl-4H- pyran-4-one, a degradation product of a hexose. Carbohydr. Res. 16, 207-21 1.

    Shibamoto. T. 1980a. Heterocyclic compounds found in cooked meats. J. Agric. Food Chem. 28,

    Shibamoto, T. 1980b. Mutagenicity of 1,5(or7)-dimethyl-2,3,6,7-tetrahydro-lH,5H-biscyclopen- tapyrazine obatined from a cyclotene/NH3 browning model system. J . Agric. Food Chem. 28,

    Shibamoto, T. 1982. Occurrence of mutagenic products in browning model systems. Food Technol. March, 59-62.

    Shibamoto, T., and Bernhard, R. A. 1976. Effect of time, temperature, and reactant ratio on pyrazine formation in model systems. J. Agric. Food Chem. 24, 847-852.

    Shibamoto, T., and Bernhard, R. A. 1977a. Investigation of pyrazine formation pathways in glucose-ammonia model systems. Agric. Biol. Chem. 41, 143- 153.

    Shibamoto, T., and Bemhard, R. A. 1977b. lnvestigation of pyrazine formation pathways in sugar- ammonia model systems. J . Agric. Food Chem. 25, 609-614.

    Shibamoto, T., and Bemhard, R. A. 1978. Formation of heterocyclic compounds from the reaction of L-rhamnose with ammonia. J. Agric. Food Chem. 26, 183-187.

    Shibamoto, T., and Russell, G. F. 1976. Study of meat volatiles associated with aroma generated in a o-glucose-H2S-NH3 model system. J. Agric. Food Chem. 24, 843-848.

    Shibamoto, T., and Russell, G. F. 1977. A study of the volatiles isolated from a o-glucose-hydrogen sulfide-ammonia model system. J. Agric. Food Chem. 25, 109-1 12.

    Shibamoto, T., Akiyama, T., Sakaguchi, M., Enomoto, Y., and Masuda, H. 1979. A study of pyrazine formation. J. Agric. Food Chem. 27, 1027-1031.

    Shibamoto, T., Nishimura, 0.. and Mihara, S. 1981. Mutagenicity of products obtained from a maltol-ammonia browning model system. J . Agric. Food Chem. 29, 643-646.

    Shigematsu, H., Shibata, S., Kurata, T., Kato, H., and Fujimaki, M. 1975. 5-AcetyI-2,3-di- hydro-1H-pyrrolizines and 5,6,7,8-tetrahydroindilizin-8-ones, odor constituents formed on heating L-proline with o-glucose. J. Agric. Food Chem. 23, 233-237.

    SocietB des Roduits Nestle, S. A. 1969. Synthesis of 2-ethyl-3.5.6-trimethylpyrazine and its use as a cocoa flavor enhancer. French Patent 1,576,580 (Chem. Absrr. 72, 120279).

    Spingarn, N. E., and Garvie, C. T. 1979. Formation of mutagens in sugar-ammonia model systems. J . Agric. Food Chem. 27, 1317-1321.




    Spingam, N. E., Slocum, L. A,, and Weisburger, J. H. 1980a. Formation of mutagens in cooked foods. 11. Foods with high starch contents. Cancer Len. 9, 7-12.

    Spingam, N. E., Kasai, H., Vuolo, L. L., Nishimura, S., Yamaizumi, Z., Sugimura, T., Mat- sushima, T., and Weisburger, J. H. 1980b. Formation of mutagens in cooked foods. 111. Isolation of a potent mutagen from beef. Cancer Len. 9, 177-183.

    Spingam, N. E., Garvie-Could, C. T., and Slocum, L. A. 1983. Formation of mutagens in sugar- amino acid model systems. J . Agric. Food Chem. 31, 301-304.

    Stephens, C. R., and Torres, A. 1971. Flavoring agents containing 2-hydroxy-3-ethyl-2-cyclo- penten-I-one. US. Patent 3,628,970 (Chem. Absrr. 76, 152329).

    Stephens, C. R. Torres, A., and Leir, C. M. 1970. 2-Hydroxy-3-alkyl-2-cyclopentenones as flavor- ing compounds. German Patent 2,005,160 (Chem. Absrr. 73, 98467).

    Stewart, G. F., and Kline, R. W. 1948. Factors influencing the rate of deterioration in dried egg albumen. Ind . Eng. Chem. 40,916-919.

    Stoll, M., Dietrich, P., Sundt, E., and Winter, M. 1967a. Sur Iarome du cacao. 11. Helv. Chim. Acra 50, 2065-2067.

    Stoll, M., Winter, M.. Gautschi, F., Flament, I., and Willhalm, B. 1967b. Sur Iarome de cafk. I. Helv. Chim. Acta 50, 628-694.

    Sulser, H. 1973. Importance of fructose lysins and their decomposition products, furosine and pyridosine for the quality control of foods. Lebensm. Wiss. Technol. 6, 66-69.

    Takken, H. I., van der Linde, L. M., Boelens, M., and van Dort, J. M. 1975. Olfactive properties of a number of polysubstituted pyrazines. J. Agric. Food Chem. 23, 638-642.

    Tanaka, M., Lee, T.-C., and Chichester, C. 0. 1977. Effect of Maillard browning reaction on nutritional quality of protein. Adv. Exp. Med. Eiol. 86B, 321-341.

    Thomas, B., and Rothe, M. 1957. Die Bildung von aldehydaromatischen Verbindungen wahrend des Backenprozess des Brotes. Erniihrungsforschung 2, 427-443.

    Toda, H., Sekizawa, J. , and Shibamoto, T. 1981. Mutagenicity of the L-rhamnose-ammonia- hydrogen sulfide browning reaction mixture. J. Agric. Food Chem. 29, 381-384.

    Tonsbeek, C. H. T., Plancken, A. J. , and van der Weerdhof, T. 1968. Components contributing to beef flavor. Isolation of 4-hydroxy-5-methyl-3(2H)-furanone and its 2,5-dimethyl homolog from beef broth. J . Agric. Food Chem. 16, 1016-1021.

    Tonsbeek, C. H. T., Koenders, E. B., von der Zijden, A. S. M., and Losekoot, Y. A. 1969. Components contributing to beef flavor. Natural precursors of 4-hydroxy-5-methyl-3(2H)- furanone in beef broth. J. Agric. Food Chem. 17, 397-400.

    Underwood, J. C., and Filipic, V. J. 1963. Gas chromatographic identification of compounds in maple syrup flavor extract. J. Assoc. Off. Agric. Chem. 46, 334-337.

    Underwood, J. C., and Filipic, V. J. 1964. Source of aromatic compounds in maple flavor. J. Food Sci. 29, 814-818.

    Underwood, I . C., and Filipic, V. J. 1965. Determination of syringaldehyde and dihydroconiferyl alcohol in maple syrup. J . Assoc. Off. Agric. Chem. 48, 689-693.

    Underwood, J. C., Willits, J. C., and Lento, H. C. 1961a. Maple syrup. XVI. Isolation and identification of compounds contributing to the flavor of maple syrup. J. Food Sci. 26, 288- 290.

    Underwood, J. C., Willits, C. O., and Lento, H. G. 1961b. Browning of sugar solutions. VI. Isolation and characterization of the brown pigment in maple syrup. J . FoodSci. 26, 397-400.

    van der Wal, B., Sipma, G., Kettenes, D. K., and Semper, A. T. J. 1968. Some new constituents of roasted cocoa. Rec. Trav. Chim. 87, 238-240.

    van der Wal, B., Kettenes, D. K., Stoffelsma, J. , Sipma, G., and Semper, A. T. J. 1971. New volatile components of roasted cocoa. J. Agric. Food Chem. 19, 276-280.


    van Praag, M., and Bidmead, D. S. 1969. Cocoa and chocolate flavor enhancers. French Patent 1,559,547 (Chem. Absrr. 72, 41960).

    van Praag, M., Stein, H. S., and Tibbetts, M. S. 1968. Steam volatile aroma constituents of roasted cocoa beans. J . Agric. Food Chem. 16, 1005-1008.

    Velikk, J., Davidek, J., Cuhrova, J., and Kubelka, V. 1976. Volatile heterocyclic compounds in the reaction of glyoxal with glycine. J. Agric. Food Chem. 24, 3-7.

    Vitzthum, 0. G., Werkhoff, P., and Hubert, P. 1975. Volatile components of roasted cocoa. Basic fraction. J. Food Sci. 40, 91 1-916.

    Waller, G. R., and Feather, M. S., eds. 1983. The Maillard Reaction in Foods and Nutriton. ACS Symp. Ser. (215).

    Wallradt, J. P., Lindsay, R. C., and Libbey, L. M. 1970. Popcorn flavor: Identification of volatile compounds. J. Agric. Food Chem. 18, 926-928.

    Wang, P., and Odell, G. V. 1973. Formation of pyrazines from thermal treatment of some amino- hydroxy compounds. J. Agric. Food Chem. 21, 868-870.

    Wasserman, A. E. 1972. Thermally produced flavor components in the aroma of meat and poultry (review). J . Agric. Food Chem. 20, 737-741.

    Wasserman, A. E., and Gray, N. 1965. Meat flavor. 1. Fractionation of water-soluble flavor precur- sors of beef. J . Food Sci. 30, 801-807.

    Wasserman, A. E., and Spinelli. A. M. 1970. Siigar-amino acid interactions in the diffusate of water extract of beef and model systems. J. Food Sci. 35, 328-332.

    Wasserman, A. E., and Spinelli, A. M. 1972. Effect of some water-soluble components on aroma of heated adipose tissue. J. Agric. Food Chem. 20, 171-174.

    Watanabe, K., and Sato, Y. 1972. Shallow-fried beef Additional flavor components. J. Agric. Food Chem. 20, 174-176.

    Wei, C.-L., Kitamura, K., and Shibamoto, T. 1981. Mutagenicity of Maillard browning products obtained from a starch-glycine model system. Food Cosmet. Toxicol. 19,749-751; cf. Chem. Absrr. 96, 157053.

    Westphal, G., and Cieslik, E. 1981. Studies on the Maillard reaction. 11. Reaction of glucose with phenylalanine in water. Nahrung 25, 749-757.

    Westphal, G., and Kroh, L. 1981. Studies on the Maillard reaction. 1. Structure and reactivity of glucosylthiourea. Nahrung 25, 31 1-319.

    Wick, E. L., de Figuereido, M., and Wallace, D. H. 1964. The volatile components of white bread prepared by a preferment method. Cereal Chem. 41, 300-315.

    Wilson, R. A. 1975. A review of thermally produced imitation meat flavors. J. Agric. Food Chem.

    Wilson, R. A,, and Katz, I. 1972. Review of literature on chicken flavor and report of isolation of several new chicken flavor components from aqueous cooked chicken broth. J. Agric. Food Chem. 20, 741-747.

    Wilson, R. A, , Mussinan, C. J., Katz, I., and Sanderson, A. 1973. Isolation and identification of some sulfur chemicals present in pressure-cooked beef. J . Agric. Food Chem. 21, 873-876.

    Wiseblatt, L. 1960a. Aromatic compounds present in oven gases. Cereal Chem. 37, 728-733. Wiseblatt, L. 1960b. The volatile organic acids found in dough, oven gases, and bread. Cereal

    Wiseblatt, L. 1961. Bread flavor research. Bakers Dig. (5 ) . 60-63, 174. Wiseblatt, L., and Kohn, F. E. 1960. Some volatile aromatic compounds in fresh bread. Cereal

    Wiseblatt, L., and Zoumut, H. F. 1963. Isolation, origin, and synthesis of a bread flavor constituent.

    23, 1032-1037.

    Chem. 37,734-739.

    Chem. 37,55-66.

    Cereal Chem. 40, 162-9.


    Wood, T. 1961. The browning of ox-muscle extracts. J . Sci. Food Agric . 12, 61-69. Yoshida, D., and Okamoto, H. 1980. Formation of mutagens by heating the aqueous solutions of

    amino acids and some nitrogenous compounds with addition of glucose. Agric . Biol. Chem. 44,

    Zaika, L. L., Wasserman, A. E., Monk, C. A,, and Salay, J . 1968. Meat flavor. 2. Procedures for 252 1-2522.

    the separation of water-soluble beef aroma precursors. J . Food Sci. 33, 53-58.




    *Department of Biochemistry. Royal Holloway College,

    University of London, Eghum Hill, Egham, Surrey T W Z O OEX, England

    I. 11.

    Introduction . . . . . . . . . . . . . . . . . . . . . . . . Components of Primary Cell Wall . . . . . . A. Pectic Polysaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Hemicellulose C. Cellulose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Glycoprotein . . . . . . . Structure of Primary Cell Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Polymer Interconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Cell Wall Models . . . . . . .

    D. A Suggested Modified Model

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


    C. Aspects of the Models . . . . . . . . . . . . . . . . . . . . . . . . . . .

    A. Stages in Fruit Development . . . . . . . . . . . . . . .

    V. Concluding Remarks ...........................

    IV. Fruit Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    B. Metabolic Aspects of Fruit C. Cell Wall Breakdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . .

    139 140 141 145 147 147 149 149 152 156 163 168 168 171 174 178 180


    From a physiological point of view, a fruit can be defined as the structural entity resulting from the development of the tissue that supports the ovule. The

    tPresent address: Department of Biochemistry, University of the West Indies, Mona, Kingston 7, Jamaica.

    139 Copyright 8 1986 by Academic Press, Inc.

    All rights of reproduction in any form reserved.


    function of a fruit is, in principle, to assist in distribution of seeds. This is often achieved by consumption of the ripe fruit and scattering of the residual seeds or by passage of seeds through the digestive tract of the animal consuming them. Premature scattering of seeds is thought to be discouraged by the indigestibility of unripe fruits.

    Fruits are generally harvested at the mature stage when growth has ceased, and full development may be achieved independent of the parent plant with negligi- ble impairment to quality. In fleshy fruits softening occurs after the mature stage, under favorable conditions, whether the fruit is attached to or detached from the parent plant, and is caused by breakdown of the cell wall structure in the pulp. Unripe fruits generally have rigid, well-defined structures, whereas ripe fruits have soft and diffused cell walls. This change is brought about by the coordi- nated action of hydrolytic enzymes on the cell wall and occurs together with biochemical and physiological activities within the cell which convert the fruit from an inedible to an edible state.

    Fruits and their products constitute a commercially significant food commodity. One of the limiting factors to the economic value of fruit crops is the relatively short duration of ripening and shelf life, with adverse effects on storage. Fruit rot increases with overripeness and senescence and is a major source of commercial loss to the fruit-growing industry. This is especially true of many tropical fruits which are cultivated in climatic conditions that facilitate rapid ripening and infection.

    Elucidation of the mechanism of cell wall breakdown that causes softening in fruits is thus of potential value to food technologists seeking chemical or physio- logical means of delaying or retarding the ripening process. The achievement of this is dependent upon an understanding of cell wall structure and an appreciation of the biochemistry of its breakdown.


    The plant cell consists of cytoplasm surrounded by a cell wall, as shown in Fig. 1. Each cell is connected to adjacent cells by a pectin-rich middle lamella. The cytoplasm of the cells is interconnected by a plasmadesmata which in the ripe fruit is thought to give a degree of cohesion. As long as a cell contrives to undergo growth, its wall remains relatively thin. In this stage of development it is called a primary cell wall and consists of roughly 90% polysaccharide and 10% protein. This is distinct from secondary walls which, in some regions of the plant, such as the stem, are deposited after growth has stopped and are responsi- ble for giving a woody texture. Cells present in fruit pulp are generally thought to contain only primary cell walls.

    The primary cell wall constituents can be divided into pectic polysaccharides


    FIG. I . Diagrammatic representation of a plant cell (after Jewell, 1979). A, Amyloplast; C, chro- matin; Ch, chloroplast; CW, cell wall; Cy, cytoplasm; D, dictyosome, several of which constitute the Golgi apparatus; ER, endoplasmic reticulum with attached ribosomes, R; I , intercellular space; Mc, mitochondrion; Mf, microfibril; MI, middle lamella; Mt, microtubule; N, nucleus; Pd, plas- madesmata; PI, plasmalemma; PV, pinocytotic vesicle, an infolding of the cell membrane; S, spherosome; T, tonoplast; V, vacuole; and Ve, vesicle.

    (34%), hemicellulose (24%), cellulose (23%), and hydroxy proline-rich glyco- protein (19%). The ratio of these constituents varies with cell walls from differ- ent plant sources. The percentage values given are for the walls of suspension- cultured sycamore cells (McNeil et al., 1979). Structures have been proposed for various regions of each of the above fractions which have received varying degrees of acceptance and will be discussed in the following sections.


    Pectic polysaccharides are generally considered to be those portions of cell walls which can be extracted by a variety of mild methods such as hot water, ammonium oxalate solution, weak acids, chelating agents, and endopolygalac-


    turonase. These reagents also extract varying amounts of other fractions from the cell wall, which creates a certain amount of ambiguity. The quantity extractable varies from 20 to 30% of the walls of meristematic and parenchymatous tissues. It is accepted that pectin is essentially composed of D-galacturonic acid, L- rhamnose, L-arabinose, and D-galactose, the proportions of these varying with the source. Other associated sugars are D-xylose, 2-O-methyl-~-fucose, D-api- ose, and D-glucornic acid.

    Fragments of pectic polysaccharide molecules have been isolated following removal from cell walls. Analysis of the pectic fraction suggests the presence of structurally different regions in the polymer. These are referred to as rham- nogalacturonan I, rhamnogalacturonan 11, homogalacturonan, arabinans, galac- tans, and arabinogalactans.

    1. Rhamnogalacturonan I

    Rhamnogalacturonan I accounts for -7% of primary cell walls. These regions of pectin are large, with estimated degrees of polymerization of 2000. The proposed structure for this fragment is shown in Fig. 2, which depicts a rham- nogalacturonan backbone to which side chains containing neutral sugars, L- arabinose, and D-galactose are linked via L-rhamnosyl residues. This has been demonstrated by Aspinall and co-workers (Aspinall er al. , 1967b, 1968a,b; Aspinall and Molloy, 1968; Aspinall and Jiang, 1974) with work on pectin from rapeseed hull, soybean cotyledon, lucerne leaves, lucerne stems, and lemon peel. Galacturonic acid residues in the backbone are attached to each other by a-1 ,4-glycosidic linkages. Analysis of rapeseed (Aspinall and Jiang, 1974) and lucerne (Aspinall and Molloy, 1968) pectin demonstrated that half of the rham- nosy1 residues in the polymer are linked via their C-2 (Fig. 2) and the rest via C-2 and C-4 (branch points; Fig. 2). As no aldobiuronic acid with a galacturonosyl residue attached to C-4 of a rhamnosyl residue has been identified, it has been assumed that this is the point of attachment of the side chains of arabinans, galactans, or arabinogalactans. L-Rhamnose, D-galacturonic acid, L-arabinose,

    4 t

    s ide chain

    4 4

    s ide chain

    FIG. 2. A possible structure of pectic rhamnogalacturonan I of primary cell walls of dicots. Linear sequences of galacturonic acid (GA) are interrupted by rhamnosyl (Rha) residues, half of the latter are attached to neutral sugar side chains. The side chains may also be attached to C-3 of some galacturonic acid residues. (Based on work described by Darvill et al., 1980a.)


    and D-galactose were found to occur in the ratio of 1:2: 1.5: 1.5 (McNeil et al., 1980). No information is available on whether the rhamnosidic bonds are in the a- or P-configuration.

    2 . Homogalacturonan

    Homogalacturonan accounts for -6% of the cell wall. Pectic polysaccharides contain regions of unbranched a- 1 ,Clinked galacturonosyl residues in addition to high-molecular-weight rhamnogalacturonan I (English et al., 1972; Talmadge el al., 1973). This is apparent from experiments showing that endopolygalac- turonase can successfully hydrolyze regions of pectin accounting for -5% of cell wall to give fragments of mono-, di-, and trigalacturonic acids. These products presumably arise from the hydrolysis of domains within large pectic polymers containing a- 1 ,Clinked galacturonans. Also classified as homogalacturonans are regions of pectin which are stable to hydrolysis by endopolygalacturonase due to esterification of the uronosyl carboxyl groups, and which do not contain side chains. Homogalacturonan fragments have been isolated from the walls of sus- pension-cultured sycamore cells (Talmadge et al., 1973), apple pectin (Barret and Northcote, 1965), and sunflower seeds (Zitko and Bishop, 1966). The syc- amore polymer had a degree of polymerization greater than 25 (Darvill et al., 1980b).

    3. Rhamnogalacturonan I1

    Rhamnogalacturonan I1 accounts for -3% of the primary cell wall and con- stitutes a small region of pectic polysaccharides. It contains 25-50 glycosyl residues (Aspinall and Cottrell, 1971) and monosaccharides that are rarely ob- served in other polymers, for example, 2-O-methyl-~-xylose and D-apiose. A highly branched structure for this polymer is apparent from the wide variety of terminal glycosyl residues it possesses. The terminal residues detected are D- galacturonic acid, D-galactose, L-arabinose, L-rhamnose, 2-O-methyl-~-fucose, and 2-O-methyl-~-xylose. Also present are 2-linked glucuronosyl, 3-linked api- osyl, 3-linked rhamnosyl, 2,4-linked galactosyl, 3,4-linked rhamnosyl, and 3,4- linked fucosyl residues (Aspinall and Cottrell, 197 1).

    4 . Arabinans

    Arabinans account for -9% of the primary cell walls of dicotyledonous plants, and a purified arabinan has been isolated from the walls of suspension- cultured sycamore cells (Darvill et al., 1980b). All the plant arabinans investi- gated from different sources appear to have similar structures (Hirst and Jones,


    L-A raf - 1 4 2

    - 3 +5) -a-L-Arf- ( 1-5 )-a-&-Araf- ( 1+5) -a-&-Araf- ( 1 +5 )-a-lfAraf- - ( I-$

    ;P 1 1 - L-Araf - L-Araf = FIG. 3. The main structural features of an arabinan from cabbage (after Selvendran, 1983).

    1947; Rees and Richardson, 1966; Aspinall and Cottrell, 1971; Siddiqui and Wood, 1974; Karacsonyi et al., 1975; Joseleau et al., 1977). While regions of unbranched chains with 1,5-linked residues were shown to occur in sycamore pectic fraction (Talmadge et al., 1973), arabinan from mustard cotyledon had branched structures (Rees and Richardson, 1966). An arabinan from willow bark displayed a degree of polymerization of 90 (Karacsonyi et al., 1975), whereas two such polymers found in Rosa glauca bark had degrees of polymerization of 34 and 100 (Joseleau et al., 1977). Linkages that have been detected in plant arabinans include Winked, 3,5-linked, 2,5-linked, and 2,3,5-linked residues (Aspinall and Cottrell, 1971; Jiang and Timmell, 1972; Joseleau et al., 1977). A partial structure of arabinan from cabbage is shown in Fig. 3 (Stevens and Selvendran, 1980).

    5. Galactans

    Galactans, together with arabinogalactans account for -9% of the cell wall. Although galactans have been derived from citrus pectin (Labavitch et al., 1976), white willow (Toman et al . , 1972), and beech (Meier, 1962), homogalac- tans are relatively rare. However, the well-characterized galactans have been from nux vomica seeds (Hirst et al., 1947) and potato tubers (Wood and Siddi- qui, 1972; Ring and Selvendran, 1978, 1981). Galactan preprations that have been obtained generally contain p-1,4 linkages; however, some 1,6 linkages have also been identified (Wood and Siddiqui, 1972). The degrees of poly- merization have been found to range from 33 in white willow to 50 in sycamore (Darvill et al., 1980b). Several pectic polysaccharides have been shown to contain both 1,3- and 1,6-linked galactosyl residues (Talmadge et al., 1973; Darvill et al., 1980b). Many of the galactosyl residues found in pectic polysac- charides are probably not part of homogalactans (Aspinall et al., 1967b; Aspinall and Molloy, 1968; Aspinall and Cottrell, 1970; Talmadge eta!. , 1973; Aspinall and Jiang, 1974; Toman et al., 1975; Siddiqui and Wood, 1976). In rapeseed hull, the galactosyl residues attached to the uronic acid backbone have been


    shown to occur as p-1 ,Clinked dimers rather than as oligosaccharides or poly- mers (Aspinall and Jiang, 1974). There are not sufficient data available at present to propose a structure for primary cell wall galactan.

    6. Arabinogalacran

    It is uncertain whether arabinogalactans generally occur in the primary cell wall of dicotyledonous plants. Evidence supporting their existence in sycamore cell wall came from the work of Albersheims group (Talmadge et al., 1973). This involved analysis of a pectic fraction released by endopolygalacturonase from the walls of suspension-cultured sycamore cells. The main arabinosyl- and galactosyl-containing components of these fractions appeared to come from a p-1 ,Clinked galactan and a highly branched arabinan, but other linkages charac- teristic of arabinogalactans were also found. These included terminal, 3-, 6-, and 3,6-linked galactosyl residues in addition to terminal, 3-, 5 - , and 2,5-linked arabinofuranosyl linkages which were all detected in substantial quantities. Ear- lier work by Northcote has indicated the presence of arabinogalactans in apple (Barrett and Northcote, 1965) and sycamore suspension-cultured cells (Stoddart and Northcote, 1967); however, these had not been rigorously characterized. There appear to be wide variations both in the composition and linkage structure of arabinogalactans. These polysaccharides isolated from rapeseed flour (Larm et al., 1976) contain 90% arabinosyl residues, whereas those from larch contain 88% galactosyl residues (Aspinall et al., 1968~). Of the seven arabinogalactans that have been studied, three have been found to contain rhamnosyl residues (cf. Darvill et al. , 1980a). Arabinogalactans are generally characterized by the pres- ence of significant amounts of 3,6-linked galactosyl residues and terminal a-L- arabinofuranosyl residues. Soybean arabinogalactan, however, has a linear p-1 ,Clinked galactosyl backbone with a-L-arabinosyl-( 1,5)-O-~-arabinose resi- dues linked to C-3 of some of the galactosyl residues (Morita, 1965a,b; Aspinall et al., 1967a). Other arabinogalactans also have a galactan backbone, but with longer arabinosyl side chains.


    Hemicellulose accounts for 24% of the primary cell wall and is considered to be made up of xyloglucans and glucuronoarabinoxylans.

    1. Xyloglucans

    Xyloglucans account for 19% of primary cell walls. Much more is known about the structure of these polysaccharides compared to the relatively more


    kAra 1 b ? 2

    a-g-Xyl p - - I e 6


    a-I=Buc p

    FIG. 4. beans (after Selvendran, 1983).

    The main structural features of the xyloglucan from parenchymatous tissues of runner

    complex pectic polymers. Xyloglucans from different sources share similar structural features (Kooiman, 1961; Hsu and Reeves, 1967; Aspinall et al . , 1969, 1977; Siddiqui and Wood, 1971, 1977; Bauer et al., 1973; Wilder and Albersheim, 1973; Courtois and Le Dizet, 1974; Barnoud et al . , 1977). This includes a backbone of p- 1 ,Clinked D-glucosyl residues with single D-xylosyl residues bonded via cu-l,6 linkages to the backbone, as shown in Fig. 4. Some of the xylosyl residues are substituted with p-1 ,2-linked D-galactosyl residues (Bauer et al., 1973; Wilder and Albersheim, 1973; Darvill et al., 1980a). Terrni- nal fucosyl residues are linked to these galactosyl residues. This linkage is believed to be a-L-fucosyl attached to C-2 of the galactose. Attached to a few glucosyl residues of the backbone are arabinopyranose residues regarded to be linked by 1,2 linkages; the anomeric nature of the linkages is not known. Ac- cording to Darvill et a1. (1980a), the present structure of xyloglucans suffer from the following shortcomings:

    1. The arabino-1,2 linkage is based solely on the findings that equimolar amounts of terminal arabinopyranosyl and 2,4,6-substituted glucosyl residues are present.

    2. It is possible that a few glucosyl residues are attached via C-6 of other glucosyl residues and similar numbers of xylosyl residues are attached to C-4 of

    3. The anomeric configurations of the glycosidic linkages in cell wall xy- loglucan have been assumed but not proved to be the same as those occurring in seed xyloglucans (Lamport, 1970).

    4. Fucosyl linkages have been assumed to be in a-anomeric configuration because the linkage is hydrolyzed by an enzyme mixture known to contain an a-1,2-fucosidase (Bahl, 1970; Valent et al., 1980).

    glucosyl .


    2 . Glucuronoarabinoxylan

    Glucuronoarabinoxylan has only recently been identified in the cell wall of suspension-cultured sycamore cells (Darvill et al., 1980b) and acounts for 5% of it. It has been shown to contain terminal 4-, 2,4-, and 3,Clinked D-xylosyl residues, terminal and 2-linked L-arabinofuranosyl residues, terminal D-gluco- soy1 residues, and terminal 4-O-methyl-~-g~ucuronosyl residues. However, the overall structure is not known.


    It is believed that the structure of cellulose which constitutes -23% of the primary cell wall is similar in all plant sources (Preston, 1974; Kolpak and Blackwell, 1976). It is made up of long unbranched chains of p-1,4-linked glucose residues. The chains exist in parallel sheet structures stabilized by in- terchain hydrogen bondings to form cellulose fibers (Gardner and Blackweil, 1974a; Preston, 1974). Estimates from electron microscopic studies give values of 4.5 X 8.3 mm (Preston, 1974) for cross-sectional dimensions of these fibers which are estimated to contain 60-70 glucan chains. Aggregations of the glucan chains within a fiber are so ordered that they are crystalline (Gardner and Black- well, 1974a,b; Preston, 1974; Sarko and Muggli, 1974; Kolpak and Blackwell, 1975, 1976) and can be subjected to X-ray diffraction studies which indicate that they may have parallel orientation with reducing ends facing the same direction. However, antiparallel orientation cannot be excluded (Tucker and Grierson, 1982). Purified cellulose invariably contains, in addition to a preponderance of glucosyl residues, minor amounts of other glycosyl residues (Preston, 1964; Muhlethaler, 1967). The possibility exists that these may be normal constituents of the glucan chains rather than noncovalently but tightly held impurities such as hemicellulose, which is known to form hydrogen bonds to cellulose in vivo.


    Proteins account for 5-10% of dicotyledonous primary cell walls and are generally present in glycosylated form (Blake and Richards, 197 1; Northcote, 1972; Dea et al., 1977; McNeil and Albersheim, 1977). The proteins conatin about 20% of hydroxyproline and relatively large amounts of alanine, serine, and threonine. These are characteristic amino acids of structural proteins of animals (Lamport, 1970) which suggests that they may have a similar role in plant cell walls. Great difficulty is often encountered in extracting the protein from cell walls under nondegradative conditions (Lamport, 1969, 1973). In effect, it has not been possible to isolate the wall glycoprotein without the use of drastic


    a - G A r a L





    4 p-g-Ara - f -


    p - L A r a f = - I

    2 JI

    p - L A r a f - - - 1

    J 4

    Hyp Ser - HYP -

    ? 4 1 1 t

    a - D - G a l p - - - p - L A r a f - - ?

    s 1

    p - L - A r a f - - -


    8 - G A r a f - - - FIG. 5. Proposed structural features of a portion of hydroxyproline-rich glycoprotein of dicot primary cell walls (based on works of Akiyama and Kato, 1976; Lamport, 1977; ONeill and Selvendran, 1980; and Lamport and Catt, 1981).

    methods, which has made it difficult to assess their relationship with other cell wall polymers in the intact wall and to determine their full structures. Most of the detailed structural studies on glycoproteins have utilized cell walls isolated from plant suspension cultures which have higher concentrations of hydroxyline-rich proteins (Lamport, 1965, 1970; Heath and Northcote, 1971). Glycoproteins with low hydroxyproline content have also been known to exist (Brown and Kimmins, 198 1; Ring and Salvendran, 198 1). The structure of hydroxyproline-rich glyco- proteins is characterized by the following features: ( l ) L-Arabinofuranose


    oligosaccharides are 0-glycosidically attached via most hydroxyl groups of hy- droxyproline residues. Little or no unglycosylated hydroxyproline appears to be present (Lamport, 1967; Lamport and Miller, 1971) (see Fig. 5). These side chains vary in length from 1 to 4 sugar residues (Akiyama and Kato, 1976, 1977; ONeill and Selvendran, 1980). (2) Single a-D-galactosyl residues are linked to the hydroxyl groups of serine.

    Analysis of cell wall glycoproteins from a variety of sources shows that hydroxyprolinyl 0-arabinosides and serinyl 0-galactosides are of universal oc- currence (Lamport, 1973, 1980). This conservation of structure points to the wall protein having a unique fundamental role in dicotyledonous plant cell wall.

    Fry (1982b) has implicated a novel phenolic amino acid, isodityrosine, in providing interpolypeptide cross-links in plant cell wall glycoproteins (see also, Lamport and Epstein, 1983; Epstein and Lamport, 1984; Smith et al . , 1984). Such linkages contribute to glycoprotein insolubility (Cooper and Vamer, 1983).


    The way in which the various components of the cell wall are linked together is largely not known. This is due mainly to the limited amount of information currently available on their structures. Present evidence indicates that intercon- nections between some of the fractions exist and on this basis various models have been proposed. These interconnections are described in the following sections.


    1. Links between Pectic Fractions

    It is believed that the various domains of pectic polysaccharides discussed in previous sections are covalently linked together to form complex molecules (see Fig. 6). All of these segments are known to be released from the primary cell wall by the action of polygalacturonase (Aspinall and Cottrell, 1971; Talmadge et al., 1973; Darvill et al., 1980b). There is little doubt that arabinans and galactans are bonded to rhamnogalacturonan I as the whole complex can be isolated by ion-exchange column chromatography. Fragments containing these three characteristic components migrate as a single spot upon electrophoresis (Barrett and Northcote, 1965). Further, no arabinan or galactan has been extract- ed from primary cell walls that is free of galacturonic acids. An egg-box model of noncovalent bonding between pectic polysaccharides in the cell wall involving clacium ions has been proposed (Rees, 1972; Grant et a f . , 1973). This


    +, Arabi-togalactan 1 -+r Highly branched rhamnogalacturonan Arabinan or Galactan - A branched ~alacturonan FIG. 6. Structural features of pectin (after Selvendran, 1983).

    depicts the latter to be fitted between two or more chains of unesterified poly- galacturonan molecules so that they are held by ionic attraction between the calcium ions and the oxygen atoms of 4 galacturonosyl residues. This arrange- ment results in greater rigidity and cross-linking of galacturonan chains (Stoddart ef al., 1967).

    Other types of noncovalent associations may also exist, but their relative importance in the arrangement of the pectic fractions in the cell wall is not known (see also Rees and .Welsh, 1977).

    2 . Links between Xyloglucan and Cellulose

    It has been suggested that xyloglucan, which is the predominant hemicellulose in primary cell walls, is strongly held to the surface of cellulose fibers by hydrogen bondings. This is supported by the following evidence:

    1. The amount of xyloglucan is quantitatively sufficient to form a monolayer coating of the cellulose fibrils (Bauer ef al., 1973; Keegstra et al., 1973).

    2 . It has been shown that xylogucan is capable of forming hydrogen bonds to cellulose (Bauer et al., 1973).


    3. Xyloglucan can be extracted from xyloglucan-cellulose complexes of the cell wall by hydrogen bond-breaking reagents such as dilute base or 8 M urea (Bauer et al., 1973).

    4. Xyloglucan can strongly bind to isolated cellulose in the absence of cata- lysts (Ray and Rottenberg, 1964; Bauer et al., 1973).

    5 . Xyloglucan can bind to the cell wall or to isolated cellulose in a reversible manner (Bauer et al., 1973).

    6. Small fragments of xyloglucan can be extracted from cell walls and sepa- rated from cellulose fibrils by enzymes which degrade xyloglucan into frag- ments. It is assumed that these small polymers are not long enough to form stable hydrogen-bonded complexes with cellulose (Bauer et al . , 1973).

    7. Enzymatically produced short fragments of xyloglucan can be induced to form complexes with cellulose by lowering the water activity of the solvent (Valent and Albersheim, 1974).

    Bonding of xyloglucans to cellulose has been suggested to prevent aggregation of cellulose fibers (Darville et al., 1980a).

    3. Links between Xyloglucan and Pectin

    Attempts at preparing significant amounts of xyloglucan covalently attached to pectic polysaccharides have not been successful (McNeil and Albersheim, 1980), even though it has been indicated that these two fractions are covalently bonded together (Bauer et al . , 1973; Keegstra et al., 1973; Talmadge et al . , 1973). The relative importance of this sort of interconnection within the cell wall is thus uncertain (see also Chambat et al., 1984; Joseleau and Chambat, 1984).

    4 . Glucuronoarabinoxylan

    Glucuronoarabinoxylan is structurally related to arabinoxylans and xylans. The latter two are capable of hydrogen bonding to cellulose (Northcote, 1972; Bauer et al . , 1973; McNeil and Albersheim, 1977). The observation that ara- binoxylans aggregate in solution (possibly forming a mixture of random coils and linear chains) has prompted the suggestion (Darvill et al . , 1980a) that glucuro- narabinoxylan may do likewise and may bind to themselves as well as to cel- lulose in cell walls. Such interaction would assist in the cross-linking of poly- mers in primary cell walls.

    5 . Role of Hydroxyproline-Rich Glycoproteins

    As yet there is no evidence to indicate that hydroxyproline-rich glycoproteins are covalently attached to other fractions within primary cell walls. It is possible


    that linkages that exist are broken under the drastic conditions employed in the extraction of this polymer. In some tissues, alkali extraction of walls releases hydroxyproline-rich glycoproteins (Stoddart et al . , 1967; Monro et al., 1974), whereas in others they are tenaciously held to the cell wall complex (Heath and Northcote, 1971; Selvendran, 1975; Selvendran et al . , 1975). Recently, ONeill and Selvendran ( 1980) released hydroxyproline-rich glycoproteins with mini- mal modification using chlorite/acetic acid solution. They argued that such conditions were likely to hydrolyze phenolic cross-links, but were considerably less effective against glycosidic or peptide linkages. They also proposed that glycoproteins could be held in the cell wall by phenolic cross-links which may be glycoprotein-protein or glycoprotein-polysaccharide in nature (see also Cooper and Varner, 1983).

    The hydroxyproline-rich glycoprotein has regions of helical conformation (Lamport, 1977; Homer and Roberts, 1979) which is likely to give a rodlike molecule (Lamport and Catt, 1981), and thus serve a structural function in the cell. Deglycosylation of the protein causes loss of the ordered conformation (Holtz and Varner, 1984), leading probably to the loss of structural role.

    6. Role of Phenols

    The existence of phenolic materials in the primary cell wall of dicotyledonous plants has been reported recently (Fry, 1979, 1982a, 1983). Very little data are available from which general conclusions about their widespread occurrence can be drawn. Ferulic acid, p-cournaric acid, and other unidentified phenols were shown to be present in the cell wall of spinach cells in suspension culture. Ferulic acid accounts for 0.5% of the cell wall and was suggested to be linked to the nonreducing termini of neutral arabinose and/or galactose-containing regions of the pectic fractions. It was suggested that such residues can cross-link in vivo to form diferuloyl bridges (Fry, 1979, 1983) which would cause a lowering of the walls extensibility and could play a role in resistance to fungal pathogens. However, the significance of the presence of such a relatively small amount of phenolic materials in plant cell walls is open to speculation.


    Any model of primary cell wall structure must be able to account for the following properties: (1) their great strength in withstanding turgor pressure; (2) their ability to grow without loss of strength; and (3) their behavior under chem- ical and enzymatic attack.

    In addition, proposals on cell wall structure should fit in with current observa- tions on wall synthesis. Models have been proposed by Albersheims group


    FIG. 7. (A) Suggested scheme for the structure of the primary wall of sycamore callus cells (drawn by Robinson, 1977, from the works of Keegstra er al., 1973, and Albersheim, 1975, and reproduced with permission). Postulated linkages: (0), covalent bonding; (E), hydrogen bonding. (B) Sug- gested scheme showing interconnections between the polysaccharide fractions in the primary cell wall of dicots (based on work of Albersheim, 1975, 1978). Postulated linkages: (o), covalent bonding; ( 111 ), hydrogen bonding.


    (Keegstra et al., 1973; Albersheim, 1975, 1978) and by Monro et al. (1976b) which are discussed in this section. Other models are not considered because they are in essence embodied by that of Albersheims group. It should be kept in mind that the proposed models were based upon a limited number of plant studies and should therefore be considered as only theoretical models.

    1. Model of Albersheim

    Albersheims model of the primary cell wall of dicotyledonous plants, put forward in 1978, is shown in Fig. 7B. This model is a modification of an earlier model proposed in 1973 (Fig. 7A) and depicts a network of cellulose fibers covered by a monolayer of xyloglucan, the latter being covalently cross-linked to each other by pectic polysaccharides via their neutral side chains.

    2 . Model of Monro

    This model was proposed by MONO and co-workers in 1976 (see Fig. 8) and is similar to that proposed by Albersheim in that it consists of cellulose microfibrils interconnected by a network of polysaccharides. However, it differs from Al- berhseims model in the following features:

    1. A polyuronide bridge is not used in the binding of hemicellulose to other

    2. Cellulose fibers are not ncecessarily covered by a layer of xyloglucan. 3. A large proportion of the hemicellulose is proposed to be bound to the wall

    4. It depicts the interaction of wall protein and hemicellulose with cellulose

    5 . Cellulose microfibrils are at right angles to the direction of elongation.

    components in the wall.

    by alkali-labile covalent bonds.


    3. Appraisal of Albersheims Model

    The model of Albersheim ( 1978) does not include hydroxyproline-rich glycoprotein which makes up a significant part of the primary cell walls. This polymer was excluded from the model on the basis that it has not been shown to be covalently attached to cell wall polysaccharides. The earlier model proposed in 1973 by Albersheims group (Keegstra et al., 1973) depicted hydroxyproline- rich glycoproteins to be covalently bonded to the cell wall via arabinogalactan side chains of the pectic fraction (Fig. 7A). This linkage was proposed on the evidence that a protease was able to solubilize pectic fragments from suspension- cultured sycamore cell walls. However, since then, evidence from other sources


    A B

    FIG. 8. Partial model of primary cell wall in lupin hypocotyl, proposed by Monro er al. (1976b). The half of the figure labeled (A) represents the extensin-hemicellulose network, and the half labeled (B) represents the separate, pectic network, which is believed not to involve the wall glycoprotein (extensin). Thus, the cellulose microfibrils (M) are separately cross-linked by two networks of polymers, the first (A) being composed of the wall glycoprotein and polysaccharide (probably hemicelluloses), and the second (B) being composed of the pectic polymers. These two networks have been separated in the figure for clarity. This model is tentative and incomplete, as the nature of the linkages between the polymers in these two networks has not yet been identified. The symbols used represent extensin (- - -), polysaccharide chain (-), and junction zone (=) between polysac- charide chains.

    has suggested that there is no covalent linkage between cell wall glycoprotein and cell wall polysaccharides (Heath and Northcote, 1971; Bailey and Kauss, 1974; Selvendran et al., 1975; Monro er al., 1976a,b; Mort and Lamport, 1977; ONeill and Selvendran, 1980).

    Although evidence for hydrogen bonding between xyloglucans and cellulose fibers is convincing, there is very little to support the proposed covalent bonding between xyloglucans and neutral side chains of pectin (see Section III,A,3). However, the observation that polygalacturonase-catalyzed removal of pectin facilitates the extraction of a small amount of xyloglucans from cell walls by 8 M urea (Bauer el al., 1973) suggests that there is some degree of interaction (proba- bly noncovalent) between the pectic fraction and xyloglucan.


    The model does not offer an explanation on how the fractions are distributed within the wall. For example, it does not distinguish between the locations of methylated and unmethylated regions of pectin nor does it account for the pres- ence of relatively greater amounts of pectic polysaccharides in the middle lamella (Preston, 1974; Selvendran, 1975). The latter has been studied by ferric hydroxa- mate staining and electron microscopy (Albersheim and Killias, 1963). It is of utmost importance to know how the cellulose fibers within the wall are oriented, as they play a major role in elongation during growth. It is accepted that in the primary cell walls of dividing, nondifferentiated higher plant tissues, the orienta- tion of microfibrils tends to be random. However, at the onset of elongation, the innermost newly formed microfibrils tend to become parallel in a direction transverse to the axis of growth.

    Much of the information used in developing the cell wall model has been derived from analysis of suspension-cultured cells which cannot be totally com- pared with growing plant cells, e.g., the presence or absence of a middle lamella in the former.

    4 . Appraisal of Monro's Model

    This model is based primarily on work on lupin hypocotyl cell walls which involves removal of fractions using varying concentrations of NaOH (Monro er al. , 1976a,b). Such conditions are expected to simultaneously result in trans- elimination of uronic acids (Neukom and Deuel, 1958; Albersheim, 1959), hy- drolysis of methyl galacturonates, p-elimination of seryl glycosides (Spiro, 1970), disruption of hydrogen bondings, and hydrolysis of glycosidic bonds. All these are nonspecific reactions and make interpretations of data extremely diffi- cult. Further, NaOH may cause conformational rearrangement of cell wall com- ponents affecting their extractibility.

    No evidence has been provided to support the proposal that extensin is covalently bonded to cellulose microfibrils at one end and the other polysaccha- rides at the other end.

    No evidence has been provided for the suggested covalent linkage between pectic polysaccharides and cellulose microfibrils as depicted by the model.


    1. Growth

    Cell elongation is highly sensitive to temperature (Ray and Ruesink, 1962; Rayle er al. , 1970b) and can be blocked by metabolic inhibitors such as KCN (Bonner, 1933; Ray and Ruesink, 1962); therefore it cannot be caused by simple


    mechanical stretching. Rapid and apparently normal cell elongation can be in- duced with a lag of less than 1 min in Avena and corn coleoptiles by CO, (Brauner and Brauner, 1943; Evans, 1967; Ray, 1969), low pH (Brauner and Brauner, 1943; Ray, 1969), and indole acetic acid and its ester (Polevoi, 1967; Rayle et al., 1970a). It is believed that cell elongation occurs as a series of independent extension steps. This is apparent from the observation that although cells can elongate under normal conditions for up to 24 hr at a constant rate (Schneider, 1938; Bonner and Foster, 1955), cell walls can only be induced to extend mechanically at a constantly diminishing rate (Cleland, 197 1). Each step of cell wall extension probably involves a biochemical modification of the cell wall in addition to physical extension. Extension cannot result simply from cleavage of polysaccharide molecules interconnecting the cellulose fibers be- cause walls that elongate manyfold would lose most of their strength. Elongated walls have essentially the same strength per unit length as walls that have not elongated. This means that during elongation, polysaccharides interconnecting cellulose fibers must either be augmented by the insertion of new polysaccha- rides or existing cross-links must be broken and the freed ends rejoined to new partners (transglycosylation) (see Albersheim, 1974). It is doubtful, however, that wall expansion is due simply to an increase in the rate of wall synthesis.

    Time-course studies by Baker and Ray (1965) have shown that auxin stimula- tion of wall synthesis is detectable only after a lag of nearly 1 hr; however, as already mentioned, wall elongation can be observed after a lag of 1 min or less. More recently it has been shown that auxin-induced elongation occurs in two biochemically distinct phases (Vanderhoef and Stahl, 1975; Kazama and Kat- sumi, 1976; Vanderhoef e? al., 1976a,b; Vanderhoef, 1979, 1980), an initial rapid stage (probably stimulated by lowering the pH from 6.0 to 4.0), and a slower, more steady phase (probably involving wall synthesis). In support of this, Vanderhoef and Dute (1981) have demonstrated that soybean hypocotyl cell walls kept in a loose state at pH 4.0 undergo only the second stage of elongation when exposed to an exogenous source of auxin (the first stage having already been induced by low pH). A scheme of events involved in cell wall extension postulated by these authors in 1981 is shown in Fig. 9, which proposes that auxin regulates and coordinates both wall loosening and the supply of wall material.

    Two models of cell wall extension have been proposed (Masuda and Sat- omura, 1970; Cleland, 197 1) which accommodate biochemical and physical processes (see Fig. 10). Both assume that growth occurs as a continuous series of independent extension steps. However, they differ in one major aspect, namely, whether the extension itself is irreversible or whether it must be rendered irre- versible by a subsequent biochemical process. In this context it is important to mention that glycoside hydrolases have been demonstrated in the primary cell


    FIG. 9. Proposed events in cell wall extension during elongation growth of pea epicotyls (after Vanderhoef and Dute, 1981). Auxin is postulated to regulate and coordinate both wall loosening and wall synthesis during extension. (A) Elongation in the intact seedling. A continuous supply of auxin keeps the wall loose by maintaining a low wall pH and keeps the cells growing by maintaining the supply of material(s) essential for wall growth. Thus, there is a steady rate of elongation. (B) Growth in an excised, elongating segment. Some 30-90 min after excision, the elongation rate decreases to a low value in the absence of endogenous auxin. Wall pH increases, so that the wall is not maintained in a loosened state, and the synthesis of materials for wall growth is terminated. (C) Acid-induced growth in auxin-depleted, excised segments. Acid is added at the arrow and mimics the wall- loosening component of auxin-regulated elongation, causing a burst of growth. Thus, acid does not induce a steady-state elongation rate; rather, the rate rises after addition of acid and then begins to decline. (D) Auxin-induced growth in auxin-depleted, excised segments. The first observable effect


    HODEL A:

    Rigid wall Biochemicalp Loosened wall ViscoelastiC+ Irreversible extension- Rigid wall

    modification extension

    HODEL B:

    Rigid wall biochemical^ Ioosened wnll , antic Reversible extension modification :tension (loosened wall) modification

    Irreversible+ Rigid wall extension

    FIG. 10. Two proposed models of cell wall extension (after Cleland, 1971).

    walls of plants (Keegstra and Albersheim, 1970; Nevins, 1970; Jaynes et al., 1972; Klis et af., 1974; Parr and Edelman, 1975; Pierrot and Van Wielink, 1977). However, it is not known whether any of these enzymes catalyze trans- glycosylation reactions in viva Ray (1962) has suggested that observed loosen- ing of cell walls caused by auxins may be due to a shift from apposition (deposi- tion of new wall only at the cell membrane) to intussuception (deposition throughout the wall). Intussucepted polysaccharides would then cause losening by forcing the cellulose microfibrils apart or by providing a lubricant to facilitate slippage. It has been shown (Ray, 1967) that in pea stem and Avena coleoptile tissues, wall synthesis was entirely by apposition in the absence of auxin, but that after treatment with auxin a sizable amount of the deposition of hemicellulose, but not cellulose, was found throughout the wall. However, evi- dence of this kind, based on correlation between wall synthesis and growth, cannot be very conclusive. It is difficult to be certain whether a positive correla- tion means a causal relationship between the two processes or whether they are affected in a parallel manner by some other agent. r\

    Both models on cell wall structure can be considered to be compatible with proposals on the mechanism of cell wall extension. Initial extension could be caused by modification of interpolymer bonding which may or may not involve cleavage of polysaccaride molecules. This could be effected by a system that induces a change (e.g., in pH) in the whole or localized regions of the wall. Stimulation of enzyme activity (hydrolytic and transferase) may consequently occur and/or a more direct influence may be brought on noncovalent polymer interactions. The modified wall would then be able to expand under turgor

    of auxin added at the arrow is the burst of growth caused by wall loosening. The elongation rate rises and then begins to fall, with kinetics very similar to those for acid-induced growth. However, the auxin-induced insertion of newly synthesized wall materials begins -50 min after auxin addition, and the rate rises again, eventually reaching a steady-state rate. Thus, the two auxin-regulated phases of elongation growth can be individually observed only when exogenous auxin is added to auxin- depleted segments. Their separation occurs because the lag times for the two phases are different; i.e., auxin-regulated wall acidification occurs with a lag near 15 min, whereas supply of auxin- regulated wall materials begins with a lag near 50 min.


    pressure. Keegstra et al. (1973) have suggested that the hydrogen bonds between xyloglucan and cellulose fibrils in dicotyledonous primary cell walls may be the bonds which are broken during cell wall extension (induced by low pH). This would allow the moving of cellulose fibers relative to each other.

    In 197 1, the relationship between wall acidification and cell elongation was first independently proposed by Hager et al. (1971) and Cleland (1971). The response of growing plant tissues to low pH closely resembles the auxin-induced growth response (Lamport, 1965; Letham, 1967; Cleland, 1971). Polysaccharide hydrolases are known to play an essential role in wall extension in bacteria (Schwarz et a l . , 1969). Evidence for a similar role in extension in all higher plant tissues is considerable. The enzymes p- 1,3-~-glucanase (Datko and Maclachlan, 1968; Heyn, 1969), cellulase (Fan and Maclachlan, 1966; Heyn, 1969), p - 1 , 6 - ~ -glucanase (Heyn, 1970), exogalacturonase (Keegstra and Albersheim, 1970), and nonspecific polysaccharide hydrolases (Katz and Ordin, 1967; Lee et a l . , 1967) have been located bound to wall preparations from various higher plant sources. In tomato fruit (Wallner and Walker, 1975; Gross and Wallner, 1979), p- 1,3-~-glucanase and polygalacturonase are bound to the primary cell wall. Auxin enhances the activities of most of these enzymes in one or more plant tissues (Fan and Maclachlan, 1966, 1967; Datko and Maclachlan, 1968; Davies and Maclachlan, 1969; Heyn, 1970; Keegstra and Albersheim, 1970). Low pH activation, however, has not been established for either of these enzymes in situ.


    2. Chemical Breakdown of Cell Wall

    It is difficult to generalize the composition of cell wall fractions reported in the literature for the basic reason that a variety of preparative procedures have been used. The heterogeneity of the chemically extracted wall fractions (Ray, 1963; Stoddart et a l . , 1967; Monro et a l . , 1976a) is an added problem in this context. In addition, some of the procedures used for the preparation of cell wall materials do not consider water-soluble polymers of the wall which may leach out and are discarded. On the other hand, preparative procedures used may permit con- tamination of the cell wall preparation with intracellular water-soluble polymers. Chemical extraction procedures which have long been used to solubilize classic wall fractions may also cause diverse effects. For example, acid extraction of pectic polymers results in the hydrolysis of arabinosyl and rhamnosyl glycosidic bonds as well as demethylation of the methyl esters of polygalacturonates (Davidson, 1967; Lamport, 1970). Similarly, alkali extraction causes degrada- tive effects.

    3. Endopolygalactouronase Breakdown of Cell Wall

    Most of the information used by Albersheims group in the construction of the cell wall model has been derived from work using endopolygalacturonase, pu-


    rified from Collitotrichum lindemuthianum (Keegstra et al. , 1973; Talmadge et al., 1973; Wilder and Albersheim, 1973; Darvill et al., 1980a). This enzyme hydrolyzes internal a- I ,Clinked galacturonosyl bonds. Incubation of the enzyme with walls of suspension-cultured sycamore cells resulted in its solubilization (- 16%). This action resulted in the removal of 75% of the total galacturonic acid present in the walls. By comparison, Knee (Knee et a / . , 1975), working with a similar enzyme from Sclerotinia fructigena, was able to remove 50% of the total uronide content of apple cell walls. Roughly half of the released material from sycamore cell walls was made up of mono-, di-, and trigalacturonides, the rest consisting of polysaccharides containing both acidic (26%) and neutral (74%) sugars. The neutral sugars, excluding rhamnose, account for 5.5% of the cell wall (or 34.5% of carohydrate solubilized by endopolygalacturonase). Looking at the cell wall model of Albersheim (Fig. 7B), hydrolysis of the rhamnogalac- turonan backbone of the pectic fractions should not release polysaccharide mole- cules containing neutral sugars unless removal of some of the pectic fraction facilitated the dissociation of some noncovalently held xyloglucan from cellulose fibrils. It seems unlikely that this would occur under the mild conditions (50 mM acetate buffer, pH 5.2) employed in the incubation of endopolygalacturonase with the cell walls. In this regard, it is worth noting that 8 M urea was able to release less than 2% of carbohydrate material from cell walls pretreated with endopolygalacturonase (Bauer et al . , 1973). These observations therefore con- flict with the Albersheim model. The ratio of arabinose and galactose (presum- ably from the proposed arabinogalactan side chains of the model) to xylose and glucose (presumably from the xyloglucan fraction) in the endopolygalac- turonase-released fraction is 8: 1. This is not consistent with the observation that pectic polysaccharides account for 35% of sycamore primary cell walls, whereas xyloglucans account for 24%.

    The 1973 model of Albersheim (Fig. 7A) does not propose the linking of each arabinogalactan side chain to each xyloglucan polymer and therefore does not suffer from either of the above criticisms. However, it predicts the solubilization of significant amounts of hydroxyproline-rich glycoproteins by endopolygalac- turonase which should be extracted in conjunction with the removal of 75% of cell wall uronic acid. Such solubilization has not been reported. Further, Stevens and Selvendran (1984) did not observe any protein present in pectic fractions removed from cabbage cell walls, which accounted for 45% of the preparation.

    4. Endoglucanase Breakdown of Cell Wall

    Endoglucanase isolated from Trichoderma viride has been used to hydrolyze xyloglucans from the walls of apple cells (Knee et a l . , 1975) and suspension- cultured cells (Bauer et a l . , 1973). It can only solubilize 1% of unmodified sycamore walls compared to 10-15% of walls pretreated with endopolygalac-


    turonase isolated from C. lindemurhianum. This suggests that the enzyme is unable to penetrate the outer matrix of the wall, which is rich in pectin, in order to reach the site of action further inside the wall. Knee er al. (1975) reported that endoglucanase from T. viride and S . frucrigena released material of varying molecular weights from apple mesocarp cell walls that had been pretreated with endopolygalacturonase. Lower molecular weight fragments accounted for -44% of carbohydrate removed. These contained predominant amounts of galactose (39%) and glucose (33%). The high-molecular-weight polymers, which ac- counted for the other 56% of solubilized carbohydrate from the cell wall, con- tained predominant amounts of galacturonic acid (34%) and arabinose (47%).

    In considering the Albersheim model of 1978 (see Fig. 7B), hydrolysis of the cell wall by endopolygalacturonase followed by endoglucanase (which presum- ably hydrolyzes the xyloglucan fraction) should result in the release of small fragments of xyloglucans consisting essentially of glucose and xylose. The high- molecular-weight polymers released should come from arabinogalactan side chains with attached pieces of rhamnogalacturonan and should contain greater proportions of arabinose, galactose, and galacturonic acid. The presence of large quantities of galactose in low-molecular-weight fractions removed by endo- glucanase from apple cell walls is inconsistent with the model proposed by Albersheim. Further, the small amount of galactose (9%) compared to arabinose (47%) present in the high-molecular-weight peak contradicts other data show- ing that the apple cell walls contain from 1 to 4 times more galactose than arabinose.

    5. Biosynthesis of Primary Cell Wall

    Very little is known about the synthesis of the primary cell wall as a whole (Kauss, 1974; Delmer, 1977, 1983; Robinson, 1977). However, an overall hy- pothetical scheme for the assembly of cell walls may be projected; the sugar nucleotides, synthesized from various pathways (see Fig. 1 I), serve as donors for the synthesis of polysaccharides, probably via glycolipid and glycoprotein intermediates (see Table I). The polysaccharides, except for cellulose, migrate outside the plasma membrane by a mechanism probably involving the Golgi system from where they are incorporated into the cell wall. Cellulose is synthe- sized in the membrane itself by an enzyme complex system that migrates from Golgi vesicles (Delmer, 1983). This system can be transferred to the plasma membrane by fusion of Golgi vesicles to it. Primary (and later secondary) cel- lulose-synthesizing complex appears to move freely within the lipid bilayer of the plasma membrane as fibril deposition proceeds, and the pattern of deposition appears to be random. Cellulose and other polysaccharides can be bonded to- gether covalently or noncovalently by mechanisms presently unknown. It is



    // I I W ) --6.P ( n l p m - 1 9

    t-i /

    / -- y o - I N o ~ I r4

    GOWCOSE aucumm



    II I-

    W X Y L O S E


    In considering the various criticisms raised in the discussion of the models previously described, an attempt is made here to present a more updated cell wall model. This is shown in Fig. 13. The main features of this model are as follows:

    1. Several layers of pectin form an outer network around the cell wall con- stituting the middle lamella (cf. Section IIl,B,3). The pectic molecules are inter- connected by covalent bonds (cf. Section Ill, A,]) and Ca2+ bridges (cf. Fig. 6). A high proportion of the bridges are arranged parallel to the direction of elonga- tion. No covalent connection between pectin and any other wall component is proposed, although some regions of the pectic fractions are strongly hydrogen bonded with xyloglucans. The pectic fraction, however, has some covalently linked glucose and xylose (see also Chambat el al.. 1984).



    Nucloetide sugar donor Product formed References ~~ ~ ~ ~~~~

    UDP- or TDP-galacturonic acid

    [ S- Adenosylmethionine] UDP-arabinose, UDP-xylose

    UDP-xylose, UDP-glucuronic acid UDP-galactose


    UDP-xylose, UDP-glucose

    GDP-mannon, GDP-glucose

    UDP-apiose, UDP-galacturonic acid


    Polygalacturonic acid

    [Methyl esters of polyuronides] Araban, xylan

    Glucuronoxylan Galactan





    Mixed linkage glucan, p-l,3-glucan

    Villemez er al. (1965); Liu er 01. (1966); Bolwell er al. (1985)

    Kauss (1974) Bailey and Hassid (1966);

    Odzuck and Kauss (1972); Ben-Arie er al. (1979); Bolwell and Northcote (1983a.b)

    Waldron and Brett (1983) Panayotatos and Villemez

    ( 1973) Franz (1973); Smith er al.

    ( 1976) Ray (1975); Villemez and

    Hinman (1975) Elbein (1969); Villemez

    (1971) Mascaro and Kindel (1977);

    Pan and Kindel (1977) Ordin and Hall (1968);

    Chambers and Elbein (1970); Pc?aud-Lenkl and Axelos (1970); Smith and Stone (1973); Delmer er al. (1977); Heiniger and Delmer (1977); Anderson and Ray (1978); Raymond er al. (1978)

    2. Under the outer pectin layer are situated several layers of cellulose fibers which are noncovalently associated with xyloglucan, and such assemblies are also noncovalently attached to each other. Cross-linking of xyloglucan-associ- ated cellulose is also provided by hydroxyproline-rich glycoproteins in a novel noncovalent manner involving isodityrosine bridges similar to that first sug- gested by Lamport and Epstein (1983). The region of the xyloglucan-associated cellulose layer nearest the cell membrane is arranged transversely to the direction of elongation, whereas the part nearest the pectin fraction is more randomly arranged. The appraisal of the model is discussed in the following subsections.


    FIG. 12. A model for the biosynthesis of cellulose (after Delmer, 1983). Numbers refer to the reactions catalyzed by the following enzymes: I , invertase; 2, sucrose synthetase; 3, hexokinase; 4, phosphoglucomutase; 5 , UDP-glucose pyrophosphorylase; 6 , 7, and 8, hypothetical reactions.

    1. Growth

    Under normal physiological conditions, the side chains of pectin molecules and possibly the rhamnogalacturonan backbone of pectin are randomly turned over at a constant rate (Labavitch, 1981). This has been supported by both pulse- labeling experiments (Matchett and Nance, 1962; Maclachlan and Duda, 1965) and by gravimetric measurements (Ray, 1963; Nelmes and Preston, 1968). Growth could therefore be considered to result from stimulation of synthesis or a slowing down of degradation. Wall extension induced by auxin has been sug- gested to result by the activation of two systems (Rayle and Cleland, 1970; Vanderhoef and Dute, 1981): one involving proton pump across the membrane


    FIG. 13. A suggested modified model of primary cell walls of dicots. Some xylose and glucose are postulated to be part of the pectic fraction which is cross-linked only to itself. Cohesion of the pectic layer with cellulose layer is presumed to be via hydrogen bondings. Cellulose fibrils are covered by xyloglucan layer (bonded via hydrogen bondings) and are cross-linked by hydroxyproline-rich glycoproteins.

    into the cell wall, and the other involving activation of wall synthesis. It is suggested that H+ ions directly weaken the Ca2+ bridges (cf. Van Cutsem and Gillet, 1983) in the cell wall, thus allowing the parallel pectin molecules to slide relative to each other under turgor pressure. This would correspond to the rapid early stage of auxin-induced expansion (cf. Section III,C,l). The cell wall loosening at low pH has been previously proposed by Cooil and Bonner (1957) and Tagawa and Bonner (1957). Evidence in support of this has recently been provided by Sol1 and Bottger (1981) who found that H+ loosening of cell walls could be mimicked by Ca2 + chelating EDTA, K + , and Na + . In addition, they found that externally applied Ca2+ could increase rigidity of cell walls. It must be noted, however, that Cleland (1960) was unable to detect a redistribution of 45Ca2+ between cell walls and the outer solution after auxin treatment. This


    could have been because the bonds involving Ca2+ and pectin were not suffi- ciently weakened to allow appreciable exchange.

    In the suggested model, wall loosening is not regarded to be caused by low pH stimulation of hydrolytic enzyme activities (cf. Section III,C, 1) within the wall, as this action, together with loosening of Ca2+ bridges, may cause the wall to come apart under turgor pressure. In this regard, a significant proportion of the arabinogalactan side chains of pectic polymers is needed to stabilize the loosened cell wall and also to play a role in checking wall expansion. Thus, the loosened wall is only able to expand up to a certain point, after which the side chains have to be broken to permit further expansion. Recently, Sol1 and Bottger (198 1) were unable to find indications of acid-activated cell wall-loosening enzymes (see also Dey and Del Campillo, 1984).

    Once the pH of the wall is raised back to its physiological value, the Ca2+ bridges become stronger and fix the pectic molecules firmly in place. The rapid early-stage cell wall elongation is therefore reduced and insertion of new wall material continues beyond this point. The whole process can be sequentially repeated in response to further auxin stimulation. Elongation of the xyloglucan- associated cellulose layer is suggested to occur by direct insertion of newly synthesized material in a direction transverse to that of elongation. Such insertion would presumably be easier after pH loosening of the wall. Extensive cross- linking of the xyloglucan-associated cellulose layer by hydroxyproline-rich glycoproteins could inhibit growth by preventing insertion of new material. However, peroxidase, which is presumably involved in immobilizing hy- droxyproline-rich glycoproteins by forming isodityrosine cross-linkages (Cooper and Varner, 1983; Lamport and Catt, 1981; Fry, 1982; Lamport and Epstein, 1983), tends not to occur in regions of highest growth rates (Lamport and Catt, 1981).

    2. Chemical and Enzymatic Actions

    As xyloglucans are not covalently attached to the cell wall, the suggested model predicts their removal by reagents which disrupt this type of bonding. Further, this process would be expected to be enhanced by degradation of the pectin network which covers the whole of the cell wall (cf. Section III,C,2). On the other hand, pectic molecules would be removed by the breaking of covalent bonds. These predictions are consistent with current observations with mango cell wall preparations (Dey et al . , unpublished work).

    The difficulty encountered in solubilizing hydroxyproline-rich glycoproteins from cell walls can be explained by it being cross-linked to itself via iso- dityrosine bridges (Lamport, 1980; Fry, 1982; Cooper and Varner, 1983; Lam-


    port and Epstein, 1983). Partial solubilization of this polymer with acidic sodium chlorite (ONeill and Selvendran, 1980) can thus be explained by the hydrolysis of chlorite-labile isodityrosine linkages.

    The suggested cross-linking of pectic molecules to each other (and not to xyloglucan) explains why it has not been possible to isolate significant quantities of xyloglucans covalently attached to pectic polymers (cf. Section III,A,3). The model also accounts for the ability of endopolygalacturonase to solubilize neutral sugars (e.g., arabinose, galactose, glucose, and xylose) as well as fragments of rhamnogalacturonans (Karacsonyi et al., 1975; Knee ef al . , 1975).


    From a physiological point of view, a fruit results from the development of the tissue that supports the ovule of a plant. This definition encompasses dissimilar organs, such as the floral axis of pineapple, receptacle of strawbeny and apple, and syconium of fig, in that an ovule is present in all these fruits.


    The life of a fruit starts with fertilization followed by a phase of growth to maturity, which in turn is followed by ripening and senescence.

    1 . Fruit Set

    This is the early phase in the life of a fruit, characterized by rapid growth of the ovary that usually follows pollination and fertilization, and is accompanied by changes such as wilting of petals and stamens. It is considered that fruit growth begins in the floral primordium (Nitsch, 1953). The pericarp of the fruit develops from the ovary wall and may differentiate into three distinct regions: the exocarp, the mesocarp, and the endocarp. However, fruit development is not restricted to the ovary and often involves noncarpellary parts of the flower (Esau, 1965). Initial development occurs mainly through cell multiplication.

    2 . Fruit Enlargement

    This is the stage following fruit set where an increase in the size of fruit occurs. It is marked by cell enlargement, although cell division also continues (Esau, 1965). In some fruits, such as the apple, expansion of intercellular spaces may be a contributing factor to enlargement. Generally, cell division predomi-


    nates in the early stages of growth, whereas cell expansion predominates during the later stages. However, there is much varietal variation, and the cell division stage usually overlaps the cell enlargement stage. In one species of tomato, Lycopersicon pimpirellijiolium, some cell division continues even to maturity, whereas in Lycopersicon esculentum, division ceases at anthesis (Nitsch, 1952). More complicated patterns of development occur in other fruits in which cell division ceases at different times in different parts of the fruit. The period of fruit growth varies from one week to several years, although periods of several months are more usual.

    3. Maturation

    This stage is reached in the life of a fruit when full development (ripening and senescence) may be achieved independent of the parent plant. After maturation, there is no further increase in the size of fruits. Fruits are normally harvested at this stage, after which they live an independent life by utilizing substrates accu- mulated during maturation.

    4 . Ripening

    During ripening the fully mature fruit converts to a more palatable state. Specific flavors are developed in conjunction with increased sweetness and de- creased acid content. Softening of the fruit occurs and is often accompanied by a change in coloration. Chlorophyll in the chloroplasts of the outermost cells decreases, while carotenoids and anthocyanins develop.

    According to the respiratory behavior late in their developmental sequence, fleshy fruits can be loosely classified as climacteric and nonclimacteric (Kidd and West, 1930; Biale and Young, 1947; Biale, 1961). Climacteric fruits, such as the mango, tomato, and apple, undergo an upsurge in respiration at the onset of ripening which is coincident with other activities resulting in changes charac- teristic to the ripening process. Known factors that influence the onset of the climacteric in fruits are temperature (generally, lowering the temperature delays onset of the climaceteric), 0, and CO, tension (generally, lowering the 0, tension below that of air or raising the CO, tension delays the climacteric), and the presence of ethylene. An upsurge in respiration occurs in climacteric fruits allowed to ripen while still attached to the plant and in fruits detached after maturity has been reached. Nonclimacteric fruits, such as the strawberry, citrus fruits, and pineapple, however, ripen gradually over a longer period of time and show no upsurge of respiration. The many changes that occur during the ripening process appear to be synchronized and are probably under genetic controls. This



    Degradative Synthetic

    Destruction of chloroplast Breakdown of chlorophyll Starch hydrolysis Destruction of acids Oxidation of substrate Inactivation by phenolic compounds Solubilization of pectins Activation of hydrolytic enzymes Initiation of membrane leakage Ethylene-induced cell wall softening

    Maintenance of mitochondria1 structure Formation of carotenoids and of anthocyanins Interconversion of sugars Increased TCA cycle activity Increased ATP generation Synthesis of flavor volatiles Increased amino acid incorporation Increased transcription and translation Preservation of selective membranes Formation of ethylene pathways

    a After Biale and Young (1981).

    contention is supported by the fact that the interval between anthesis and ripening under similar environmental conditions is roughly constant for any given fruit. A summary of changes, suggested by Biale and Young (1981), that occur during the ripening process is presented in Table 11. A strong demand for energy is placed upon the system for the continuation of processes which include transcrip- tion, translation, and synthesis of ethylene and flavor compounds. The energy is supplied by some of the degradative processes, in particular the hydrolysis of starch. Glucose produced by this process is consequently utilized during the ripening process (Hulme, 1958, 1961; Biale, 1961; Rolz et al., 1972). The interrelation and mechanism whereby these changes are coordinated is presently unknown. One of the difficulties in determining this is in trying to discern causative factors from their effects. Ripening may be considered to occur as a chain of events which are dependent upon the completion of preceding steps. However, Hobson (1979) has suggested that ripening should be considered as a number of key processes taking place simultaneously, each one having its own control mechanism which is loosely coordinated with those of the other pro- cesses.

    5. Senescence

    This is the stage that begins somewhere during the ripening process and continues until the end of the life of the fruit. It is characterized by a general and increasing failure of many synthetic processes and susceptibility of the fruit to fungal attack.



    1 . Protein Synthesis

    There is considerable evidence for the involvement of protein synthesis (de novo enzyme synthesis) at the climacteric stage during fruit development (Jones et al., 1965; Sachler, 1966; Lance et al . , 1967; Frankel et al . , 1968; Sacher, 1973). Increases in the ratio of protein nitrogen to total nitrogen have been reported to occur (Hulme, 1954; Rowan et al . , 1958; Biale, 1961). It is not certain whether de novo synthesis of protein catalyzes the climacteric rise. How- ever, recent reports of ripening-related changes in the levels of different transfer RNAs (Mettler and Romani, 1976) and messenger RNAs (Rattanpanone et al . , 1977, 1978) in the tomato fruit support this contention. Further support is pro- vided by the observation of ribosomal RNA synthesis prior to the climacteric peak (Jones et al., 1965; Looney and Patterson, 1967; Richmond and Biale, 1967; Ku and Romani, 1970; De Swart et al . , 1973). Hobson (1975) suggested that the synthesis of enzymes required for ripening may be at the expense of other proteins, as there is little evidence suggesting drastic alteration in total protein content during ripening.

    2 . Ethylene Production

    It has long been recognized that ethylene acts as a hormone in plants and can have profound effects on the ripening process, particularly in climacteric fruits (Beyer, 1981). Its true role in fruit ripening has not been fully determined. It is widely accepted that treatment of fruits with low concentrations of ethylene brings forward the time of onset of the climacteric upsurge without altering the shape of the cycle. It is only effective if applied before the climacteric phase has begun in the plant when it is not already influenced by endogenous ethylene production. There is no return to a preclimacteric stage once an adequate ex- posure to this hormone has been achieved. This is in contrast to the effect of ethylene on nonclimacteric fruits where a response can be achieved throughout postharvest life (Biale and Young, 1962). Further, once the hormone is removed, its effect on this type of fruit (nonclimacteric) is reversed.

    Detailed studies by Pratt and Goeschl (1968) and Pratt et al. (1977) showed that in the honeydew muskmelon, reduction in firmness and upsurge in respira- tion appeared to be the only ripening parameters that were directly related to ethylene action. Other physiological changes, such as decrease in growth rate and increase in soluble solid content, preceded accelerated ethylene production. It was also observed that while ethylene was able to induce a respiratory climac-


    teric in the immature honeydew muskmelon, this response did not result in the accumulation of sugars and flavor components characteristic of mature fruit ripening. From work on Coxs Orange Pippin apples, Rhodes and Reid (1975) concluded that a factor other than a change in ethylene level determines the time of the onset of ripening in the apple. Kosiyachinda and Young (1975) came to the same conclusion while working on avocado and cherimoya fruits. The rela- tionship between ethylene production and the respiratory rise in fruits appears to be varied. The onset of the rise precedes ethylene production in some fruits such as the avocado, banana, guava, and honeydew muskmelon, while in fruits such as Coxs Orange Pippin apple, apricot, and cantaloupe muskmelon, it coincides with ethylene production, and in other fruits such as the fuerte avocado, chafley cherimoya, haden mango, and rutger tomato, it occurs after ethylene production (Biale and Young, 1981). From such observation, Biale and Young (1981) suggested that upsurge in ethylene synthesis is not the initial factor in inducing the respiratory climacteric, although they questioned the reliability of some of these studies. McGlasson et al. (1975) concluded that the onset of ripening in normal tomato is not controlled by endogenous ethylene, although an increase in its production is probably an integral part of the ripening process. They found that treatment of 40-80% mature tomatoes with propylene stimulated respira- tion, but did not bring about ethylene production. Normally, treatment of climac- teric fruits with ethylene stimulates an increase in its production (Mapson and Robinson, 1966; Quazi and Freebairn, 1970).

    3 . Starch Hydrolysis

    Increased hydrolysis of stored reserves of starch, often in the form of granules, is closely related to the ripening process. In banana, starch content of the pulp varies between 20 and 30% in the unripe fruit and between 1 and 2% in the ripe fruit. The consequence of starch hydrolysis is the formation of sucrose, glucose, and fructose. Young et al. (1975) found that a sharp decrease in starch content in Valery bananas occurred when the respiratory rise was well on the way toward the peak. They suggested that hydrolytic enzymes involved were either activated or synthesized de novo at the onset of ripening and reported the presence of two forms of a-amylase, two forms of P-amylase, and three phorphorylases which were present at all stages of ripening. Several inhibitors were found that pre- vented starch hydrolysis in extracts of preclimacteric bananas.

    4 . Glycolysis

    Fructose- 1,6-diphosphate, unlike glucose-6-phosphate and fructose-6-phos- phate, increases (20-fold) during ripening of bananas (Young et al., 1975).


    Crossover plot analysis on the same fruit indicates that a regulatory site occurs in phosphofructokinase. Barker and Solomon ( 1962) supported the contention that cellular fructose- I ,6-diphosphate was a major controlling factor of the respira- tion rate, while Pearson and Robertson (1954) proposed that it was controlled by the ADP:ATP ratio.

    5. Phosphorylation

    Millerd and co-workers (1953) considered that the climacteric rise could be brought about by the uncoupling of phosphorylation. However, the enhancement of phosphorylative capacity with ripening does not support this proposal (Biale and Young, 1981). Hobson (1965) demonstrated that in tomatoes, which were subjected to the action of uncoupling agents, production of enzymes necessary for the ripening process continued. He suggested the possibility that loose coupling of phosphorylation during the climacteric rise may result in a net increase in the synthesis of energy-rich bonds at an early stage, leading to the formation of additional enzymes necessary for the furtherance of ripening. Biale and Young (1981) indicated from studies using isolated mitochondria and tissue discs that tighter coupling of oxidative phosphorylation may occur in prepara- tions from ripe avocados compared to unripe fruits.

    6. Mitochondria1 Involvement

    Various responses are obtained, with intermediates of the Krebs cycle as substrates, on the mitochondrial activity. Some of these are summarized in Table 111. Using succinate as substrate, a rise in respiratory control, together with an




    4 0 d m g mitochondrial


    Substrate Cofactor Unripe Ripe

    Malate None 185 145 TPP 500 1105

    TPP 480 450 Pyruvate None 68 740

    TPP 222 1140

    a-Ketoglutarate None I10 400

    a After Biale and Young (1981).


    increase in ADP:O ratio (an index of degree of esterification of ADP to ATP per oxygen atom consumed in substrate oxidation), is observed during ripening. Using malate, addition of ADP elicits a small response with mitochondria from preclimacteric avocados. Addition of thiamine pyrophosphate (TPP) is necessary to obtain respiratory control and maximal oxidation. In mitochondria from the ripe fruit, however, no TPP is needed to obtain the response. With a-ketogluta- rate as substrate, the requirement for TPP is much more pronounced with mito- chondria from unripe than ripe fruit. With pyruvate as substrate, a similar effect is obtained as for malate, except that differences in values for ripe and unripe fruits are more pronounced. The in vivo effect of TPP is not known.


    I . Pectic Fraction

    It is widely accepted that softening of fruit which accompanies ripening is essentially caused by the conversion of insoluble wall-bound protopectin of high molecular weight to water-soluble pectin. Dolendo et al. (1966) reported that softening of fruit which is indicated by pressure measurements coincided with increases in soluble pectins and decreases in protopectin. Solubilization of pectin during ripening has been demonstrated in the mango (Dennison and Ahmed, 1967; Krishnarmurthy and Subramanyam, 1973), apple (Doesburg, 1957; Knee, 1973b, 1978a,b), tomato (Hobson and Davies, 1971; Gross and Wallner, 1979), avocado (McCready and McCoumb, 1954), peach (McCready and McCoumb, 1954; Pressey er al., 1971), pear (Ahmed and Labavitch, 1980a), date (Rouhani and Bassiri, 1976), and strawbeny (Knee er a l . , 1977). Doesburg (1957) has proposed that movement of calcium in cell walls may assist solubilization of pectin during ripening. Reported losses of pectin from cell walls are consistent with observed decreases in cell wall galactose, galacturonic acid, and in some cases arabinose content during the ripening process (Knee, 1973a,b; Bartley, 1976; Knee et a l . , 1977; Gross and Wallner, 1979; Ahmed and Labavitch, 1980a). Ahmed and Labavitch (1980a) have shown that there is solubilization of a high-molecular-weight branched arabinan from the cell walls of ripening pears. This polysaccharide consists of a backbone of a-1 S-linked L-arabinosyl residues with a-linked L-arabinosyl side chains at C-2 and/or C-3 and has a similar structure to pectic polymers present in the primary cell wall of cultured sycamore cells (Darvill and McNeil, 1980). There is also solubilization of an acidic frac- tion of lower molecular weight galacturonan free of arabinosyl residues.

    Ultrastructural studies have shown that wall breakdown is accompanied by dissolution of the middle lamella region of parenchyma cells, which is rich in


    pectin, leading to cell separation in apple (Arie et al., 1979), tomato (Hobson, 1981), pear (Arie et al., 1979), and strawberry (Knee et al., 1977). The mecha- nism whereby such drastic changes are brought about is not fully understood.

    Enzymes capable of degrading pectin which have been identified in fruits are endopolygalacturonase (PG), pectinmethylesterase (PME), exopolygalacturo- nase, P-galactosidase, and a-L-arabinosidase. The enzyme which is generally known to be directly involved in wall breakdown is PG (Grierson et al., 1980, 1981), although it seems certain that other enzymes also take part (see also Rexova-Benkova and Markovic, 1976; Hobson, 1981 ; Dey and Brinson, 1984).

    a. PME. Pectinmethylesterase catalyzes the deesterification of galacturo- nosy1 residues (present in rhamnogalacturonan), in which the carboxyl group is methyl esterified. It does not hydrolyze methyl esters in short-chain galac- turonans. Blocks of free carboxyl groups are produced following the action of PME (Deuel and Stutz, 1958), which suggests that deesterification occurs in a linear manner. Lee and Macmillan (1970) have shown that the enzyme acts at both the reducing ends and interior regions of highly esterified pectin chains. It has been suggested that hydrolysis, catalyzed by the enzyme, only occurs adja- cent to free carboxyl groups ( S o h and Deuel, 1955).

    PMEs are present at the immature stage of most fruits and generally reaches maximal activity immediately preceding or early in the climacteric rise. This enzyme has been found in peaches (Schewfelt, 1965), pears (Nagel and Patter- son, 1967), tomatoes (Dennison et al., 1954; Hobson, 1963; Besford and Hob- son, 1972), bananas (Hultin and Levine, 1965), avocados (Awad and Young, 1979), and mangoes (Mattoo and Modi, 1969). Correlation between the level of enzymatic activity and cell wall softening is presently not clear. For example, Hamson (1952) reported that the level of PME activity is higher in firm than in soft tomatoes, whereas Hobson (1963) reported 40% greater activity at the ripe stage. Pressey (1977) suggested that part of the confusion may be due to the complexity of the enzyme, which exists in multiple forms in tomato and whose level varies with stages of ripeness and with different varieties. The level of PME activity in peaches also varies with variety and shows no particular trend with fruit softening (Schewfelt, 1965). In banana, however, the activity of the enzyme increases 10-fold during ripening (Hultin and Levin, 1965). Three forms exist in this source, one of which increases continually during ripening.

    A role for PME in cell wall softening is not apparent, although it has been considered that its action may be needed prior to the action of PG, which has a preference for deesterified pectate in some fruits (McCready et al. , 1955; Luh et al., 1956; Jansen et al., 1960; Patel and Phaff, 1960; Reymond and Phaff, 1965; Brady, 1975; Bartley, 1978).


    b. PG. Endopolygalacturonase catalyzes random hydrolysis of the rham- nogalacturonan backbone of pectin. It is not present in all fruits. However, it has been identified in many fruits, such as tomato (Hobson, 1963, pear, date (Hasegawa et al., 1969; Pressey and Avants, 1976), cranberry (Knee and Bart- ley, 1981), peach (Pressey and Avants, 1973), cucumber (Pressey and Avants, 1975; Saltveit and McFeeters, 1980), and avocado (Awad and Young, 1979). PG cleaves pectate randomly, first to oligogalacturonates and ultimately to galac- turonic acid, but the rate of hydrolysis decreases with smaller polymers. The rates of hydrolysis of tetra-, tri-, and digalacturonate are 7, 1.6, and 1 % of the rate of a long-chain polymer (Pate1 and Phaff, 1960).

    It is generally accepted that PG is primarily responsible for dissolution of the middle lamella during ripening in some fruits. Pressey and Avants (1973) showed that the enzyme was effective in solubilizing pectin from washed peach cell walls. McCready and McComb (1954) reported that PG activity increases during ripening of avocados. The enzyme was shown to be capable of hydrolyz- ing pectate to intermediate oligogalacturonates which were then slowly hydro- lyzed to galacturonic acid (McCready et al., 1955; Grant et al., 1973). Ahmed and Labavitch (1980b) found that treatment of unripe pear cell walls with pu- rified PG solubilized an acidic unbranched arabinan with similar characteristics to a polymer solubilized during normal ripening. Wallner and Bloom (1977) also found that tomato PG accomplished solubilization of green tomato walls which appeared to be nearly as extensive as that which occurred during in vivo ripening. However, the enzyme did not cause the 4 0 4 0 % decrease in cell wall galactose which occurs during the natural process. Wallners group (Wallner and Walker, 1975; Wallner and Bloom, 1977; Gross and Wallner, 1979) has shown that a sharp increase in PG activity in the tomato is accompanied by an increase in the solubility of a rhamnogalacturonan fraction (which can be extracted from iso- lated cell walls by 4-hr incubation with water at 30C). This fraction (average molecular weight 20,000) was found to be almost free of neutral sugar residues and could not be extracted by incubating cell walls of unripe fruits. It was suggested that conversion of high-molecular-weight pectic polysaccharides with- in the cell wall to water-soluble rhamnogalacturonan of lower molecular weight occurred during ripening. This was proposed to be accomplished by a two-stage mechanism involving removal of neutral pectic side chains followed by PG- catalyzed hydrolysis of the exposed backbone.

    Much attention has been focused on the involvement of PG in the ripening process, particularly in tomato fruits, and much progress has been made recently in the understanding of its involvement in tomato ripening. Its activity appears about 2 or more days after the onset of ethylene production and sharply increases thereafter as ripening proceeds (Hobson, 1964; Ahmed and Labavitch, 1980b; Brady et al . , 1983). This has been shown to be caused by de novo synthesis of


    the enzyme rather than by activation (Tucker and Grierson, 1982). Three forms of PG, designated as FGI, PGIIA, and PGIIB, have been identified in tomato (Pressey and Avants, 1973; Rexova-Benkova and Markovic, 1976; Tucker er al., 1980; Brady et al., 1983; see also Moshrefi and Luh, 1983, 1984).

    Reports on the molecular weight of PGI vary from 84,000 to 115,000, and it has been suggested that it may be a dimer of PGIIA (with molecular weight 43,000). PGIIB has a molecular weight of 46,000. During the early ripening process, PGI is the only form present (Tucker er al., 1980; Brady er al., 1983). As the tomato ripens, PGII activity increases until it becomes the dominant form (Tucker et al., 1980; Brady ef al. , 1983). It is not certain whether this is due to the two forms of the enzyme being produced simultaneously or to the conversion of PGI to PGII. Evidence suggests, however, conversion may not cause an increase in enzymatic activity as the two purified forms have similar specific activities. The roles of the various forms of the enzyme in cell wall softening are still unclear.

    c. Exopolygalacturonase. Exopolygalacturonase catalyzes the hydrolytic removal of small oligosaccharide fragments from the ends of pectin molecules. It has been found in pears (Pressey and Avants, 1976), peaches (Pressey and Avants, 1973), cucumbers (Pressey and Avants, 1975), and apples (Bartley, 1978). The apple enzyme is capable of degrading cortical cell wall preparations in v i m , releasing low-molecular-weight uronic acid residues and polyuronide. Peach exopolygalacturonase showed optimum activity at pH 5.5 and required calcium ions for activity (Pressey and Avants, 1973). It also displayed maximal activity against polymers containing 20 or more residues releasing galacturonic acid. Evidence indicated that cleavage occurred at the nonreducing end of pectate molecules, which were progressively shortened. Exopolygalacturonase from cu- cumber also showed optimum activity at pH 5.5 (Pressey and Avants, 1975) and displayed a similar mode of action to the enzyme from peach. It was also activated by calcium ions, but showed most activity against substrates containing 6-12 residues. There is very little evidence availabe at present to implicate exopolygalacturonase in cell wall softening.

    d. P-Galacrosidase. Increases in the activity of P-galactosidase with ripen- ing have been reported in many fruits (Bartley, 1974; Yamaki and Matsuda, 1977; Brinson, 1984). Its involvement in cell wall breakdown has.not been established, although Bartley (1974) has reported that it is capable of hydrolizing a potato P- 1 ,Cgalactan similar to that found in apple cell walls as well as whole apple cell walls. Pressey (1983) has recently reported the presence of multiple forms of P-galactosidase, designated I, 11, and 111, in tomato fruit. The molecular weights calculated from gel filtration chromatography were 144,000, 62,000,


    and 71 ,OOO, respectively. It was suggested that 111 may be a dimer of I. Only I1 was able to hydrolyze (58%) a galactose-rich polysaccharide isolated from toma- to cell walls releasing free galactose. During ripening, the levels of activity of I and I11 decreased whereas the activity of I1 increased. Knee and Bartley (1981) have suggested that P-galactosidase may be involved in removing galactose from cell walls during ripening. Pressey (1983) has suggested a similar role for the enzyme in tomato.

    e . a-L-Arabinosidase. As in the case of P-galactosidase, a-L-arabinosi- dase has been reported in many fruits, and its level of activity increases with ripening (Brinson, 1984). Ahmed and Labavitch (1980b) were unable to generate reducing power from a purified arabinan using an enzyme preparation that con- tained a-L-arabinosidase activity. It may be that glycosidase activities deter- mined by incubation with p-nitrophenyl substrates give an inaccurate picture of in vivo enzyme specificity (Pharr er al., 1976). Although the involvement of this enzyme in cell wall breakdown has not been established, it is likely to remove a- L-arabinosyl residues from the structurally important hydroxyproline-rich cell wall glycoprotein. Such an action is known to disrupt the secondary conforma- tion of the glycoprotein (Holtz and Varner, 1984) and thus may cause softening of cell wall.

    f. Hemicellulase. Loss of cell wall hemicellulose has not been established as part of the ripening process. The relative amounts of monomers characteristic of these polysaccharides, such as xylose and glucose, have not been shown to decline in the cell wall during ripening of apples (Bartley, 1976), strawberries (Bartley, 1974), tomatoes (Wallner and Bloom, 1977), and pears (Jermyn and Isherwood, 1956). In addition, enzymes capable of degrading hemicellulose have not been identified in fruits. Pear (Ahmed and Labavitch, 1980b) and tomato (Wallner and Walker, 1975) lack xylanase activity, although both contain P-D-xylosidase and P-D-glucosidase activities. Tomato contains P- 1 , 3 - ~ - glucanase activity (Wallner and Walker, 1975), but the likely natural substrates of this enzyme have not been shown to be present in the cell wall of fruits. P- 1,4- Glucanase activity has been found in strawberries, tomatoes, and pears (Hobson, 1981; Barnes and Patchett, 1976). However, doubt that this enzyme has a prima- ry role in fruit softening arises from the observation that its activity is normal in nonripening mutants of tomatoes (Poovaiah and Nukaya, 1979).


    It is certain that tissue softening during the process of fruit ripening is related to breakdown of the organized structure of primary cell wall. In order to gain


    greater understanding of the process, it is therefore important to have in-depth knowledge of the primary cell wall. This includes isolation and detailed charac- terization of the components of the wall and a determination of the precise nature of their cohesive interlinking forces. Presently, our knowledge of the primary cell wall structure is based upon only a limited number of studies, in many of which suspension-cultured cells have been used instead of the intact plant tissues.

    Much is known of the important constituent, the pectic fraction, of the cell wall; however, the position of the hemicellulose, xyloglucan, in the cell wall architecture is not entirely clear. Multiple forms of xyloglucan are known to occur (Joseleau and Chambat, 1984); some are linked by covalent bonds and/or hydrogen bonds to the pectic fractions. One evidence showing covalent linkage comes from isolation of a pectic fraction from mango pulp cell wall by ion- exchange chromatography (Dey et af., unpublished results). Undoubtedly, xy- loglucans have an important ultrastructural position as junction polysaccharides forming hydrogen-bonded connections between cellulose and pectic fractions of cell wall (see Fig. 14). The physiological significance of xyloglucans is probably in cell elongation and the swelling property of cell walls; thus, further knowledge about its self-association and ability to cross-link with other cell components is important.

    The polygalacturonase-catalyzed degradation of pectic rhamnogalacturonan resulting in the dissolution of the middle lamella is probably the major contrib- utor to tissue softening during fruit ripening. However, in the absence of poly- galacturonase in the tissue, alternative mechanisms involving other enzymes must be considered. Such enzymes would act on particularly pectic side chains. Nevertheless, the role of a relatively less-known pectic enzyme, polymeth- ylgalacturonase, should not be ignored. Unlike polygalacturonase, the aforesaid enzyme does not require the pectin to be demethylated prior to cleavage of glycosidic bond. The occurrence of the enzyme was first predicted by Rexova-

    Arabinogalactan or

    Arabinan Rhamnogalacturonan or pectic fraction

    Galactan (main chain) (side chain)

    RhnrnnoEfllncturonnn - pectic fraction

    (main chain)

    Cellulose I I I I

    Minor xylose/glucose corn onents

    ?side chain) I I I

    Xyloglucan I I I


    FIG. 14. (- - -), hydrogen bond.

    Suggested links between some fractions of primary cell walls. (-), Covalent linkage;


    Benkova and Markovic (1976) and later demonstrated by other authors (Mizrahi et al., 1976, 1982; Kopeliovitch et al., 1980; Malis-Arad and Mizrahi, 1983). The loss of L-arabinose and D-galactose during ripening of some fruits, e.g., mango (Dey et al. , unpublished results), may be due to a-L-arabinofuranosidase and P-D-galactosidase, respectively, the activity of which increase severalfold. The true roles of these enzymes will become clear only by the use of natural substrates for their assay, as opposed to commonly employed p-nitrophenyl substrates. Specific enzymes acting on such linkages as P-glucosidic-, a-xy- losidic-, and a-L-fucosidic- of xyloglucans may also play important roles in the detachment of the rhamnogalacturonans from other polymers.

    The sequence of appearance of various cell wall-degrading enzymes during ripening seems also important in tissue softening. For example, demethylation of pectin by pectinmethylesterase will facilitate the action of polygalacturonase. Similarly, it is probable that the action of specific enzymes is required for at least partially detaching the pectic backbone of rhamnogalacturonans from cross- linking polymers prior to the action of galacturonase (exo- and/or endo-).

    Finally, in order to fully understand tissue softening, it will be necessary to gain greater knowledge of the biosynthesis of wall-degrading enzymes, mecha- nisms controlling their activities, and the mechanism of their transport to their sites of action.


    M. A. J. is thankful to Professor J. B. Pridham and Dr. E. Percival for their constant interest and valuable suggestions during the Ph.D. work.


    Ahmed, A. E., and Labavitch, J. M. 1980a. Cell wall metabolism in ripening fruit. I. Cell wall changes in ripening Bartlett pears. Plant Physiol. 65, 1009-1013.

    Ahmed, A. E., and Labavitch, J. M. 1980b. Cell wall metabolism in ripening fruit. 11. Changes in carbohydrate-degrading enzymes in ripening Bartlett pears. Plant Physiol. 65, 1014- 1016.

    Akiyama, Y., and Kato, K. 1976. Hydroxyproline-arabinosides from suspension cultured tobacco cell wall. Agric. Biol. Chem. 40, 2343-2348.

    Akiyama, Y., and Kato, K. 1977. Structure of hydroxyproline-arabinoside from tobacco cells. Agric. Biol. Chem. 41, 79-81.

    Albersheim, P. 1959. Instability of pectin in neutral solutions. Biochem. Biophys. Res. Commun. 1, 253-256.

    Albersheim, P. 1974. The primary cell wall. In Plant Biochemistry (J. Bonner and J. E. Varner, eds.), pp. 225-274. Academic Press, New York.


    Albersheim, P. 1975. Walls of growing plant cells. Sci. Am. 232, 81-95. Albersheim, P. 1978. Concerning the structure and biosynthesis of primary cell walls of plants. Inr.

    Albersheim, P., and Killias, U. 1963. Histochemical localization at the electron microscope level.

    Anderson, R. L., and Ray, P. M. 1978. Labeling of the plasma membrane, of pea cells by a surface

    Arie, R. B., Kislev, N., and Frenkel, C. 1979. Ultrastructural changes in the cell walls of ripening

    Aspinall, G. 0.. and Cottrell, 1. W. 1970. Lemon-peel pectin. 11. Isolation of homogeneous pectins

    Aspinall, G. O., and Cottrell, I. W. 1971. Polysaccharides of soybeans. VI. Neutral polysaccharides

    Aspinall, G. O., and Jiang, K. S . 1974. Rapeseed hull pectin. Curbohydr. Res. 38, 247-255. Aspinall, G. 0.. and Molloy, J. A. 1968. Pectic substances from lucerne (Medicugo surivu). Part 111.

    Fractionation of polysaccharides extracted with water. J. Chem. SOC. (C) 2994-2997. Aspinall, G. 0.. Begbie, R., Hamilton, S . . and Whyte, J. N. C. 1967a. Polysaccharides of soy-

    beans. Part 111. Extraction and fractionation of plysaccharides from cotyledon meal. J. Chem.

    Aspinall, G. O., Cottrell, 1. W., Egan, S . V., Momson, 1. M., and Whyte, J. N. C. 1967b. Polysaccharides of soy-beans. Part IV. Partial hydrolysis of the acidic plysaccharide complex from cotyledon meal. J . Chem. SOC. ( C ) 1071-1080.

    Aspinall, G. 0.. Craig, J. W. T., and Whyte, J. N. C. 1968a. Lemon-peel pectin. Part I. Fractiona- tion and partial hydrolysis of water-soluble pectin. Curbohydr. Res. 7, 442-452.

    Aspinall, G. 0.. Gestetner, B., Molloy, I . A., and Uddin, M. 1968b. Pectic substances from lucerne (Medicugo surivu). Part 11. Acidic oligosaccharides from partial hydrolysis of leaf and stem pectic acids. J. Chem. SOC. ( C ) 2554-2559.

    Aspinall, G. O., Fainveather, R. M., and Wood, T. M. 1968c. Arabinogalactan A from Japanese larch (Lurk leprolepis). J . Chem. SOC. (C) 2174-2179.

    Aspinall, G. 0.. Molloy, J. A., and Craig, J. W. T. 1969. Extracellular polysaccharides from suspension-cultured sycamore cells. Can. J. Biochem. 47, 1063- 1070.

    Aspinall, G. 0.. Krishnamurthy, T. N., and Rosell, K. G. 1977. A fucogalactoxyloglucan from rapeseed hulls. Curbohydr. Res. 55, 11-19.

    Awad, M., and Young, R. E. 1979. Postharvest variation in cellulase, polygalacturonase, and pectinmethylesterase in avocado (Persea umericunu Mill, cv. Fuerte) fruits in relation to respi- ration and ethylene production. Plunr Physiol. 64, 306-308.

    Bahl, 0. P. 1970. Glycosidases of Aspergillus niger. 11. Purification and general properties of 1.2-a- L-fucosidase. J. Biol. Chem. 245, 299-304.

    Bailey, R. W., and Hassid, W. Z. 1966. Xylan synthesis from uridinediphosphate-D-xylose by particulate preparations from immature corncobs. Proc. Nor!. Acud. Sci. U.S.A. 56, 1586- 1593.

    Bailey, R. W., and Kauss, H. 1974. Extraction of hydroxyproline-containing proteins and pectic substances from cell walls of growing and non-growing mung bean hypocotyl segments. Pluntu

    Baker, D. B., and Ray, P. M. 1965. Direct and indirect effects of auxin on cell wall synthesis in oat

    Barker, J., and Solomon, T. 1962. Mechanism of climacteric rise in respiration in banana fruits.

    Barnes, M. F., and Patchett, B. J. 1976. Cell wall-degrading enzymes and the softening of senescent

    Rev. Eiochem. 16, 127-150.

    Am. J. Eor. 50, 732-745.

    localized glucan synthetase. Plant Physiol. 61, 723-730.

    apple and pear fruit. Plunr Physiol. 64, 197-202.

    and examination of some associated polysaccharides. Can. J. Chem. 48, 1283-1289.

    from cotyledon meal. Can. J. Chem. 49, 1019-1022.

    SOC. (C) 1065-1070.

    119, 233-245.

    coleoptile tissue. Plunr Physiol. 40, 345-352.

    Nurure (London) 196, 189.

    strawbeny fruit. J. Food Sci. 41, 1392-1395.


    Bamoud, F., Mollard, A., and Dutton, G. G. S. 1977. A P-1,4-xyloglucan present in the culture

    Barrett, A. J., and Northcote, D. H. 1965. Apple fruit pectic substances. Biochem. J . 94,617-627. Bartley, I. M. 1974. f3-Galactosidase activity in ripening apples. Phytochemistry 13, 2107-21 11. Bartley, I. M. 1976. Changes in the glucans of ripening apples. Phytochemistry 15, 625-626. Bartley, I. M. 1978. Exo-polygalacturonase of apple. Phyrochemistry 17, 213-216. Bauer, W. D., Talmadge, K. W., Keegstra, K., and Albersheim, P. 1973. The structure of plant cell

    walls. 11. The hemicellulose of the walls of suspension-cultured sycamore cells. Plant Physiol. 51, 174-187.

    Ben-Arie, R., Kislev, N., and Frenkel, C. 1979. Ultrastructural changes in the cell walls of ripening apple and pear fruit. Plant Physiol. 64, 197-202.

    Ben-Arie, R., Ordin, L., and Kindinger, J. L. 1973. A cell-free xylan synthesizing enzyme from Avenu sativa. Plant Cell Physiol. 14, 427-434.

    Besford, R. T., and Hobson, G. E. 1972. Pectic enzymes associated with the softening of tomato fruit. Phytochemistry 11, 2201-2205.

    Beyer, E. M., Jr. 1981. Ethylene action and metabolism. In Recent Advances in the Biochemistry of Fruits and Vegetables (J. Friend and M. J. C. Rhodes, eds.), pp. 107-121. Academic Press, New York.

    Biale, J. B. 1961. Post-harvest biochemistry of tropical and subtropical fruits. Adv. Food Res. 10,

    Biale, J. B., and Young, R. E. 1947. Critical oxygen concentrations for the resparation of lemons.

    Biale, J. B., and Young, R. E. 1962. Biochemistry of fruit maturation. Endeavour 21, 164-174. Biale, J . B., and Young, R. E. 1981. Respiration and ripening in fruits. In Recent Advances in the

    Biochemistry of Fruits and Vegetables (J. Friend and M. J. C. Rhodes, eds.), pp. 1-39. Academic Press, New York.

    Blake, J. D., and Richards, G. N. 1971. Evidence for molecular aggregation in hemicelluloses. Carbohydr. Res. 18, 11-21.

    Bolwell, G. P., and Northcote, D. H. 1983a. Arabinan synthase and xylan synthase activities of Phaseolus vulgaris. Subcellular localization and possible mechanism of action. Biochem. J.

    Bolwell, G. P., and Northcote, D. H. 1983b. Induction of growth factors of polysaccharide synthases

    Bolwell, G. P., Dalessandro, G., and Northcote, D. H. 1985. Loss of polygalacturonic acid synthase

    Bonner, 3. 1933. The action of the plant growth hormone. J. Gen. Physiol. 17, 63-76. Bonner, J . , and Foster, R. J. 1955. The growth-time relationships of the auxin-induced growth in

    Brady, C. J. 1976. The pectinesterase of the pulp of banana fruit. Aust. J. Plant Physiol. 3, 163- 172. Brady, D. J. , Meldrum, W. K., McGlasson, W. B., and Ah, Z.-M. 1983. Differential accumulation

    of the molecular forms of polygalacturonase in tomato mutants. J . Food Biochem. 7 , 7-14. Brauner, L., and Brauner, M. 1943. Relations between water-intake and oxybiosis in the living plant

    tissue. 11. Tensility of cell wall. Rev. Fac. Sci. Univ. Istanbul. 8, 30-75. Brinson, K. 1984. Changes in carbohydrates and glycosidases in mango fruits during ripening. Ph.D.

    thesis, University of London. Brown, R. G., and Kimmins, W. C. 1981. Recovery of glycoproteins from plant tissues. Recent

    Adv. Phytochem. 15, 115-134. Chambat, G., Bamoud, F., and Joseleau, J. P. 1984. Structure of the primary cell walls of suspen-

    sion-cultured Rosa glaucu cells. I. Polysaccharides associated with cellulose. Plant Physiol. 74,

    medium of Rosa glaucu cell suspensions. Physiol. Veg. 15, 153-161.


    Am. J . Bor. 34, 301-309.

    210, 497-507.

    in bean cell suspension cultures. Biochem. J. 210, 509-515.

    activity during differentiation in sycamore. Phytochemistry 24, 699-702.

    Avena coleoptile sections. J . Exp. Bor. 6, 293-302.



    Chambers, J., and Elbein, A. D. 1970. Biosynthesis of glucans in mung bean seedlings. Formation of p-1 ,4-glucans from GDP-glucose and p- I ,3-glucans from UDP-glucose. Arch. Eiochem. Eiophys. 138, 620-631.

    Cleland, R. 1960. Effect of auxin upon loss of calcium from cell walls. Plant Physiol. 35,581-584. Cleland, R. 1971. Cell wall extension. Annu. Rev. Plant Physiol. 22, 197-222. Cooil, B. J . , and Bonner- J. 1957. The nature of growth inhibition by calcium in the Avena

    coleoptile. PIanta 48, 696-723. Cooper, J . B., and Varner, J. E. 1983. Insolubilization of hydroxyproline-rich cell wall glycoprotein

    in aerated carrot root slices. Eiochem. Eiophys. Res. Commun. 112, 161-167. Courtois, J. E., and Le Dizet, P. 1974. Comparative study of the structure of three galactosyl-glucans

    (amyloids) of seeds. An. Quim. 70, 1067-1072. Cutsem, P., van, and Gillet, C. 1983. Proton-metal cation exchange in the cell wall of Nitella

    flexilis. Plant Physiol. 73, 865-867. Darvill, A, , McNeil, M., Albersheim, P., and Delmer, D. P. 1980a. The primary cell walls of

    flowering plants. I n Biochemistry of Plants (N. E. Tolbert, ed.), Vol. I , pp. 91-162. Academic Press, New York.

    Darvill, A., McNeil, M., and Albersheim, P., unpublished results cited in Darvill, A , , McNeil, M., Albersheim, P., and Delmer, D. P. 1980b. The primary cell walls of flowering plants. In Biochemistry of Plants (N. E. Tolbert, ed.), Vol. 1, pp. 91-162. Academic Press, New York.

    Datko, A. H., and Maclachlan, G. A. 1968. Indoleacetic acid and the synthesis of glucanases and pectic enzymes Plant Physiol. 43, 735-742.

    Davidson, E. A. 1967. Carbohydrate Chemistry. Holt, New York. Davies, E., and Maclachlan, G. A. 1969. Generation of cellulase activity during protein synthesis by

    pea microsomes in vitro. Arch. Biochem. Eiophys. 129, 581-587. Dea, I. C. M., Morris, E. R., Rees, D. A., Welsh, E. J. , Barnes, H. A., and Price, J. 1977.

    Association of like and unlike polysaccharides: Mechanism and specificity in galactomannans, interacting bacterial polysaccharides, and related systems. Carbohydr. Res. 57, 249-272.

    Delmer. D. P. 1977. The biosynthesis of cellulose and other plant cell wall polysaccharides. Recent Adv. Phytochem. 11, 45-77.

    Delmer, D. P. 1982. Hemicellulose synthesis. I n CRC Handbook of Biosolar Resources (A. Mitsui and C. C. Black, eds.), Vol. 1 , Part 1, pp. 357-361. CRC Press, Boca Raton, Florida.

    Delmer, D. P. 1983. Biosynthesis of cellulose. Adv. Carbohydr. Chem. Eiochem. 41, 105-153. Delmer, D. P., Heiniger, U., and Kulow, C. 1977. UDP-glucose: Glucan synthetase from develop-

    ing cotton fibers. I. Kinetic and physiological properties. Plant Physiol. 59, 713-718. Dennison, R. A,, and Ahmed, E. M. 1967. Irradiation effects on the ripening of Kent mangoes. J.

    Food Sci. 32, 702-705. Dennison, R. A,, Hall, C. B., and Nettles, V. F. 1954. Pectin esterase activity of tomato fruits at

    different stages of maturity. Proc. Ann. Meet. Am. SOC. Hortic. Sci., Florida 51, 17. De Swardt, G. H., Swanepoel, J. H., and Duvenage, A. J. 1973. Relation between changes in

    ribosomal RNA and total protein synthesis, and the respiration climacteric in pencarp tissues of tomatoes. Z. Pflanzenphysiol. 70, 358-363.

    Deuel, H., and Stutz, E. 1958. Pectic substances and pectic enzymes. Adv. Enzymol. 20, 341-382. Dey, P. M., and Brinson, K . 1984. Plant cell walls. Adv. Carbohydr. Chem. Eiochem. 42,265-382. Dey, P. M., and Del Campillo, E. 1984. Biochemistry of the multiple forms of glycosidases in

    Doesburg, J. J. 1957. Relation between the solubilization of pectin and the fate of organic acids

    Dolendo, A. L., Luh, B. S., and Pratt, H. K. 1966. Relation of pectic and fatty acid changes to

    plants. Adv. Enzymol. 56, 141-249.

    during maturation of apples. J. Sci. Food Agric. 8, 206-216.

    respiration rate during ripening of avocado fruits. J. Food Sci. 31, 332-336.


    Elbein, A. D. 1969. Biosynthesis of a cell wall glucomannan in mung bean seedlings. J. B i d . Chem.

    English, P. D., Maglothin, A., Keegstra, K., and Albersheim, P. 1972. A cell wall degrading endopolygalacturonase secreted by Collifotrichum lindemuthiunum. Plunr Physiol. 49, 293- 297.

    Epstein, L., and Lamport, D. T. A. 1984. An intramolecular linkage involving isodityrosine in extensin. Phyfochemisrry 23, 124 I - 1246.

    Esau, K. 1965. Plant Anatomy. Wiley, New York. Evans, M. L. 1967. Kinetic studies of the cell elongation phenomenon in Avenu coleoptile segments.

    Fan, D. F., and Maclachlan, G. A. 1966. Control of cellulase activity by indoleacetic acid. Can. J.

    Fan, D. F., and Maclachlan, G. A. 1967. Studies on the regulation of cellulase activity and growth in

    Franz, G. 1973. Biosynthesis of Salep mannan. Phyrochemisrry 12, 2369-2373. Frenkel, C., Klein, I., and Dilley, D. R. 1968. Protein synthesis in relation to ripening of pome

    Fry, S. C. 1979. Phenolic components of primary cell wall and their possible role in the hormonal

    Fry, S. C. 1982a. Phenolic components of primary cell wall: Feruloylated disaccharides of D-

    Fry, S. C. 1982b. Isodityrosine, a new cross-linking amino acid from plant cell-wall glycoprotein.

    Fry, S. C. 1983. Feruloylated pectins from the primary cell wall: Their structure and possible

    Gardner, K. H., and Blackwell, J. 1974a. Hydrogen bonding in native cellulose. Biochim. Biophys.

    Gardner, K. H., and Blackwell, J. 1974b. Structure of native cellulose. Biopolymers 13, 1975-2001. Grant, G. T., Moms, E. R., Rees, D. A., Smith, P. J. C., and Thorn, D. 1973. Biological

    interactions between polysaccharides and divalent cations: the egg-box model. FEBS Len. 32,

    Grierson, D., Tucker, G., and Robertson, N. 1980. The regulation of gene expression during the ripening of tomato fruits. In Quality in Stored and Processed Vegetables and Fruits (P. W. Goodenough and R. K. Atkin, eds.), pp. 179-191. Academic Press, New York.

    Grierson, D., Tucker, G., and Robertson, N. 1981. The molecular biology of ripening. I n Recent Advances in the Biochemistry of Fruits and Vegetables (J. Friend and M. J. C. Rhodes, eds.), pp. 149-160. Academic Press, New York.

    Gross, K. C., and Wallner, S. J. 1979. Degradation of cell wall polysaccharides during tomato fruit ripening. Plant Physiol. 63, 117-120.

    Hager, A., Menzel, H., and Krauss, A. 1971. Experiments and hypothesis concerning the primary action of auxin in elongation growth. Planru 100, 47-75.

    Hamson, A. R. 1952. Factors which condition firmness in tomatoes. Food Res. 17, 370-379. Hasegawa, S., Maier, V. P., Kaszycki, H. P., and Crawford, J. K. 1969. Polygalaturonase content

    Heath, M. F., and Northcote, D. H. 1971. Glycoprotein of the wall of sycamore tissue-culture cells.

    Heiniger, U., and Delmer, D. P. 1977. UDP-glucose: Glucan synthetase in developing cotton fibers.

    244, 1608-1616.

    Ph.D. thesis. University of California, Santa Cruz.

    Bot. 44, 1025-1034.

    excised pea epicotyl sections. Can. J . Bot. 45, 1837-1844.

    fruits. Plant Physiol. 43, 1146-1 153.

    regulation of growth. Plant 146, 343-35 1.

    galactose and L-arabinose from spinach polysaccharide. Biochem. J. 203, 493-504.

    Biochem. J . 204, 449-455.

    functions. Plunta 157, 1 11-123.

    Acta 343, 232-237.

    195- 198.

    of dates and its relation to maturity and softness. J. Food Sci. 34, 527-529.

    Biochem. J . 125, 953-961.

    11. Structure of the reaction product. Plunr Physiol. 59, 719-723.


    Heyn, A. N. J . 1969. Glucanase activity in coleoptiles of Avena. Arch. Biochem. Biophys. 132,442-

    Heyn, A. N. J. 1970. Dextranase activity in coleoptiles of Avena. Science 167, 874-875. Hirst, E. L., and Jones, J. K. N. 1947. Pectic substances. Part VI. The structure of araban from

    Hirst, E. L., Jones, J. K. N., and Walder, W. 0. 1947. Pectic substances. Part VII. The constitution

    Hobson, G. E. 1963. Pectinesterase in normal and abnormal tomato fruit. Biochem. J . 86,358-365. Hobson, G. E. 1964. Polygalacturonase in normal and abnormal tomato fruit. Biochem. J . 92, 324-

    Hobson, G. E. 1965. The ripening of tomato fruit as affected by the injection of certain chemicals. J.

    Hobson, G. E. 1979. What factors are involved in the onset of ripening in climacteric fruit? Comment. Plant Sci. 2, 225-236.

    Hobson, G. E. 1981. Enzymes and texture changes during ripening. In Recent Advances in the Biochemistry of Fruits and Vegetables (J. Friend and M. J. C. Rhodes, eds.), pp. 123-132. Academic Press, New York.

    Hobson, G. E., and Davies, J. N. 1971. Biochemistry and physiology of commercially important fruits. 13. Tomato. In The Biochemistry of Fruits and Their Products (A. C. Hulme, ed.), Vol. 2, pp. 437-482. Academic Press, New York.

    Holtz, G. V., and Varner, J. E. 1984. Reinforced polyproline I1 conformation in a hydroxy-proline- rich cell wall glycoprotein from carrot root. Plant Physiol. 74, 247-25 1.

    Homer, R. B., and Roberts, K. 1979. Glycoprotein conformation in plant cell walls. Circular dichroism reveals a polyproline I1 structure. Plant 146, 217-222.

    Hsu, D. S. , and Reeves, R. E. 1967. The structure of Nasturatium amyloid. Carbohydr. Res. 5 ,

    Hulme, A. C. 1954. Studies in the nitrogen metabolism of apple fruits. The climacteric rise in respiration in relation to changes in the equilibrium between protein synthesis and breakdown. J .

    Hulme, A. C. 1958. Some aspects of the biochemistry of apple and pear fruits. Adv. Food Res. 8,

    Hulme, A. C. 1961. Some enzymes concerned in ripening of fruits. Soc. Chem. I n d . (London)

    Hultin, H. O . , and Levine, A. S. 1965. Pectin methyl esterase in the ripening banana. J. Food Sci.

    Jansen, E. F., Jang, R., and Bonner, J. 1960. Binding of enzymes to Avena coleoptile cell walls.

    Jaynes, T. A,, Haskins, F. A, , Gort, H. J., and Kleinhofs, A. 1972. Solubility of P-glucosidase in

    Jermyn, M. A., and Isherwood, F. A. 1956. Changes in the cell wall of pear during ripening.

    Jewell, G. G. 1979. Fruits and vegetables. In Food Microscopy (J. G. Vaughan, ed.), pp. 1-34.

    Jiang, K. S. , and Timell, T. E. 1972. Polysaccharides in the bark of aspen (Populus tremuloides). 11.

    Jones, J. D., Hulme, A. C., and Woolforton, L. S. C. 1965. The respiration climacteric in apple

    Joseleau, J. P., and Chambat, G. 1984. Structure of primary cell walls of suspension-cultured Rosa


    Arachis hypogea. J . Chem. Soc. 1221-1225.

    of galactan from Lupinus alba. J. Chem. Soc. 1225-1229.


    EXP. Bot. 16, 411-422.


    EXP. Bot. 5, 159-172.


    Monogr. 11, 93-106.

    30, 917-921.

    Plant Physiol. 35, 567-574.

    homogenates of sweetclover leaves and bean hypocotyls. Plant Physiol. 49, 277-279.

    Biochem. J . 64, 123-132.

    Academic Press, New York.

    Isolation and structure of an arabinan. Cell. Chem. Technol. 6, 499-502.

    fruits. New Phytol. 64, 158-167.

    glauca cells. 11. Multiple forms of xyloglucans. Plant Physiol. 74, 694-700.


    Joseleau, J. P., Chambat, G., Vignon, M., and Bamoud, F. 1977. Chemical and I3C N.M.R. studies on two arabinans from the inner bark of young stems of Rosa glauca. Carbohydr. Res. 58, 165- 175.

    Karacsonyi, S., Toman, R., Janecek, F., and Kubackova, M. 1975. Polysaccharides from the bark of the white willow (Salix alba L.): Structure of an arabinan. Carbohydr. Res. 44, 285-290.

    Katz, M., and Ordin, L. 1967. A cell wall polysaccharide-hydrolysing enzyme system in Avena sariva L. coleoptile. Biochim. Biophys. Acta 141, 126- 134.

    Kauss, H. 1974. Biosynthesis of pectins and hemicelluloses. In Plant Carbohydrate Biochemistry (J. B. Pridham, ed.), pp. 191-205. Academic Press, New York.

    Kazama, H., and Katsumi, M. 1976. Biphasic response of cucumber hypocotyl sections to auxin. Plant Cell Physiol. 17, 467-473.

    Keegstra, K., and Albersheim, P. 1970. The involvement of glycosidases in the cell wall metabolism of suspension-cultured Acer pseudoplatanus cells. Plant Physiol. 45, 675-678.

    Keegstra, K., Talmadge, K. W., Bauer, W. D., and Albersheim, P. 1973. The structure of plant cell walls. 111. A model of the walls of suspension-cultured sycamore cells based on the interconnec- tions of the macromolecular components. Plant Physiol. 51, 188-196.

    Kidd, F., and West, C. 1930. Physiology of fruit. I. Changes in the respiratory activity of apples during their senescence at different temperatures. Proc. R . SOC. London Ser. B 106, 93-109.

    Klis, F. M., Dalhuizen, R., and Sol, K. 1974. Wall-bound enzymes in callus of Convolvulus arvensis. Phytochemistry 13, 55-57.

    Knee, M. 1973a. Polysaccharides and glycoproteins of apple fruit cell walls. Phytochemistry 12,

    Knee, M. 1973b. Polysaccharide changes in cell walls of ripening apples. Phytochemistry 12, 1543- 1549.

    Knee, M. 1978a. Properties of polygalacturonate and cell cohesion in apple fruit cortical tissue. Phytochemistry 17, 1257-1260.

    Knee, M. 1978b. Metabolism of polymethylgalacturonate in apple fruit cortical tissue during ripen- ing. Phyrochemisrry 17, 1261-1264.

    Knee, M., and Bartley, I. M. 1981. Composition and metabolism of cell wall plysaccharides in ripening fruits. In Recent Advances in the Biochemistry of Fruits and Vegetables (J. Friend and M. J. C. Rhodes, eds.), pp. 133-148. Academic Press, New York.

    Knee, M., Fielding, A. H., Archer, S. A, , and Laborda, F. 1975. Enzymic analysis of cell wall structure in apple fruit cortical tissue. Phyrochemisrry 14, 2213-2222.

    Knee, M., Sargent, J . A., and Osborne, D. J . 1977. Cell wall metabolism in developing strawberry fruits. J. Exp. Bor. 28, 377-396.

    Kolpak, F. J . , and Blackwell, J. 1975. Mercerization of cellulose. Textile Res. J. 45, 568-572. Kolpak, F. J . , and Blackwell, J . 1976. Determination of the structure of cellulose 11. Macromole-

    Kooiman, P. 1961. The constitution of Tamarindus amyloid. Rec. Trav. Chim. 80, 849-865. Kopeliovitch, E., Mizrahi, Y., Rabinowitch, H. D., and Kedar, N. 1980. Physiology of the tomato

    mutant alcobaca. Physiol. Plant. 48, 307-31 1. Kosiyachinda, S. , and Young, R. E. 1975. Ethylene production in relation to the initiation of

    respiratory climacteric in fruit. Plant Cell Physiol. 16, 595-601. Kirshnamurthy, S., and Subramanyam, H. 1973. Pre- and post-harvest physiology of the mango fruit

    (Mangifera indica L.). Trop. Sci. 15, 167-181. Ku, L. L., and Romani, R. J. 1970. The ribosomes of pear fruit. Their synthesis during the

    climacteric and the age-related compensatory response to ionizing radiation. Plant Physiol. 45,

    Labavitch, J. M. 1981. Cell wall turnover in plant development. Annu. Rev. Plant Physiol. 32, 385-


    cules 9, 273-278.

    40 1-407.



    Labavitch, J . M., Freeman, L. E., and Albersheim, P. 1976. Structure of plant cell walls. Purifica- tion and characterization of a P-1,4-galactanase which degrades a structural component of the primary cell walls of dicots. J . Eiol. Chem. 251, 5904-5910.

    Lamport, D. T. A. 1965. The protein components of primary cell walls. Adv. Bor. Res. 2, 151-218. Lamport, D. T. A. 1967. Hydroxyproline-0-glycosidic linkage of the plant cell wall glycoprotein

    extensin. Nurure (London) 216, 1322- 1324. Lamport, D. T. A. 1969. The isolation and partial characterization of hydroxyproline-rich glycopro-

    teins obtained by enzymic degradation of primary cell wall. Biochemistry 8, 1155-1 163. Lamport, D. T. A. 1970. Cell wall metabolism. Annu. Rev. Plant Physiol. 21, 235-270. Lamport, D. T. A. 1973. Glycopeptide of extensin. 0-o-Galactosyl serine and 0-L-arabinosyl

    hydroxyproline. In Biogenesis of Plant Cell Wall Polysaccharides (F. A . Loewus, ed.), pp. 149-164. Academic Press, New York.

    Lamport, D. T. A. 1977. Structure, biosynthesis, and significance of cell wall glycoprotein. Recent Adv. Phytochem. 11, 79-1 15.

    Lamport, D. T. A. 1980. Structure and function of plant glycoproteins. I n The Biochemistry of Plants (J. Preiss, ed.), Vol. 3, pp. 501-541. Academic Press, New York.

    Lamport, D. T. A, , and Catt, J . W. 1981. Structure and Function of plant glycoproteins. I n Encyclopedia of Plant Physiology (W. Tanner and F. A. Loewus, eds). Vol. 13B, pp. 133- 165. Springer-Verlag, Berlin and New York.

    Lamport, D. T. A,, and Epstein, L. 1983. A new model for the primary cell wall: A concatenated extensin-cellulose network. Curr. Top. Plunr Biochem. Physiol. 2, 73-83.

    Lamport, D. T. A,, and Miller, D. H. 1971. Hydroxyproline arabinosides in the plant kingdom. Plant Physiol. 48, 454-456.

    Lance, G. , Hobson, G. E. , Young, R. E., and Biale, J . B. 1967. Metabolic processes in cytoplasmic particles of the avocado fruit. IX. The oxidation of pyruvate and malate during the climacteric cycle. Plunr Physiol. 42, 471-478.

    Larm, O., Theander, 0.. and Aman, P. 1976. Structural studies on a water-soluble arabinogalactan isolated from rapeseed (Brussicu nupus) Acru Chem. Scund. Ser. B 30, 627-630.

    Lee, M., and Macmillan, J. D. 1970. Mode of action of pectic enzymes. 111. Site of initial action of tomato pectinesterase on highly esterified pectin. Biochemistry 9, 1930- 1934.

    Lee, S., Kivilaan, A , , and Bandursk, R. S. 1967. In vitro autolysis of plant cell walls. Plant Physiol. 42, 968-972.

    Letham, D. S. 1967. Chemistry and physiology of kinetin-like compounds. Annu. Rev. Plant Physiol. 18, 349-364.

    Liu, T. Y., Elbein, A. D., and Su, J. C. 1966. Substrate specificity in pectin synthesis. Biochem. Eiophys. Res. Commun. 22, 650-657.

    Looney, N. E., and Patterson, M. E. 1967. Changes in total ribonucleic acid during the climacteric phase in yellow transparent apples. Phytochernistry 6, 1517- 1520.

    Luh, B. S., Leonard, S. J., and Phaff, H. J . 1956. Hydrolysis of pectic materials and oligouronides by tomato polygalacturonase. Food Res. 21, 448-455.

    McCready, R. M., and McComb, E. A. 1954. Texture changes in brined cherries. West. Cunner Packer 46, nos. 13, 17, 19.

    McCready, R. M., McComb, E. A., and Jansen, E. G . 1955. Action of tomato and avocado polygalacturonase. Food Res. 20, 186- 191.

    McGlasson, W. B., Dostal, H. C . , and Tigchelaar, E. C. 1975. Comparison of propylene-induced responses of immature fruit of normal and rin mutant tomatoes. Plant Physiol. 55, 218-222.

    Maclachlan, G. A,, and Duda, C. T. 1965. Changes in concentration of polymeric components in excised pea-epicotyl tissue during growth. Eiochirn. Biophys. Acfu 97, 288-299.

    McNeil, M., and Albersheim, P. 1977. Chemical ionization mass spectrometry of methylated hexitol acetates. Curbohydr. Res. 56, 239-248.


    McNeil, M., Darvill, A, , and Albersheim, P. 1979. The structural polymers of the primary cell walls of dicots. Forrschr. Chem. Org. Nurursr. 37, 191-249.

    McNeil, M., Darvill, A., and Albersheim, P. unpublished results cited in Darville, A., McNeil, M., Albersheim, P., and Delmer, D. P. 1980. The primary cell walls of flowering plants. In Biochemistry of Plants (N. E. Tolbert, ed.), Vol. 1 , pp. 91-162. Academic Press, New York.

    Malis-Arad, S. , and Mizrahi, Y. 1983. Stress-induced ripening of the non-ripening tomato mutant nor. Physiol. Plunr. 59, 213-217.

    Mapson, L. W., and Robinson, J. E. 1966. Relation between oxygen tension, biosynthesis of ethylene, respiration, and ripening changes in banana fruit. 1. Food Technol. 1, 215-225.

    Mascaro, L. J . , and Kindel, P. K. 1977. Characterization of [T]apiogalacturonans synthesised in a cell-free system from Lemnu minor. Arch. Biochem. Biophys. 183, 139-148.

    Masuda, Y . , Oi, S.. and Satamura, Y. 1970. Studies on the role of cell wall-degrading enzymes in cell wall loosening in oat coleoptiles. Plunr Cell Physiol. 11, 631-638.

    Matchett, W. H., and Nance, J. F. 1962. Cell wall breakdown and growth in pea seedling stems. Am. 1. Bor. 49, 311-319.

    Mattoo, A. K., and Modi, V. V. 1969. Ethylene and ripening of mangoes. Plunr Physiol. 44, 308- 310.

    Meier, H. 1962. Studies on a galactan from tension wood of Beech (Fugus silvuricu L.). Acm Chem. Scand. 16, 2275-2283.

    Mettler, I. J . , and Romani, R. J. 1976. Quantitative changes in tRNA during ethylene-induced ripening (ageing) of tomato fruits. Phyrochemistry 15, 25-28.

    Millerd, A,, Bonner, J., and Biale, J. B. 1953. The climacteric rise in fruit respiration as controlled by phosphorylative coupling. PIanr Physiol. 28, 521-531.

    Mizrahi, Y . , Dostal, H.-C., and Cherry, J. H. 1976. Descriptive physiology and biochemistry of the abnormally ripening tomato fruit (Lycopersicon esculenrum) cv. Snowball. Physiol. Plant. 38,

    Mizrahi, Y., Zohar, R., and Malis-Arad, S. 1982. Effect of sodium chloride on fruit ripening of nonripening tomato mutants nor and rin. Plant Physiol. 67, 497-501.

    Monro, J. A,, Bailey, R. W., and Penny, D. 1974. Cell wall hydroxyproline-polysaccharide associa- tion in Lupinus hypocotyls. Phyrochemisrry 13, 375-382.

    Monro, J. A., Bailey, R. W., and Penny, D. 1976a. Hemicellulose fractions and associated protein of lupin hypocotyl cell wall. Phyrochemisrry 15, 175- 18 I .

    Monro, J. A., Penny, D., and Bailey, R. W. 1976b. The organization and growth of primary cell walls of lupin hypocotyls. Phyrochemisrry 15, 1193-1 198.

    Morita, M. 1965a. Polysaccharides of soybean seeds. I. Polysaccharide constituents of hot-water- extract fraction of soybean seeds and an arabinogalactan as its major component. Agric. Biol. Chem. (Tokyo) 29, 564-573.

    Morita, M. 1965b. Polysaccharides of soybean seeds. 11. Methylated arabinogalactan isolated from methylated product of hot-water-extract fraction of soybean seed polysaccharides. Agric. Biol. Chem. (Tokyo) 29, 626-630.

    Mort, A. J., and Lamport, D. T. A. 1977. Anhydrous hydrogen fluoride deglycosylates glycopro- teins. Anal. Biochem. 82, 289-309.

    Moshrefi, M . , and Luh, B. S. 1983. Carbohydrate composition and electrophoretic properties of tomato polygalacturonase isoenzymes. Eur. J . Biochem. 135, 51 1-514.

    Moshrefi, M., and Luh, B. S. 1984. Purification and characterization of two tomato polygalacturo- nase isoenzymes. 1. Food Biochem. 8, 39-54.

    Miihlethaler, K. 1967. Ultrastructure and formation of plant cell walls. Annu. Rev. PIunr Physiol. 18,

    309-3 12.



    Nagel, C. W., and Patterson, M. E. 1967. Pectic enzymes and development of the pear (Pyrus

    Nelmes, B. J., and Preston, R. D. 1968. Wall development in apple fruits: A study of the life history

    Neukom, H., and Deuel, H. 1958. Alkaline degradation of pectin. Chem. Ind. (London) 683. Nevins, D. J. 1970. Relation of glycosidases to bean hypocotyl growth. Planr Physiol. 46,458-462. Nitsch, J. P. 1952. Plant hormones in the development of fruits. Q. Rev. Biol. 27, 33-57. Nitsch, J. P. 1953. The physiology of fruit growth. Annu. Rev. Plant Physiol. 4, 199-236. Northcote, D. H. 1972. Chemistry of the plant cell wall. Annu. Rev. Planr Physiol. 23, 113-132. Odzuck, W., and Kauss, H. 1972. Biosynthesis of pure araban and xylan. Phyrochemisrry 11,2489-

    2494. ONeill, M. A., and Selvendran, R. R. 1980. Glycoproteins from the cell wall of Phnseolus

    coccineus. Biochem. J . 187, 53-63. Ordin, L., and Hall, M. A. 1968. Cellulose synthesis in higher plants from UDP-glucose. Plant

    Pan, Y . T., and Kindel, P. K. 1977. Characterization of particulate D-apiosyl and o-xylosyltransfer-

    Panayotatos, N., and Villemez, C. L. 1973. The formation of a P-1.4-galactan chain catalysed by a

    Parr, D. R., and Edelman, J. 1975. Release of hydrolytic enzymes from the cell walls of intact and

    Patel, D. S . . and Phaff, H. J. 1960. Properties of tomato polygalacturonase. Food Res. 25, 47-57. Pearson, I . A., and Robertson, R. N. 1954. The physiology of growth in apple fruits. VI. The control

    of respiration rate and synthesis. Ausr. J. Biol. Sci. 7, 1-17. P6aud-Leniiel. C., and Axelos. M. 1970. Structural features of the P-glucans enzymically syn-

    thesised from uridine diphosphate glucose by wheat seedlings. FEBS Lert. 8, 224-228. Pharr, D. M., Sox, H. N., and Nesbitt, W. B. 1976. Cell wall-bound nitrophenyl glycosidases of

    tomato fruits. J. Am. SOC. Horric. Sci. 101, 397-400. Pierrot, H., and van Wielink, J. E. 1977. Localization of glycosidases in the wall of living cells from

    cultured Convolvulus arvensis tissue. PIanra 137, 235-242. Polevoi, V. V. 1967. The action of auxin on electropotential, growth, respiration, and synthesis of

    RNA in sections of corn seedlings. Order of reactions. Wiss. Z. Univ. Rosrok, Math. -Narunviss. Reihe 16, 477-478.

    Poovaiah, B. W., and Nukaya, A. 1979. Polygalacturonase and cellulase enzymes in the normal rutgers and mutant rin tomato fruits and their relationship to the respiratory climacteric. Planr

    Pratt, H. K., and Goeschl, I . D. 1968. Role of ethylene in fruit ripening. In Biochemistry and Physiology of Plant Growth Substances (F. Wightman, ed.), pp. 1295-1302. Runge, Ottawa.

    Pratt, H. K.. Goeschl, J. D., and Martin, F. W. 1977. Fruit growth and development, ripening and the role of ethylene in the Honey-Dew muskmelon. J. Am. SOC. Horric. Sci. 102, 203-210.

    Pressey, R. 1977. Enzymes involved in fruit softening. Am. Chem. SOC. Symp. Ser. 47, 172-191. Pressey, R. 1983. P-Galactosidases in ripening tomatoes. Planr Physiol. 71, 132-135. Pressey, R., and Avants, J. K. 1973. Separation and characterization of endopolygalacturonase and

    Pressey, R., and Avants. J. K. 1975. Cucumber polygalacturonase. J. Food Sci. 40, 937-939. Pressey, R., and Avants, J . K. 1976. Pear polygalacturonases. Phyrochemisrry 15, 1349-1351. Pressey, R., Hinton, D. M., and Avants, J. K. 1971. Development of polygalacturonase activity and

    Preston, R. D. 1964. Structural and mechanical aspects of plant cell walls with particular reference to

    communis). J . Food Sci. 32, 294-297.

    of a parenchyma cell. J. Exp. Bor. 19, 496-518.

    Physiol. 43, 473-476.

    ase from Lemnu minor. Arch. Biochem. Biophys. 183, I3 I- 138.

    Phuseolus aureus enzyme. Biochem. J . 133, 263-271.

    disrupted carrot callus tissue. Planra 127, 1 I 1- I 19.

    Physiol. 64, 534-537.

    exopolygalacturonase from peaches. Planr Physiol. 52, 252-256.

    solubilization of pectin in peaches during ripening. J. Food Sci. 36, 1070-1073.


    synthesis and growth. In The Formation of Wood in Forest Trees (M. A. Zimmerman, ed.), pp. 169-188. Academic Press, New York.

    Preston, R. D. 1974. The Physical Biology of Plant Cell Walls, Ch. 1. Chapman & Hall, London. Quazi, M. H., and Freebairn, H. T. 1970. The influence of ethylene, oxygen, and carbon dioxide on

    Rattanapanone, N., Grierson, D., and Stein, M. 1977. Ribonucleic acid metabolism during the

    Rattanapanone, N., Speirs, J., and Grierson, D. 1978. Evidence for changes in messenger RNA

    Ray, P. M. 1%2. Cell wall synthesis and cell elongation in oat coleoptile tissue. Am. J. Bor. 49,

    Ray, P. M. 1963. Sugar composition of oat coleoptile cell wall. Biochem. J. 89, 144-150. Ray, P. M. 1967. Radioautographic study of cell wall deposition in growing plant cells. J. Cell Biol.

    Ray, P. M. 1969. In Communications in Development (A. Lang, ed.), pp. 172-205. Academic

    Ray, P. M. 1975. Golgi membranes form xyloglucan from UDP-glucose and UDP-xylose. Planr

    Ray, P. M., and Rottenberg, D. A. 1964. Uronic acid constituents of oat coleoptile cell walls.

    Ray, P. M., and Ruesink, A. W. 1962. Kinetic experiments on the nature of the growth mechanism

    Rayle, D. L., and Cleland, R. 1970. Enhancement of wall loosening and elongation in acid solutions.

    Rayle, D. L., Evans, M. L., and Rainer, H. 1970a. Action of auxin on cell elongation. Proc. Narl.

    Rayle, D. L., Haughton, P. M., and Cleland, R. 1970b. In virro system that stimulates plant cell

    Raymond, Y . , Fincher, G. B . , and Maclachlan, G. A. 1978. Tissue slice and particulate P-glucan

    Rees, D. A. 1972. Shapely polysaccharides. Biochem. J. 126, 257-273. Rees, D. A., and Richardson, N. G. 1966. Polysaccharides in germination. Occurrence, fine struc-

    ture, and possible biological role of the pectic arabinan in white mustard cotyledons. Biochemis- rry 5 , 3099-3107.

    Rees, D. A., and Welsh, E. J. 1977. Secondary and tertiary structures of polysaccharides in solutions and gels. Angew. Chem. Int. Ed. Engl. 16, 214-224.

    Rexova-Benkova, L., and Markovic, 0. 1976. Pectic enzymes. Adv. Carbohydr. Chem. Biochem.

    Reymond, D., and Phaff, H. J. 1%5. Purification and certain properties of avocado polygalacturo-

    Rhodes, M. J. C., and Reid, M. S. 1975. Factors in regulation of maturation of fruits. CNRS Rep.

    Richmond, A., and Biale, J. B. 1967. Protein and nucleic acid metabolism in fruits. RNA synthesis

    Ring, S. G., and Selvendran, R. R. 1978. Purification and methylation analysis of cell wall material

    Ring, S. G., and Selvendran, R. R. 1981. An arabinogalactoxyloglucan from the cell wall of

    Robinson, D. G. 1977. Plant cell wall synthesis. Adv. Bor. Res. 5 , 89-15],

    the ripening of banana. Bor. Gaz. 131, 5-14.

    development and ripening of tomato fruits. Phyrochemistry 16, 629-633.

    content related to tomato fruit ripening. Phyrochemistry 17, 1485- 1486.


    35, 659-674.

    Press, New York.

    Physiol. 56, S-84.

    Biochem. J . 90, 646-655.

    in oat coleoptile cells. Dev. Biol. 4, 377-397.

    Planr Physiol. 46, 250-253.

    Acad. Sci. U.S.A. 65, 184-191.

    extension growth. Proc. Narl. Acad. Sci. U.S.A. 67, 1814-1817.

    synthetase activities from Pisum epicotyls. Planr Physiol. 61, 938-942.

    33, 323-385.

    nase. J. Food Sci. 30, 266-273.

    Paris 238, 189-192.

    during the respiratory rise of avocado. Biochim. Biophys. Acra 138, 625-627.

    from Solanum tuberosum. Phyrochemistry 17, 745-752.

    Sohnum ruberosum. Phyrochemisrry 20, 25 1 1-25 19.


    Rolz, C., Deshpande, S . , Piaz, L., Ortiz, J., Flores, M. C., Sanchez, M., and De Ortega, M. 1972. Chemical changes and fruit quality during the ripening of tropical fruits. Turriulba 22, 65-72.

    Rouhani, I., and Bassiri, A. 1976. Changes in the physical and chemical characteristics of shahani dates during development and maturity. J . Hortic. Sci. 51, 489-494.

    Rowan, K. S., Pratt, H. K., and Robertson, R. N. 1958. The rleation of high-energy phosphate content, protein synthesis, and climacteric rise in the respiration of ripening avocado and tomato fruits. Aust. J . Biol. Sci. 11, 329-335.

    Sacher, J. A. 1966. Permeability characteristics and amino acid incorporation during senescence (ripening) of banana tissue. Plant Physiol. 41, 701-708.

    Sacher, J. A. 1973. Senescence and post-harvest physiology. Annu. Rev. Plant Physiol. 24, 197- 224.

    Saltveit, M. E., Jr., and McFeeters, R. F. 1980. Polygalacturonase activity and ethylene synthesis during cucumber fruit development and maturation. Plunc Physiol. 66, 1019-1023.

    Sarko, A., and Muggli, R. 1974. Packing analysis of carbohydrates and polysaccharides. 111. Valonia cellulose and cellulose 11. Mucromolecules 7, 486-494.

    Schewfelt, A. L. 1965. Changes and variation in the pectic constitution of ripening peaches as related to product firmness. J . Food Sci. 30, 573-576.

    Schneider, C. L. 1938. Interdependence of auxin and sugar for growth. Am. J. Bor. 25, 258-270. Schwarz, U., Asmus, A,, and Frank, H. 1969. Autolytic enzymes and cell division of Escherichiu

    coli. J . Mol. Biol. 41, 419-429. Selvendran, R. R. 1975. Cell wall glycoproteins and polysaccharides of parenchyma of Phuseolus

    coccineus. Phyrochemisrry 14, 2175-2180. Selvendran, R. R. 1983. The chemistry of plant cell wall. In Dietary Fibre (G. G. Birch and K. J.

    Parker, eds.), pp. 95-147. Applied Science, London. Selvendran, R. R., Davies, A. M. C., and Tidder, E. 1975. Cell wall glycoproteins and polysaccha-

    rides of mature runner beans. Phytochemistry 14, 2169-2174. Siddiqui, I. R., and Wood, P. J. 1971. Structural investigation of water-soluble rapeseed (Brussicu

    cumpesrris) polysaccharides. Part I. Rapeseed amyloid. Curbohydr. Res. 17, 97-108. Siddiqui, I. R., and Wood, P. J. 1974. Structural investigation of oxalate-soluble rapeseed (Brussicu

    cumpesrris) polysaccharides. Part 111. An arabinan. Curbohydr. Res. 36, 35-44. Siddiqui, I. R., and Wood, P. J. 1976. Structural investigation of oxalate-soluble rapeseed (Brussicu

    compestris) polysaccharides. Part IV. Pectic polysaccharides. Curbohydr. Res. 50, 97- 107. Siddiqui, I. R., and Wood, P. J. 1977. Structural investigation of sodium hydroxide-soluble rapeseed

    (Brussicu cumpesrris) polysaccharides. Part V. Fucoamyloid. Curbohydr. Res. 53, 85-94. Smith, J. J., Muldoon, E. P., and Lamport, D. T. A. 1984. Isolation of extensin precursors by direct

    elution of intact tomato cell suspension cultures. Phyrochernistry 23, 1233- 1239. Smith, M. M., and Stone, B. A. 1973. P-Glucan synthesis by cell-free extracts from Lolium

    multiflorum endosperm. Biochim. Biophys. Accu 313, 72-94. Smith, M. M., Axelos, M., and P6aud-Lentie1, C. 1976. Biosynthesis of mannan and mannolipids

    from GDP-mannose by membrane fractions of sycamore cell culture. Biochimie 58, 1195- 1211.

    Soll, H. J., and Bottger, M. 1981. Physicochemical properties of cell wall and pectates in relation to acid-induced elongation. I n Cell Walls 81 (D. G. Robinson and H. Quader, eds.), pp. 206- 219. Wissenschaftliche Verlagsgesellschaft, Stuttgart.

    Solms, J., and Deuel, H. 1955. Uber den Mechanismus der enzymatischen Verschifung von Pek- tinstoffen. Helv. Chem. Acfu 38, 321-329.

    Spiro, R. G. 1970. Glycoproteins. Annu. Rev. Biochem. 39, 599-638. Stevens, B. J. H., and Selvendran, R. R. 1980. Structural investigation of an arabian from cabbage

    (Brassicu oleruceu var. Capitata). Phytochemistry 19, 559-561.


    Stevens, B. J . H., and Selvendran, R. R. 1984. Pectic polysaccharides of cabbage (Brussicu oleruceu). Phyrochemisrry 23, 107-1 15.

    Stoddart, R. W., and Northcote, D. H. 1967. Metabolic relationships of the isolated fractions of the pectic substances of actively growing sycamore cells. Biochem. J. 105, 45-59.

    Stoddart, R. W., Barrett, A. J., and Northcote, D. H. 1967. Pectic polysaccharides of growing plant tissues. Biochem. J . 102, 194-204.

    Tagawa, T., and Bonner, J. 1957. Mechanical properties of the Avenu coleoptiles as related to auxin and to ionic interactions. Plunr Physiol. 32, 207-212.

    Talmadge, K. W., Keegstra, K., Bauer, W. D., and Albersheim, P. 1973. The structure of plant cell walls. I . The macromolecular components of the walls of suspension-cultured sycamore cells with a detailed analysis of the pectic polysaccharides. Plant Physiol. 51, 158-173.

    Toman, R., Karacsonyi, S., and Kovacik, V. 1972. Polysaccharides from the bark of the white willow (Sulk ulba L.). Structure of a galactan. Curbohydr. Res. 25, 371-378.

    Toman, R., Karacsonyi, S., and Kubackova, M. 1975. Studies on the pectin present in the bark of white willow (Salix ulbu L.): Fractionation and acidic polymerization of the water-soluble pectin. Curbohydr. Res. 43, 1 11-1 16.

    Tucker, G. A., and Grierson, D. 1982. Synthesis of polygalacturonase during tomato fruit ripening. Plunru 155, 64-67.

    Tucker, G. A., Robertson, N. G., and Grierson, D. 1980. Changes in polygalacturonase isoenzymes during the ripening of normal and mutant tomato fruit. Eur. J. Biochem. 112, 119-124.

    Valent, B. S., and Albersheim, P. 1974. The structure of plant cell walls. V. On the binding of xyloglucans to cellulose fibres. PIunr Physiol. 54, 105- 108.

    Valent, B. S., McNeil, M., and Albersheim, P. unpublished results cited in Darvill, A., McNeil, M., Albersheim, P., and Delmer, D. P. 1980. The primary cell walls of flowering plants. In Biochemistry of Plants (N. E. Tolbert, ed.), Vol. 1, pp. 91-162. Academic Press, New York.

    Vanderhoef, L. N. 1979. Auxin-regulated cell enlargement: Is there action at the level of gene expression? In Genome Organization and Expression in Plants (C. J. Leaver, ed.), pp. 159- 173. Plenum, New York.

    Vanderhoef, L. N. 1980. Auxin-regulated elongation: A summary hypothesis. In Plant Growth Substances (F. K. Skoog, ed.), pp. 90-96. Springer-Verlag, Berlin.

    Vanderhoef, L. N., and Dute, R. R. 1981. Auxin-regulated wall loosening and sustained growth in elongation. Plant Physiol. 67, 146- 149.

    Vanderhoef, L. N., and Stahl, C. A. 1975. Separation of two responses to auxin by means of cytokinin inhibition. Proc. Nurl. Acud. Sci. V.S.A. 72, 1822- 1825.

    Vanderhoef, L. N., Stahl, C. A., Williams, C. A,, Brinkman, K. A., and Greenfield, J. C. 1976a. Additional evidence for separable responses to auxin in soybean hypocotyl. Plum Physiol. 57,

    Vanderhoef, L. N., Stahl, C. A, , and Lu, T. S. 1976b. Two elongation responses to auxin respond differently to protein synthesis inhibition. Plunr Physiol. 58, 402-404.

    Villemez, C. L. 1971. Rate studies of polysaccharide biosynthesis from guanosine diphosphate-am- glucose and guanosine diphosphate-a-o-mannose. Biochem. J. 121, 15 1- 157.

    Villemez, C. L., and Hinman, M. 1975. UDP-Glucose-stimulated formation of xylose containing polysaccharides. Plunr Physiol. 56, S-79.

    Villemez, C. L., Liu, T. Y., and Hassid, W. Z. 1965. Biosynthesis of polygalacturonic acid chain of pectin by a particulate enzyme preparation from Phuseolus uureus seedlings. Proc. Nurl. Acud. Sci. U.S.A. 54, 1626-1632.

    Waldron, K. W., and Brett, C. T. 1983. A glucuronyltransferase involved in glucuronoxylan synthesis in pea (Pisum surivum) epicotyls. Biochem. J . 213, 115-122.

    817-8 19.


    Wallner, S. J., and Bloom, H. L. 1977. Characteristics of tomato cell wall degradation in vifro. Implications for the study of fruit-softening enzymes. Plant Physiol. 60, 207-2 10.

    Wallner, S. J., and Walker, J . E. 1975. Glycosidases in cell wall-degrading extracts of ripening tomato fruits. Plant Physiol. 55, 94-98.

    Wilder, B. M., and Albersheim, P. 1973. The structure of plant cell walls: A strucural comparison of the wall hemiceullulose of cell suspension cultures of sycamore (Acer pseudoplatanus) and red kidney bean (Phaseolus vulgaris). Plant Physiol. 51, 889-893.

    Wood, P. J. , and Siddiqui, I. R. 1972. Isolation and structural studies of a water-soluble galactan from potato (Solanurn tuberosum) tubers. Carbohydr. Res. 22, 212-220.

    Yamaki, S., and Matsuda, K. 1977. Changes in the activities of some cell wall-degrading enzymes during development and ripening of Japanese pear fruit (Pyrus serotina Rehder var. culta Rehder). Plant Cell Physiol. 18, 81-93.

    Young, R. E., Salminen, S., and Somsrivichai, P. 1975. Factors in regulation of maturation in fruits. CNRS Rep. Paris 238, 271-282.

    Zitko, V., and Bishop, C. T. 1966. Structure of a galacturonan from sunflower pectic acid. Can. J . Chem. 44, 1275-1282.

  • This Page Intentionally Left Blank







    V. VI.


    Kikkoman Corporation, Noda-shi. Chiba-ken 278, Japan

    Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Japanese Shoyu B. The Soy Sauce Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Koikuchi Shoyu B. Usukuchi Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Tamari Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    E. Saishikomi Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recent Research and Technological Advances in Shoyu Manufacture A. Comparison between Whole and Defatted Soybeans as

    Raw Materials . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Treatment of Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Koji Molds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    D. Shiro Shoyu . . . . . . . . . . . . . . . . . . .

    D. Koji Making . . . . . . . . . . . . . . . . .

    F. Refining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    A. Color Compounds of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Measurement of Shoyu Color . , . . . . . . . . . . . . . . . . . . . . . . . C. Browning Mechanism of Shoyu Flavor Evaluation of Koikuchi Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Volatile Flavor Ingredients of Koikuchi Shoyu . . . . . . . . . . . . . . . . . . . . . . . A. Organic Acids B. Alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Esters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    E. Phenolic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Lactones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    H. Sulfur-Containing Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Color of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


    D. Carbonyls and Related Compounds . . . .

    G. Pyrazines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    196 197 202 204 204 206 206 201 208 209

    209 211 220 222 227 235 24 1 242 243 244 251 26 I 264 265 267 261 278 280 28 1 282

    195 Copyright 0 1986 by Academic Press. Inc.

    All rights of reproduction in any form reserved.


    I. Terpenes ................................. J. Flavor Constituents of the Topnote Aroma of the

    K. Methods of Quantitative Analyses of the Volatile Flavor Constituents

    L. Contribution of Volatile Flavor Constituents to Overall Flavor Evaluation ......................

    VII. Safety Problem of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonproductivity of Mycotoxins by Japanese Industrial Molds . . . . . . . Fluorescent Compounds Produced by Aspergillus Molds with R f Values Resembling Those of Aflatoxins . . . . . . . . . . . . .

    of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    A. B.

    C. Mycotoxins Other Than Aflatoxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mutagenic Substances in Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    F. Biological Tests of Shoyu . . . . . . . . . . . . . . . . .

    D. E. Bactericidal Action of Shoyu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    VIII. Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . B. Koji Molds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Reduction of Fermentation Period of Mash ...................... D. Application of Enzyme Preparations . E. Refining and Pasteurization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Flavor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    283 283


    286 287 287

    290 297 298 299 300 30 1 30 I 30 I 302 303 303 304 304 313


    There are major differences between the saccharification process in traditional food preparation of Western countries and that of the Orient. The amylolytic enzymes used for saccharification in Western countries have been derived from malt, while in the Orient, Aspergillus or Rhizopus molds have been utilized as the main source of amylolytic enzymes.

    For example, in the West, beer is prepared by first saccharifying the starch of barley with malt, while in the preparation of alcoholic beverages from rice or wheat in the Orient, Aspergillus or Rhizopus molds are cultured on some parts of these raw materials to produce amylolytic enzymes. These mold-cultured mate- rials, called koji in Japan, are mixed with the other remaining parts of rice or wheat and water to make mash, which is concurrently subjected to enzymatic saccharification, lactic fermentation, and yeast fermentation.

    Throughout the world, fermented foods and beverages are prepared by con- verting sugar into lactic acid by the action of lactobacilli. Examples are lactic acid drinks, cheese, pickles, and some important fermented foods from mid- Asia, the Middle East, and Africa, such as idli, kishk, ogi, and mahewu. Lactic acid fermentation is important not only in the manufacture of these fermented foods and beverages, but in the manufacture of alcoholic beverages in the Orient


    as well because it prevents the undesirable acetic acid fermentation in the early stage of the manufacture of these foods.

    The techniques of enzymatically hydrolyzing certain protein foods into amino acids and small peptides to make them more nourishing and flavorful have long been known. For example, since ancient times people in the West have enriched the flavor of cheese by fermenting it with some Penicillium molds. The people in the Orient have enriched the flavor of fish and meat, proteinous beans, pulses, and some cereals by fermenting them with the proteolytic and amylolytic en- zymes produced by Aspergillus or Rhizopus molds, or with lactobacilli and yeasts in the presence of high salt concentrations. These foods, formerly called chiang in China and soy, sho, or hishio in Japan, can still be found in every Asian country, including Japan. They are believed to be the antecedents of shoyu and miso now in use, and their recorded use dates back 3000 years in China. The prototypes of these foods are believed to have been introduced from China to Japan 1300 or more years ago. The history of fermented soybean foods in China and Japan is summarized in Table I.

    In an article entitled Aroma and Flavor of Japanese Soy Sauce in this serial publication (Vol. 10, 1960), the methods of preparing Japanese shoyu and relat- ed research were discussed (Yokotsuka, 1960). The present article reviews the research conducted on shoyu and the technological advances in its preparation made since that time.


    Shoyu is the Japanese name for soy sauce, a popular liquid condiment used in oriental cuisine. Many varieties of shoyu are produced in Japan and other oriental countries. Their characteristics depend on the various kinds and different ratios of raw ingredients used, the kinds of microbes employed, and the conditions of preparation. Although most varieties are made from vegetable materials, fish soy is popular in Southeast Asian countries and is sometimes produced in Japan in small amounts. Fish soy is not included in the Japanese Agricultural Standard (JAS) definition of shoyu, however.

    According to JAS, in the production of genuine fermented shoyu, heat-treated raw materials, soybeans and wheat, are inoculated with koji mold (Aspergillus oryzae or A. sojae) to make koji, which is then mixed with salt water to make mash or moromi. Moromi is fermented with lactobacilli and yeasts and then well aged. The JAS recognizes five kinds of shoyu. Their names and production levels are shown in Table 11, and their typical chemical compositions are present- ed in Table 111.

    Of all shoyu consumed in Japan, 85% is of the koikuchi type, which means dark in color and made from approximately equal parts of soybeans and wheat.



    China Japan

    Shu-Ching (700 B . c . ) ~ Chuc Chiang (made from fish, bird, or meat)

    Chu (made from crushed wheat, or wheat flour made into balls or cakes, or cooked rice)

    Chiang (made from soybeans or wheat) Shi and shi-tche

    Chi-Min-Yao-Shu (532-549)b

    Tang dynasty (618-906) Ben-Chao-Gong-Mu ( 1590)b

    Chiang-yu Tao-yu

    Manyo-shu (350-759) Koji (same as chu) Hishio (same as chiang, made from fish, meat.

    or soybeans) Koma-hishio and miso Taiho-Law (70 1 )b

    Soybean-hishio, miso, kuki (same as shi) taremiso, usudare, misodamari

    Ekirinbon-Setsuyoshu ( 1598)b Shoyu (same Chinese characters as chiang-yu)

    Honcho-Shokukan ( 1962)b Shoyu, miso, tamari

    Industrial production of koikuchi-shoyu in Noda (1561) and Choshi (1616). that of usukuchi- shoyu in Tatsuno (1666). that of miso in Sendai (1645); export of shoyu from Naga- saki, Japan (1668); visit of C. Thunberg to Japan from Sweden (1775)

    0 Arranged from "35 Years History of Noda Shoyu Company (1955)"; The History of Kikkoman (1977); Sakaguchi Kinichiro (1981); Wang and Fuang (1981); and Bo (1982, 1984).

    Names of old references. Note: Chu: mold-cultured cereals; chiang; a mixture of chu, proteinous foodstuffs, and salt; shi:

    mold-cultured soybeans with or without salt; shi-tche: the saltwater extract of shi; chiang-yu: the liquid separated from chiang; tao-yu: the liquid separated from soybean chiang.

    The koikuchi mash is subjected to vigorous lactic and alcohol fermentations, and the finished product is pasteurized at a rather high temperature (about 80C) to give it a characteristic dark reddish brown color and strong heat flavor.

    Good-quality koikuchi shoyu contains 1.5- 1.8% (grams per volume) total nitrogen, 3-5% reducing sugar (mainly glucose), 2-2.5% ethanol, 1-1.5% polyalcohol (primarily glycerol), 1-2% organic acid (predominantly lactic acid of pH 4.7-4.8), and 17-18% sodium chloride. In order for a shoyu to have palatable taste, about one-half of its nitrogenous compounds must be free amino acids, and more than 10% of the nitrogenous compounds must be free glutamic acid.

    Usukuchi shoyu is made from a mixture containing more soybeans and less wheat than the koikuchi type. The saccharified rice koji with water, which is called amasake, or enzymatically saccharified starch or glucose, is sometimes



    Total production: Total sales:

    Koikuchi Usukuchi Talllari Saishikomi Shiro


    1,187, 148a I , 184,3060

    902,862 (84.4)b 138,261 (12.9) 20,885 (2.0) 3,130 (0.3) 5,042 (0.5)

    1,070,180 (100.1)

    0 Bureau of Foods, Japan. In kiloliters. Numbers in parentheses are per-

    centages. From Japan Shoyu Inspection Associa- tion. This amount was checked by Japan Agri- cultural Standards.


    (JANUARY. 1983)"

    Kind Kind of Number of shoyu special grade of sample Baume NaCl TNb FN RS Alc pH Ex Col

    Koikuchi Ordinaryc 6 21.8 17.1 1.56 0.90 3.10 Super= 3 22.4 17.1 1.69 0.96 3.81 Ultrasuperc 1 23.5 17.4 1.83 0.98 3.87 Less saltd 5 19.9 13.5 1.57 0.91 3.53 Reduced saltd 2 16.3 8.9 1.55 0.87 3.41

    Usukuchi Ordinarye 5 22.2 18.5 1.19 0.72 4.04 Supere 1 22.3 18.1 1.49 0.92 3.83 Less saltd 1 19.5 14.9 1.20 0.71 4.79

    Tamari 1 23.1 17.1 1.89 0.99 3.05 Saishikomi 2 27.5 14.1 2.24 1.06 9.43 Shiro 1 24.9 17.9 0.53 0.29 16.7

    2.23 4.85 2.17 4.84 1.84 4.79 3.20 4.88 3.41 4.86 2.57 4.89 2.65 3.97 3.73 4.95 - 3.03f 4.97 1.47 4.89 0.08 4.74

    19.2 21.1 22.2 21.4 22. I 16. I 18.8 18.0 22.1 34.9 20.5

    1 1 9 3 9 8

    28 26 28 7 2>


    a Arranged from J. Japan Soy Sauce Res. Inst. 9(2), 90 (1983). TN: Total nitrogen (%) (g/lOO ml); FN: formyl nitrogen; RS: reducing sugar; Alc: alcohol; Ex: extract

    without salt; Col: number of shoyu color standard issued by Japan Shoyu Inspection Association, the smaller the darker.

    Total nitrogen content: Ordinary, more than I .50%; super, more than 1.65%; ultrasuper, more than 1.80%. Salt content: Less salt, less than 20%; reduced salt, less than 50% of the standards,which are koikuchi,

    Extract without salt: Ordinary, more than 14.0%; super, more than 15.4%. The classification of special 17.5%; usukuchi, 19.9%; tamari, 17.9%; saishikomi, 15.6%; and shiro, 17.9%.

    grade is provided in the bylaws of the Japan Soy Sauce Association. f The too high content of alcohol for tamari (underlined) may be the one added after fermentation.


    added to usukuchi mash to ameliorate the salty taste. The nitrogen content of the finished product does not exceed 1.2%. Usukuchi shoyu is used mainly for cooking when one wishes to preserve the original color and flavor of the food- stuff. Koikuchi shoyu imparts a strong aroma and dark color during cooking.

    Tamari shoyu is made mostly from soybeans with only a small amount of wheat. Its nitrogen content is sometimes more than 2% and there is only a trace of alcohol.

    Shiro shoyu is very light in color and is made mostly from wheat with very little soybean. It is said that shiro shoyu was invented about 150 years ago in Aichi Prefecture, in central Japan.

    Saishikomi shoyu is made by enzymatically hydrolyzed soybeans and wheat in shoyu instead of the commonly used salt water. It was prepared for the first time about 200 years ago in Yamaguchi Prefecture, in the western part of Japan. Its original name was kanro shoyu, and it is characterized by its heavy taste due to its high content of extractable materials, nitrogenous compounds, and sugars. It is favored by some consumers as a dipping shoyu for some typical Japanese dishes such as raw fish and broiled eel.

    The JAS specifies three grades for each variety of shoyu: special, upper, and standard. The grade is determined by organoleptic evaluation, total nitrogen content, soluble acids (without sodium chloride), and alcohol content. Only high-quality shoyu made by fermentation can qualify for the special grade. About 65% of Japanese shoyu was qualified as special grade in 1980. The JAS for the special grade of koikuchi shoyu is more than 1.5% total nitrogen, more than 16% extract, and more than 0.8% alcohol.

    Blending fermented shoyu with 50% or less of chemical hydrolysate of plant protein or 30% or less of enzymatic hydrolysate of plant protein on a nitrogen basis is permitted for making products of upper and standard grades as long as the characteristic flavor of fermented shoyu is maintained.

    The yearly consumption of shoyu per capita in Japan is about 10 liters; 4.4 out of 10 liters is consumed in homes and the remaining 5.6 liters is consumed by institutions and industry. The shoyu producers in Japan are assumed to be less than 3200 in number; the five largest manufacturers produce 50% of the total and some 50 other companies contribute 25% of the total produced.

    The Japanese consume 34.1 g shoyu daily, which contains 14.0 g carbohy- drate, 2.4 g protein, 0.2 g fat, and 5.8 g salt. Since average daily intake of protein in Japan is about 80 g, the role of shoyu as a source of protein or amino acid is not significant (Bureau of Foods, Japan, 1976).

    The primary role of shoyu in the Japanese diet is as a source of salt, flavor, and color, especially for a bland and basic diet of rice, fish, bean curd (tofu), fermented beans, and boiled vegetables.

    Worldwide, shoyu has long been recognized as a flavorful compliment to meat



    Item Fermented soy sauce Protein chcmical hydrolysate

    Amino acid N/total N 45-50% 60-65% Amino acid contents

    Dominant organic acids

    Ethanol Higher alcohols

    Volatile S compounds CH3SH, (CH3)zS

    Buffer capacity Total polyol (glycerol) a-Diketone compounds

    Pyrazines Characteristic volatile

    Tryptophane (+), methionine,

    Lactic acid ( I - l S % ) , acetic acid

    1-3% Isobutyl, n-Butyl, isoamyl alco-

    hols Trace amount 0.08, 0.02 mgltotal N g Usually higher +++ Diacetyl, acetylpropionyl, acetyl-

    Less Ethyl lactate, HEMF, methionol,

    cystine; higher



    Tryptophane (-), glutamic acid,

    Levulinic acid (1.2- I .4%), for-

    ND' ND

    aspartic acid; higher

    mic acid (0. I-0.5%)

    +++ 0.22, 2.10 mg/total N g Usually lower + Diacetyl

    More Methional, y-valerolactone,

    flavor components 4-ethylguaiacol, 2-phenyl- methyl sulfide, etc. ethanol, etc.

    Color stability Lower Higher

    ND, No data.

    and high-fat dishes. It is thought by some to increase the appetite and to promote digestion; others claim that it has beneficial and medicinal effects. Its role in promoting the secretion of gastric juice has been compared with caffeine and histamine by some physicians.

    The original soy sauce product is considered to have been originated from Asian continents, but there are some aspects of production in Japan that differ- entiate the current Japanese shoyu from other oriental soy sauce products:

    1. Greater amounts of wheat are mixed with soybeans as raw materials. 2. Protein from the raw materials is highly degradated by the enzymes from

    3. Mash is subjected to vigorous lactic and alcoholic fermentations. 4. Pasteurization is done at higher temperatures to give strong aroma, flavor,

    These characteristics of Japanese shoyu production also explain the differences between chemical hydrolysate of plant protein, or so-called HVP, and Japanese genuine fermented shoyu in terms of the chemical components and the organo- leptic evaluations, as indicated in Table IV.

    A. soyae or A. oryzae.

    and color to the product.



    The fermented soy sauce industrially produced in Korea is of the Japanese koikuchi type, and in 1970, the annual industrial production was estimated to be about 220,000 kl. Assuming that the per capita daily consumption is about 20 ml, the amount cited above is equivalent to one-half the demand for soy sauce in Korea, and it is estimated that the same amount is produced at home (Li, 1970). The homemade soy sauce is prepared by a traditional method in which cooked soybeans are smashed and made into small balls, then subjected to natural inoculation of Aspergillus and Rhizopus molds, a process taking several months in winter. When spring comes, these mold-cultured materials are extracted with salt water. The liquid part is boiled and fermented under the sun to make soy sauce. The residue of extraction is mixed with salt and stored to make miso, sometimes along with red pepper.

    The annual production of soy sauce in Taiwan was estimated to be 130,000 kl in 1976, which is equivalent to 10 liters per capita comsumption per year. The largest four producers manufacture about 60% of the total produced. Of Taiwan soy sauce, 5-10% is estimated to be inyu, which is made only from soybeans and very much resembles Japanese tamari. The remarkable characteristics of inyu are that it is prepared from black soybeans instead of yellow soybeans, and that the black bean koji is washed with water before it is mixed with salt water to make mash. There are three national standard grades of soy sauce in Taiwan, and in 1980, their total nitrogen percentages were 1.4, 1.2, and 1 .O%, respectively.

    Fermented soy sauce similar to inyu and tamari is still being produced in the southern part of China, and it seems to be the prototype of the soy sauce prepared only from soybeans. In Japan, tamari mash is usually fermented in wooden kegs, but the soy sauce in Taiwan, Singapore, and the southern part of China is fermented in big china pots placed in the sun. However, most of the soy sauce made in Peking and Shanghai today is prepared differently: The koji is prepared by using a large-scale method to culture A . oryzae with a mixture of steamed soybeans and wheat or wheat bran (6:4), and the koji is mixed with salt water to make hard mash, the moisture content of which is about 80% and the salt concentration about 6-7%. This hard and low-salt mash is kept at 45-50C for about 3 weeks for enzymatic digestion. The digested mash is extracted with hot salt water and then with plain hot water. The residue without salt is good for animal feed. There is no alcoholic fermentation of mash or the pressing of mash as there is in the case of Japanese shoyu manufacture. The nitrogen basis of the soy sauce yield in 1979 was 75-80% because the defatted soybean as raw material was cooked by the NK method. The highest governmental standard of soy sauce is as follows: total nitrogen, 1.6%; reducing sugar, 4%; and sodium chloride, 19% or more.


    Soybeans (400 kg)

    Soaking in water Roasting of that of raw materials

    Autocliving Crushing I Wheat (340 kg) Seed culture, 0.1- 0.2% weight

    I I

    Mixing Moisture content 40-45% I I

    Culturing mold

    (Koji making)

    Moisture content of koji

    25-3036, pH 6.5-7.0

    48-72 hr

    Mixing I

    (Mash making) I

    Fermenting tank 2000 liters, 6-8 months

    -4 Salt (276 kg) Water (1 200 liters) 1. Enzymatic degradation of materials

    2. Lactic acid fermentation

    3. Yeast fermentation

    Aging I

    Soy cake ~ Raw shoyu pH 4.8-5.0

    (-220 kg, 30% moisture) I 1500-1600 liters Pasteurization TN 1.6-1 .a%


    I I Sediment Refined shoyu

    I I pH 4*7-4.9 * Or defatted soybean (330 kg), moistened with 420 liters hot water


    FIG. 1. (1976); Fukuzaki (1972) and others.

    Koikuchi shoyu fermentation. Prepared by Yokotsuka from the Bureau of Foods, Japan




    Japanese-fermented shoyu of the koikuchi type involves five main processes: the treatment of raw materials, the making of koji, the making and aging of mash, pressing, and refining. One example of the preparation of koikuchi shoyu is schematically indicated in Fig. 1 .

    1 . Treatment of Raw Materials

    Whole soybeans, or more commonly, defatted soybean grits, are moistened and cooked with steam under pressure. This process greatly influences the di- gestibility of soybean protein. Details will be provided in a later section. Wheat kernels, the other half of the raw materials, are roasted at 170-180C for a few minutes, then coarsely crushed into four or five pieces.

    2. The Making of Koji

    These two materials are inoculated with a small amount of seed mold or pure culture of A. oryzae or A. sojae. This mixture is spread to a depth of 30-40 cm on a large perforated stainless-steel plate having a rectangular shape that is 5 m in width and 12 m in length, for example, or a doughnut shape with a diameter of 15-30 m. The heat-treated raw materials are aerated for 2-3 days with controlled temperatures and moisture-controlled air, which comes up from the bottom holes through the ingredients to create the proper conditions for mold cultivation and enzyme formation. The temperature of the materials is kept at -3OC, and the moisture content of the materials, which is 40-43% at the beginning of cultiva- tion, decreases to 25-30% after 2 or 3 days. This allows the mold to grow throughout the mass and provides the enzyme necessary to hydrolyze the protein, starch, and other constituents of the raw materials. This mold-cultured material is called koji.

    3. The Making and Aging of Mash

    In making mash, the koji is mixed with saline water which has a %3% salt content and a volume 120-130% that of raw materials. The mash, or moromi, is transferred to deep fermentation tanks. Approximately 5- to 10-kl wooden kegs or 10- to 20-kl concrete tanks for shoyu fermentation are now being re- placed by resin-coated iron tanks of 50-300 kl. The moromi is held for 4-8 months, depending upon its temperature, with occasional agitation with com-


    pressed air to mix the dissolving contents uniformly and to promote the microbial growth. During the fermentation period, the enzymes from koji mold hydrolyze most of the protein to amino acids and low-molecular-weight peptides. Approx- imately 20% of the starch is consumed by the mold during koji cultivation, but almost all of the remaining starch is converted into simple sugars. More than half of this is fermented to lactic acid and alcohol by lactobacilli and yeasts, respec- tively. The initial pH value drops from 6.5-7.0 to 4.7-4.9. The lactic acid fermentation produced in the beginning stage is gradually replaced by yeast fermentations. Pure-cultured Pediococcus halophylus and Saccharomyces rouxii are sometimes added to the mash. The salt concentration of mash remains at 17- 18% (weight per volume) after 1 or 2 months. The high concentration of mash effectively limits the growth to only a few desirable types of microorganisms.

    4 . Pressing of Mash

    An aged mash is filtered through cloth under high hydraulic pressure. Usually 12-13 liters of shoyu mash is put on a square sheet of cloth, 100 X 100 cm, which is then folded into a square, 70 X 70 cm. A second, smaller square sheet of cloth, 65 X 65 cm, is placed on top to wrap the mash. Successive layers are added and placed in a wooden box until there are 300-400 sheets of folded cloth containing the mash. These are then pressed for 2-3 days under hydraulic pres- sure. The pressure is increased in two or three steps, sometimes reaching 100 kg/cm2 in the final stage, making the moisture content of the presscake less than 25%. A diaphragm-type of pressing machine has recently been used for shoyu mash filtration instead of the batch-type hydraulic press, resulting in a presscake with a moisture content of more than 30%. The residue from the pressing of the shoyu mash, or shoyu cake, is used for animal feeds for cows and ducks.

    5 . Refining

    The liquid part of the mash obtained by pressing is stored in a tank and divided into three layers: the sediment on the bottom, the clear supernatant of the middle layer, and the oil layer floated on top. The middle layer is sometimes further clarified by filtration with Kieselgel as a filter aid in order to get the raw shoyu. After adjusting the salt and nitrogen concentrations to the standard, the clarified raw shoyu is pasteurized at 70-80C and stored in a semiclosed tank. The clear middle layer is bottled or canned, or sometimes spray dried. The oil layer separated from the heated shoyu consists of free fatty acids, and their ethyl esters derived from the yeat metabolism of soybean and wheat oils, and it is sometimes mixed with paint as a antifreezing agent.



    The principles of usukuchi shoyu preparation are almost the same as those of koikuchi shoyu, except that all of the procedures are directed at getting lighter color and aroma in the final product (Fukuzaki, 1972) by the following:

    1. Making a mixture containing more soybeans and less wheat than koikuchi shoyu;

    2. Using a strain of mold belonging to A . oryzue, which is a better producer of a-amylase, whereas the strain for koikuchi shoyu is A . sojue or A . oryzue, which is a good producer of both protease and a-amylase. Aspergillus oryzue tends to impart milder aroma and flavor and lighter color to the final product as compared to A. sojue;

    3. Making a more diluted mash with a lower nitrogen content. The volume of water used to made koikuchi mash is 120-130% that of the raw ingredients, while that of usukuchi shoyu is 130-150%;

    4. Keeping the higher salt concentration of mash. With usukuchi shoyu, the salt is about 17-18% weight per volume, while in the case of koikuchi, it is about 16- 17%;

    5. Culturing the koji and fermenting the mash for a shorter period of time than is used for koikuchi shoyu;

    6. Avoiding excessive heat in treating raw materials during mash fermenta- tions and aging of mash, and in pastuerizing the final product in the preparation of usukuchi shoyu; and

    7. Adding a dextrin-like substance, such as enzymatically hydrolyzed rice koji, in order to make the color stable and to ameliorate the salty taste.


    The basic principles of tamari shoyu preparation are almost the same as those of koikuchi and usukuchi, except for the following items (Yoshii, 1960):

    1. Autoclaved soybeans or defatted soybean grits with a small amount of roasted and crushed wheat (20:3) are treated with an extruder to make pellets 12- 16 mm in diameter. These pellets are inoculated with the seed mold and the powder of roasted barley, the amount of which is less than 1.5% that of the raw materials.

    2. The strain used for tamari koji is originally Aspergillus tumarii, but A . sojae, A . oryzue, or the mixture of these strains are also used. The raw material used for tamari contains smaller amounts of carbohydrate material than those used for koikuchi. The strain of mold may not necessarily be the best producer of amylase, but a good producer of protease and lipase is required.


    3. Tamari koji is usually dried for several days so as to decrease the weight of koji by 7-8% before preparing mash. This dried koji is mixed with salt water, the volume of which is 50-80% that of the raw material. Tamari mash cannot be agitated with compressed air, as can koikuchi or usukuchi mash; only the liquid part of mash is repeatedly siphoned off and poured onto the surface of mash.

    4. Major fermentation that occurs in mash is lactic acid formation by P. halophylus, and there is almost no alcoholic fermentation by yeasts. The di- gestibility of nitrogenous materials in tamari mash is much less than in koikuchi or usukuchi.

    5 . The liquid part of mash cannot be separated by the pressing of mash, but is obtained by dripping, followed by one or two extractions with salt water to get lower grade products. Tamari shoyu usually is not pasteurized and is usually heated at a low temperature to avoid the burnt odor derived from its high concentration of extractable substances with heating.

    6. The residue from dripping the tamari mash used to be sold as tamari miso for making soup, but consumers have come to prefer traditional miso or fer- mented soybean paste, which, because of its preparation, retains the delicious liquid part of the mash. Consequently, there is a tendency to prepare tamari mash by mixing a 100-130% volume of salt water with raw materials to shorten the fermentation period, to increase lactic fermentation, and to make it possible to get the liquid part of mash by pressing instead of dripping.

    A flowchart for the preparation of tamari shoyu is shown in Fig. 2.


    Shiro shoyu is made mostly from wheat with very little soybean, the volume of which is 10-20% that of wheat. The specific characteristics of shiro shoyu preparation are as follows: Wheat is polished to remove about 5% of the outer layer of grain, which is rich in pentose; the soybeans are roasted and then crushed, followed by dehulling to decrease the pentose content; these compo- nents are then steamed under an atmospheric pressure. The light-colored product is derived from the decreased pentose levels of the raw materials.

    The recommended strain of mold for shiro shoyu koji is A. oryzae which has a long stalk and was originally used for making miso. The amount of salt water needed to make mash is 120-130% of the volume of the raw materials. The nitrogen content of shiro shoyu is about one-half that of koikuchi shoyu, and the reducing sugar content is very high, ranging from 15 to 20%. The salt content is usually 18-19%.

    Like usukuchi shoyu, shiro shoyu is used mainly for cooking, and a gradual increase in the production level of shiro shoyu has recently been observed in


    Defatted soybean Crushed wheat (2000 kg) (300 kg) +

    Steaming I

    Moistening (90C. 95%)

    I Autociaving

    I Pe@ting

    Seed mold ,-I (A. famari) Koji making

    (72 hr, 2,5-2EoC)

    Salt - ( 1 000

    Water (3600

    T-J liters)

    Dry'ing (2269 kg, 30-50

    Baumkl Mixing

    (Mash making) I


    Siphoning of liquid part

    Aging (8-24 months)

    Dripping of mash


    I I

    Salt water (1700 liters.Baum6 15)

    1-2 weeks

    (Second crop, (3400 kg) 1700 liters. TN 1.40%. nitrogen yield, 16.4%)


    Salt water (2500 liters, Baumd 13)

    - 1-2 weeks


    Ban tamari 1 Residue 11200 kg) (Third crop, 3800 liters, TN 0.65%. nitrogen yield, 16.9%)

    1 Residual mash

    (3560 kg) I Kibiki tamari (First crop, 1700 liters, TN 2.1% nitrogen yield for raw materials, 24.4%)

    FIG. 2. Tamari shoyu fermentation. From H. Yoshii (1960). Brewing Industry. p. 198.

    Japan. There are good introductions to shiro shoyu by H. Yoshii (1960) and K. Fukuzaki (1972).


    Saishikomi shoyu is made by enzymatically degrading koikuchi koji in shoyu instead of salt water. The volume of shoyu used for this purpose is 110- 120% that of the raw material of koji. Mash is stored for 3 months at 26-28"C, followed by 5-6 months at room temperature. There is almost no microbial fermentation during mash storage. Average chemical composition of six mar- keted products in 1960 were Baumt, 29.26%; sodium chloride, 15.3%; total nitrogen, 2.250%; reducing sugar, 10.76%, and pH, 4.6 (Bureau of Foods, Japan, 1976.)



    Recent technological improvements in shoyu industry are summarized as

    1. The use of more defatted soybean instead of whole beans. 2. Increase in protein digestibility of raw materials from 65 to 90% as the

    result of improved methods of cooking soybeans and wheat, the selection and mutation of starter molds, improved conditions for culturing molds or making koji, and the control of mash in terms of the temperature, pH, types, and behavior of lactobacilli and yeasts, and the chemical compo- nents.


    3. Reduction of the time for koji cultivation from 72 hr to 48 hr. 4. Decrease of fermentation period of mash from 1 to 3 years to about 6

    5 . The use of pure cultured starters of lactobacilli and yeasts. 6. Mechanization of the equipment and expanding of the production scale. 7. Improvement of quality and reduction of cost.



    Until 50 years ago, only whole soybeans were used as the raw material for shoyu. Today, defatted soybean grits are prepared by extracting the dehulled and crushed whole soybean with a solvent. Hexane at a lower boiling point is widely used for this purpose. In Japan during 1978, of the total number of soybeans used for the production of shoyu, only 3.2% were whole beans. Years ago, the yield and the quality of shoyu made from defatted soybean used to be inferior to those produced from whole beans, but today the disadvantages of using defatted soy- beans as the raw material of fermented shoyu, compared with whole beans, have been largely overcome by advances in technology. Yokotsuka (1972) compared whole and defatted soybeans with respect to cost, enzymatic digestibility of the protein, fermentation period, the relative difficulty in manufacture (especially in koji making, mash controlling, and mash pressing), and the quality of the shoyu produced in terms of its chemical components, organoleptic properties, and stability:

    The defatted soybeans used years ago were much more difficult for enzymatic digestion than those of today, and the quality of the defatted soybean used in the fermentative production of shoyu has been much improved. This is due to the improvement in the pressing method which uses an expeller with a battery system or a continuous extractor to remove the solvent. Furthermore, the extract-


    ing temperature has been reduced by using hexane instead of benzene as the solvent. These changes have only slightly reduced the quality of the protein, altered during processing. A nitrogen solubility index (NSI) value of about 20 for defatted soybean is generally believed to be adequate for shoyu production.

    The enzymatic digestibility of the proteins contained in whole and defatted beans in the course of shoyu production is reported to be 62 and 60%, respec- tively (Kawamori, 1940), but these figures have been almost the same, and sometimes higher, for defatted beans by 1-2% since the invention of the NK method of cooking, which will be discussed later.

    Based on protein content, the cost of whole soybean is 10% higher than defatted soybeans.

    Whole soybeans are reported to have a slower rate of fermentation, about 15 months for whole beans at room temperature versus only 10 months for defatted beans (Yokotsuka, 1960). The fermentation period of shoyu mash is principally dependent upon the enzymatic digestibility of cooked soybeans and on the en- zymatic activities of koji. These two factors are now considered to be associated with the manufacturing technology and not with the differences between whole and defatted soybeans. The average fermentation period of shoyu nowadays is about 6 months.

    It has been traditionally easier to make good koji with whole beans than with defatted ones, since the conventional method involves cooling the materials by hand mixing, and the larger particles of whole beans are cooled more easily than are the smaller particles of the defatted bean. Recently, it has become easy to prepare a good koji from defatted beans by using mechanical koji equipment in which the temperature is controlled by mechanical aeration.

    Shoyu made from whole beans has been reported to have a lighter color and better color stability, a higher alcohol and glycerol content, a smaller amount of lactic acid and reducing sugar, and a better organoleptic evaluation than the shoyu made from defatted beans (Okuhara and Yokotsuka, 1958; Moriguchi and Ishikawa, 1960a,b; Moriguchi and Kawaguchi, 1961; Moriguchi and Ohara, 1961). The glycerol contents of shoyu made from whole and defatted beans were reported to be 1-1.2% and 0.6-0.7%, respectively (Okuhara and Yokotsuka 1958, 1962, 1963). The amount of glycerol in a shoyu mainly derived from the degradation of soybean oil is calculated to be about 0.5%. However, glycerol has also been produced in mash by the yeast fermentation of glucose in the presence of high salt concentrations.

    Shoyu mash is now subjected to much more vigorous yeast fermentation than before, resulting in a higher concentration of glycerol, sometimes reaching I .5- 1.7%. Thus, the advantage of using whole soybeans to give shoyu a higher glycerol content has diminished (Sakurai and Okuhara, 1977). The lactic acid content of shoyu is now easily adjusted by controlling the degree of lactic acid


    fermentation in mash regardless of the kind of soybean used. Almost the same can be said for the alcohol content of shoyu, although the aerobic condition of the mash made from the whole beans was in the past thought to make alcoholic fermentation easier.

    The differences between the chemical components of those shoyus made from whole beans and those derived from defatted beans can be minimized by making the physical structure of whole beans similar to that of defatted beans by pressing (Okuhara and Yokotsuka, 1963).

    Both the color intensity and the color stability of a shoyu seem to be funda- mentally related to the degree of digestion of raw materials and not to the kind of soybean used.

    Because whole soybeans have a higher protein content and are therefore more costly than defatted beans, a shoyu made from defatted beans of the same price has a higher content of free amino acids, including glutamic acid, which gives it a more delicious taste. Nevertheless, it is true that a shoyu made from whole beans has some characteristic flavor. This may account for the fact that some shoyu producers are still making their products from the mixture of whole and defatted soybeans.

    From environmental viewpoints, whole beans have several waste problems, e.g., soaking water and sticky liquid from cooking. Both contain carbohydrates and proteins from whole beans, which should be removed before disposing of the wastes.

    A large portion of the oil contained in soybeans and wheat is metabolized into the ethyl esters of higher fatty acids and glycerol in the course of yeast fermenta- tion. The ethyl esters make the pressing of mash difficult, as these must then be separated from the upper layer of the liquid obtained by pressing. The separated oil is called shoyu oil, and its major chemical constituents are ethyl linolate and ethyl oleate accompanied by free higher fatty acids and sitosterols. This by- product is troublesome to shoyu producers.


    1 . Soybeans

    The protein in raw soybeans is present in an undenatured state and is not hydrolyzed by the enzymes of koji mold. Therefore it is necessary to denature the soybean protein so that it can be digested by the enzymes of koji mold to make shoyu. Steam cooking has generally been used to denature the soybean protein. Years ago, soybeans were steamed or boiled at atmospheric pressure, but Ka- wan0 (1938) found that when the soybeans were cooked at the gauge pressure of 0.5 kg/cm as compared with 0, 1 .O, 1.5, and 2.0 kg/cm, the highest enzymatic



    Digestibility (%) of proteins

    in mash, salt 18%, Ratio Ratio Cooking mom temperature, between formyl N between glutamic N method I year and total N (8) and total N (76)

    Conventionalb 68.7 NK methodc 73. I Increasing ratio 106.4

    49.4 53.8


    5.5 1.3


    From Tateno and Uneda (1955). Kikkoman Shoyu Co., Ltd. Cooked at 0.8 kg/cm2 for I hr, soybeans left in autoclave for additional 12 hr. Cooked at 0.8 kg/cm2 for 1 hr, soybeans taken out of autoclave immediately.

    digestibility of cooked soybeans and the highest free amino acid content of the shoyu prepared from cooked soybeans were obtained.

    Until 25 years ago, soybeans were cooked at a gauge pressure of 0.8 kg/cm for several hours. Since then, the time has been shortened to less than I hr under the same pressure. Thoroughly moistened soybeans were cooked in a rotary cooker and the materials immediately cooled to below 40C by reducing the inside pressure with the aid of a jet condenser. This method was called the NK method (Tateno et af., 1955) and is given in Table V.

    The protein digestibility in shoyu manufacture (i.e. , the ratio between the total nitrogen of a shoyu and that of the raw materials) was increased from 69 to 73% by the NK method. In the conventional cooking method, the soybeans are cooked at the same pressure and for the same time as in the NK method, but the cooked soybeans remain in the autoclave after steaming for an additional 12 hr without opening the seal. It is important that there be enough water in cooked soybeans (about 58% of the volume for whole beans and about 62% for defatted beans) because the utilization of total and amino nitrogen increases with an increase in the moisture content of the beans. It is also important that the steaming soybeans be uniform and that no undenatured protein is left in the cooked soybeans.

    The treatment of soybeans with water containing methanol, ethanol, or pro- panol was found to markedly increase the enzymatic digestibility of protein (Yamaguchi, 1954; Fukushima et af., 1955, 1957). These treatments are given in Tables VI and VII, respectively. These methods have not been employed indus- trially mainly because of the difficulty of making koji and the possibility of bacterial contamination during koji cultivation, which results in a final shoyu product of inferior organoleptic quality.




    Protein Total N Amino N NaCl Reducing sugar digestibility

    Denaturing method (96) (%) (8) (a) (8)

    Methanolb I .92 0.98 17.50 2.92 90.69 Conventional cookingc I .39 0.63 17.63 4.53 65.16

    a Yamaguchi (1954). Japanese patent 219,545. Kikkoman Shoyu Co., Ltd. Boiled with methanol for 2 hr, methanol removed, and then steamed for 1 hr without pressure. Steamed with 1 kg/cm2 pressure for 1 hr and kept in the autoclave without reducing pressure for

    several hours.

    Yokotsuka er al. (1966) found it useful to increase the enzymatic digestibility by cooking at a higher temperature for a shorter time than the NK method, as given in Table VIII. This method indicated the possibility of having 92-93% protein digestion in shoyu production, with a final product of better organoleptic quality (Yasuda et al., 1973a,b). Similar research results were reported by Ha- rada et ul. (1968) in which defatted soybean was cooked at the elevated pressure of 4 kg/cm for 3 min.

    In the above cases, thoroughly moistened soybeans were cooked by using saturated steam. Aonuma et al. (1970) reported a new method of cooking soy- beans and wheat used for brewing without adding or with adding 10-20% of moisture before cooking by using superheated steam at a gauge pressure of 4-8




    Protein Total N Amino N Amino N/ digestibility

    Treating method of defatted soybean (a) ( w h ) (%) (wlv) total N (a)

    Boiling with 85% (w/v) ethanol, 40 min 1.80 I .80 59.6 89.3 Boiling with 70% (w/v) isopropanol, 60 min 1.77 I .oo 56.5 84.3 Control 1: NK cooking I .62 0.90 55.3 80.6 Control 2: NK cooking 1.69 0.93 55.6 80.8

    a Fukushima er al. (1955, 1957). Japanese patent 236,368,237,805 (1955) 248,103 (1957). Kik- koman Shoyu Co., Ltd.




    Digestibility of protein

    (kglcm') (min) (salt 0%, 37"C, 7 days) Steam pressure Cooking time in enzyme solution

    0.9 I .2 1.8 2.0 3.0 4.0 5.0 6.0 7.0

    45 10 8 5 3 2 1 '/2 '/4

    86% 91 91 92 93 94 95 95 95

    a Yokotsuka er al. (1966). Japanese patent 929,910. Kik- koman Shoyu Co., Ltd.

    kg/cm or at 200-289C for not less than 15 sec. They confirmed almost the same protein digestibility as that obtained by saturated steam under the above-men- tioned conditions. This method has the advantages of making it possible to stock the heat-treated raw materials.

    This new high temperature-short time (HTST) method of cooking raw mate- rials for shoyu brewing spurred the development of several types of continuous cookers, shown in Fig. 3.

    At the same time, the NK method was also greatly improved. Protein di- gestibility of 87.80% was achieved by cooking soybeans at 1.7 kg/cm for 8 min by using an NK cooker compared with 81.80% obtained by the conventional NK cooking conditions at 0.9 kg/cm for 40 min (Iijima et al., 1973). The time for cooling of autoclaved soybeans in an NK cooker is greatly associated with their proteolytic digestibility, which is given in Table IX (Yasuda et al., 1973a).

    By enlarging the diameter of both the entry and exhaust steam pipes of an NK cooker to make the cooking time precise and to make the cooling time of cooked soybeans as fast as possible, the protein digestibility increased by about 3% under the same conditions (Eguchi, 1977).

    If portions of soybean protein remain undenatured, they mix with the final shoyu product which becomes turbid when it is diluted with water and heated, thus diminishing its commercial value. The relationship between steaming pres- sure and time for soybean cooking to the denaturation of soybean protein is indicated in Fig. 4 (Yokotsuka et al . , 1966a). The influence of the cooling speed


    FIG. 3. (charge), (2) steam, (3) cooker, (4) rotary valve (discharge).

    Continuous soybean cooker. (A) Screw type; (B) net conveyer type. (1 ) Rotary valve

    is not great at low cooling temperatures, but slow cooling following higher cooking temperatures and shorter cooking time gives rise to overdenaturation.

    The recent trends of raw material treatment for fermented foods are toward the application of very high gauge pressure of 20-90 kg/cm2 for less than 5 sec and that of intermediate moisture contents of materials, which are between the wet and dry methods. These are summarized in Table X.

    Yokotsuka et al. (1965) and Hayashi et al. (1968a,b) demonstrated a 2-3% increase of enzymatic digestibility of soybean protein by treating it with a small




    Experiment Pressure Time Cooling time Digestibility no. (kg/cm2) (min) (min)b (%)

    1 2.0 5 I 91.65 2 2.0 5 5 91.32 3 2.0 5 20 85.38 4c I .o 45 1 87.25

    From Yasuda er al. (1973a). Time required to attain atmospheric pressure after cooling. Control.



    ' t 0 ' ' ' " ' ' ' ' I '

    0 1 2 3 4 5 6 7 8 9 10

    Steaming time (min)

    FIG. 4. Denaturation of soybean protein by steaming at 130% moistening. (A) Overdenaturation region; (B) underdenaturation region; (C) proper denaturation region for shoyu production. From Yokotsuka et al. (1966).

    amount of sulfite (0.04-0.4%), peroxide (0.1-0.2%), or perchlorite (0.2-0.4%) before or after cooking the soybeans. These procedures were presumed to de- stroy the S-S linkages of hydrophobic bonds in the remaining protein molecules of cooked soybeans or those newly produced during cooking, which do not interact with the proteases, thereby hindering digestion.

    2 . Wheat

    Wheat is the major source of carbohydrate among the raw materials used to make shoyu, but its protein content cannot be overlooked because it constitutes about one-fourth of the protein in shoyu. If wheat is insufficiently roasted, its raw starch or P-starch cannot be digested by the mold amylase and becomes white particles in the presscake of mash. However, if wheat is overroasted, the protein digestibility decreases. The P-starch in wheat kernels must be changed


    into a-starch by adequate roasting in order to be digested by mold amylase. The content of a-starch in roasted wheat is determined by calculating the ratio of the amount of glucose digested from roasted wheat to the glucose digested from wheat throughly boiled with amylase produced by Aspergillus mold, which serves as the control.

    Canadian wheat was mainly utilized in the Japanese shoyu industry previously because of its relatively high protein content. Haga et al. (1970) found almost no differences between Japanese wheat and wheat imported from Canada, the United States, and Australia with regard to the content of a-starch after roasting, the loss of carbohydrate during koji cultivation, the difficulty of pressing of mash, and the sensory evaluation of the final product.

    Wheat is continuously roasted in a rotary oven with sand which recycles in the oven and is kept separate from the roasted wheat. The content of a-starch in the roasted wheat is used as the criterion for the extent of roasting. The preferred amounts of a-starch in roasted wheat used for the preparation of koikuchi and usukuchi shoyu were reported to be 40% and 20-30%, respectively, by Mor- iguchi and Nishiyama (1960).

    The factors to increase the enzymatic digestibility of starch and protein, re- spectively, of roasted wheat by recycling heated sand are contradictory to each other, which is indicated in Fig. 5 (Aiba, 1982). It is effective to increase the a- starch content of roasted wheat kernels by making the moisture content of wheat before roasting 15-25% (Yamaguchi et al., 1961). The same HTST method used with soybeans, adding some 10% moisture, gives good results as well for roast- ing wheat which has the same amount of water added (Uchiyama and Mat- sumura, 1974). According to Aiba (1982), good results were obtained when the wheat containing more than 8% moisture was treated with hot air of more than 150C for less than 45 sec at atmospheric pressure. The highest digestibilities of starch and protein of the roasted wheat were 86 and 97%, respectively. In actual shoyu production, the utilization of starch in raw materials was improved by 2%.

    When roasted wheat is crushed into four or five pieces accompanied by small- er particles of wheat flour, the preferred amount of wheat flour which passes through the 32 meshes is about 30%, which covers the cooked soybeans to reduce its surface moisture content and to minimize the bacterial contamination during koji cultivation (Umeda, 1967).

    According to Aouma et al. (1971), the wheat treated in a current of super- heated steam having a gauge pressure of 8.0 kg/cm and a temperature of 260C for 8 sec followed by rapid cooling by explosion puffing produced a higher content of a-starch, a superior ability to absorb and retain moisture during koji cultivation, and a better yield of shoyu compared with the roasted wheat obtained by the conventional method. They also reported on a process of heating the


    I I I I

    150 160 170 Temperature ("C)

    FIG. 5. From Aiba (1982).

    mixture of crushed soybeans and wheat by explosion puffing for the use of shoyu manufacture.

    Fujita and Kishi (1975), reported that during the milling process, a mixture of wheat flour and wheat bran was produced. This mixture was moistened and subjected to extrusion steam cooking to obtain a heat-denatured wheat product for shoyu manufacture. The a-starch content in this product ranged from 30 to 42%, while that of roasted wheat from 12 traditional shoyu producers ranged from 17 to 58%. The heat-denatured wheat for koikuchi and usukuchi shoyu production was prepared by adjusting the ratio of wheat flour to wheat bran. The greater the wheat bran, the higher the nitrogen content of shoyu and the darker its color, making it suitable for koikuchi shoyu. This heat-treated wheat was moistened and then cultured with A. oryzae to make koji. The protein di- gestibility of this koji in 18% salt water was 93.6%, while that of the wheat roasted by the conventional method was 79.7%. This difference in protein di- gestibility as a result of this preparation is equivalent to a 3-4% increase in the protein digestibility of final shoyu produced from the mixture of wheat and soybeans. This heat-treated wheat also has the advantage of being better able to absorb and retain moisture, which is good for koji cultivation.

    Conventional roasting wheat with recycling heated sand. (A) Total nitrogen; (B) starch.



    Mold strains used for food fermentation are selected on the basis of the following characteristics:

    1. Providing good flavor to the final products 2. Readiness with which they become mold starters with a sufficient amount

    3. Ease and speed of growth, making them easy to handle in koji making 4. Providing enzymatic activity, especially high proteolytic activity and

    macerating power to decompose the tissues of soybeans and wheat 5. Consuming a small quantity of carbohydrate in the raw materials during

    koji cultivation, yielding more sugar and alcohol in the mash 6. Having a shorter stalk (conidiophore), which makes possible the mechan-

    ical cultivation of materials with greater thickness. It is known that a long- stalk koji mold tightens the materials during koji cultivation and makes the aeration of the materials difficult.

    of spores

    7. Having genetic stability with little back mutation 8. Providing final products with desirable color (light or dark) as required. 9. Producing no toxic substances such as aflatoxins, cyclopiazonic acid,

    aspergillic acid, koji acid, P-nitropionic acid, oxalic acid, and other kinds of so-called mycotoxins

    10. Yielding a mash that is easy to press.

    According to Murakami (1973), among 327 strains of Aspergillus mold used in food fermentation in Japan, 159 were used for sake brewing. These included 157 strains of A. oryzae; 43 strains were used for miso brewing, which included I strain of A. sojae, 2 strains of A. tamarii, and 38 strains of A. oryzae; and 125 strains were used for shoyu brewing, which included 29 strains of A. sojae and 92 strains of A. oryzae.

    Shoyu koji cultured with A. oryzae has a lower pH value, lower carbohydrate content, higher activity of cu-amylase, higher activity of acid proteases, high activity of acid carboxypeptidase, and lower activity of polygalacturonase than does koji cultured with A. sojae (Tereda ef al., 1980, 1981). Furthermore, Hyashi et al. (1981) reported that koji cultured with A. sojae had a higher pH value because it contains less citric acid, and more carbohydrate because of lower consumption during koji cultivation. Shoyu koji cultured with A. sojae resulted in lower viscosity of mash, a higher content of sugar, lactose, and ammonia, a lower pH value of raw shoyu, and less coagulant produced by pasteurization of raw shoyu because of fewer active enzymes derived from koji.

    Rhizopus molds are widely used in food fermentations in China, Taiwan, and


    Indonesia. In Japan, the koji is sometimes contaminated with Rhizopus molds such as R. nigricans when the temperature is too low.

    Ebine et a f . (1968) compared the proteolytic activities produced by 36 strains of Rhizopus with that of Aspergillus molds. Rhizopus tamari and R. thermosus were found to grow well on wheat and soybeans and to provide a very high proteolytic activity at pH 3.0, but almost none at pH 6.0. In this regard, the Rhizopus molds were distinctly different from the Aspergiffus molds in that they gave a somewhat lower protein digestibility than did the Aspergillus molds on a small scale of experimental brewing. Improvements in the proteolytic activity of koji molds have been achieved by induced mutation, crossing, or cell fusion (Nasuno et al., 1971, 1972; Nasuno and Nakadai, 1971; Nasuno and Ohara, 1971, 1972a,b). In one case, a 2-6% increase in protein digestibility in shoyu production was produced by the use of an induced mutant of A. sojae in which protease increased six times above that of the mother strain.

    Furuya er al. (1983) achieved a strain of koji mold having both strong pro- teolytic activity and good spore formation through protoplast fusion, and Ushi- jima and Nakadai (1983, 1984) achieved a strain having strong proteolytic ac- tivity and good glutaminase formation through the same method as above, since these two factors in each case tended to be contradictory to each other.

    It is generally recognized that the total proteolytic activity of koji is well correlated with its alkaline protease activity, i.e., the major protease produced by koji molds is an alkaline protease. But besides this, three kinds of acid protease, two kinds of neutral protease, and one semialkaline protease have been isolated (Nakadai, 1977; Nasuno and Nakadai, 1977), as presented in Table XI.


    BY Aspergillus sojae THROUGH SEPHADEX G-100"

    Protease M W Units/g koji

    Acid I Acid I1 Acid 111 Neutral I Neutral I1 S e m i a I k a I i Alkali

    39,000 100,000 31,000 41,000 19,300 32,000 23,000

    41.1h 10.0h 4.6h 80.0C 8.7c

    55.4= 929.OC

    0 From Nakadai et ol. (1977). J . Japon Soy Sauce Res. Inst. 3(3), 99.

    Activity on casein at pH 3.0. c Activity on casein at pH 7.0.


    The strong soybean protein digesting ability of the neutral protease I and 11, and especially of the former, has been reported (Sekine, 1972a,b; 1976), but other investigators have observed that the proportion of neutral protease to the total protease produced by 109 kinds of koji mold ranges from 10 to 20% (Tagami et al . , 1977). Three kinds of aminopeptidase and four kinds of carbox- ypeptidase have also been isolated; it is these peptidases, especially leucine aminopeptidase, which are greatly associated with the enzymatic formation of formol nitrogen and glutamic acid in shoyu mash (Iguchi and Nasuno, 1978).


    The following guidelines should be followed in koji cultivation:

    1. Grow as much mold mycelia and mold enzymes as possible. 2. Prevent the inactivation of the enzymes produced. 3. Minimize the carbohydrate consumption in raw materials during cultiva-

    4. Avoid as much bacterial contamination in the starting materials and during

    5 . Shorten the cultivation time with minimum use of water, electricity, and

    The mixture of cooked soybeans and roasted crushed wheat kernels is mixed with 0.1-0.2% of starter mold, A. oryzue or A. sojue. The mixed materials are usually cultured for 72 hr in small boxes or trays in a warm room, in which the temperature is controlled by windows. About 1 ton of raw material is divided into about lo00 wooden trays with a thickness of 3-5 cm.

    The materials are cooled twice by hand mixing when their temperature rises to about 35C or more because of the growth of molds. One example of the temperature change during koji cultivation in wooden trays is presented in Fig. 6.

    The temperature for culturing mold on raw materials is lowered from the traditional level with protease formation in the koji, although 35C or more is


    the cultivation of mold as possible.

    fuel oil.

    ~. - 9 18 27 36 45 54 63 72


    FIG. 6. Temperature change of materials during 4-day koji cultivation by the conventional method using wooden trays. From Shibuya (1969).


    I I I 1 I

    14pm 1 8 24 Barn 16prn 24 Barn 16 24 S a m

    FIG. 7. (A), second cooling. From Haga (1968).

    Four-day koji cultivation, keeping the temperature of materials at 25C. (0). First cooling;

    FIG. 8. (1968).

    Preferable temperature change of materials during 3-day koji cultivation. From Haga

    considered to be adequate (Yamamoto, 1957). Temperatures as high as 30-35C have been found to be preferable for the growth of mycellium and the prevention of bacillus as a contaminant in the beginning stages of koji cultivation. A lower temperature, 20- 25"C, is necessary both before and during spore formation in the latter stage or after the second cooling when protease develops in the koji (Ohara et al., 1959). It has been suggested that koji be prepared at a constant temperature of 23-25C for 66 hr and cooled twice (Fig. 7) to produce more protease and to avoid the inactivation of peptidase, which occur above 25C (Miyazaki et al., 1964; Tazaki et al., 1966; Imai et al., 1967). The preferred temperature change during koji cultivation is shown in Fig. 8, but it is difficult to maintain this temperature change in the conventional hand-operated method of koji making. However, development of mechanical equipment for koji cultiva- tion has made it possible to provide the desired temperature and humidity of the materials to be cultured with koji mold, to reduce the time required for koji cultivation from 72 to 48 hr, to increase the enzymatic activities of koji, and to reduce the undesirable bacterial contamination in koji. The typical temperature change of materials during mechanical koji cultivation with a throughflow sys- tem of aeration is shown in Fig. 9.


    FIG. 9. flow system of aeration. (0). First cooling; (A), second cooling. From Shibuya (1969).

    Temperature change of materials during 3-day mechanical koji cultivation with through-

    The appropriate mixing ratio of soybeans and wheat to be cultured with mold generally ranges between (4:6) and (6:4). According to Shibata et al. (1967), a lower C/N ratio results in a smaller amount of mycelia and a greater amount of alkaline protease in koji, while a higher C/N ratio gives a greater amount of mycelia and a predominance of acid protease in koji. There appears to be no correlation between the C/N ratio of materials to be cultured with the mold and the formation of neutral protease. At a higher cultivation temperature, the forma- tion of mycelia and acid protease increases and that of alkaline protease de- creases.

    The average consumption of starch in raw materials during koji cultivation is about 20%, depending upon the moisture content of the materials. According to Abe et al. (1975) and Katagiri et al. (1976), a remarkable decrease in the moisture content of materials takes place, sometimes as much as 50%, from hour 17 of cultivation (first cooling) to hour 30. During that time, a remarkable increase in the activities of protease and amylase, and of NSI (water-soluble protein/total nitrogen), sometimes reaching 50%, are observed, and the water- soluble protein becomes 31% as compared with that of the final koji.

    Although the level of bacterial contamination in koji has reached as high as 107-9/g in Japan, this level does not constitute a health hazard. The major bacterial contamination of koji cultured in wooden trays years ago was from the genus Bacillus because of the difficulty in cooling the materials by hand, but in the modem throughflow system of mechanical cultivation, the dominant bacte- rial contaminant is from the genus Micrococcus, which is more aerobic and grows at a lower temperature than Bacillus. In addition, Leuconostoc and Lac- tobacillus are sometimes found in Koji. Too much Bacillus contamination in koji not only reduces the proteolytic activities of koji, but also makes the flavor of the shoyu inferior. The presence of too much Micrococcus also lowers its pH value, which leads to inferior protein digestion in mash; and if the dead cells of Micro- coccus remain in shoyu, its filtration is sometimes more difficult. It is possible to reduce the bacterial contamination in koji to 106 or less by starting with a bacterial count of lo2 or less in the materials in the beginning stage of koji


    cultivation. To avoid the bacterial contamination during koji cultivation, the total moisture content of the starting materials should be 40-50% at most. For the same reason, it is advisable to reduce the moisture content of the surface of the cooked soybeans by initially wrapping them with roasted, then with finely crushed wheat kernels. Some investigators have found that a temperature below 34C is best for avoiding bacterial contamination (Ishigami et af . , 1965, 1967; Fujita et a f . , 1977).

    The inhibition effects of acetic acid, lactic acid, citric acid, and hydrochloric acid on the growth of various microorganisms were reported by Hayashi et af . (1979). By keeping the acetic acid concentration beyond the range of 0.4-0.8%, based on the water content of the koji substrate, the growth of a strain of Micrococcus sp., which had been isolated from shoyu koji, was effectively suppressed, but the growth of various strains of koji molds was not inhibited. Acetic acid had a remarkable inhibition effect on the growth of some strains of bacteria belonging to Staphylococcus species, Gram-negative aerobes, and En- terobacteriaceae which were artificially added to the koji substrate, but the growth of lactic acid bacteria was not retarded or was retarded only slightly. In addition, the use of sulfite, glycine, ammonium acetate, or a combination of these substances to retard bacterial contamination has been reported by some researchers.

    The following is a list of the mechanical equipment used in koji cultivation:

    1. Throughflow system of aeration (Fig. 10) a. Batch-type with a rectangular, perforated plate b. Batch-type with a doughnut-shaped moving perforated plate c. Continuous-type with a doughnut-shaped moving perforated plate

    2. Rotarydrum 3. Surface-flow system of aeration: The temperature- and moisture-controlled

    4. Liquid cultivation air flows over the materials which are placed in numerous trays

    Systems la and lb can handle about 5-10 tons of raw materials in one batch. When System lb is used, the plate is moved only when the raw materials are ready to begin cultivation, the materials are mixed, and the finished koji is taken out.

    A new system for continuous koji cultivation in the shoyu industry (System lc above) has been developed by Aka0 et a f . (1972). The principal apparatus consists of a perforated and doughnut-shaped circular plate with an outer diame- ter of 38 m. The plate rotates slowly once every 48 hr. The temperature of the solid mixture of cooked soybeans and roasted wheat, which spreads thickly over the plate, can be controlled by circulating humid air through the culture bed and the housing area in which the apparatus is installed. With continuous production,


    A B C

    FIG. 10. Koji culturing machines with throughflow system of aeration. (A) Rectangular type; (B) circular type (batch); (C) circular type (continuous). ( 1 ) Perforated plate; (2) feeding conveyer; (3) discharging conveyer.

    any desired decrease in the pressure through the bed can be established in the direction of plate rotation, and the product can be discharged without difficulty. A washer and a dryer are located near the place where the koji product dis- charges. An annular space bounded by the concentric and cylindrical walls of the rotary disk is subdivided by radial plates into a total of 96 compartments, similar to an echelon in shape. The lower edges of the two walls and the radial plates slide on a stationary floor. The perforated plate covers the upper surface of the disk, serving as the bottom of the solid culture spread. The stationary floor, supporting the radial plates and the walls of the disk, has 85 vents which permit the inflow of air; the remaining compartments are used to discharge the koji product and to clean the disk.

    The rate of airflow through the culture bed is checked easily by measuring the change in pressure. The culture becomes solidified considerably by water evap- oration due to the heat released from the mold growth and by an inextricable network of the mycelia. As solidification continues, the airflow resistance through the culture medium increases and causes a rise in the temperature of the solid culture, suggesting the possibility that the mold will become nonviable. By crushing the solidified culture, airflow resistance can be reduced without inflict- ing serious damage to the mycelia. Four cutters and two mixers are used for this purpose.

    Table XI1 presents one example of an air supply arrangement used during koji



    G W P Number Temperature Humidity Superficial Ressure drop no. ofcomponents ("c) (W air velocity (ma) (cm H20)

    I 20 34 90 I 1 10 2 10 34 90 13 30 3 10 27 97 17 25 4 20 25 97 15 20 5 15 20 97 I 1 15

    @ From Akao et 01. (1972).

    cultivation. In miso manufacturing, the rotary drum is widely used only for rice koji cultivation. The surface-flow system of aeration used in shoyu koji cultiva- tion (System 3 above) is not popular today. The use of liquid cultivation of koji mold for shoyu production has not been successful yet because of the high cost of facilities and the resulting lack of flavor of the final product.

    Aka0 and Okamura (1983) reported on the cultivation of A. sojae in an air- solid fluidized bed. By use of a bench scale and a pilot plant, ground wheat bran having a 40% moisture content was fluidized by sterile air at 33C for 50 hr. Cell yield in this method increased two- or threefold, and activities of alkaline pro- tease and peptidase increased 5- to Sfold, as high as those obtained in conven- tional solid culture.


    I . Temperature of Mash

    Based on many years of experience, the Japanese have long known that the best quality shoyu results when koji and mash are prepared in February or March at a room temperature of 5-15"C, and the mash is fermented and aged from spring until autumn. Alcohol fermentation takes place in summer when the room temperature rises to around 30C. It has also been known that shoyu prepared during the summer has less total nitrogen, amino nitrogen, and glutamic acid, a high level of organic acids, and an inferior organoleptic evaluation than shoyu prepared during the winter. In Japan, the ideal change in temperature for shoyu fermentation occurs during the 8-month period from winter to autumn. Today mash is usually made by mixing koji with brine solution of about 0"C, keeping the temperature of the new mash below 15C for several days, and gradually raising it to 28-30C after 20-30 days (Ebine e? ul., 1976). A I-3% increase in the protein digestibility of the new mash is expected with cooling because the


    lower temperature prevents a rapid decrease in the pH value caused by too rapid lactic fermentation and the inactivation of alkaline protease (Komatsu et al. , 1968; Tazaki et af., 1969; Goan, 1969; Ueda er a f . , 1958; Haga et a f . , 1967; Imai et a f . , 1969).

    According to Machi (1966), the volume of glutamic acid content per lg of total nitrogen in two kinds of shoyu was 0.82 mg and 0.66 mg, respectively. In the first instance, the shoyu was prepared over a period of 330 days, beginning in January when the temperature was 10C under naturally occurring temperature changes; the other was prepared over a period of 220 days, also beginning in January, but with the warming of mash artificially to a constant temperature of 20C. Kuroshima et a f . (1969) have pointed out that glutaminase, which is derived from koji molds, is very sensitive to heat, and its activity rapidly de- creases in new mash. Glutamine is converted into glutamic acid due to the action of glutaminase, but when the glutaminase is inactivated, glutamine nonenzymati- cally changes into pyroglutamic acid, which is not flavorful compared to glu- tamic acid. Kuroshima reported that glutamic acid present in the average shoyu on the market consists of 60% free glutamic acid, 10% pyroglutamic acid, and 30% a conjugated form. Shikata et a f . (1978) separated the glutaminase in koji molds into two fractions, water soluble and insoluble. The latter, which remains in the cells, is more resistant to heat and salt and is the major contributor to the production of free glutamic acid in shoyu mash. Adding glutaminase resistant to heat and salt produced by some yeasts to the new mash has been found to be effective in increasing the glutamic acid content of the final product as long as the temperature of the mash is below 60C (Yokotsuka et a f . , 1968d, 1970, 1972; Iwasa et af., 1972a,b.).

    2 . Period of Mash Fermentation

    The remarkable increase in protein digestibility of shoyu due to recent im- provements in the process of soybean cooking and to koji cultivation has also helped to reduce the fermentation period from the 1-3 years required in the past to less than 1 year.

    According to Udo (1931), the time needed to produce the highest level of glutamic acid in shoyu mash prepared from wheat and soybeans and fermented under natural temperature was 15 months. Umeda (1953) reported a period of 10 months is required for mash prepared from wheat and defatted soybean. Today, however, 3-4 months are required, although a few more months are necessary to complete fermentation and aging. Completing the fermentation process within 6 months without damaging the final product can be accomplished by keeping the shoyu mash at a temperature of about 30C, but heating it to temperatures of 35- 40C reduces its organoleptic qualities. In addition, the amount of water to be


    0 L 3b ' 410 ' 510 ' sb Temp (M

    FIG. 1 1 . Safety zone for enzymatic digestion of shoyu koji. Protein digestibility and amino acid content in Zone A are better than those in Zone B . Anaerobic bacteria: less than lW/g in koji, 10z/g in final koji; aerobic bacteria: less than 108/g in koji, 102/103/g in final broth. From Yokotsuka er af. (1977); Japanese patent No. 1,042, 917; Takamatsu er al. (1975): Japanese patent No. 1,120,428.

    mixed with koji and the salt content of the mash are other major factors which determine the fermentation period. A ratio of 1.2-1.3 parts of water to 1 part of raw materials and 17-18% (wh) of salt in the mash after 1 or 2 months seems to be the average figures in actual industrial production. By keeping the mixture of shoyu koji and water with 0% salt at 55C for 24 hr or by keeping the same mixture with 8% salt at 43C for 48 hr, in both cases with strong agitation and in the presence of heat and salt-resistant glutaminase, about a 90% protein di- gestibility with a more than 5.5 pH value and with a higher glutamic acid content than that of average fermented shoyu was achieved without microbial contamina- tion. The relationship between the salt concentrations and the temperature to avoid microbial contamination in the enzymatic degradation of shoyu koji is indicated in Fig. 11 (Takamatsu et al., 1975; Yokotsuka et al., 1977). It usually takes 2-3 months to finish the lactic and yeast fermentations of salty shoyu mash at 2O-3O0C, but these fermentations of the liquid separated from the enzymatic degradate of shoyu koji at elevated temperatures can be finished within 5 to 10


    days (Yokotsuka and Asao, 1969). Moreover, these fermentations of the liquid degradate of shoyu koji containing more than 8% salt of pH 3.0-7.0 can be finished within 2 or 3 days by passing the liquid through two or three columns which are packed with immobilized lactobacilli, Saccharomyces rouxii and Can- dida versatillis if necessary, respectively (Akao er al., 1982).

    3. Microbes in Mash

    In newly produced mash, salt-intolerant lactobacilli and wild yeasts derived from koji are destroyed rapidly, and Bacillus subtilis remains only as spores. Salt-tolerant Micrococci also rapidly disappear because of anaerobic conditions of mash. The predominant active microbes in shoyu mash are salt-tolerant lac- tobacilli and yeasts such as Saccharomyces rouxii and Candida (Torulopsis) versatillis or C . etchellsii.

    Sakaguchi Kenji (1958) found that major lactobacilli were Pediococcus soyae and Buchanan et al. (1974) determined that they were P . halophylus mor- phologically. Good results have also been obtained by adding pure cultured lactobacilli to the new mash (Watanabe et al . , 1970; Nagase et al., 1971; Jose and Sugimori, 1973). In one typical lactic fermentation of shoyu mash, the intitial inoculum of 102-103 of lactobacilli reached lo8 after 3 months (Jose et al., 1976). Lactic and alcohol fermentations of shoyu mash are presented in Fig. 12. Caution must be taken not to add too much lactic starter as this is correlated with a decrease in the pH value and a decrease in protein digestibility. Some researchers have noted that the diversity of lactobacilli in shoyu mash relates to the aroma, pH, and color of shoyu (Fujimoto et al., 1980), to the metabolic roles of organic acids (Terasawa et al . , 1979), and to the presence of sugars and some amino acids, such as arginine, histidine, tyrosine, and aspertic acid (Uchida, 1978). Various metabolic patterns by lactobacilli in shoyu mash are summarized in Table XIII.

    The initial pH value of mash, 6.5-7.0, gradually decreases as the raw mate- rials are degraded and lactic acid fermentation proceeds, and at around pH 5.5 , yeast fermentation takes the place of lactic acid fermentation. The predominant yeast of shoyu fermentation, S. rouxii, grows and reaches to a viable count of 106-107/ml. To accelerate the alcoholic fermentation and to shorten its develop- ment time, pure cultured yeasts, s. rouxii, are sometimes added to the shoyu mash when its pH value reaches about 5.3, usually 3-4 weeks after the mash making (Watanabe et al., 1970). The addition of Torulopsis yeasts along with S . rouxii have been recommended as a way of obtaining good volatile flavors in the finished product (Suzuki et al.. 1972). The changes in the viable counts of these two kinds of yeast in a natural shoyu mash kept at room temperature are present- ed in Table XIV (Keitaro er al., 1968).




    I '- I /ernperatwe of Mash

    FIG. 12.

    0 30 60 90 120 150 Days d Fermentation

    Lactic and alcoholic fermentation of shoyu mash. From Jose et al. (1976).


    1. Homofermentation: Glucose -D 2 mol lactic acid 2. Heterofermentation:

    3. 4.

    Glucose -D 1 mol lactic acid, ethanol, acetic acid, CO2, Hz, acetone, butanol

    67 patterns of metabolic manners for arabinose, lactose, melibiose, manitol, and sorbitol Metabolic manners for amino acids and citric acid:

    Histidine -D Histamine + COz Tyrosine + Tyramine + CO2 Arginine Citric acid Aspartic acid -D Alanine + COz

    + Ornithine + 2 NH3 + COz -D Acetic acid + malic acid -D Lactic acid + COz

    Adapted from Fujimoto er al. (1978, 1980); Iizuka (1973); Terasawa (1979); and Uchida (1978, 1982).




    Months Saccharomyces Other kinds of aging rouxii ( Torulopsis)

    0.3 0.6 I 1.5 3 6 1 8

    10 12

    0 x 104 136 38 1 530 22 1

    0 0

    21 0 0

    436 X 104 70

    323 100 131 399 I82 96 66


    " From Mogi Keitaro er al. (1968). J . Agric. Chem. Soc. Japan 42(8), 466.

    The factors that most hinder the activities of lactobacilli and yeasts in shoyu mash are, in the case of lactobacilli, its salt content and, in the case of yeasts, ingredients such as guaiacol and vanillin and alcohol, which can be extracted with ether (Sakasai et al . , 1975a,b; Noda et al . , 1975, 1976a-c). Lactic acid fermentation is affected by the yeasts derived from koji and others (Kusumoto et al., 1977; Fujimoto er al., 1980).

    The effect of oxygen supply, initial pH, and inoculum size on growth and fermentation of P. halophylus and S . rouxii was examined. In mixed culture with an initial pH of 6.0, the growth of P. halophylus was inhibited by S . rouxii under aerobic conditions, and the growth of S . rouxii was inhibited by P. halophylus under anaerobic conditions. With an initial pH of 5.6, the growth of P. halophy- lus declined irrespective of the aeration condition (Inamori et a l . , 1984).

    4. Ingredient Change during Mash Fermentation

    Proteins, carbohydrates, and oil from soybeans and wheat are degraded by protease, peptidase including glutaminase, amylase, and lipase, pectinase, and phosphatase derived from koji. The activities of protease and amylase remaining in mash as relates to the progress of fermentation are shown in Table XV.

    The quantities of glycine, alanine, valine, and leucine increase as mash fer- mentation advances, and the quantities of aspertic acid, serine, proline, hys- tidine, arginine, and tyrosine decrease, mostly because of decarboxylation by lactobacilli. The quantity of glutamic acid decreases after reaching its peak



    Proteaseb Amylase

    Days of mash PH 3 PH 7 PH 9 PH 5 fermentation ( 15'WC (25C) ( 15C) (25'C) ( 15C) (25C) ( 15C) (25OC)

    7 20 40 60 90

    135 150 180

    100% 100% 132 103 106 78 86 73 79 78 68 65 74 59 79 63

    100% 88 77 72 41 24 17 18

    100% 100% 100% 100% 100% 91 94 100 82 107 72 82 80 94 55 60 76 59 81 73 24 42 24 63 71 19 21 11 34 34 16 12 11 33 29 17 15 14 44 33

    a Arranged from Komatsu (1968). Seuson. Sci. 15(2), 18. Determined at pH values indicated. Starting temperature of mash.

    because of its nonenzymatic conversion into pyroglutamic acid. Citric acid and malic acid are derived from the raw materials after -60 days. In a shoyu mash for which salt concentration is less than 15%, some lactobacilli such as Lac- tobacillus pluntarium grow and totally decompose glutamic acid by decarboxyla- tion (Hanaoka, 1976). Mono- and disaccharides rapidly decrease as a result of


    Days NaCl of mash @ / I 0 0 ml)

    7 17.7 20 17.4 40 16.9 60 16.7 90 16.6

    135 16.6 I50 16.5 180 16.4

    T N b FN NH3-N FN/TN (g/100 ml) (gl100 ml) (g/100 ml) pH (a)

    0.98 0.36 0.06 5.7 37.1 I .29 0.53 0.09 5.6 41.0 1.55 0.13 0.15 5.0 46.7 1.61 0.80 0.20 4.8 49.3 I .67 0.85 0.21 4.7 51.1 1.69 0.89 0.21 4.7 52.5 I .96 0.91 0.21 4.7 55.7 1.69 0.94 0.20 4.1 45.7

    TNUR (%)

    44.7 55.4 74.4 78.1 81.4 82.4 82.7 83.1

    Arranged from Komatsu (1968). Season. Sci. 15(2), 13. Note: TN, total nitrogen; FN, formyl nitrogen; TNUR, TN in shoyu/TN in raw materials.

    Temperature of mash: 15C (0-30 days); 25C (31-150 days); 25C (31-150 days); and 28C (151- 180 days).


    the lactic and yeast fermentations. The changes in the major chemical ingredients in shoyu mash are presented in Table XVI.

    5 . Agitation of Mash

    Shoyu mash is occasionally agitated with compressed air for the following purposes:

    1. To control the uniform salt concentration and pH value of mash as relates to the enzymatic solubilization of the raw materials in order to prevent bacterial putrification by salt-intolerant bacteria in those parts of mash where there is a low concentration of salt. Accordingly, frequent agitation becomes necessary, es- pecially in newly produced mash.

    2. To provide sufficient oxygen to promote the growth of yeasts in the middle stage of fermentation.

    3. To mix the Torulopsis yeasts which grow aerobically on the surface of the mash in the latter stage of fermentation.

    4. To prevent the growth of too many film-forming yeasts on the surface of mash in the middle and latter stages of fermentation, since overabundance ad- versely affects the aroma of the final shoyu.

    Koizumi and Takahashi (1974) studied the effects of varying frequencies of mash agitation during fermentation. In the control group, the newly produced mash was agitated once every 4 days for 1 month, once every 3 days for the next month, and then three or four times a month. The frequency of agitation in the experimental group was made about one-half that of the control group. The experimental groups frequency of agitation resulted in shoyu with better total nitrogen solubility, better alcoholic fermentation, and a deeper shade of red. By contrast, mixing mash by syphoning the liquid part from the bottom and then pouring it on the surface, as is done in the preparation of tamari shoyu, yields far inferior products based on the above quality characteristics.

    The salt content in water greatly affects its oxygen solubility, which decreases to 60% in 10% salt water and to 40% in 18% salt water as compared with that in fresh water. There was no oxygen solubilized in shoyu mash 15 min after its agitation with compressed air. Yeasts were found surviving within 10 cm, but not within 40 cm from the surface of the shoyu mash I month after mash making (Miyauchi er al., 1981).

    The balance between lactic and yeast fermentation is also greatly affected by the water-absorbing ability of the solid in new mash, which relates to the kinds of koji and to the initial frequency of mash agitation. When a new mash absorbs much water and a smaller water layer in the mash remains, lactic fermentation tends to increase and yeast fermentation is suppressed. A too frequent agitation


    makes mash too sticky and promotes lactic fermentation but suppresses yeast fermentation. The salty mash made from a koji of less water-absorbing ability tends to make more yeast fermentation and less lactic fermentation, if the mash is not so severely agitated (Inamori et al., 1977).


    I . Pressing of Mash

    An aged mash is press-filtered through cloth under increasing hydraulic pres- sure, sometimes reaching 100 kg/cm2, for 1-3 days. The difficulty of pressing shoyu mash has lessened with its increased protein digestibility in recent years. The viscosity of aged shoyu mash used to be more than 3000 cp; today it is generally less than 2000 cp. The activity of plant tissue-degrading enzymes in koji is highly correlated with a decrease in both the viscosity and the amount of shoyu mash presscake.

    Tazaki ef ul. (1962) and other researchers have reported on the effectiveness of cellulase in increasing the extractibility and digestibility of soybean proteins. Further, according to Nakayama ef al. (1965) and Harada ef al. (1966), the addition of macerating enzymes to cellulase enhances the efficacy.

    Ishii et al. (1972) isolated A. sojae no. 48, which processes strong degradating activity in the tissues of soybeans and wheat. Two kinds of shoyu koji were cultured using two different molds: (1) A. sojae no. 48, and (2) A. sojae X816, which was used as the control, and produced the stongest proteolytic activity among the available strains of mold. Both of these koji were digested in 25% saltwater solution. As indicated in Table XVII, the Koji cultured with A. sojae no. 48 exhibited a 3.7% higher protein digestibility and yielded twice as much reducing sugar as compared to the control mold. A significant difference in viscosity was observed in the two mashes after 3 months, and the mash of A. sojue no. 48 was much easier to filtrate. The enzyme activities produced by these two molds in shoyu koji are compared in Table XVIII. Remarkable differences exist in the activities of cellulase C, and the pectinase system such as macerating activity, pectin transeliminase, and endopolygalacturonase. While the cellulase C, preparation from Tricoderma viride which was added to shoyu koji exhibited almost no effect on protein digestibility and only a slight increase in the yield of reducing sugar, these results suggest that the pectolytic enzyme of A. sojae no. 48 in the presence of hemicellulose is effective in promoting the degradation of soybean and wheat tissues and in lowering the viscosity of shoyu mash. But the activity of pectin transeliminase is remarkably hindered in a 17-18% salt solution.

    The relationships between enzyme activities of shoyu koji and mash viscosity



    PRODUCED BY Aspergillus sojae X-816

    AND Aspergillus sojae NO. 48 IN SHOYU KOJI"

    Activity (unitslml of extract)


    Enzyme system X-816 NO. 48

    P-Glucanase Cellulase C, CM-cellulase P-Glucosidase P- 1,3-Glucanase

    Macerating activity Pectin transeliminase Endopol ygalacturonase

    Hemicellulase Xylanase Arab an as e Galactanase

    Acid protease Alkaline protease



    0 4.38 2.77 2.10

    >O 0.53


    26.56 0.52 1.70

    1.35 4.99

    71.4 5.69

    14.01 6.50

    85.7 26.84


    30.90 0.92 1.70

    1.32 2.04

    a From Ishii et al. (1972). Each strain was grown on a culture medium which was composed of 15 g of defatted soybean and 15 g of wheat in an Erlenmeyer flask at 30C. After 3 days the enzyme in shoyu koji was extracted with 150 ml of water.



    OF NaCP

    A. sojae X-816 A. sojue No. 48

    NaCl 17.2% 16.9% Total Nlml 1.66 1.70 Degradation rate of total N 82.8 86.5 Reducing sugar/ml 0.46 0.90

    From Ishii et ul. (1972). Shoyu koji (30 g) of each strain was allowed to stand with 60 ml of 25% NaCl solution at 30C for 37 days.


    and filtration rates were analyzed by using the stepwise multiple regression analysis method. The correlation coefficient between mash viscosity and filtra- tion rate was -0.772 (significant at I % level). Contributing proportions of pectin-liquifying activity, pectin lyase activity, and carboxymethyl cellulose (CMC) saccharifying activity for mash viscosity were 21.3, 19.7, and 17.4%. respectively. Contributing proportions of pectin-liquifying activity, pectin lyase activity, pectic acid-liquifying activity, and ACM saccharifying activity for the filtration rate were 22.0, 18.3, 9.7, and 8.4%, respectively.

    Kikuchi et af . (1975, 1976, 1977; Kikuchi, 1976) investigated those chemical compounds in shoyu mash which make the pressing of shoyu mash difficult and concluded that the problem is largely due to the presence of acidic polysaccha- ride. The insoluble solid contained in the presscake made from shoyu mash was estimated to consist of 10% microbial cells, 30% protein, and 20-30% nonpro- teinous substances derived from soybeans and wheat, respectively. Among these, the content of noncellulose polysaccharides was 7%, but its contribution to the filtration resistance of the insoluble solid was 70%, more than 40% of which was attributed to the acidic polysaccharides. One kind of acidic polysaccharide found in the shoyu contained more than 90% glacturonic acid, which forms a strong gel in high concentrations of aqueous salt solution and is presumed to make the filtration of shoyu mash difficult. Among the three kinds of polysac- charides present in the cell wall of soybeans (arabinogalactan. cellulose, and acidic polysaccharide), acidic polysaccharide goes into shoyu presscake at the ratio of 2: 1. The amount of acidic polysaccharide in shoyu was determined to be only 0.7%, but its contribution to the viscosity of shoyu was 20%.

    2 . Pasteurization

    The filtrate of an aged mash is heated at 70-80C in order to retard the greater part of microbial and enzymatic reactions. The major changes resulting from this heating are the formation of an agreeable flavor and dark brown color, the separation of heat-coagulant substances, an increase in acidity, clarity, and anti- yeast potency, a decrease in the reducing sugar and amino acid content, and the evaporation of volatile compounds (Yokotsuka, 1954; Yokotusuka et al . , 1956, 1958; Okuhara et a f . , 1961; Onishi 1970a, 1971, 1972, 1975, 1976). It is sometimes necessary to remove or destroy the heat-tolerant bacterial spores either by the HTST method or by filtration.

    Retarding alcohol evaporation during pasteurization of shoyu improved its organoleptic acceptance, but only when the coverage is tightly attached to the surface of shoyu, and caution must be exerted to avoid producing too much heat flavor. Adequately cooling the shoyu after heating is necessary. Until recently, the heated shoyu was stored in open tanks for 6-7 days to promote clarification, and it was generally believed that covering the containers of heated shoyu during


    this period produced shoyu of an inferior quality. In Japan, it is legal to add benzoic acid or butyl-p-hydroxybenzoate to the refined shoyu as a preservative, but the trend seems to be toward using aseptic bottling or the addition of ethanol as a preservative. There is a general tendency in Japan to lower the heating temperature of shoyu in the final stage of production in order to produce a product with a milder flavor and a lighter color.

    According to Hashimoto et al. (1971, 1972, 1973, 1974, 1976), the heat- coagulating substances produced by heating raw shoyu are equivalent to 10% of its volume and 0.025-0.05% of its weight. They consist of 89.1% protein, 9.7% carbohydrate, and 1.2% ash. Their nitrogen content is 0.2-0.4% that of the shoyu. The amino acid composition of these heat-coagulating substances in shoyu is significantly different from that of soybeans or wheat in the ratio of aspartic acid to glutamic acid or of proline to leucine. The major ingredients of the heat-coagulating substances in shoyu, determined by immunological identifi- cation, are the undenatured proteins derived from the enzymes produced by koji mold.

    The speed with which coagulation occurs is inversely related to the heating temperature and is thought to be due to the inactivation of proteases which do not tolerate heat, such as acid proteases, neutral protease, and alkaline protease, and to heat-tolerant neutral protease 11, which is relatively stable at 60 or 80C. This fact may explain why the coagulation of shoyu achieved by heating is not caused merely by the coagulation of the undenatured protein derived from the raw materials. Adding a small amount of raw shoyu to pasteurized shoyu remarkably promotes coagulation, which means that there are some factors in raw shoyu which promote coagulation when heated. Coagulation of shoyu is also promoted by the addition of protease isolated from koji, especially when the optimal pH value is 5.0, particularly at higher heating temperatures. The addition of the heat- resistant acid protease, which is produced by Penicillium duponti K1014, is stable at 6O"C, and exhibits its highest activity at pH 4.6 and 75C remarkably promoted the coagulation of heated shoyu. It has been suggested that the protein molecules associate with each other through hydrophobic bonds by the action both of heat and proteases.

    Contributions to the sedimentation of coagula during the shoyu pasteurization process were investigated by measuring the amount, density, and particle sizes of the coagula. The results indicated that a-amylase had no effect, acidic protease had some promoting effect, and alkaline protease had a remarkable retarding effect on the sedimentation of coagula in shoyu.

    3. Chemical Composition of Koikuchi Shoyu

    Koikuchi shoyu produced at the United States Kikkoman plant was analyzed by Okuhara and Yokotsuka (1977), with results presented in Table XIX. The



    Percent (w/w) Percent (w/w) Component of shoyu, as is of shoyu, dry basis

    Soluble solids (dry matter) Alcohol Water (by difference) Inorganic components

    Sodium Chlorine Calcium Potassium Phosphorus Magnesium Sulfur Iron Manganese

    Total Organic components

    Polyols Glycerol Mannitol

    Total Ether-soluble compounds Ether-soluble volatile matter Amino acids

    Lysine Histidine Cystine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Ornithine


    34.00 I .47


    6.10 8.82 0.02 0.40 0.15 0.07 0.06 0.002 0.001


    1 S O 0.17 I .67 0.14 0.005

    0.56 0.21 0.07 0.22 0.90 0.36 0.45 1.92 0.59 0.34 0.38 0.47 0.12 0.41 0.62 0.08 0.36 0.49 8.55

    17.94 25.94 0.06 1.17 0.44 0.21 0.17 0.006 0.003


    4.41 0.50 4.91 0.41 0.01

    1.65 0.62 0.21 0.65 2.65 1.06 I .32 5.65 I .74 I .OO 1.12 1.38 0.35 I .21 I .82 0.24 1.06 1.44




    TABLE XIX (Continued)

    Percent (w/w) Percent (w/w) Component Of s h o p , as is of s h o p , dry basis

    Ammonia 0.30 0.88 Organic acids

    Formic Acetic Citric Succinic Lactic

    Total sugars

    Monosaccharides Mannose Arabinose Galactose Xylose Glucose Unidentified

    Total Disaccharides Oligosaccharides Polysaccharides

    Total organic components Solids accounted for With ammonia calculated

    as amino acids

    Total sugars (as glucose)

    0.02 0.16 0.04 0.05 0.68 0.95

    0.06 0.08 0.17 0.06 2.05 0.23 2.65 0.65

    1.15 4.45


    16. I 31.7 32.69

    0.06 0.47 0.12 0.15 2.00 2.80

    0.18 0.24 0.50 0.18 6.03 0.68 7.81 1.91

    3.38 13.10 47.3 93.2 96.1


    From Okuhara and Yokotsuka (1977).

    middle column gives the component percentages based on the liquid condiment as it is used by the consumer. The right-hand column lists the components on a dry basis. The soluble solids were divided almost equally between inorganic (46%) and organic components (47%). Sodium and chlorine were the principal inorganic constituents. Polyalcohols comprised almost 5% of the soluble solids, amino acids 25%, organic acids nearly 3%. and carbohydrates 13%. Amino acids were determined both before and after acid hydrolysis to obtain free and total values for each. Table XIX gives the values after hydrolysis except for methionine and tyrosine, which are unstable in acid hydrolysis. The values for these two amino acids were obtained before acid hydrolysis and therefore do not reflect any methionine and tyrosine that might have been bound in peptides. Furthermore, Table XIV shows no values for tryptophan. Recently, Hugli and Moore (1972) found 0.002% by a different method. Even with these deficiencies


    the values reported in the table account for 93.2% of the soluble solids in shoyu. The ammonia found after acid hydrolysis probably resulted from decomposition of amino acids and should be calculated as amino acid. When this is done, the figure for total amino acids becomes 9.8% on a wet basis. This calculation accounts for 32.69 g, or 96.1% of the soluble solids in the sample shoyu. However, shoyu also contains browning pigments in addition to the compounds described in the analytical tables. The components listed do not include those many compounds present in trace amounts. Approximately 300 components of the ether-soluble volatile fraction, which constitutes less than 0.005% of fer- mented shoyu, have been identified to date.

    Among the components, the free amino acids have been of most interest because of their characteristic taste and appreciable quality. The free amino acids usually account for 40-50% of the total soluble nitrogen and about 40% of the residual nitrogenous substances which can be hydrolyzed by acid to form addi- tional free amino acids (Oka and Nagata, 1974a). As the latter nitrogenous substances are oligopeptides, it has long been thought that some peptides may contribute to the flavor of fermented food products, although few concrete data have been reported.

    Oka and Nagata (1974a,b) fractionated a shoyu sample by gel filtration on a Shephadex G- 15 column, with subsequent subfractionation on the basis of acid- ity by ion-exchange chromatography. After preliminary fractionation, the com- ponents in the subfractions were transformed into copper salts, and these chro- matographed to separate out neutral peptide subfractions. The peptide fractions were further fractionated on a preparative amino acid analyzer and by paper chromatography. Thus, three glycopeptides and eight dipeptides were isolated and characterized as the major neutral peptide components in shoyu. However, the practical contributions of these components to the flavor of shoyu were judged to be negligible. Four dipeptides and sugar derivatives of ten dipeptides and two tripeptides were isolated by further fractionation of the acidic subfrac- tions and characterized as the major acidic peptides in shoyu. However, it was difficult to attribute any direct contribution of these peptides to the flavor of shoyu on the basis of their quantity and taste response.


    The color of shoyu is an important attribute to Japanese dishes, although it has become lighter in recent years. The color and flavor of shoyu are very closely related, as both are affected by the aging of mash and the pasteurization of raw shoyu. During the brewing process, the development of shoyu color derives mainly from nonoxidative and nonenzymatic browning reactions. Enzymatic


    reactions, which occur between amino compounds and sugars, are rare. When koikuchi or usukuchi shoyu is packed in glass bottles or cans, the color is relatively stable, but it darkens rather quickly after the seal is broken due to the oxidative and nonezymatic browning reaction. These reactions cause the organoleptic quality of shoyu to be inferior. In the preparation of usukuchi shoyu, considerable effort is directed toward minimizing the intensity of color development by decreasing the amount of protein and total solid in the mash, increasing its salt concentration, and by avoiding too long a period of fermenta- tion and aging as well as extended heating of the raw shoyu during pasteuriza- tion. In these respects, it differs from the production of koikuchi shoyu.


    Kurono (1927) reported that the color of shoyu was a type of melanoidin pigment and consisted mainly of two compounds: C,,H,,N,O,, and C,,H,,N,O,,. Omata et af. (1955d) separated the color substances of shoyu into two fractions, acidic and basic, by column and paper chromatography and then spectrometrically determined the increases of these fractions during the brewing of mash and the storage of shoyu. The quantitative increase in the acidic fraction was higher than that in the basic fraction. Mitsui and Kusaba (1957) also isolated two kinds of shoyu pigment, one of which was the same as that isolated by Kurono.

    Hashiba (1971) isolated the browning compounds present in shoyu by gel filtration with Sephadex G-25 into three peaks, PI, PII, and PIII, according to the rate of elution. The quantity of PI increased during oxidative storage, while the quantity of PI11 increased remarkably during the pasteurization process. The increase of PI gave the shoyu a dark brown color, while that of PI11 gave it a red tone. Hashiba (1973a) purified the melanoidin produced during the storage of shoyu at 37C for 50 days by dialysis, DEAE-cellulose chromatography, and Sephadex G-100 gel filtration until a single band on the disc electrophoresis appeared. The color of melanoidin thus obtained was not affected by heating or oxidation. When hydrolyzed, the melanoidin liberated sugars such as glucose, xylose, galactose, and arabinose, and all of the amino acids found in shoyu.

    Motai et al. (1972) and Motai and Inoue (1974a) fractionated the material which gives shoyu its color into eight color components by DEAE-cellulose chromatography with stepwise elution. The color intensity of each peak became darker, and E450 and molecular weight became higher with successive orders of elution. When the shoyu was heated, the color components became brighter, while with oxidation they became darker in tone. The melanoidin pigments prepared by heating an aqueous solution of glycine and xylose at 100C for 2 hr were chromatographically fractionated into eight color components. The frac- tionated color components from the glycine-xylose model system exhibited


    similar changes when heated and oxidized, as did those of shoyu. Based on spectral measurements, elemental analysis, and amino acid analysis, all the color components appeared to be very similar in chemical structure, having stepwise different molecular weights. The infrared absorption spectra of eight peaks had the same pattern, suggesting that they were melanoidins. These results indicate that shoyu is made up of at least eight kinds of melanoidin pigment having different degrees of polymerization.


    The color of shoyu represented by the International Commission on Illumina- tion (CIE) system was reported by Omata and Ueno (1953a) to have a dominant wavelength of 590-620 nm, an excitation purity of 86-88%, and a luminous transmittance of 0.14-0.17. Using the CIE system, Umeda and Saito (1956) analyzed the color of shoyu prepared from mash aged for different periods of time. The color standard of shoyu was prepared from known chemical pigments so as to match the color changes described above, which consisted of 30 degrees of color of the same visual distance. This method of using a color standard to assess the color of a shoyu is very simple and convenient when the shoyu has a single or consistent color tone (i.e., when it is separated from mash or just after pasteurization). However, shoyu having different color tones produced by oxida- tion during storage is difficult to analyze by this method.

    Motai (1976) has reported a linear relationship between the logarithm of absorbance (log A) and the wavelength (450-650 nm) in the color distribution of shoyu and melanoidin prepared from the model system. (A similar relationship has been observed in whiskey, cola drinks, beer, caramel, and miso.) There was no change in log A per 100 nm (designated as AA) in each of eight pigments which were fractionated from shoyu or melanoidin using the model system, either during heating or oxidation. Therefore, the log A per 100 nm in this case can be used as the parameter for expressing the color tone of shoyu.

    That shoyu becomes more red in color when heated and takes on a darker brown tone when oxidized is generally acknowledged (Okuhara et al., 1969).

    As shown in Fig. 13, Motai (1976) observed three types of browning reactions of shoyu in relation to the increase in color intensity (E450). His findings are summarized below:

    Type a: The color tone darkens along with an increase of AA, which occurs during the storage of shoyu in the open air as a result of non- enzymatic oxidative browning.

    Type b: The color tone is unchanged along with the unchanged AA, which occurs in bottled or canned shoyu as a result of the heat-dependent browning of shoyu.


    m C m 9,

    n In m

    0 +

    E 5 m m 0 -1

    .- L


    I I I I I

    450 500 5 5 0 600 650

    Wavelength (nm)

    FIG. 13. Change of A4 during the browning reaction. From Motai et al. (1972).

    Type c: The color tone becomes bright along with an increase in AA during the aging of the mash or pasteurization of shoyu, which occurs as a result of the heat-dependent browning.

    The AA values of koikuchi and usukuchi shoyu available on the market were reported to be 0.63-0.70 and 0.56-0.60, respectively. The smaller AA value of usukuchi shoyu is due to the shorter aging period of the mash and to the shorter pasteurization time used in the preparation of usukuchi as compared to koikuchi shoyu .


    1. Color Formation during the Brewing Process

    About 50% of the color of koikuchi shoyu is formed during the fermentation and aging of mash, and the remaining 50% during pasteurization of shoyu. Both



    Period of mash fermentation Color degree Percentage

    (months) M (E450) of color formation

    1b 0.45 2.48 13.7 26 6.08 6.08 33.7 6b 8.54 8.54 47.2 6 (pasteurized)c 18.07 18.07 100.0

    a From Motai (1976). The color of the liquid part of mash was determined. The liquid part of mash was pasteurized and the color was determined.

    are considered to be due primarily to heat-dependent browning, or the so-called Maillard browning reaction between amino compounds and sugars. One example of the development of color during shoyu preparation is given in Table XX (Motai, 1976). The amount of hexose in shoyu, which is mainly glucose, is from 6 to 10 times greater than the proportions of pentose. However, some researchers have considered that the sugar, which is involved primarily in the browning of shoyu, is a pentose such as xylose and arabinose (Kamata and Sakurai, 1964; Kato and Sakurai, 1962; Shikata et al., 1971a).

    About 30% of the pentosan contained in the raw materials is degradated into the water-soluble form, and the resultant color intensity of the mash is propor- tional to the amount of pentosan dissolved in the mash. The degradation of pentosan in the course of koji cultivation increases with the elevation of the temperature of koji, which results in an increase in the color intensity of the shoyu obtained. According to Okuhara et al. (1969), to three kinds of shoyu were added 0.025, 0.5, 0.75, 1.0, 1.25, 1.5, 1.75, and 2.0% xylose, respec- tively, and these were heated at 80C for 5 hr. The results are presented in Fig. 14. There was a linear relationship between color intensity and the utilization of pentose, but the degree of color change per 1 mg utilization of xylose for the three kinds of shoyu was different. When the mixture of xylose, glucose, glycine, lactic acid, 18% salt water, and shoyu was heated after adjusting its pH to 4.8, the increase in color intensity was smaller and occurred more slowly than when the shoyu or chemical protein hydrolysate was heated. The participation of xylose in the heat-dependent browning of shoyu was calculated to be only 10- 20% (Okuhara et al., 1975).

    Okuhara er al. (1970) statistically analyzed the relationship between the com- position of raw shoyu separated from mash and its rate of browning. The correla- tion and multiregression models were calculated as indicated in Table XXI. In order to obtain various concentrations of individual shoyu components, in Pro-


    F 0 0 tn

    W u C m n - L


    n m

    1 .o


    5 10 15

    Pentose consumption ( m g / m l )

    FIG. 14. al. (1969).

    Relation between xylose (pentose) consumption and pigment formation. From Okuhara et

    portion I, varying weights were given to the koji per unit amount of raw mate- rials, and in Proportion 11, small differences were given to the concentration of salt mash. Those components that correlated significantly with the browning rate were the concentration of koji in mash (CK), the weight of Koji from a unit amount of raw material (WK), the amount of shoyu obtained from a unit amount of raw material (L), nonamino nitrogenous compounds (TN - FN), the titratable acidity (TA), reducing sugar of shoyu (RS), the reducing power of raw shoyu [R(N)], and above all (TN - FN) and (RS). Such correlations indicate that the intensity of browning pigment can be calculated from (TN - FN) and (RS). Okuhara et al. (1971) also speculated that (TN - FN) might be significantly correlated with the browning of shoyu, based on the following experimental findings:

    1. The browning of shoyu takes place more rapidly than does a solution of

    2. The browning of shoyu is dependent upon the degree of mash fermentation sugar and amino acids.

    and the amount of (TN - FN).



    CK WK L TN - FN FN - NH3 - N NH3 - N TA RS Alcohol Org. A PH RN R(HN) Red(N) Red(HN) NaCl C N b C W b C(HN)b

    * * e

    * ** *

    * * -* **

    (-) * *


    -** **(+I *

    (-) **(++) **

    *(-) (- -)

    * ** * **

    ** **(++) **

    (+I * **(+) ** ** **(++)

    -** -*

    ** ** * *

    ** * **


    ** ** **

    ** **(+) **(++)


    **(+) ** **(++) **

    -** -** - * *( - -)

    ** **

    a From Okuhura et al. (1970). b C(N): Color intensity of raw shoyu; C(H): color intensity of pasteurized shoyu; C(HN): sub-

    tracted C(N) from C(H); C(0XH): subtracted C(H) from color intensity of oxidized shoyu; C(HF): subtracted color intensity of faded shoyu from C(H).

    c Shoyu types I or 11. CK: The ratio of koji to mash water; WK: weight of koji (kg) per material unit; L: volume of

    shoyu per material unit; TN - FN: subtracted form01 nitrogen (%) from total nitrogen (56); FN - NH3 - N: subtracted ammonium nitrogen (%) from folmol nitrogen; NH3 - N: ammonium nitrogen (a); TA: titratable acidity (meq/lO ml); RS: residual reducing sugar (8); Alcohol: % (v/v); Org. A: organic acid (meqlml); pH: pH of raw shoyu; RN: reducing power of raw shoyu; R(HN): subtracted RN from reducing power of pasteurized shoyu; Red(N): reduction (pg/ml) of raw shoyu; Red(HN): subtracted Red(N) from reductone (pg/ml) of pasteurized shoyu; NaCI: % (wlv).

    e *, Linear correlation is significant at 5% level; **, linear correlation is significant at 1% level; (+) or (-), partial correlation is significant at 5% level; (++) or (--), partial correlation is significant at 1% level; +, positive correlation; -, negative correlation.


    0 LD

    f Lu







    0 0.1 0.2 0.3 0.4 0.5 0.6

    [F.N] X[Pentose]% of the d igested solution

    FIG. 15. The relation between (FN) X (pentose) and color degree. (x), Polished rice; (a), corn; (v), com gluten; (V), domestic wheat; (W), wheat gluten; (O), dehulled domestic wheat; (A), imported wheat; (O), Whole soybeans; (A), wheat bran; (0), defatted soybean. From Motai et al. (1975).

    3. The browning of a solution of xylose and glycine is accelerated by adding a small amount of shoyu.

    4. The browning of a solution of xylose and glycine is accelerated by adding the enzymatic hydrolysate of soybean protein formed by using the crude extract of shoyu koji as an enzyme source.

    Shikata et al. (1971b) enzymatically hydrolyzed the raw materials of shoyu with an enzyme mixture of cellulase, diasatase, and protease and observed that the color intensity of the hydrolysate solution was highly correlated with formol nitrogen X pentose% (Fig. 15).

    According to Motai et al. (1973, when shoyu was brewed varying the ratio of concentration of soybeans and wheat, the heated shoyu became darker in color with greater concentrations of soybean. The color tone of shoyu formed by heating became lighter with an increasing ratio of the soybean concentration; shoyu produced from wheat alone exhibited a darker color. These investigators isolated the amino fraction and the sugar fraction from the enzymatic hydrolysate of defatted soybean and wheat, respectively, and determined the contributions of these fractions to the browning of shoyu under heat. It was suggested that the contributions of the amino fractions of soybeans and wheat were 75 and 25%,




    Koji (72 hr cultivation) 0 Liquid part of 3-day mash 0.9

    6-day mash 3.2

    33-day mash 14.0 20.0

    10-day mash 10.0

    Raw shoyu after 4 months

    From Kato er 01. (1961).

    respectively, and those of the sugar fractions were 44 and 56%, respectively. These data suggest that the contribution of soybeans and wheat to changes in color observed in shoyu when subjected to heat is 60 and 40%, respectively.

    Omata et al. (1955a,b) found that some ether-soluble carbonyl compounds of shoyu, including furfural and acetaldehyde, darken the color of shoyu.

    Kato (1958, 1959) concluded that aromatic amine-N-xylosides decompose to form red pigments of melanoidin when catalyzed by a weak acid and that furfural is not an intermediate in melanodin production. Kato (1960) isolated 3-deoxy- xylosone and 3-deoxy-~-g~ucosone from this reaction mixture as bis-2,4-di- nitrophenylhydrazones and pointed out the significance of their role as intermedi- ates in the browning reaction which occurs during the development of melanoi- din. Kato et al. (1961) also identified 3-deoxy-~-glucosone in fermented shoyu. Its quantities were 8 mg% in koikuchi shoyu, 3 mg% in usukuchi shoyu, and 17 mg% in tamari shoyu, respectively, but it was not found in the chemical hydroly- sate of plant protein. These amounts of 3-deoxy-~-glucosone in fermented shoyu are much greater than that of furfural in shoyu, which was reported to be 0.2-0.7 mg% by Omata (1955b). Moreover, it was pointed out that 3-deoxy-~-glucosone is more reactive with amino acids than with furfural. The amount of 3-deoxy-~- glucosone formed during the preparation of shoyu is presented in Table XXII (Kato et al., 1961). 3-Deoxypentosone has also been isolated from pasteurized shoyu, but in smaller quantities than 3-deoxy-~-glucosone, a finding which has been attributed to its lack of stability. When xylose was added to pasteurized shoyu and kept at 37"C, a rapid increase of 3-deoxypentosone content was observed. When pasteurized shoyu was kept at 40C or heated at 80"C, the amount of hexose and pentose of shoyu decreased, the intensity of the color (as determined by measuring the absorbance of 470 nm) increased, and the 3-deoxy- D-glucosone gradually increased and then decreased after.reaching a peak. This

  • 250


    y o

    H-C-OH y 7


    c=o I


    I CH,OH


    Furf urals H .................................................................

    -2 H,O H o H c ~ c H o Pasteurization

    Aldose (Pentose) w

    Amino compounds

    Water soluble pigments

    reaction was accelerated by the oxygen in air, but also proceeded in the absence of air (Kato and Sakurai, 1962). From these results, the reaction mechanism by which the color intensity of pasteurized shoyu increases is presumed to proceed as follows: aldose + 3-deoxyosones ---* color pigments.

    When the chemically synthesized 3-deoxyosones, including 3-deoxypen- tosone, 3-deoxy-~-glucosone, and 3-deoxygalactosone, were added to shoyu, large amounts of water-soluble browning color substances were produced, but only a small amount of browning color substances was produced from furfural. Kato and Sakurai (1963) concluded that in amino-carbonyl reactions most 3- deoxyosones are reacted with the excess amount of amino acids instead of being converted into hydroxymethylfurfural (HMF), resulting in the formation of water-soluble pigments (see Fig. 16).

    According to Burton et al. (1963), almost all carbonyl compounds react with amino radicals, and a-ketoaldehydes (pyruvic aldehyde, 3-deoxyosones), di- ketons (diacetyl), and a-p-unsaturated aldehydes (crotonaldehyde, furfurals) are the most reactive species among them. a-Hydroxyaldehydes change into a-p- unsaturated aldehydes after dehydration and react with amino compounds. Re- ducing sugars react with amino radicals and produce a-ketoaldehydes (3-deox- yosones) and a-p-unsaturated carbonyls (unsaturated osones) and take part in the browning reactions. Xylose is 10 times more reactive with amino radicals than glucose.

    Water hard soluble pigments


    According to Motai (1976), there is a linear relationship between the increase in intensity of color atid the elevation of temperature during pasteurization of shoyu: This relationship is expressed by the following equation:

    D = x 10-0.04'

    where D is the time attaining to definite color intensity, a is a constant, and t is the temperature of pasteurization. The a value varies with the concentrations of pentose and total nitrogen in shoyu; generally, the higher the pentose X total nitrogen, the lower the a value. The intensity of browning increased 2.5-3 times for koikuchi shoyu, corresponding to a 10C elevation of temperature within the range of 50-90C (Motai, 1976; Onishi, 1970b). Generally, the higher the pH value tends to be, the greater the extent of the browning reaction, but within the average range of pH values of shoyu (4.6-4.9), there is no practical difference in the extent of heat-dependent browning.

    Moriguchi and Ohara (1961) observed that when soybeans to which 0.8% K2S20, had been added were steamed and then subjected to enzymatic diges- tion, as in the usual method of usukuchi shoyu preparation, the color of shoyu obtained was lighter by 37% than a control group. Okuhara and Saito (1970) reported a slight effect on the depression of heat-dependent browning of shoyu when it was decolorated by heating with reducing agents such as ascorbic acid or systine, metals such as Zn, Al, Fe, Na, and Mg, or by electrolysis.

    Shoyu made from mash which has been well fermented with yeast is less susceptible to heat-dependent browning because the pentose is assimilated by the yeast along with glucose, and the decrease in rH value caused by yeast fermenta- tion prevents the browning reaction (Okuhara et al., 1975).

    The color of shoyu is also dependent on the kinds of lactobacilli or P. halophylus in salty mash (Fujimoto et al., 1980). Kanbe and Uchida (1984) reported that the rH value of the shoyu mash naturally inoculated with lactobacilli decreased to 7.5 around the time of maximum growth, about 50 days from mash making, while the shoyu mash inoculated with 1 X 1W/g of P. halophylus no. 34, which was isolated as having a strong reducing potency, showed an rH value of 6.0 at the peak of its growth. The color of the shoyu mash inoculated with no. 34 after 180 days of storage was more than 35% lighter than that of the naturally inoculated control mash. The raw shoyu was separated from these two kinds of salty mashes. They were pasteurized under the same conditions and their heat-dependent and oxidative browning rates were determined on the same basis of NaCl at 17.2% and TN at 1.57% of shoyu. These rates were 24 and 18% less, respectively, than those of the shoyu from the control mash. It was also observed with the test shoyu that the reducing power for potassium ferricyanide and the contents of hy- droxymethylfurfural, reductones, and 3-deoxyosones, all of which belong to the so-called browning intermediate compounds, were 73, 73, 78, and 9 1%, respec- tively, of those of the control shoyu.

  • 252

    a 1

    - E &


    0 u-2 0











    0 2 0 4 0 6 0 8 0 0 2 4 6 8

    Time ( d a y )

    FIG. 17. Change of s h o p color on storage at 30C. Opened condition: 300 ml of shoyu was placed in the beaker with the same height, but different surface area, covered with cellophane film, and stored at 30C. From Motai and Inoue (1974).

    2. Color Formation on the Storage Shelfafter Opening the Seal

    According to Omata and Ueno (1953b), the change in color which occurs in pasteurized shoyu during storage is not due to enzymatic reactions, but is greatly affected by the action of air and, to a lesser degree, temperature and light. When the color of shoyu is deepened by aeration, its transmittance curve does not change significantly.

    When the pasteurized shoyu is sealed and stored in a glass bottle or a can, the color intensity increases as the result of browning by heat, but without a change in the AA value. The increase in color intensity of shoyu stored open to the air is much greater and is caused by nonenzymatic oxidative browning in which the AA value decreases as the ratio between the surface area and the volume of the shoyu increases. These changes are shown in Fig. 17 (Matai and Inoue, 1974).

    The effective participation of reductones in the oxidative browning reactions was pointed out by Hodge (1953). The nonenzymatic oxidative browning of shoyu which occurs during storage has been attributed to the participation of such intermediates of Maillard reaction as reductones, Amadori reaarangement com-


    pounds, and melanoidins. The ascorbic acid in shoyu, which belongs to the catalog of reductones, changes into dehydroascorbic acid with oxidation, which in turn reacts with amino acids to deepen the color of shoyu (Omata et al., 195%). Okuhara et al. (1972) heated raw shoyu at 60, 70, and 80C, respec- tively, to obtain different intensities of color. These pasteurized shoyu were then shaken in the air for 24 hr at 40C. The oxidative browning and the reductone formation which took place during the heat treatment was almost proportional to the starting color intensity of each pasteurized shoyu. The oxidative browning reaction took place simultaneously with the utilization of the reductones during the oxidation. It seems likely that the pigment formed from the oxidative brown- ing reaction was not from an amino-carbonyl reaction of the dehydro com- pounds of the reductones, but that it was an oxide of the reductones. The oxidative browning mechanism was analyzed by multiple correlation analysis; the results confirmed the correlation between the oxidative browning and the oxidation of reductones formed during the heating of shoyu and the BaumC of the shoyu.

    The amount of oxidative browning determined by OD at 600 nm with both raw and pasteurized shoyu was correlated only with the initial color intensities and not with the temperature of heating, as shown in Fig. 18. The shoyu of the same color intensities browned by oxidation to the same color degree, regardless of the contents of reducing sugars, peptides, amino acids, and other compounds, if the shoyu was made from the same raw materials and the concentrations of both total nitrogen and sodium chloride were adjusted to be the same (Okuhara et al., 1977).

    Hashiba (1973b) separated shoyu into three fractions-cationic, neutral, and anionic-by using ion-exchange resins. When these three fractions were stored separately, only the cationic one darkened considerably. When they were com- bined and stored, the color of the mixture increased at nearly the same rate as that of the original shoyu. The effects of the anionic fraction containing organic acids and the ashed cationic fraction on the overall browning of shoyu were calculated

    0.1 0.1 5 02 Init ial color intensity

    FIG. 18. Relationship between oxidative browning and initial color intensity of shoyu. Pasteuriza- tions were carried out at (x) 80C. (A) 70C. and (0) 60C. Pasturized shoyu was oxidized by shaking at 40C for 24 hr. ha, Increment in OD at 600 nm by oxidation. From Okuhara etal. (1977).


    to be 10-20% and 20%, respectively. The sum of the contribution rates on the anionic fraction, the neutral fraction, the amino acids, and the ashed cationic fraction in the browning of shoyu was calculated to be -40%. Compounds responsible for the residual 60% are thought to be present in the cationic fraction. It was suggested that such compounds have strong reducing powers and oxygen- uptaking ability.

    Hashiba (1974) prepared a simulated shoyu, which was an amino acid solution containing glucose (5%) , xylose (l%), NaCl (17%), and lactic acid (2%), and adjusted the final pH to 5.0. This sugar-amino acid model system was stored for aging for 3 months at 30C under anaerobic or aerobic conditions and subse- quently for another 2 weeks at 37C under aerobic conditions in order to analyze the extent of oxidative browning. The oxidative browning of the model systems increased as the length of the aging period increased; the model system aged under anaerobic conditions darkened less than did those aged under aerobic conditions. Adding 40 ppm Fe2+ to the model system, which is the average amount of Fez+ in shoyu, accelerated this oxidative browning reaction. An Amadori product, 1 -deoxy- 1-glycine-D-fructose, was isolated from the aged glucose-glycine model system and played a role in causing a marked increase in the rate of oxidative browning. Hashiba (1975) also isolated an Amadori com- pound, 1 -deoxy- 1 -diglycine-D-fructose, from the glucose-diglycine model sys- tem. This Amadori compound promoted the oxidative browning of the aqueous solution of glucose and diglycine, which was further accelerated 30-40 times by the presence of Fez+. In the browning reaction between glucose and triglycine, similar intermediates were detected.

    The Amadori compounds, composed of aromatic or heterocyclic amino acids such as fructose-tyrosine, fructose-phenylalanine, fructose-histidine, and fruct- ose-tryptophan, were especially reactive in oxidative browning, which was syn- ergistically accelerated by the presence of Fez + and Mn2 + . Oxygen is thought to accelerate the breakdown of Amadori compounds to liberate amino acids and glucose (Hashiba, ,1976). The Amadori compounds derived from pentose, such as xylose-glycine, browned more rapidly than those from hexose, such as fruct- ose-glycine. In a reaction between glucose and seven peptides, the liberation of C-terminal amino acids by the cleavage of peptide bonds was observed. There is some evidence that amino acids are liberated from the peptide in Amadori com- pounds, as the peptide bond in Amadori compound has been found to be more labile than that of free peptide (Hashiba et al., 1977).

    Amadori compounds have been isolated from shoyu by ion-exchange chro- matography, gel filtration, and paper chromatography (Hashiba, 1978). Five compounds-fructose-glycine, fructose-alanine, fructose-valine, fructose-iso- leucine, and fructose-leucine-were identified and their relative quantities in shoyu estimated to be approximately 0.2, 0.3, 1.2, 1.3, and 1.5 mM, respec-



    Browning after 14 days at 37C (E559)

    Content in shoyu Amadori compound" Nonoxidative Oxidative (mm molc/liter)

    F-Gly 0.012 0.266 0.2 F-Ala 0.009 0.360 0.3 F-Val 0.009 0.360 I .2 F-isoLeu 0.006 0.400 I .3 F-Leu 0.010 0.310 I .5 Mixture of amino acids and sugard O.OO0 0.003 -

    From Hashiba (1978). 0.1 M Amadori compounds were added to amino acid mixture; F, fructose. 0.1 M glucose was added to amino acid mixture. Contents of Amadori compounds in shoyu.

    tively. These Amadori compounds caused increases in browning with the pres- ence of oxygen and iron (Table XXIII). In addition, amino acids promoted the oxidative browning process. Amadori compounds from pentose or peptides were considered to be so unstable that they would have decomposed while passing through the chromatographic resin. For this reason they were not isolated from shoyu. Amadori compounds reacted with iron and produced red pigments, from which colorless compounds (Fig. 19), were separated (Hashiba and Abe, 1984).

    According to Hashiba (1981), the participation of peptides in the browning process during the aging of shoyu mash was remarkable, but amino acids are more active than peptides in the oxidative browning of shoyu. The respective contributions of pentose and hexose to oxidative browning were estimated to be 75 and 25%.

    According to Motai and Inoue (1974b), the color compounds of shoyu consist of polymerized melanoidins at different degrees, and the oxidative color increase in shoyu as a result of this polymerization occurs according to the following equation:

    E = K x M a

    where E is the color intensity determined by Ei8=,, (450 nm), M is the molecular weight, and K and a are constants. The a value is almost totally independent of the length of heating and the kinds of sugars used in the melanoidin reaction, hut is dependent on the kinds of amino compounds used and, above all, on their molecular weights. Table XXIV presents a comparison between K and a values


    OH R ; H ( F - G l y )


    CH3 i F - A l a C H 2 0 H

    I I RC H COO H

    FIG. 19. Colorless compounds separated from red pigments which were produced by oxidative browning between F-Gly or F-Ala and iron. From Hashiba and Abe (1984).

    of miso color, shoyu color, and various other melanoidins, and further suggests that the melanoidins of shoyu originate from di- or tripeptides.

    It is generally ackowledged that deterioration of the flavor of shoyu is related to its oxidative browning; it is more highly correlated with increased darkening (or a decrease in AA caused by oxidative browning) than with increased color intensity (or the heightened red color which occurs with heating). Changes in volatile flavor components, especially a decrease of ethyl acetate and an increase of acetaldehyde, have been observed with oxidative browning (Onishi, 1970).

    3. Other Factors Which Afect Browning

    The elevation of temperature (but only up to 65C) promotes oxidative brown- ing of shoyu (Motai, 1976). Lactic acid and citric acid also promotes the process (Hashiba, 1973). However, phosphoric acid has no effect, although it does



    Melanoidin K a

    Miso Shoyu Gly-xylose system Lys- Glu- GlY2- G 1 y - k ~ - GlY3-

    4.57 x 1.32

    2.75 0.29 1.45 0.39 0.30 0.56 0.11 0.70 0.01 15 0.95 2.70 x 1.45

    4.47 x 10-4 1.30

    From Motai and Inoue (1974).


    promote heat-dependent browning in the Maillard browning reaction (Kato, 1956).

    The influence of iron on the browning reaction has long been known. Furuta and Ohara (1954) added 30-60 ppm Fe3+ to shoyu mash and observed an immediate 12-20% increase in browning. By comparison, a 21-30% increase was noted in a control sample after storage for 20 days at 30C. Adding Fe3 + to shoyu heated to temperatures of 80-100C had little effect, unlike its addition to mash. According to Hashimoto et al. (1970), the average amount of iron in shoyu is 20-30 ppm and is calculated to be derived from raw materials: soy- beans, wheat, salt, and water. Most of the iron in raw and heated shoyu is in the form of Fez+. The addition of tannic acid or potassium ferrocyanide effectively removed 60-70% of the iron in shoyu without affecting its organoleptic quality. When these chemicals were added during shoyu heating, about 93% of the iron was removed. The influence of Fe2+ on darkening during storage was less than that of Fe3 + and Cu2+. However, the rate of browning during storage of shoyu containing 7-10 ppm Fe and which had been prepared by the above procedure was about equal to that of shoyu that typically contains 20-23 ppm Fe. The oxidative browning increase in the color intensity of shoyu containing 2 ppm Fe was slightly less than that of untreated shoyu containing 32 ppm Fe.

    According to Hashiba et al. (1970), Fez+ and Mn2+ contribute to shoyus increased color intensity during oxidation, while the other trace metal ions, Cu2 + , Zn2 + , Co2 + , and Cd2 -t , do not. Iron ions (Fe2 + , Fe3 + ), however, have no effect on the darkening of shoyu with treatment.

    According to Hashiba (1973c), when raw shoyu was ultrafiltrated, it lost 10- 40% of its initial color; the color intensity of treated shoyu when heated was 33- 50% that of untreated shoyu. No sedimentation was found in the course of heating the treated raw shoyu. When pasteurized shoyu was ultrafiltrated, the intensity of its color decreased to about half that of untreated shoyu. The sub- stances related to the browning of shoyu, such as Fe2+, 3-deoxyglucosone, hydroxymethylfurfural, reductone, carbonyl compounds, and ferricyanide-re- ducing substance, were removed by ultrafiltration; 5-7% of the total nitrogen and 20% of the reducing sugar contained in shoyu were removed by the same procedure.


    Using a multivariate analysis, Tanaka et al. (1969a) indicated that among the factors by which preference for a given shoyu was formed, its chemical composi- tion as a whole contributed only 46.3%. Among the individual chemical compo- nents, listed in descending order of their degree of correlation to a preferred




    Partial correlation

    Component coefficient@

    Alcohol B a u d NaCl Reducing sugar Color Formyl nitrogen Total nitrogen Glutamic acid Titratable acidity

    Ammonium nitrogen PH

    0.35 -0.30 0.21 0.21

    -0.19 0.11 0.09 0.09 0.07 0.05 0.02

    a From Tanaka et nl. (1969a). Contributing proportion: 46.3%.

    rating, are alcohol, BaumC, sodium chloride, reducing sugar, color, formol nitrogen, total nitrogen, glutamic acid, titratable acidity, pH, and ammonium nitrogen (see Table XXV). The 17 aspects relating to odor contributed 96.5%. There were no predominant factors, but fragrance and the aroma of alcohol were the major desirable factors; the major negative factors were the smell of chem- ically hydrolyzed proteins, an oily smell, a Natto smell, an abnormal smell, a butyric acid smell, a warm brewing smell, a steamed soybean smell, and a moldy smell (see Table XXVI). Factors related to taste contributed 97.6% to preference judgments. A good aftertaste, a pure, a palatable, and a moderate salty taste were the major desirable factors; a too sweet, too sour, and an abnormal taste were the major negative factors (see Table XXVII).

    The relationship between the organoleptic evaluation of a shoyu and its chem- ical constituents was investigated with 59 brands of shoyu available on the Japanese market (Tanaka et al., 1970). Nineteen kinds of chemical and physical analyses were conducted. The results indicated that the fragrance of a fermented shoyu was roughly proportional to its ethanol and extract content ( r = +0.700 and +0.425, respectively). .The correlation coefficient in a linear regression estimation between the smell of the chemical hydrolysate of defatted soybean or some other plant protein and the levulinic acid content of fermented shoyu blended with the chemical hydrolysate was found to be +0.942. The preferred




    T ratio of major partial correlation coefficients

    Component (11 out of 17)b


    Smell of chemical hydrolysate of proteins Oily smell Deteriorated smell Fragrance Natto smell Abnormal smell Butyric acid smell Warm brewing smell Steamed soybean smell Alcohol smell Moldy

    -1.59 -1.34

    1.29 1.18

    -0.97 -0.97 -0.76 -0.73 -0.67

    0.66 -0.61

    From Tanaka el al. (1969a). Contributing proportion: 96.5%.



    T ratio of partial Component correlation coefficientsb

    Aftertaste 2.98 Pure 2.46 Sweet -1.86 Sour -1.34 Salty 1.30 Palatable 1.25 Abnormal -1.14

    Good body -0.53 Harmony - 1 . 1 1

    a From Tanaka et al. (1969a). b Contributing proportion: 97.6%.


    aroma of a shoyu was negatively correlated with the levulinic acid content (r = -0.538). The preferred pH value of a shoyu in terms of aroma was found to be between 4.6 and 4.8. The higher pH value of a shoyu suggests the possible blending of some quantity of chemical hydrolysate of plant proteins or an unde- sirable bacterial contamination during koji cultivation and/or mash fermentation, all of which relate to the fact that few of these samples had a good aroma. On the basis of these findings the investigators concluded that in order for shoyu to have a good aroma it should (1) be prepared by a genuine fermentation process without the addition of the chemical hydrolysate of plant proteins; (2) be free from undesirable bacterial contamination during koji and mash fermentations; (3) be made from mash thoroughly fermented with yeasts; (4) have the pH value 4.6-4.8; and (5) be appropriately balanced in terms of its chemical components.

    The palatability of shoyu as a function of its salt and total nitrogen content was investigated by Tanaka et al. (1969b). Nine kinds of shoyu were prepared by the combination of total nitrogen, 1.5, 1.6, and 1.7% (w/v), and sodium chloride, 14, 16, and 18%. Degree of saltiness, palatability, and overall taste preferences were organoleptically evaluated. While it is expected that the degree of perceived saltiness would be proportional to actual salt concentrations, it is interesting that this perception of saltiness also increases with an increase in the total nitrogen content; however, this is readily understandable as this relates to an increase in the free amino acid content. Palatability also increases with decreases in salt concentrations, but the greatest palatability was correlated with a salt concentra- tion of 16%.

    The interrelationships among salt, sugar, and organic acid in shoyu were also investgated. The 27 different test samples were prepared by adding (1) salt, 1 or 2%, (2) glucose, 0.5 or 1.0%, and (3) sodium lactate, 90 or 180 meq/liter to a shoyu which contained total nitrogen 1.535%, sodium chloride 15.5%, reducing sugar 4.0%, and organic acid (90% of which is lactic acid) 150 meq/liter. The degree of saltiness, sweetness, Sourness, and overall taste preferences were determined by a series of sensory tests. At a 4.5% sugar level, the degree of perceived saltiness increased with an increase in organic acid content. The shoyu was rated as saltier with an increase in the sugar content when the salt concentra- tion was about 15%, but as less salty with an increase in the sugar content when the salt concentration was about 18%. An increased salty taste was also reported with an increase in organic acid concentration from 105 to 375 meq/liter when the sugar content was less than 4.25% and the salt content was 16.5%. Shoyu was judged to be more sour in taste with higher salt concentrations given a 4.5% sugar content. The highest preference rating was awarded to shoyu with 16% salt, 4.25% sugar, and 240 meqjliter organic acid. But irrespective of the sugar level (3.5, 4.5, and 5.5%, respectively), preference in taste increased with increased amounts of organic acid.


    The results of multivariate analysis indicated that both a harmonious balance of various taste components, such as salty, acidic, sweet, bitter, and delicious, and a good aroma are important to an organoleptically preferred shoyu. Accord- ing to Mori (1979), 4 out of 12 factors which contribute 60% to judgments about the quality and taste of shoyu, in decreasing order of importance, are as follows: ( 1) nitrogenous constituents, such as amino nitrogen, formyl nitrogen, glycine, total nitrogen, and glutamic acid; (2) sugar constituents, such as extract, total sugars, glucose, and reducing sugar; (3) potency of delicious taste; salty, acidic, and bitter taste in positive direction, and sweet and delicious taste in a negative direction; and (4) taste factors relative to lactic acid fermentation and others; lactic acid, acetic acid, and ammonium nitrogen contribute in a positive direc- tion, whereas malic acid and citric acid contribute in a negative direction.

    Since a linear correlation was found between the sensory test used in assessing shoyu flavor and the gas-liquid chromatographic (GLC) data from a stepwise multiple regression analysis, an effort was made to use the GLC data to correlate an objective evaluation of shoyu flavor. The multiple correlation coefficient (r) increased with higher step numbers, exceeded 0.9 at step 10, and reached 0.968 at the last step, 43. The standard error of estimate reached a minimum value at step 28 and then increased gradually. The regression model most predictive of the test panels sensory ratings was calculated for each step, and the resulting calculated models were tested by substituting the GLC data. The results indicated that GLC data provide a reliable estimate of quality ratings obtained by subjec- tive sensory tests (Aishima et al., 1976, 1977).

    Next, the contributing proportions of all the peaks of a GLC pattern were calculated to determine the importance of each peak for the whole aroma of shoyu. In one study, eight principal components were identified from 39 GLC peaks as significant factors in shoyu aroma, contributing a cumulative proportion of 87% to the total variance. The second peak contributed the greatest propor- tion, 57.6% (Aishima, 1979a-c).


    More than 20 Japanese investigators had isolated about 130 flavor compounds from fermented shoyu by the time Goto first introduced the gas chromatography mass spectrometry (GCMS) method into this area of research in 1973, adding six new volatile compounds using this method. Instrumental analysis, which in- volved using gas or liquid chromatography with ultraviolet (UV), infrared (IR), nuclear magnetic resonance (NMR), and mass spectrometry (MS), rapidly in- creased the number of volatile flavor constituents isolated from shoyu. Nearly 300 kinds of such compounds have been identified to date as contributors to the


    fragrance of koikuchi shoyu. These include 37 hydrocarbons, 32 alcohols, 41 esters, 15 aldehydes, 4 acetals, 19 ketones, 24 acids, 17 phenols, 16 furans, 8 lactones, 6 furanones, 5 pyrones, 27 pyrazines, 7 pyridines, 6 other nitrogenous compounds, 16 sulfur-containing compounds, 4 thiazoles, 3 terpenes, and 3 others (Yokotsuka, 1953a, 1975; Asao et al., 1958a,b, 1967; Sasaki, 1975; Asao and Yokotsuka, 1977; Sasaki and numomura, 1978, 1981; Nunomura et al., 1976a,b, 1977a,b, 1978, 1979, 1980; Nunomura and Sasaki 1981, 1982; Yokot- suka er al., 1980).

    Solvent extraction or steam or vacuum distillation of shoyu or shoyu cake (the residue from the pressing of shoyu mash) was applied to the concentration of volatile flavor constituents of shoyu in the past when a fairly large number of isolated flavor compounds was necessary for the determination of their chemical structures.

    Sasaki (1975) investigated methods of obtaining a volatile flavor concentrate of shoyu which is most similar to that identified by a sensory evaluation of a given shoyu. The best way, he concluded, is by distilling the shoyu under vacuum of 15 mm Hg at 40C, collecting the distillate through successive cold traps consisting of a mixture of ice and sodium chloride, dry ice, and ethanol, and liquid nitrogen, extracting the distillate with dichloromethane, and then evaporating the solvent.

    Three popular brands of koikuchi shoyu, A, B, and C, obtained on the Ja- panese market, were treated by this method. Fifty milliliters of each were divided into the volatile flavor concentrates (40 ml) and the distillate and were filled to 50 ml with distilled water. Three kinds of original shoyu, the volatile flavor concen- trates, and the residue, and nine mixtures of each flavor concentrate and residue were subjected to a sensory evaluation of volatile flavors by a ranking Hedonic method (see Table XXVIII). The average scores of samples A and B were not markedly different from each other, but sample C was judged to be distinctly inferior to both A and B. The relative rankings of the flavor concentrates a, b, and c closely paralleled those of the original shoyus, but minor differences were observed among the three distillates, a, b, and c. The flavor concentrate c, obtained from the sample shoyu receiving the lowest score (C), was also ranked last when added to the distillation residues a, b, or c. These findings suggest that volatile flavor concentrates prepared in this way are nearly identical to the volatile flavors of the original shoyus. Further evidence is the importance of a shoyus distillate to its ultimater favor, lending support to the validity of deter- mining the quality of ones preference for a particular shoyu by its aroma based organoleptically on checking its volatile odor.

    Japanese researchers of shoyu flavor have endeavored to identify the fol- lowing:

    1. As many volatile flavor constituents as possible 2. The nature of those compounds which impart flavor to fermented shoyu



    4. 5.

    6. 7.


    Ranking Code no.6 I 2 3

    I B 1.36c

    2 a 1.46

    3 a' 1.75

    4 a a'

    I .44 5 a

    b' 1 . 1 1

    6 b C'

    I .33 7 a

    b' I .27

    A 1.64

    b 1.53 b'

    2.06 b b'

    1.66 b b'

    1.88 a C'

    1.67 a a'


    C 3.00

    3.00 C'

    2. I9



    b' 3.00


    3.00 b C'






    a From Sasaki (1975). bCode 1 : Shoyu (A, B, C) (samples are koi-

    kuchi heated shoyu); code 2: distillate (a, b, c); code 3: residue (a', b', c'); codes 4-6: distillate + residue; code 7: comparison among the highest in code numbers.

    Average of ranking.

    Some of the chemical flavor constituents which improve the flavor of fermented shoyu as well as those which have a negative effect The preferred combination of flavor constituents The differences between the flavor constituents of raw and pasteurized shoyu The stability of volatile flavor compounds in shoyu The differences between the volatile flavor constituents of the chemical hydrolysate of plant protein and fermented shoyu

    The most important component of the flavor of fermented shoyu seems to exist

    1. When the volatile fraction of a shoyu is further fractionated into its func-

    in its weak acidic fraction, which is recognized in the following ways:

    tional groups, the strongest flavor is observed in its phenolic fraction.


    2. When a shoyu is neutralized with alkali, its flavor immediately disappears and does not return in full strength when it is acidified.

    3. At lower pH value, i.e., within the range of 4.6-5.0, sensory tests of shoyu flavor yield better ratings.

    4. Some of the most important flavor compounds, such as maltol and 4- hydroxyfuranones, are in weak acidic fraction. These were isolated from the peaks of gas chromatography. The flavor characteristic of fermented shoyu was strongest among all of the peaks obtained.

    5. Some other isolated flavor ingredients, such as phenols, lactones, cyclothene, and phenol esters, seem to be essential to the flavor of fermented shoyu.


    The organic acids found in shoyu are presented in Tables XXIX and XL. The pasteurized shoyu (I) in Table XXIX appears to contain more levulinic and formic acid and less succinic acid than does (11), a genuine fermented shoyu, while (I) is blended with a fairly large amount of a chemical hydrolyzed of plant protein, usually from defatted soybean.

    Most of the organic acids found in shoyu have fairly high threshold values, but


    Unpasteurized Chemical hydrolysate Pasteurized Pasteurized Organic acid shoyu of plant protein shoyu (I) shoyu (11)

    n-Butyric Isobutyric Unknown Propionic Levulinic Acetic PyNViC Formic a-Ketobutyric Lactic Succinic Pyroglutamic Glycolic Malic Citric


    0.1 0.5 0.9 5.2

    25.4 87.9 36.5 2.3 1.9

    887.8 25.2 30.0 8.8

    Trace Trace 11 12.5

    3 .O 0.8



    194.6 0.7





    13.8 4.2 3.3

    25.3 1191.4

    1.4 2.4 0.3

    13.0 237.6 134.0 10.5 53.1

    1 .o 852.6 21.2 44.3 4.4 1.7 8.8


    0.5 -

    - 4.0 4.4

    126.2 11.9 6.2 0.2

    1156.6 49.8

    110.6 9.9

    Trace Trace 1480.3

    From Ueda et al. (1958).


    that of isovaleric acid is relatively low-0.7 mglliter in water (Stahl, 1973)- and its high content is derived from the contamination of Bacillus nutto during koji cultivation. Adding cinnamic acid to shoyu gives it a foul odor, the result of bacterial contamination in shoyu mash; however, this compound has not as yet been isolated from shoyu. It is difficult to estimate the contribution of each organic acid to the flavor because, as Salo et al. (1972) pointed out, they react synergistically with each other.

    a-Ketobutyric acid is sometimes isolated from the chromatographic fraction of shoyu having a strong shoyu-like aroma, and the chemically synthesized a- ketobutyric acid also has a strong aroma reminiscent of a very important compo- nent of shoyu flavor. This compound was found in the chemical hydrolysate of a protein containing threonine (Wieland and Wiegandt, 1955) and was recognized as an important flavor component (Blockman and Frank, 1955). Sulser et al. (1967) reported that newly synthesized a-ketobutyric acid has no taste or odor, and that a-hydroxy-p-methyl-A,a,P-hexenolactone (an isomer of a-keto-P- methyl-y-carprolactone which is produced by the condensation of two molecules of a-ketobutyric acid, followed by cyclization and decarboxylation) has a very strong odor, characteristic of chemical protein hydrolysate. The threshold value of a-ketobutyric acid was reported to be 0.04 ppm; its odor easily attaches to skin and clothes, and its taste remains for several hours.


    The kinds of alcohols isolated from shoyu and their relative quantities are indicated in Tables XXXIX and Table XL, respectively. The ethanol content of shoyu ranges from 1 to 3.5% (v/v). The important alcohols, because of their relative threshold values and high proportion, are n- and isobutyl alcohol, iso- amyl alcohol, 2-phenyl alcohol, and furfuryl alcohol. The latter three are found in typical alcoholic beverages and resemble yeasts in their metabolic action and synthesis of amino acids. The so-called fuse1 alcohols are mainly derived from the fermentation of hexoses and, to a lesser degree, from the degradation of the corresponding amino acids following Ehrlichs pathway (Webb et al., 1969).

    Since the amount of furfuryl alcohol increases during pasteurization of shoyu in the final processing, its quantity indicates the degree of pasteurization. Hexyl alcohol is derived from raw soybeans (Nakajima and Takei, 1949), and meth- ylnonylcarbinol is derived from soybean oil (Shoji, 1936).

    The differences in the contents of the major flavor ingredients and various alcohols of eight brands of shoyu selected from the Japanese market in 1980 are indicated in Table XXX (Nunomura and Sasaki, 1981); their relative proportions in four brands are compared schematically in Fig. 20 (Sasaki, 1975).



    Peak no. Compound Content (ppm)

    1. 2. 3. 4. 5. 6. 7. 8. 9.

    10. 11. 12.

    2-Methyl-I -propano1 (isobutyl alcohol) 3.07- 18.35

    3-Methyl-] -butanol (isoamyl alcohol) 4.47-22.45 I-Butanol (n-butyl alcohol) 1.41- 11.48

    3-Hydroxy-2-butanone (acetoin) 5.05-8.44 Ethyl 2-hydroxypropionate (ethyl lactate) 7.35-27.12 Furfuryl alcohol 4.35-10.07 3-(Methy1thio)- I-propanol (methionol) 2.60-4.47 2-Phenylethanol 3.71 - 10.25

    4-Ethyl-2-methyoxyphenol (4-ethylguaiacol) (4-EG) 1.12-3.67 4-Hydroxy-2(or 5)-ethyl-5(or 2)-methyl-3(2H)-furanone (HEMF) 177.78-418.67

    4-Hydroxy-2,5-dimethyl-3(2H)-furanone (HDMF) 1.83-5.39

    4-Hydroxy-5-methyl-3(2H)-furanone (HMMF) 84.54- 1 53.58

    From Nunomura and Sasaki (1981), unpublished. Kikkoman Corporation, Japan.

    ' 8

    Koikuchi A Koikuchi B

    Koikuchi C Usukuchi

    FIG. 20. Contents of major flavorous ingredients in four brands of shoyu in Japan as compared to (A) as the standard. ( I ) , Isobutyl alcohol; (2). n-butyl alcohol; (3). isoamyl alcohol; (4). acetoin; (3, ethyl lactate; (6), furfuryl alcohol; (7), methionol; (8). 2-phenylethanol; (9). 4-ethylguaiacol. From Sasaki (1975).



    Pasteurized shoyu Pasteurized shoyu Compound (A) (B)

    Ethyl acetate I .3 I .80 Ethyl propionate 0.14 - Ethyl isovalerate 0.08 - Butyl acetate 0.31 Trace Isoamyl acetate 0.06 Trace Ethyl lactate 3.60 3.90

    Ethyl levulinate 3.36 1.30 Ethyl benzoate 3.00 0.99 Ethyl succinate 1.89 0.34 Ethyl maleate 1.44 Trace

    Ethyl malonate 4.00 I .94

    a From Morimoto and Murakami (1966). J . Ferment. Tech- nol. 44(8), 461-475.


    Forty various types of esters have been isolated from shoyu; the results of qualitative analyses of 11 major ones are listed in Table XXXI. Almost all, except from butyl acetate and isoamyl acetate, are ethyl esters, but the existence of many others, produced by the combination of alcohols and organic acids found in shoyu, is also considered. N-, and isobutyl benzoate or vanillate have been found to be more important to shoyu flavor than either ethyl benzoate or ethyl vanillate. The presence of different amounts of these esters lent a dis- tinctive character to different kinds of shoyu (Yokotsuka, 1975). Since ethyl levulinate is produced by the metabolism of yeasts in the shoyu mash to which the chemical hydrolysate of defatted soybean is blended, the amount of ethyl levulinate is an indicator of the amount of chemical hydrolysate blended to genuine shoyu mash.


    I . Maillard Reaction and Strecker Degradation

    It is generally acknowledged that heating raw shoyu increases the proportion of acetaldehyde and other lower aldehydes. The so-called caramel reaction only takes place in the course of heating sugars, but this reaction is accelerated in the presence of amino residues, which is generally known as the Maillard reaction.



    Amino acid Corresponding aldehyde

    Glycine CH2(NH2)COOH a- Alanine CH~CH~(NHZ)COOH a-Aminobutyric acid CH3CH2CHz(NHz)COOH Valine (CH~)~CHCH~CH(NHZ)COOH Leucine (CH3)2CHCH2CH(NH2)COOH Isoleucine CH~CH~CH(CH~)CH(NHZ)COOH Serine HOCH2CH(NHz)COOH Threonine CH,CH(OH)CH( NH2)COOH Methionine CH~SCH~CHZCH(NHZ)COOH Cysteine HSCHzCH(NH2)COOH Cystine (-SCH2CH(NHz)COOH)z Glutamic acid HOOCCHzCHzCH(NH2)COOH Phenylglycine C6HsCH( NH2)COOH Phenylalanine C~HSCH~CH(NH~)COOH

    Formaldehyde HCHO Acetaldehyde CH3CHO Propionaldehyde CH3CHzCHO Isobutyraldehyde (CH3)zCHCHO Isovaleraldehyde (CH3)2CHCHzCHO 2-Methylbutyraldehyde CH3CH2CH(CH3)CH0 Glycol aldehyde HOCHzCHO Lactic aldehyde CH,CH(OH)CHO Methional CH~SCH~CHZCHO Mercaptoacetaldehyde HSCHzCHO Dithioacetaldehyde OHCCHzSSCH2CHO Succinic monoaldehyde HOOCCHzCH2CHO

    Phenylacetaldehyde C6HSCH2CH0 Benzaldehyde C~HSCHO

    aFrom Yokotsuka (1975).

    a-Dicarbonyl compounds which are produced by this reaction degradate into one carbon-less aldehydes or ketones by Strecker degradation (see Table XXXII). These compounds further degradate into other odorous compounds: acetaldehyde and propionaldehyde from lactic aldehyde, dimethyl sulfide from methional, and sulfur dioxide from dithioacetaldehyde, for example. These compounds are clearly rich in fermented shoyu which contains a large amount of free amino acids.

    2. a-Dicarbonyl Compounds

    Diacetyl, CH,COCOCH,, is produced by the oxidation of acetoin, CH,CH (OH)COCH,CH, (Yamada, 1928, 19291, and acetylpropionyl, CH,COCOCH, CH,, is derived from acetylethylcarbinol, CH,CH(OH)COCH,CH, (Asao and Yokotsuka, 1963). However, acetylbutyryl, CH,COCOCH,CH,CH,, which was first isolated from shoyu (Asao and Yokotsuka, 1961a), was reported to be produced by heating C7H,,05, a lactol compound isolated from shoyu (Asao and Yokotsuka, 1961b; Yokotsuka and Asao, 1961). The 3% sodium chloride-soluble part of the chloroform extract of unpasteurized shoyu was adjusted to pH 7.2 and then extracted with ether. From this ether extract, 4-ethylguaiacol and tyrosol were isolated, but the residue still had a very strong shoyu-like flavor. It was then subjected to column chromatography of aluminum oxide, and the crude crystals


    obtained were purified by sublimation until the mp was 114-1 15C. The yield of the crystals was 20 mg from 40 liters of shoyu.

    By elementary analysis and molecular weight, the molecular formula C,H,,O, was derived. The oxidation of this compound by periodate yielded a-ketobutyric acid, pyruvic acid, propionaldehyde, and carbon dioxide. Oxidation by potassium permanganate yielded acetic acid and propionic acid, and when oxidized by chromic acid, it was suggested that the sample has one - C H , and one -CH,CH, at the end of the structure. The sample was very unstable at room temperature and decomposed into propionaldehyde, pyruvic acid, acetaldehyde, carbon dioxide, a-hydroxybutyric acid, and acetylbutyryl. When heated with diluted sulfuric acid, the sample decomposed into 1 mol carbon dioxide, 1 mol a-hydroxybutyric acid, and acetic acid, yielding a good amount of acetylbutyryl. The acetylbutyryl was identified as 2,4-dinitrophenylhydrazone (mp 242-243"C), which was compared with the authentic compound. The IR spectrum of the sample indicated the existence of a lactol linkage and OH in its structure. Based on these experimental results, the chemical structure (A) was tentatively assigned to the C,H,,O, compound, but it was later corrected to be (B) by Nunomura et al. (1976).

    (A) (B )

    Acetylbutyrl lends a fruity fragrance, which is different from diacetyl, and is detected by a sensory test at the concentration of (l:1Op7). However, the fra- grance of acetylpropionyl resembles that of fermented rice wine. The total a- diketone compounds increase from 0.05-0.1 to 0.2-0.3 mg% during the pasteurization of shoyu conducted at 80C for 5 hr. This increase was calculated to be much greater than that produced by the degradation of the C,H,,O, com- pound contained in shoyu. The total a-diketon compounds in pasteurized shoyu was first isolated by steam distillation and then converted into dioxime deriva- tives, which were fractionated through column chromatography containing Dowex 1 .X-8 of borate type into three peaks, P1, P2, and P3. These three peaks were purified by sublimation in a vacuum, followed by rechromatography and recrystallization. Three kinds of purified crystal thus obtained were compared with authentic samples with respect to their melting points, their IR spectra, and so on. P1, P2, and P3 were identified as dioximes of diacetyl, acetylpropionyl, and acetylbutyryl, respectively. Their proportions in shoyu were found to be in the ratio of 100:20:3. The total glyoxal content, including glyoxal and meth-


    ylglyoxal, also increased during the pasteurization of shoyu and was equivalent to about one-third of the total diketone compounds.

    Asao and Yokotsuka (1963) investigated the formation of these a-diketone compounds in shoyu mash. They reported that small amounts of methylglyoxal and diacetyl were produced by the cooking of soybeans, and that some increase in diacetyl and in trace amounts of acetylpropionyl occurred during koji cultivation. Pediococcus hulophylus did not affect the formation of dicarbonyl compounds, but their presence doubled with the fermentation of mash by S. rouxii. However, the increase in a-diketone compounds during the heating of shoyu was positively correlated with the amount of reducing sugar present in the shoyu. It was demonstrated that glyoxal and methylglyoxal are produced from xylose in the course of heating shoyu, while acetylpropionyl is oxidatively produced from acetylethylcarbinol, which was identified in shoyu. The other precursor of acetyl- propionyl corresponding to the C,H120, compound, which is the precursor of acetylbutyryl, could not be detected.

    The formation of acetylmethylcarbinol and acetylethylcarbinol in shoyu mash takes place a little earlier than that of alcohol, and both correspond to a decrease in the amount of reducing sugar in mash which results from yeast fermentation. The researchers also noted the conversion of acetylmethylcarbinol and of acetyl- ethylcarbinol from pyruvic acid and a-ketobutyric acid, respectively, by the action of yeasts.

    Saccharomyces rouxii demonstrated the strongest conversion ability among the yeasts belonging to the Succharomyces. Also noted was an inverse correlation between the ability of different yeasts to produce alcohol and to produce acetyl- ethylcarbinol. The good producers of alcohol, such as Sacchuromyces cereviciae, had the tendency to be poor producers of acetylethylcarbinol; the good producers of acetylethylcarbinol, such as S. rouxii, were poor producers of alcohol. a - Diketone compounds and glyoxal consisting of from four to eight carbons were chemically synthesized, and their contributions to the flavor constituents of some fermented foods were organoleptically evaluated. a-Diketones exhibited a flavor resembling shoyu and rice wine, while glyoxals resembled vinegar when detected at a concentration of 10-6-10-8 (Yokotsuka and Asao, 1961).

    Various investigators have reported the isolation of some dicarbonyl com- pounds of tentative chemical structures from the steam distillate or the distillate of shoyu: C8H,,02 (Ikeda and Kawaguchi, 1922), C,H,O, and C,H,02 (Kodama, 1922), and C,H1,02 (Nakaiima and Takei, 1949). Yokotsuka and Asao (1961) pointed out the close resemblance of these compounds to acetyl- propinoyl or acetylbutyryl with regard to the characteristic yellow color of the liquid sample, the results of chemical analyses and the melting by silver oxide, the color reaction with 2,4,-dinitrophenylhydrazine, and their absorption of Br2 (see Table XXXIII).


    3. y-Pyrones

    Kihara (1940) first isolated maltol (1) from the chemical hydrolysate of defat- ted soybean, then 20 mg crude crystals of maltol were isolated from 2 liters of shoyu (1983). Maltol has been known as a characteristic flavor constituent of malt (Brand, 1894). Kihara found that maltol exists in soybeans in a conjugated form with polysaccharide, from which maltol is separated during the heating of soybeans or shoyu.

    Maltol is produced by the caramelization of maltose, or the sugars containing maltose, and only in minimal amounts from glucose and starch (Baker et al., 1953; Diemer and Hara, 1959). The formation of maltol is promoted in the presence of amino radicals at lower temperatures in neutral conditions (Patton, 1950; Hodge et al., 1963). Maltol is a typical caramel flavor compound and synergistically enhances sweetness at a concentration of 30-250 ppm. Maltol also synergistically enhances the flavor of vanillin, glutamic acid, and some other amino acids (Hayashi and Kawase, 1970). It is a typical cyclicenolone, having an enolic radical in its molecule and weak acidity. It0 (1972) observed that the aromatic flavor of the weak acidic fraction of foods in general is often due to maltol. Ethylmaltol (2) has not been isolated yet from nature, although it has four to six times the flavor intensity of maltol. Isopropylmaltol (3) was reported to have a shoyu-like flavor, but it has not been isolated from shoyu. 5- Hydroxymaltol (4) has a weak maltol-like flavor and was isolated from roasted barley (Shimizu et al., 1970) and from shoyu (Nunomura et al., 1980). 3- Methoxy-2-methyl-4H-pyran-4-one (5) was isolated from shoyu, but has no aroma (Nunomura et a f . , 1980).

    4. 4-Hydroxy-3-jiiranones and the Related Compounds

    a. Isolation of HEMF. The flavor concentrate from the chloroform extract of unpasteurized shoyu was directly subjected to gas chromatography (Nuno- mura et al., 1976b). The results are shown in Fig. 21. Each peak that was fractionated by gas chromatography equipped with a TCD detector was also subjected to a sensory test of its aroma. Peaks 36-39, and especially no. 39, had the aroma most resembling that of shoyu. GCMS analyses proved that peak 36 was 2-phenylethanol, and peak 37 was a mixture of three compounds: 142- pyrroly1)- 1-ethanone (2-acetylpyrrole) (odorless), maltol, and 3-methoxy-2- methyl-4H-pyran-4-one (odorless) (Nunomura et a f . , 1980). One example of gas chromatograms of shoyu flavor concentrate by GCMS is shown in Fig. 22. The


    12 23



    1 34 30 1


    0 1 0 2 0 30 40 5 0 60 rnin

    FIG. 21. Gas chromatograph of s h o p flavor concentrate. Instrument: Shimazu 4BM-PF(FDI). Conditions: Column: FFAP 1096, 3 mm i.d. x 2 m (glass). Injector: 240C; detector: 240C. Column oven: 50-180C (3Wmin). Carrier gas: N2, 30 ml/min. From Nunomura et al. (1976).

    major constituent of peak 39 was first isolated by distillation in a vacuum from 760 liters of shoyu, then concentrated by extraction with CH2C12, followed by fractionation and purification through gas chromatography, which gave 176 mg of oil substance. This compound, which was chemically synthesized by Lucian0 Re et al. in 1973, was designated 4-hydroxy-2(or S)-ethyl-5(or 2)-methyl-3(W) furanone (HEMF) as the result of UV, proton magnetic resonance (PMR), C- NMR, and high-resolution MS determinations (Nunomura ef al., 1976b, 1980). HEMF was first isolated from fermented shoyu, but was not detected in the chemical hydrolysate of plant protein. It exists in the form of a tautomer, (A):(B) = 3:2, which was determined by PMR as follows:



    (A) (6)

    4 - H y d rox y - 2( or 5)- e t h y I - 5( or 2) - methyl-3(2 H)-furanone (HEMF)

  • 'i

    Y 7 a l A 1 0 1 0 30 LO 50 60 70 8 0 90 t o o 1 1 1 "

    FIG. 22. Gas chromatograph (capillary column) of shoyu flavor concentrate (GCMS). Instrument: RMU-6MG. Column: FFAP, glass, 0.25 mm i.d. x 30 m. Oven temperature: 60-220C, 2"C/min. He: 0.2 kg/cm2. Ionizing voltage: 20 eV. Ion source temperature: 200C From Nunomura et al. (1980).


    HEMF seems to be the most important flavor ingredient and characteristic component of fermented shoyu in view of its high proportion (about 100-200 ppm) and its very low threshold value (less than 0.04 ppb) in water (Ohloff, 1978). Adding 0.01 ppm of HEMF to shoyu is very effective in ameliorating shoyu's otherwise salty taste.

    b. OX-HEMF. HEMF is quite stable in shoyu, but is unstable in alkali and acid. Under the basic condition, it changes into the odorless compound 4,4,5- trihydroxy-2-ethyl(or methyl)-5-methyl(or ethyl)-3-tetrahydrofuranone (OX- HEMF) .


    The IR spectrum of OX-HEMF coincided with that of a very unstable sub- limatic compound, C7H,205, which was isolated from a weak acidic fraction of unpasteurized shoyu and tentatively identified as 2-furanone, the structure of which was given previously (Asao and Yokotsuka, 1961a,b; Yokotsuka and Asao, 1961). It is presumed that the conversion of HEMF into OX-HEMF occurred in the course of the alkali treatment of the chloroform extract of un- pasteurized shoyu with 5% Na,CO,. OX-HEMF degradates by heating or by autooxidation into such compounds as acetylbutyryl (2,3-hexandione), a-ket- obutyric acid, a-ketopropionic acid, acetaldehyde, and other compounds, as shown in Fig. 23.

    c . HDMF, HMMF, and the Other 4-Hydroxy-J(2H)-furanones. The quan- tity of 4-hedroxy-2,5-dimethyl-3(2H)-furanone (HDMF) in shoyu was reported to be about 10 ppm (Nunomura et al., 1980), with a threshold value of 0.04 ppb in water (Ohloff, 1976). This compound was first isolated from pineapple and was reported to have a pineapple-like flavor, having a threshold value of 0.1-0.2 ppm (Rodin et al., 1965).

    The quantity of 4-hydroxy-5-methyl-3(2H)-furanone (HMMF) (Nunomura et al., 1979) is small, but increases when shoyu is heated, reaching about 200 ppm. HMMF was reported to have a caramel flavor similar to roasted chestnuts. HEMF, HMMF, and HDMF resemble each other in chemical structure, but have different patterns of development. HEMF is produced by the yeast fermentation of shoyu mash, while HMMF and HDMF are typical browning compounds.

  • 276

    (OX -HEMF 1

    -. Autoxidat ion C2H5COCOOH






    -. In Shoyu : Stable


    Under basic condition r


    FIG. 23. Oxidation of process of HEMF. From Nunomura et al. (1976b).

    A total of 257 strains of yeast isolated from 1 1 brewing houses of Kikkoman Company were cultured at 30C for 40 days in an aseptic filtrate of shoyu mash fermented for 25 days at pH 4.8 and adjusted with lactic acid in a flask with shaking once a day. All the strains tested produced HEMF, on the average 129.6 ppm, with ranges from trace to 28.4 ppm (Sasaki et al., 1984).

    (HEMF) (HDMF) (HMMF)

    d. Other Similar Compounds. In 1970, the Ajinomoto Company got a Japanese patent for improving the flavor of foods and condiments by adding 4- hydroxyJ-methyl-2,3-dihydrofuran-3-0ne (6). This compound was purified by sublimation from the heated product of sugar and amino acids, and was reported to form colorless, needle-shaped cyrstals, mp 126- 127"C, to have maltol-like flavor, and to turn dark blue in color when mixed with FeCl,.


    Yokotsuka (1958) isolated the compound C,H603 from a weak acidic fraction of ether extract of shoyu with a yield of 10 mg from 10 liters of shoyu. It formed needle-shaped crystals, with an mp of 126-128"C, was sublimatic, had a strong ricecracker-like flavor, and turned a dark green color when mixed with FeCl,. These two compounds would seem to be identical judging from the descriptions given above; however, Sulser et al. (1967) presumes that the compound C,H60, has the structure of a-keto-y-valerolactone (7).

    Nunomura et al. (1980) isolated 2-methyl-3-tetrahydrofuranone (8) from shoyu, which had been previously isolated from coffee by Gianturco et al. (1964). with no description of its aroma. Isomaltol (9) has also been isolated from shoyu and has the fragrance of burnt sugar (Nunomura et al., 1980).

    5 . Alkylcyclopentadiones

    Several kinds of alkylcyclopentadiones, such as 2-hydroxy-3-methyl-2-cyclo- pentene- 1 -one (lo), 3-ethyl-2-hydroxy-2-cyclopentene- 1 -one (1 1), 2-hydroxy- 3,4-dimethyl-2-cyclopentene-l-one (12), and 3-hydroxy-3,5 ,-dimethyl-2-cyclo- pentene- 1 -one (13), have been isolated from heat-treated foodstuffs, such as coffee beans, and the compounds of fructose degradated by heat (Gianturco et al., 1963, 1964; Gianturco and Friedel, 1963). Gyclotene is only one of the compounds belonging to this group which was isolated from shoyu (Nunomura et al., 1980). All of these alklycyclopentadiones were reported to have a flavor similar to caramel, roasted sugar, or maple syrup.


    0 0 0 0

    6. Acetals

    Yokotsuka (1950) identified a great amount of isovaler-aldehyde-diethylace- tal, (CH,),CHCH,CH(OC,H,),, in the steam distillate of shoyu or shoyu cake (the press residue of shoyu mash). Also identified in the same distillate, but with less certainty, were a-hydroxyisocaproaldehyde-diethylacetal, (CH,),CHCH,


    CHOHCH(OC,H,),, and/or a-ketoisoaldehyde-diethylacetal, (CH3),CHCH, COCH(OC,H,),. Inasmuch as these compounds were not identified in the ether extract of shoyu, and since acetals in general are unstable in an acidic condition like that of shoyu, these acetals were presumed to be synthesized in the stream of shoyu and to constitute an important part of the flavor of shoyu vapor while cooking. On the other hand, chemically synthesized n-, and isobutyracetal and isovaleracetal were claimed to have as important a volatile flavor as is organolep- tically detectable in shoyu or rice wine.

    Fujita (1960, 1961) reported that the diethylacetals of phenylglyoxal, C,H, COCH(CO,H,),-benzylglyoxal, C,H,CH,COCH(OC,H,),, methionylgly- oxal, CH,SCH,CH,COCH(OC,H,),, and sec-butylglyoxal, (CH,),CHCOCH (OC,H,),-exhibited the characteristic aroma of shoyu.

    Yoshida et a f . (1980) analyzed the topnote of aroma concentrate from shoyu and identified ethanol, ethyl acetate, isobutyraldehyde, the diethylacetals of these aldehydes, isoamyl alcohol, and a trace amount of dimethyl sulfide.


    4-Ethylguaiacol(4EG) (Yokotsuka, 1953) and p-ethylphenol (Asa and Yokot- suka, 1958) are important flavor ingredients of shoyu belonging to its weak acidic fraction, which had been isolated from shoyu prior to the finding of other weak acidic constituents of flavor. The formation of phenolic compounds in shoyu production has been studied (Yokotsuka et al., 1967a,b; Asao et al . , 1967, 1969). These phenolic compounds were reported to derive for the most part from wheat. The phenolic fraction of wheat was observed to increase in the course of roasting, and vanillin, ferulic acid, and vanillic acid were identified as the major constituents of this phenolic fraction. Shakuchirin (Sahia and Shaw, 1961), which was found in the seed leaves of wheat and in coniferyl alcohol as a part of its lignin structure, were identified as the possible precursors of these phenolic compounds (see Fig. 24). The formation of vanillin and vanillic acid from a part of ferulic acid and the formation of p-hydroxycinnamic acid and its conversion into p-hydroxybenzoic acid have been observed during the growth of Aspergillus molds during koji cultivation (Asao and Yokotsuka, 1958). The greatest amount of phenolic was reported after the first 24 hr of the 72-hr period of koji cultivation, which coincided with the maximum mycelial growth of koji mold. The major constituent of the phenolic fraction of koji was identified as ferulic acid. Ferulic acid and p-hydroxycinnamic acid are metabolized into 4EG andp-ethylphenol, respectively, in the latter half of the period of yeast fermenta- tion of shoyu mash (moromi) by the action of a Candida (Torulopsis) yeast, wuch as C . versatillis or C . etchellsii, and not by S . rouxii, which is generally considered to be the predominant yeast in shoyu fermentation. The various kinds


    Raw materials I ocn,

    on c n , o c o c n r c n f i o H

    Coml*ryl Shakuchwm alcohol

    Koji making A i p c i g d l u ~ so@e

    P-Coumarn: 4-Hydoxy- Esters of Fsrulic acid Esters of Vanillic acid Vanillin Vanillic acid acid benmic acid Bsnzoic acid

    E l l . , * 01 4-Ethyl- 4-Elhylphanol Eslarr o f 4-Elhylguamcol guamacol 4-Ethylphenol

    FIG. 24. brackets have not been identified. From Yokotsuka et al. (1967a,b); Asao et al. (1967).

    Formation of alkylphenols during the manufacturing process of shoyu. Structures in

    of yeast that convert ferulic acid into 4EG are listed in Table XXXIV. The yield of a fractional distillate with a bp of - 185C from the steam distillate of shoyu cake was small, but it had a strong shoyu flavor. Organic acids, phenols, car- bonyls, and sulfur-containing compounds were removed from this fraction, and the residue slightly hydrolyzed to obtain acetic acid, benzoic acid, 4EG, ethyl vanillate, and at least two kinds of unknown phenols by paper chromatography. These findings suggest that the unstable phenolic esters between phenols, such as 4EG and ethyl vanillate, and organic acids, such as acetic acid and benzoic acid, are present in shoyu and are the precursors of free phenolic compounds which increase during the pasteurization of shoyu. The quality of free phenols, includ- ing 4EG, doubled when shoyu was pasteurized at 80C for 5 hr.

    About 25% of the 50-70 samples of shoyu tested in 1964, 1965, and 1966 contained 0.5-2.0 ppm of 4EG. The organoleptically best 10 samples among 50 and only 1 sample among the remaining 40 contained 4EG (Yokotsuka et al., 1967a,b). Thus, 4EG is a very important ingredient of fermented shoyu, as the difference of 0.5% of 4EG in shoyu is easily detected by a sensory test and characterizes a given brand of shoyu. Moreover, it was observed that 4EG tasted




    Stage of fermentation Formation Assimilation of rnororni Strain of yeast Stock number of 4EG of nitrate




    Torulopsis famata Pichia farinosa Trichosporon behrendii Candida porimorpha Saccharomyces rouxii Saccharomyces rouxii

    var. halomembranis Saccharomyces acidifaciens Saccharomyces acidifaciens

    var. halomembranis Torulopsis halophylus Torulopsis nodaensis Torulopsis versatilis Torulopsis etchellsii Torulopsis anomala Torulopsis sake

    E29a A6



    s 9 R6


    E-7 NO. 210

    N-24, IOA-40 N-21, 29B-45 N552, 2C-5 15A-26, 19C-7 3B-42, 17C-28 5C-5, 228-2

    (I From Asao and Yokotsuka (1958).

    like fermented shoyu and ameliorated its salty taste. Noda and Nakano (1979) determined the quantity of 4EG in the three popular brands of koikuchi shoyu in Japan to be 1.0, 1.8, and 2.1 ppm, and in the three brands of usukuchi shoyu to be 0.5, 1.3, and 0.3 ppm, respectively. The yeast flora in 35 kinds of shoyu mash obtained in Hokkaido (northernmost island of Japan) was studied in 1960, and 4 was found that organoleptically good mashes contained large amounts of Candida (Torulopsis) etchellsii and C . versatilis (Sasaki et al . , 1964, 1966a,b; Yoshida, 1979). Among 257 strains of yeast isolated from shoyu mash, 17 produced 4.51 ppm of 4EG on an average, ranging from 0.31 to 8.99 ppm (Sasaki et al., 1984).


    In general, aliphatic lactones are important among the flavor ingredients of foods because of their strong and characteristic flavor. Many varieties and large amounts of y-IactQnes are found in animal foods and greatly contribute to the dairy flavor, for example. Many kinds of y-lactones are also present in vegeta-


    bles. Four kinds of y-lactones have been identified in Japanese fermented shoyu (Nunomura et al., 1980): 4-butanolide (y-butyrolactone) (14), 4-pentanolide (y- valerolactone) (13, 2-methyl-4-butanolide (16), and 2-pentene-4-olide (17). A schematic presentation of their chemical structures is presented here.

    (14) (15) (16) (17)

    There is a very small amount of 4-pentanolide in fermented shoyu, but a large amount in the chemical protein hydrolysate or its yeast-fermented product, which in Japan is called semichemical shoyu, as well as the ethyl levulinate, which is also a characteristic ingredient of semichemical shoyu (Nunomura et al., 1977a).

    Liardon and Phillipossian (1978, 1980) cultured koji A. oryzae with a mixture of cooked soybeans and wheat, and combined the koji with 18% saline water to make the mash, which was adjusted to pH 4.5 and then fermented with S. rouxii at 38-40C for 30 days to make shoyu. This process might be slightly different from the average shoyu produced in Japan in that the mash is not fermented with, but is fermented by yeasts at a very high temperature. From this shoyu, eight kinds of lactones were isolated, including the four previously cited (14-17), and four more recently isolated y-lactones: 4-hexanolide (y-caprolactone) (18), 2- methyl-2-buten-4-olide (19), 2-hydroxy-3,3-dimethyl-4-butanolide (20), and 5- hydroxy-4-hexanolide (21), presented here.

    All of these y-lactones have been widely identified in many foodstuffs, includ- ing black tea, cocoa, coffee, pineapple, tomato, peach, apricot, strawberry, plum, tobacco, fried onion, roasted peanut, beef tallow, lard, mushroom, and sherry wine.


    Pyrazines are typical components of the so-called browning flavors-corn, nut, and bread-and they play an important role in the flavor of heat-treated foodstuffs (Hodge, 1972).




    Concentration (mglliter) Ratio

    Compound Raw shoyu Pasteurized shoyu (heated/raw)

    2-Methylpyrazine 0.024 0.075 3.1

    Ethylmethylpyrazine 0.338 0.746 I .9 Trimethy lpyrazine 0.040 0.050 1.3

    Dimethy lpyrazine 0. I84 0.746 4. I

    From Nunomura er al. (1978).

    The greater part of pyrazines in foods is produced by the heat degradation of proteins and amino acids or by the chemical reactions between sugar and protein, although some are biosynthesized in plant tissues, such as 2-isobutyl-3-methox- ypyrazine in bell pepper (Buttery et al., 1969). Approximately 70 pyrazines have been identified in foods to date, but it is only since 1970 that the importance of pyrazines as food flavor ingredients has been generally recognized and utilized in the manufacture of artificial food flavorings. It is likely that shoyu contains many kinds of pyrazine compounds, but until recently, tetramethylpyrazine is the only one that has been isolated. Nunomura et al. (1978, 1980) identified 27 pyrazines in the basic fraction of shoyu by GCMS analysis. The flavor of pyrazines in shoyu is weakened by the weak acidic pH value of shoyu (4.7-4.9), but becomes dominant when the pH of shoyu is neutralized by dilution with water in cooking. When shoyu is heated, there is a substantial increase in the quantity of pyrazines (as indicated in Table XXXV), suggesting that they are one of the characteristic flavor components of pasteurized shoyu.


    When an aqueous solution of mercuric chloride (HgCI,) is added to shoyu, part of the characteristic volatile flavor disappears at once, perhaps evidence of the fact that some sulfur-containing compounds play an important role in the volatile flavor. Methionol (3-methylthio- 1-propanol), which was first isolated from shoyu by Akabori and Kaneko (1936), and methional (3-omethylthio-l- propanal), synthesized by these researchers (1937), are claimed to be important ingredients of shoyu flavor. Yokotsuka (1953) identified lower mercaptans and mercaptals in the steam distillate of shoyu cake. It is believed that these com- pounds are produced not only by fermentation, but by the heating of sulfur- containing compounds during distillation. It is generally known that methylmer-




    Sample HIS CH2SH (CH3)ZS

    Fermented shoyu 3.40 1.20 0.22

    Chemical hydrolysate of plant protein 5.30 4.70 44.60 Semichemical shoyub 3.10 I .90 4.40

    a From Ueno (1963), Report of Kikkoman Shoyu Co., Ltd., Vol. 5. The mixture of chemical hydrolysate of defatted soybean, soybean and

    wheat koji, and salt water is fermented with Succharomyces rouxii for I week at 30C in the presence of 18% salt.

    capto radicals, methyl mercaptan, and hydroxysulfide are produced by microbial metabolism or by the heat degradation of sulfur-containing compounds. 1 he chemical hydrolysate of plant protein contains more amounts of lower boiling sulfur compounds than does fermented shoyu, as shown in Table XXXVI (Ueno, 1963).

    Dimethyl sulfide, present in the chemical hydrolysate of defatted soybean, is produced by the degradation of methioninel methyl sulfonium, Me(CH,)SCH,- CH,CH(NH,)COOH, itself produced by the reaction between methionine and methyl chloride. Methyl chloride is decomposed from the methoxy group of soybeans by the action of HCI (Ogasawara, 1963). Guadagni etaf. (1963) reported threshold values of dimethyl sulfide and methyl mercaptan (methanthiol) to be 0.02 and 0.33 ppb, respectively.


    Several kinds of terpenes have been isolated from whiskey, brandy, ram, and fuse1 oil. It is interesting to note that borneol, bornyl acetate, and cis-rose oxide [4-methyl-2-( 2-methyl- 1 -propenyl)-tetrahydropyran] were isolated from shoyu (Nunomura et af., 1976a, 1979).


    Newly pasteurized fermented shoyu has a characteristic pleasant odor, most of which disappears in a short time by natural evaporation. Sasaki and Nunomura (1979) directly analyzed the topnote flavor concentrate of pasteurized shoyu by the GCMS method. The sample to be analyzed was prepared by passing helium




    Concentrations Coefficient Compounds (ppm) ( x , n = 10) of variation (96)

    Methanol 9.45 4.43 Acetaldehyde 3.76 9.58 Ethanol 5605.18 3.50 Propionaldehyde 1.70 8.52 Acetone 2.09 3.75 Ethyl formate 1.66 3.02

    Isobut yraldeh yde 6.38 3.16 Ethyl acetate 33.41 1.83

    n-Butyl alcohol 0.69 10.75 Isovaleraldehyde 8.17 2.88 2,3-Pentanedione 0.76 8.25 Isoamyl alcohol 2.36 9.38

    n-Propyl alcohol 0.82 5.64

    Isobutyl alcohol 3.79 I .75

    From Sasaki and Nunomura (1978).

    gas through the shoyu at 20C and then trapping the vapor by dry ice-ethanol, liquid nitrogen, and activated carbon, in succession. A total of 24 compounds, 3 of which were isolated for the first time, were identified; of them, 14 are listed in Table XXXVII. The respective constituents of odor in three isolated compounds were calculated, and the aroma of the headspace gas from fresh fermented shoyu was attributed primarily to isovaleraldehyde, ethanol, and isobutyraldehyde, as


    Relative Concentration Threshold odor units

    Compound (PPm) (ppm in water) Odor units (90) ~~

    Ethanol 5605.18 1.83 x 10-1 30,629.40 33.05

    Isovaleraldeh yde 8.17 1.5 x 10 -4 54,466.67 58.77 Isobutyraldehyde 6.38 9.0 x 10-4 7,088.89 7.65 Acetaldehyde 3.76 1.5 x 10-2 250.67 0.27 Propionaldehyde 1.70 9.5 x 10-3 178.95 0.19

    Ethyl acetate 33.41 6.0 x 10-1 55.68 0.06

    LI From Sasaki and Nunomura (1978).


    shown in Table XXXVIII. The researchers recognized the presence and impor- tance of volatile sulfur compounds, but could not identify them.

    Yoshida et al. (1980) analyzed the topnote aroma concentrate of shoyu and identified ethanol, ethyl acetate, isobutyraldehyde, isovaleraldehyde, the di- ethylacetals of these aldehydes, isoamyl alcohol, and a trace amount of dimethyl sulfide.

    Sasaki and Nunomura (1979) measured the loss of the aroma constituents due to the evaporation of shoyu at 23C. Propionaldehyde disappeared completely in 15 min and half of the ethanol and isovaleraldehyde dissipated in 30 min, indicating the freshness of a pasteurized shoyu rapidly exposed to the open air.


    The accuracy of the quantitative analyses is particularly critical in an investi- gation of the relationship between the gas-chromatographic pattern of a shoyu and its organoleptic evaluation. Sasaki et al. (1980) compared the coefficients of variation (CV) and the recoveries of flavor compounds of shoyu using three analytical procedures:

    I . Shoyu (50 ml) was distillated in a vacuum at 40C and the distillate (35 ml) saturated with NaCl and then extracted with CH,Cl,, which had been concentrated into 2 ml by evaporation of the solvent.

    2. Shoyu (50 ml) was saturated with NaCl and then extracted with CH,Cl,, which was concentrated into 2 ml.


    lsobutyl alcohol n-Butyl alcohol Isoamyl alcohol Acetoin Ethyl lactate Furfuryl alcohol Methionol 2-Pheny lethanol 4-Ethylguaiacol

    15.19 27.90 9.68 31.20 7.66 31.50

    10.00 15.10 12.67 6.50 13.95 25.90 6.85 26.30

    14.29 56.10 16.90 64.70

    8. I 38.00 1.39 6.4 44.93 I .44 4.2 54.12 1.39 4.6 49.73 2.23 6.4 43.13 1.29

    12.8 61.58 5.88 16.1 66.22 1.86 26.7 71.85 3.37 17.0 82.82 I .48

    102.00 103.90 108.10 96.20

    102.30 92.70 88.40 97.30 97.90

    From Sasaki er al. (1980). b Recoveries.




    Ethanol Lactic acid Glycerol Acetic acid HMMF 2,3-Butanediol Isovaleraldehyde HEMF Methanol Acetol Ethyl lactate 2.6-Dimethoxyphenol Ethyl acetate Isobutyraldehyde Methyl acetate Isobutyl alcohol

    31.501.10 14,346.57 10.208.95 2,107.74

    256.36 238.59 233.10 232.04 62.37 24.60 24.29 16.21 15. I3 14.64 13.84 11.96

    Furfuryl alcohol lsoamyl alcohol Acetoin n-Butyl alcohol HDMF Acetaldehyde 2-Phenylethanol n-Propyl alcohol Acetone Methionol 2-Acetylpyrrole 4-Ethylguaiacol Ethyl formate y -Butyrolactone 4-Ethylphenol

    I I .93 10.01 9.78 8.69 4.83 4.63 4.28 3.96 3.88 3.65 2.86 2.77 2.63 2.02


    a From Yokotsuka er al. (1980).

    3. Shoyu (5 ml), 2 ml methyl acetate, and 1 g NaCl were shaken in a closed test tube and then centrifuged at 3400 rpm for 10 min at 5C. The methyl acetate layer was directly subjected to gas-chromatographic analysis.

    As is indicated in Table XXXIX, procedure 3 gave the most reliable results. One example of an analysis of the flavor constituents of shoyu is indicated in Table XL .


    The content of 4EG and the sensory evaluation of shoyu are in a parabolic relationship, and both too great and too small an amount were not liked by consumers. The optimum content of 4EG was roughly claimed to be less than 0.5 ppm (Yokotsuka, 1967~).

    Mori et al. (1982, 1983) confirmed the correlation coefficients between each of 27 kinds of odor components and the sensory evaluation of their companys shoyu to be 0.313 at the highest. This fact suggested that it was difficult to predict the scale value of the shoyu by only one kind of odor component. By checking the effects of all combinations of each of two components, the com- bination of 4EG and methionol was found chiefly to influence the variation of


    sensory data. The optimum sum of two components was first determined to be 4.5 ppm, then the optimum content of each component was found to be 0.3 ppm for 4EG and 3.9 ppm for methionol, respectively, which was confirmed both by mathematical calculation and by an addition test with shoyu. The relationship between the content of 4EG or methionol and the sensory evaluation of shoyu was parabolic. The same authors (1984) conducted a similar experiment with 30 brands of shoyu on the Japanese market. The content of each flavor constituent of the samples ranged wider than that of a simple brand of product. The highest correlation coefficient was found for ethyl acetate to be (r = -0.551). Ethyl acetate was found to give a kind of freshness to shoyu. Both n-butyric acid and HEMF were found to be in a parabolic relationship to sensory evaluation. Most shoyu tested contained about 1 ppm of n-butyric acid, and generally, a content of more than 3 pprn of n-butyric acid yielded an inferior sensory evaluation. The average content of HEMF of the shoyu tested ranged from 100 to 200 ppm, and an inferior sensory evaluation was given to the shoyu that contained less than 50 ppm of HEMF. A total of 595 combinations of each 2 among 35 flavor compo- nents was checked for their contents. Many were found to have the sum of contents of 4EG and methionol at more than 4.5 ppm, which was reported to be optimum. The sum of contents of acetoin and isobutyric acid was found in this case to be highly associated with sensory evaluation of shoyu. Sasaki er al. (1984) compared Japanese fermented shoyu (I) and Southeast Asian soy sauces (11) in terms of flavor constituents of headspace gas and solvent extract as follows:

    1. The content of HEMF was 150-400 ppm for I, but 0-trace for 11. 2. The sum of isobutyl alcohol, n-butyl alcohol, isoamyl alcohol, methionol,

    and 2-phenylethanol of I1 was 0-20 ppm, which was about one-half of I. 3. Methional[3-(methylthio)propanal] was distributed widely in both I and 11,

    with the contents of 0.2-2.0 ppm. 4. More pyrazines were found in accordance with an increase of HVP, which

    was blended with fermented soy sauce.



    The capability of some strains of mold to produce mycotoxins has been re- ported. Examples are aflatoxins, ochratoxins, sterigmatocystin, patulin, pen- icillic acid, islanditoxin, cyclopiazonic acid, and zearalenones, including T-2 toxin. Among these, aflatoxins seem to be the most important because of their


    acute toxicity and significant carcinogenicity. Moreover, according to Sargeant et al. (1961), aflatoxins are produced by the Aspergillusflaws group, which include Japanese koji molds, such as A. oryzae and A. sojae, classified by Sakaguchi and Yamada (1944), used for food fermentation. According to the classification by Raper and Fennel (1965), the A. jlavus group includes A.flavus, A. parasiticus, and A . oryzae, while aflatoxin producers are found in A . flavus and A. parasiticus. The question of whether Aspergillus molds used for food preparation produce aflatoxins follows logically. Murakami ( 1 97 I ) studied the taxonomic classification of Aspergillus molds. He reported that industrial mold mostly belongs to the A. oryzae group, A. sojae, and A. tamurii, while all of the aflatoxin-producing molds belong to A. parasiticus and A. toxicurius Murakami, which are clearly distinguishable from industrial molds. Nevertheless, it is important to note that these Aspergillus molds are morphologically continuous with regard, for example, to the roughness of their stalks or color and surface conditions of their spores. Therefore, it is sometimes difficult to classify these molds definitively using only their morphological features. This is especially true in differentiating between A . sojae and A. parasiticus, both of which are good producers of proteolytic enzymes. Several classifications of Aspergilli are sum- marized in Table XLI. Among the 125 strains of mold used for shoyu production in Japan, there are 29 A . sojae and 92 A. oryzae (Murakami, 1973). According- ly, from the viewpoint of the food industry, it becomes extremely important to confirm by means of chemical analyses that the molds to be used do not produce aflatoxin.

    Some investigators have reported negative findings in studies of the use of Japanese industrial molds in fermentation and the production of aflatoxin. Hesseltine er al. (1966) studied 53 cultures at the Northern Regional Research Laboratory in the United States, but tests of miso, shoyu, and sake, all made with strains of A. oryzae, were negative. Aibara and Miyaki (1965) examined 180 strains, including those used in the preparation of miso and cheese, but analyses with UV absorption, excitation, and fluorescence spectra revealed no producer of aflatoxin. Masuda et al. (1965) studied 21 strains of industrial mold with the same results. Murakami et al. (1967, 1968) examined 214 kinds of Aspergillus mold by fluorometry and thin-layer chromatography (TLC), including 176 indus- trial strains, for their aflatoxin-producing ability. Thirteen strains gave fluores- cent spots on TLC corresponding to aflatoxin, but their UV absorption spectra were different from those of aflatoxin. Manabe et al. (1968) observed that 49 strains among 2 12 koji molds exhibited aflatoxin-like fluorescent spots on TLC, but that all of their UV absorption spectra were different from those of aflatoxins. Kinoshita et al. (1968) concluded from their results of TLC and UV absorption spectra that of 37 strains of mold isolated from Japanese katsuobushi (dried bonito used for seasoning), shoyu, and miso, none produced aflatoxin.



    1. Sakaguchi and Yamada (1944) Koji molds: Aspergillus oryzae (Ahlburg) Corn

    A. sojae. Sakaguchi et Yamada, prominently echinulate conidia and smooth- walled conidiophore

    2. Raper and Fennel (1965) A. jlavus groups: A. f7avus L.

    A. jlavus var colummaris R. et F. A. parasiticus A (include A. sojae) A. oryzae (A) C. A. ramari K unnamed species

    3. Murakami (1971) A. oryzae group: A. sojae S. et Y.

    A. tamari K A. oryzae (A) C. A. oryzae var. viride M. A. oryzae var. brunneus M. A. parasiticus S . A. toxicallius M. A. jlavus L.

    A. jlavus group:

    4. American Type Culture Collection (1982)

    5. Kurtzman (1983) Recognized A. sojae S. et Y. as a new species, and separated from A. parusiricus

    A 90% or more relatedness of A. jlavus, A. oryzae, A. parasiticus, and A. sojue regarding DNA structuresa

    Kurtzman (1983).

    The research techniques used in these investigations and the kinds of data thus generated limit the analysis largely to a comparison of R, values of TLC and fluorescence spectra of the spots. Fluorescent compounds with violet-to-green fluorescence resembling aflatoxin B or G, which have been reported to be pro- duced by Aspergillus molds, are flavacol (Dunn et al . , 1949), isoxanthopterin (Kaneko, 1965), ferulic acid (Asao and Yokotsuka, 1958a), aflatoxin B and G (Sargeant et al., 1961), some degradated products of ergosterol (Yokotsuka et al., 1966), and others (Kihara et al., 1944). Among the fluorescent compounds produced by Aspergillus molds, aflatoxin clearly differs from the others with respect to its R, value on TLC. Some investigators, however, have found that a fairly large number of strains of Aspergillus mold do produce aflatoxin-like fluorescent compounds having R, values on TLC, similar to the aflatoxins, but with different UV maximum absorptions. These include seven kinds of pyrazine compounds, isocoumarin compounds, lumichrome, and unknown compounds


    with aflatoxin G-like green fluorescence. The existence of these fluorescent compounds makes it difficult to determine the capability of some strains of mold to produce aflatoxin by TLC alone.


    Yokotsuka et al. (1966b, 1967c, 1968a,b) and Sasaki et al. (1967, 1968a,b) examined 73 industrial strains of Aspergillus mold used either for the production of shoyu, miso, and rice wine, or found in the stock cultures, for their production of fluorescent compounds after being cultured in a zinc-containing Czapek Dox medium (Nesbitt et a l . , 1962). About 30% of these strains showed fluorescent spots resembling those of aflatoxin B or G. Ultraviolet absorption spectra of the eluants of 14 strains whose R , values were similar to aflatoxin B 1 were divided into two groups having UV absorption maxima at 320-330 and 310-315 nm, respectively. However, no eluant had a UV absorption maximum of 363 nm, which is characteristic of aflatoxin B, . Likewise, eluants of the spots of eight strains whose R , values were similar to that of aflatoxin G, did not show the UV absorption maximum of aflatoxin GI (Yokotsuka el al., 1968~).

    The best producing strain of the fluorescent compounds whose UV absorption spectrum was at 320-330 nm was A . sojae X- 1, a wild strain cultured in pep- tone-enriched Czapek Dox medium (modified Mayers medium). Eight fluores- cent spots were observed on TLC, but their R, values were different from those of aflatoxin B, using 11 kinds of solvent systems (see Fig. 25). Eight fluorescent compounds, including flavoacol, and eight nonfluorescent compounds, includ- ing aspergillic acid and hydroxyaspergillic acid, were isolated in cyrstalline form from cultured broth according to the method depicted in Fig. 26. The chemical structure of each compound was determined by elemental analysis, melting point, NMR, UV spectrum, IR spectrum, and so forth. Aflatoxin B-like com- pounds were related to each other, with a pyrazine ring common to their struc- ture. This was indicated by similar UV spectra, with absorption at 310-320 nm, and similar IR spectra, with absorption at 1600 cm- I .

    Also confirmed was the finding that when these isolated compounds have 2- hydroxypyrazine rings, they give fluorescence, but when the first nitrogen has an oxide structure, they give no fluorescence. The maximum absorption 1R spectra at -950 cm-1 seemed to be associated with those differences. Identified fluo- rescent and nonfluorescent pyrazine compounds are listed in Tables XLII and XLIII, respectively (Sasaki et al . , 1967, 1968a-c; Yokotsuka et a l . , 1968b,c).

    These compounds are considered to be condensation products of two mole- cules of amino acid, as was suggested by J. C. MacDonald el al. (MacDonald, 1961, 1962, 1965, 1967; Micetich and MacDonald, 1965). Examples include








    E C C

    5 6

    e 0

    0 0

    . O


    0 ' 0

    - 7

    :: 1 1 Solvent number

    FIG. 25. Variation of R f values of aflatoxin B-like compounds of Aspergillus sojue X-1 with different solvent systems. Aflatoxin BI and GI; Black spots on left side of each column. Aflatoxin B- like compounds: BO to B8 from the top on right side. Absorbent: Kieselgel G, 0.5 mm. Solvent 1: Benzene-ethyl acetate (3 + I); 2: benzene-acetone (3 + I ) ; 3: chloroform-ethyl acetate (3 + 1); 4: benzene-ethyl acetate-ethanol (30 + 19 + 1); 5: chloroform-methanol (97 + 3); 6: benzene- ethanol (9 + I ) ; 7: chloroform-ethyl acetate-ethanol (30 + 19 + I ) ; 8: ethyl acetate-hexane (3 + I ) ; 9: chloroform-acetone (3 + 1); 10: ethyl acetate-methanol (3 + I ) ; 11: acetone-hexane (3 + 1). From Yokotsuka et ul. (1966).

    leucine and leucine, isoleucine and isoleucine, isoleucine and leucine, and valine and leucine, in our case. In view of the fact that thin-layer chromatographs have exhibited many other faint spots of possible fluorescent pyrazine compounds, it is reasonable to suspect that many of the pyrazine compounds produced by molds from two molecules of amino acid (e.g., from valine and valine or from valine and isoleucine) exist in nature.

    Importantly, these are not limited to A. sojae X- 1, but are equally applicable to the production of A. sojae and A. oryzae, which are actually used in the prepara- tion of fermented foods. The crystals of fluorescent pyrazine compounds BO to B6 (excluding B5 and B7) were injected intraperitoneally into mice. Because of the shortage of test samples, only three mice were tested for each dosage, 250 mg and 500 mg/kg. These compounds exhibited no acute toxicities of more than 250 mg/kg (Sasaki et al., 1968a-c). The same test for toxicity was applied to the









    T PY



    E C




    S PRO



    D B

    Y A

    spergillus sojae X-l SIM


    R T

    O A






    H R






    R, VA


    ES O

    N TLC

    I. --C




    11. --C




    111. --C








    s V

    . --C












    R N

    ame of com


    BO I

    I 2-Hydroxy-3,6-di-sec-butylpyrazineb


    I1 I

    Deoxyaspergillic acid


    I1 I1

    Flavacol B

    2 I1

    111 D

    eoxymutaaspergillic acid


    I IV



    I1 IV


    eoxyhydroxyaspergillic acid B

    5 I1


    2-Hydroxy-6-( 1 -hydroxy-2-methylpropyl)-3-isobutylpyrazineJ


    I1 V

    I 2-H

    ydroxy-6-( 1 -hydroxyisopropyl)-3-isobutylpyrazine

    a From

    Yokotsuka er al. (1967), Sasaki et al. (1967, 1968a).

    Uncertain identification.

  • 294

    Counter current




    Hexane (4) : Benzene (1):

    80% aq. methanol (5)

    Broth of A. soiae X- 7

    Modified Meyer's medium

    Surface culture, 30C, 15 days

    Extraction (chloroform)

    Condensation in vacuo, 55C I

    Extraction (5% HCI)

    Aqueous layer I p H 3 Extraction (chloroform)

    Counter current distribution I

    Hexane Koltoff buffer

    (PH 9.0)

    TLC (Kiesel G, 0.5 mm)

    Benzene (60):

    Ethylacetate (36):

    Ethanol (4)

    Ether extraction of 8 spots I

    Crude crystals I

    60, 1,2,2',4, 5.6.1

    and one uncertain compound

    Chloroform layer I

    Extraction (3% Na2C03)

    T LC

    cu Salt I


    Recrystallization AO, 2, 3,4, 5

    (flourescent) and three uncertain compunds


    FIG. 26. Separation of fluorescent and nonfluorescent pyrazine compounds from culture of As- pergillus sojae X-1. From Yokotsuka ef al. (1966, 1967, 1968a,b) and Sasaki er al. (1967, 1968a).



    FROM THE CULTURE OF Aspergillus oniki 1784"

    Acute toxicity Compound number Melting UV spectra, Fluorescence on mice

    and structure point (T) A:? nm on TLC (LDso mg/kg) (ip)

    56 246, 315 + 550- I250 BV-1 (Me1 l e i n )



    121-121.5 244.5. 315 + 1000-1500

    109-109.5 246, 315 -

    BY-4 171


    From Sasaki er al. (1970).

    nonfluorescent compounds, for the total mixture, and for AO, 2 ,3 ,4 , and 5. The toxicities of aspergillic acid and hydroxyaspergillic acid have been previously reported in the literature (Nakamura and Shiro, 1960, 1961). The results suggest the toxicity of these compounds is similar to that of aspergillic acid, LD,,, which is -100 mg/kg.

    From the culture of the strains which produce aflatoxin B-like, bluish violet fluorescent compounds of Group 2, three isocoumarin compounds, including mellein (Nishikawa, 1933), 4-hydroxymellein, and 3,4-dihydro-3 &di- hydroxy-3-methylisocoumarin, were isolated (see Table XLIV). These iso- coumarin compounds were found to be produced by some strains of Aspergillus



    Color of Melting fluorescence Excitation Fluorescence

    point on TLC uv s Ctra spectra spectra Compound ("C) (365 nm) (A,,, nm) nm) (ALC'i nm) MeOr

    G3 266 Bluish green 342 350 410 Aflatoxin G I 247-250 Green 363 365 450

    a From Yokotsuka et al. (1968~).

    ochraceus. Although A . ochraceus is not used in food industries, it is found in foodstuffs as a contaminant. Under certain experimental conditions, these com- pounds also exhibit fluorescence and R, vlaues resembling those of aflatoxin B (Sasaki et al., 1970).

    Regarding green fluorescent compounds produced by Aspergillus molds, 7 out of 72 strains tested exhibited four kinds of green fluorescent spots on TLC. From 400 liters of cultured broth of Aspergillus M4-1, which is used in making miso, four kinds of green fluorescent compounds were isolated: three kinds of crude crystals, and 6.2 mg of purified crystals with mp 266C. TLC yielded R, values resembling that of aflatoxin G, under certain experimental conditions, but analy- ses with 15 kinds of solvent systems confirmed their difference from aflatoxin G,. Other physical properties, including UV absorption, were also different from those of aflatoxin G, (Yokotsuka et al., 1968c), as shown in Table XLV.

    Approximately 200 strains of mold, including 126 newly added to the previous 73 strains, were reexamined for their productivity of aflatoxins. It was found that none produced aflatoxins. However, almost all newly tested strains produced lumichrome, C,,H,,J,O, (Karrer et al., 1934). This compound displays green fluorescence and an R, value similar to that of aflatoxin G under certain experimental conditions (Sasaki et al . , 1974).

    Two strains of A . f laws Link were reported to produce aflatoxin B , (Kurata et al., 1968, 1969). One was isolated from wheat flour imported into Japan, and the other was isolated from homemade rural miso. Sasaki et al. (1975) reconfirmed that aflatoxin was produced from the former strain, but aflatoxin B, was not detected in the latter strain. The purified fluorescent sample isolated from 300 liters of cultured broth exhibited the same R, value as did aflatoxin B , using TLC with chloroform, methanol 97:3, but a different R, value from that obtained when aflatoxin B , was analyzed with a solvent system composed of benzene:acetone, 3: 1. The sample was further purified into two compounds with a different UV absorption from that of aflatoxin B , .


    From these data it is evident that some Aspergillus molds produce fluorescent compounds with R , values resembling those of aflatoxins. Indetecting and char- acterizing samples that are contaminated with aflatoxins, R, values should be determined with two or more solvent systems, and UV and IR spectral data should also be used. This implies that the compounds must be chemically iso- lated and identified.


    A total of 69 strains of Japanese industrial mold was tested for their productivi- ty of aspergillic acid, kojic acid, P-nitropropionic acid, and oxalic acid, although these acids are not carcinogenic and their toxicity is not as great as aflatoxins (Yokotsuka et al., 1969). The following are the respective numbers of non-acid- producing strains among the 69 tested in liquid media: aspergillic acid (N = 40); kojic acid (N = 32); P-nitropropionic acid (N = 48); and oxalic acid (N = 37). Some strains of mold that proved to be good producers of aspergillic acid and kojic acid in liquid media did not produce these acids on a solid substrate composed of soybeans and wheat, at least not within the usual 2-day period required for koji cultivation. These tests confirmed that koji, or a mixture of soybeans and wheat cultured with these moles, does not contain a sufficient amount of these weak toxic compounds to constitute a hazard to humans who consume shoyu, even when it is prepared from koji cultured with the strongest acid producer of these toxic compounds among the strains tested.

    Yokotsuka et al. (1977) were unable to detect aflatoxin, patulin, ochratoxin, or sterigmatocystin in the culture of A. sojae, which is used for shoyu produc- tion. Sasaki (1980) checked the ability of 33 kinds of industrial Aspergillus mold to produce aflatoxin, sterigmatocystin, ochratoxin, patulin, cyclopiazonic acid. and penicillic acid. None of the strains tested produced these compounds, with the exception of a very few strains which produced cyclopiazonic acid. Sasaki concluded that it is feasible to avoid mycotoxin contamination from a purely cultured starter mold if the strains which do not produce these mycotoxins are selected. Manabe et al. (1985) observed that some koji molds belonging to A. oryzae or A. sojae produced cyclopiazonic acid. Shinshi et al. (1985) found that cyclopiazonic acid added to salty shoyu mash was decomposed by S . rouxii or Candida (Torulopsis) versatillis, especially by the latter. They did not find shoyu on the market which contained cyclopiazonic acid.

    According to Yokotsuka (1977), Kikkomans koji culture of A. sojae does not contain patulin, ochratoxin, or sterigmatocystin. Attempts to detect lysinoalanine in fermented shoyu have been unsuccessful. Five lots of shoyu from Kikkomans Wisconsin plant in the United State were analyzed for heavy metal. Arsenic, mercury, and selenium were not detected, lead and copper were found in trace amounts, and the figure for total heavy metals was less than 2 ppm.



    Although the noncarcinogenicity of fermented shoyu has long been known from long-term animal studies, the mutagenicity of heated products of amino acids or proteins such as Trip-P-l , Trip-P-2, Glu-P-1, and Glu-P-2 has been established more recently. Secliff and Mower (1977) reported that soy sauce produces mutagens upon the heating of glucose, galactose, and arabinose in shoyu.

    Using the salmonella/mammalian microsome mutagenicity test, Lin et al. (1978) found that when treated with nitrite at 2000 ppm, soybean sauce produced a mutagenic substance. As fermented shoyu sometimes contains a small amount of amines (e.g. histamine and tyramine), the formation of mutagenic substances as a result of the reaction between amines and an abundance of nitrite is possible. Shibamoto (1983) mixed soy sauce with 100,500,1000, and 2000 ppm of sodium nitrite, adjusting pH at 3.0, and heated the mixture for 2 hr at 25C and then for an additional 30 min at 80C. Only at the highest concentration, 2000 ppm, was mutagenicity exhibited in the Ames test. Shibamoto concluded that the formation of nitrosamines may not be significant because the quantity of nitrite used in the study was excessive compared with actual food systesm. It is generally reported that the nitrite concentration remaining in the human stomach after a meal is estimated to be about 5 ppm, or about 15 ppm at most. It was also reported that just after ingestion of cured ham, the concentration in the stomach is about 70 ppm. According to Nagahara et al. (1984), shoyu itself did not represent mutagenicity. 1 -methyl- 1,2,3,4-tetrahydro-~-carboline-3-carboxylic acid (MTCA) decreased in a buffer solution when treated with more than 10 ppm of nitrite for 1 hr at 37C and pH 3.0, but in shoyu, it decreased with more than 250 pprn of nitrite. Tyramine decreased in a buffer solution when treated with more than 50 ppm of nitrite for 1 hr at 37C and pH 1 .O, but in shoyu, it did not decrease even when treated with 2300 ppm of nitrite. Nagahori et al . (1980) reported that the addition of 5-7% fermented shoyu to a mixture of dimethylamine and nitrite at pH 3.6 suppressed the formation of N-nitrosodimethylamine by 60-80%. Moreover, the quantity of nitrosamine formation hindering substances in fermented shoyu in- creased with the advance of fermentation and aging of the mash. These substances were identified as the amino acids present in shoyu, which react more easily with nitrite than with dimethylamine. Ochiai et al. (1982) and Wakabayashi et al. (1983) isolated a nitrosable precursor of mutagens from shoyu. Its chemical structure was confirmed to be 1,2,3,4-tetrahydroharman-3-carboxylic acid ( 1 - methyl-l,2,3,4-tetrahydro-~-carboline-3-carboxylic acid, MTCA). When this compound was treated with 3450 pprn of nitrite for 1 hr at 37C and pH 3.0, the nitration product was strongly mutagenic to Salmonella typhimurium TA 100. Wakabayashi et al. (1983) determined the tyramine content of shoyu to be 17-



    pH 3, 3pc, ' hr Mutagenic (Salmonella tryphimurium TA 100) Eighty percent of nitrosation product is not muta-

    genic. From Ochiai et al. (1982)


    (-) - (IR,3S)-MTCA and its isomer Content in shoyu is 0.03-0.7 ppm.

    Reaction is one-third at pH 3. No reaction at pH 5-6. From Nagao et al. (1983) .+

    OH - OH 2000 pm

    Tyramine 3-Diazotyramine Content in shoyu is 0-lo00 ppm.

    H Yan and Lee (1984) 2-Acetylpyrrole Content in shoyu is 3 ppm.

    2250 ppm, and when tyramine was treated with 2300 ppm of nitrite for 1 hr at 37C and pH 1.0, strong mutagenicity to TA 100 was observed. Yen and Lee (1984) isolated 2-acetylpyrrole as a nitrosable premutagen from shoyu. These are given in Table XLVI.


    Ujiie et al. (1956) identified the bactericidal nature of a commercial fermented shoyu with respect to nine kinds of intestinal pathogenic bacteria, such as Esche- richia coummunis, Shigella jlexneria, Vibrio cholerate Inaba, Salmonella 9- phimurium Shikata, Bacillus subtilis (B-3 1). The kinds of bacteria present were attributed to the acidity, high osmotic pressure, and some of the chemicals contained in the shoyu. To the sample, 0.005% of butyl-p-hydroxybenzoate was added as a preservative. Sakaguchi et al. (1975) tested the fate of staphylococci during incubation in a normal shoyu and in a milder shoyu, containing 17% (w/v) and 9% (w/v) of sodium chloride, respectively. No chemical preservatives


    were added to either sample. The normal shoyu which initially contained lo6 staphylococci per milliliter was nearly free of viable staphylococci within 3 hr. In the milder shoyu, over 90% of the cells were destroyed within 22-30 min, while in the normal shoyu, only 13-14 min were required. That sodium chloride contributes to the destruction of Staphylococci in soy sauce is evident because the rate of killing in normal shoyu is greater than in milder shoyu. The fate of staphylococci in phosphate buffer saline solutions with a pH level of 4.7 contain- ing 10 and 17% sodium chloride, respectively, was tested under the same condi- tions. The time taken to destroy over 90% of the cells in the 10% solution and in the 17% solution was 980-1440 min and 460-530 min, respectively. These results suggest the participation of some factor other than sodium chloride in the destruction of staphylococci in shoyu. The activity of Closrridium botulinum in shoyu was also tested during months of storage at 30C. Neither C. botulinum 62A (Type A) nor C. botulinum Okre (Type B) grew during this time. The number of Type A spores remained the same, but those of Type B decreased slightly in number after the 3 months.

    According to Yamanoto et al. (1978), the time needed for the total destruction of Escherichia coli 215 or Staphylococcus aureus 209P (ATCC 11522) inocu- lated in fermented shoyu was dependent upon the initial number of cells in these bacteria; 4-6 hr for 103/ml, 24-48 hr for 105/ml, and 5-7 hr for 107/ml. A high salt content was the dominant factor in accelerating the speed of sterilization; the pH value and amount of alcohol, total nitrogen, and ether-soluble compounds were judged to be supplementary factors.


    The long-term effects of Japanese shoyu (Kikkoman) on the gastric mucosa of intact rats and those with a fundasectomy were studied by MacDonald and Dueck (1976) in Canada. At the end of the test period, the animals that had been fed shoyu were smaller than the controls; the 15 intact rats that received the shoyu were healthier, more active, and lived 33 months longer than did the 7 controls. Breast tumors developed in 10 control rats, but in none of the experimental animals given shoyu. These findings suggest that shoyu does not appear to be carcinogenic in rats; its prolonged use impaired neither health nor longevity. Oshita et al. (1977) studied the acute and long-term effects of large amounts of Kikkoman shoyu on mice and rats. The acute toxicity of shoyu was attributed to the toxicity of its sodium chloride component. The oral LD,, values for shoyu were 20.6 ml/kg for rats and 27.3 mg/kg for mice. In long-term feeding tests (1.5 years for mice and 6 months for rats), the food intake of animals given a diet containing shoyu was otherwise comparable to that of the control group. This was true even for animals given a diet containing 10% powdered shoyu (corre-


    sponding to -25% liquid shoyu). Although the animals that were fed shoyu were smaller than the controls, no significant differences in mortality were observed between the two groups. In addition, male rats given a diet containing 5% or 2% powdered shoyu grew faster than rats fed an equivalent amount of sodium chlo- ride alone (i.e., diets containing 2.25% or 0.9% sodium chloride). At the highest dose level, 10% powdered shoyu, there were significant differences in the uri- nary systems of experimental and control animals. While both rats and mice developed enlarged kidneys and bladders, rats developed higher concentrations of in serum, and mice gave evidence of hydronephrosis after 1.5 years. The same effects were observed in animals who received sodium chloride in the same concentrations as those fed the highest level of shoyu. There was no indication of carcinogenic effects at any level of shoyu feeding.



    The precise mixture of soybeans and wheat used as the raw materials in shoyu production is the result of technological know-how developed over hundreds of years. But the shoyu-like seasonings can be prepared from a mixture of plant proteins and starches other than soybeans and wheat, and these are available worldwide. The by-products of oil pressing and extraction, such as peanut cake, copra meal, cottonseed meal, rapeseed protein, and sesame protein, have been experimentally substituted, with good results. Many kinds of mung beans also seem to have good potential. On the other hand, wheat kernel is considered to be the best starch raw material for shoyu, but barley, rye, oats, rice, and corn are sometimes used. The superiority of the wheat kernel lies in its high source of protein (glutamine) and its high concentration of glucosides which give koikuchi its destructive bran flavor. The use of a mixture of wheat bran and starches other than wheat, such as rice, corn, and potato, would be worth exploring for those nations that do not have wheat. Steaming and puffing corn kernels, mung beans, and other raw materials which have hard plant tissues or extruding these moist- ened powders may also give good results in shoyu preparation.


    Although the amount of total nitrogen extracted from the raw materials for use in shoyu production has reached -9O%, the possibility of increased protein digestibility seems to exist. A systematic effort to identify better strains of koji mold which have a greater capability to produce protease and enzymes that


    degrade plant tissue in general is needed. Such strains of mold will not only increase protein digestibility, but will shorten the fermentation period and en- hance the flavor of shoyu.

    A strain of koji mold having enzymes which readily degrade plant tissue in the presence of high salt concentrations would reduce the viscosity of mash. If its viscosity could be reduced, shoyu mash could be press-filtered by a much sim- pler and more economic apparatus than is now in use.

    The relationship between various kinds of koji mold and the volatile flavor constituents of shoyu should be studied. While it is clear that the volatile flavor of the final shoyu is influenced by the kind of koji mold used, its biochemical details are not yet known. To date, the selection of koji mold for this purpose has been conducted only by means of a sensory test.


    Although the fermentation period of shoyu mash has been shortened by 6-8 months as a result of technological developments, a much greater reduction in time is necessary for economic reasons. The time needed for the enzymatic degradation of the raw materials and for lactic and yeast fermentation is about 3 months, but at least another 3 is required to complete the aging process. During these 3 months, the color deepens and the flavor develops fully. Gas-chro- matographic patterns reveal that the major chemical changes that take place during the last 3 months of aging are mostly due to heat-dependent reactions. But the temperature cannot be raised too high to accelerate the aging process (e.g., to more than 35"C), as this generates an unpleasant odor. In order to shorten the fermentation period, the chemical changes occurring during the aging of shoyu mash must first be fully understood.

    To shorten the early stage of mash fermentation, which usually takes 3 months, lactic and yeast fermentation can be accelerated by adding a sufficient number of pure cultured cells of lactobacilli or Succharomyces yeast to the mash and increasing the temperature of the mash 30C or more. The physiology of lactobacilli and shoyu yeasts in the presence of low and high salt concentrations should be studied, especially with respect to the chemical processes by which flavor develops.

    If shoyu mash is enzymatically degraded at more than 50C by using heat- tolerant strains of mold, no salt needs to be added to the mash, and degradation of the protein is completed within 1 or 2 days, yielding a mash whose protein is highly digestible. Caution should be taken not to reduce the free amino acid content, especially that of glutamic acid, by sapplying heat-tolerant peptidases such as glutaminase. To avoid the development of the unpleasant odor which develops when mash is subjected to high temperatures, the salt concentration can


    be reduced to less than 10% and the temperature lowered to 40C in order to complete the enzymatic degradation of starch and proteins with good yields. Thus, an enzymatic protein hydrolysate of good chemical composition in terms of amino acids and sugars can be produced. However, the enzymatic degradate must undergo lactic and yeast fermentation for -3 months (comparable to the normal shoyu fermentation period) so that it acquires a shoyu-like flavor in a slurry or pasty condition of shoyu mash, but as for the liquid separated from the enzymatic degradate, lactic and yeast fermentations can be finished in about a week in batch-type fermentation tanks, or in 2 or 3 days when the liquid is passed through the columns packed with the immobilized cells of these microbes. Sen- sory quality improvements still need to be studied with regard to this method.


    The solid koji method of shoyu manufacture seems to be the simplest, most economic, and best quality-producing method of degrading protein by the use of enzymes. Research efforts directed at substituting enzyme preparations from microbial sources other than koji molds have not yet succeeded. It is well established that the enzyme systems involved in koji molds are much better than those of Rhizopus and Bacillus with regard to their ability to decompose soybean protein in the presence of high salt concentrations. Moreover, the safety of koji molds has been proved by hundreds of years of consumer use. The search for better strains of mold for shoyu production should begin with the exploration of koji molds A. oryzae or A . sojae. Nevertheless, supplementing the enzyme systems of koji molds with other microbial sources would also help to increase the protein digestibility and amino acid content of shoyu and to reduce both the fermentation period and the viscosity of mash.


    The heat-coagulant substances produced by pasteurization during the indus- trial manufacture of shoyu today pose a difficult problem. Centrifugation is not an effective means of removing these substances from shoyu because of the high gravity of shoyu which contains salt and other solids. Filtration with some aid, such as Celite, is the best method for clarifying the clear upper layer of heated shoyu after sedimentation. But it is very difficult to recover shoyu from the sediment layer which contains more than 95% shoyu after filtration centrifuga- tion. A method using sedimentation-accelerating substances has been tried with- out success.

    It has also been established that heat coagulation in pasteurized shoyu is positively correlated with the amount of protease in raw shoyu. In addition, the


    ongoing protein activity in shoyu is positively associated with the protease ac- tivity of koji, the alcohol content of shoyu mash which inhibits the protease activity, the period of fermentation of mash during which protease activity grad- ually decreases, and with the pH value of shoyu mash, which at lower levels inactivates protease. Consequently, some method of reducing the protease ac- tivity remaining in mash needs to be found. Decomposing or removing the precursor of heat coagulation in shoyu before pasteurization, keeping the mash or raw shoyu at higher temperatures, or subjecting raw shoyu to ultrafiltration have been tried.


    The volatile and nonvolatile flavor constituents of koikuchi shoyu are mostly products derived from the metabolism of raw materials by koji molds, lac- tobacilli, and yeasts, and from their mutual chemical reactions during the pasteurizing and aging of mash. Although it is difficult to make the koikuchi flavor stronger without increasing its color intensity, it would be interesting to try to produce the flavor compounds of koikuchi shoyu biochemically for the pur- pose of a wide application to food preparation other than shoyu, or to prepare a shoyu lighter in color, but with the characteristic koikuchi flavor, or to prepare a shoyu having the color of koikuchi, but weaker in taste.

    There is also a need to develop a method for evaluating the quality of a shoyu from the contents of some key flavor components to supplement the current sensory test (see Table XLVII).

    G. COLOR

    Fermented shoyu consists of many kinds of enzymatic intermediate degradates of raw materials which are unstable under heat or oxidation and which react with each other. By comparison, the chemical hydrolysate of plant protein, to which almost all materials are ultimately degraded by strong hydrolysis with HCl at more than IOO'C, is more stable under these conditions. The stabilization of fermented shoyu under heat and oxidation is therefore a very difficult but impor- tant problem.

    One of the best ways to prevent fermented shoyu from deteriorating chem- ically is to convert it into a powdered form by dehydration. However, its low boiling point, labile chemical components, and nonvolatile hygroscopic compo- nents, such as glycerol and lactic acid, make dehydration a difficult research problem.



    Compound Molecular Molecular Reference

    weight formula numbera

    I. Hydrocarbons (37) 1. Benzene 2. Toluene 3. Styrene 4. o-Xylene 5. m-Xylene 6. p-Xylene 7. Ethylbenzene 8. Mesitylene 9. 1,2,3-Trimethylbenzene

    10. 1,2,4-Trimethylbenzene 11. I -Ethyl-2-methylbenzene 12. Cumene 13. Naphthalene 14. 4-Methylindan 15. 5-Methylindan 16. 1,2,3,4Tetrahydronaphthalene 1 7. 1 -Ethyl-2,3-dimethylbenzene 18. l-Ethyl-2,4-dimethylbenzene 19. l-Ethyl-3,5-dimethylbenzene 20. 2-Ethyl-l,3-dimethylbenzene 21. 2-Ethyl-I ,4-dimethylbenzene 22. 2-Ethyl-l,2-dimethylbenzene 23. l-Methyl-2(or 4)-propylbenzene 24. 1,2,3,5-Tetramethylbenzene 25. 1,2,4,5-Tetramethylbenzene 26. 1.2-Diethylbenzene 27. 1,3-Diethylbenzene 28. 1 ,CDiethylbenzene 29. Butylbenzene 30. Cyclohexylcyclohexane 3 I . I -Methylnaphthalene 32. 2-Methylnaphthalene 33. 2,3,5(or 6)-Trimethylnaphthalene 34. Tetradecane 35. Pentadecane 36. Hexadecane 37. 5-Phenyldodecane

    II. Alcohols (30) 1. Methanol 2. Ethanol

    78 92

    104 106 106 106 106 1 20 120 120 120 1 20 128 132 132 132 134 134 134 134 134 134 134 134 134 134 134 134 134 134 I42 142 170 198 212 226 246

    32 46

    1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

    2, 3



    TABLE XLVII (Continued)

    Compound Molecular Molecular Reference

    weight formula numberu

    11. Alcohols (30) (continued) 3. I-Propen-3-01 4. 2-Prooen-1-01 (ally1 alcohol) 5 . I-Propanol 6. 2-Propanol 7. 2-Methyl- 1-propanol

    9. 2-Methyl-2-buten-1-01 8. I-Butanol

    10. I-Penten-3-01 11. 3-Penten-2-01 12. 2-Methyl-I-butanol 13. 3-Methyl-I-butanol 14. 1-Pentanol 15. 3-Pentanol 16. 3-Buten-I ,2-diol 17. 2-Ethoxyethanol 18. ~-2,3-Butanediol 19. meso-2,3-Butanediol 20. (E)-2-Hexen-l-ol 21. 1-Hexanol 22. Benzyl alcohol 23. 2,3-Dimethyl-2-pentanol 24. 2,4-Dimethyl-3-pentanol 25. 3-Methyl-3-hexanol 26. 2-Phenylethanol 27. I-Octen-3-01 28. 5-Nonanol 29. 2-Phenyl-I-butanol 30. 2-Undecanol

    111. Esters (41) 1. Methyl acetate 2. Ethyl formate 3. Ethyl acetate 4. 2-Oxopropyl acetate (acetol acetate) 5 . Ethyl propionate 6. Butyl formate 7. 1 -Methylpropyl acetate 8. 2-Methylpropyl acetate 9. Butyl acetate

    10. Ethyl 2-hydroxypropanoate (ethyl lactate)

    1 I . 3-Methylbutyl acetate 12. Pentyl acetate

    58 58 60 60 74 74 86 86 86 88 88 88 88 88 90 90 90

    100 102 108 I16 116 116 122 128 144 150 172

    74 74 88

    100 102 I02 116 116 I16 118

    130 130

    4 9 2 5 , 6 5 , 3 2 1 2 1 2, 7 2, 3 2 2 1 8 2, 10, I 1 2, 10, 11 8 3 5 1 1 1 5 , 12 8 1 I 3

    4 13, 4 5 , 14, 15 2 2 8 2 1 6 5 , 16

    17, 5 18


    TABLE XLVII (Continued)

    Compound Molecular

    weight Molecular formula

    Reference numbera

    ~ ~ ~~

    111. Esters (41) (conrinued) 13. 2-Methylpropyl propionate 14. Ethyl 3-methylbutanoate 15. Ethyl 2-methylbutanoate 16. Ethyl pentanoate 17. 2-Ethyoxyethyl acetate 18. Ethyl 4-oxopentanoate (ethyl levulinate) 19. Ethyl hexanoate (ethyl caproate) 20. Diethyl oxalate 21. 2-Phenylethyl formate 22. Ethyl benzoate 23. Diethyl malonate 24. 2-Phenylethyl acetate 25. Ethyl phenylacetate 26. 3-Methylbutyl 3-methylbutanoate 27. Ethyl octnoate (ethyl caprylate) 28. Diethyl maleate 29. Diethyl succinate 30. Ethyl 3-phenylpmpenoate

    (ethyl cinnamate) 31. Pentyl hexanoate (amyl caproate) 32. Ethyl nonanoate (ethyl pelargonate) 33. 2-Phenylethyl butanoate 34. 4-Formyl-2-methoxyphenyl acetate

    (vanillin acetate) 35. Ethyl 4-hydroxy-3-methoxybenzoate

    (ethyl vanillate) 36. Ethyl dodecanoate (ethyl laurate) 37. Ethyl tetradecanoate (ethyl myristate) 38. Ethyl hexadecanoate (ethyl palmitate) 39. Ethyl 9.12-octadecadienoate

    40. Ethyl 9-octadecenoate (ethyl oleate) 41. Ethyl octadecanoate (ethyl stearate)

    (ethyl linoleate)

    IV. Aldehydes (15) 1. Acetaldehyde 2. Propanal 3. 2-Methylpropanal 4. Butanal 5 . 2-Methylbutanal 6. 3-Methylbutanal 7. Pentanal 8. Hexanal

    130 I30 130 130 132 144 144 146 150 150 160 164 164 172 I72 172 174 176

    186 186 192 194


    228 256 284 308

    310 312

    44 58 72 72 86 86 86


    c 14H2802 C16H3202

    c I i d 3 6 0 2 C20H3602

    1 19, 20 9 17

    6 19 48 1 5 , 19, 20 6 5 5 17 49 6 5 , 6 1

    17 17 1 2

    2. 7

    17 17, 5 21, 22 22

    22 21

    5 , 14, 23, 24 5 , 24 5 , 25, 24 24, 26 1 5 , 14, 25, 24 27 17, 2



    TABLE XLVII (Conrinued)

    Compound Molecular Molecular Reference

    weight foxmula numbera

    IV. Aldehydes (15) (continued) 9. Benzaldehyde

    10. 2.3-Dihydro-4H-pyran-2-carbaldehyde 11. Phenylacetaldehyde 12. 3-Phenyl-2-propenal (cinnamaldehyde) 13. 2,5-Dimethyl-2,3dihyd~1-5H-pyran-2-

    carbaldehyde 14. 2-Methyl-3-phenyl-2-propenal

    (a-meth ylcinnamaldeh yde) 15. 4-Hydroxy-3-methoxybeenzaldehyde

    (vanillin) V. Acetals (4)

    1. 1, I -Diethoxyethane 2. I , I-Diethoxy-3-methylbutane 3. 1 ,I-Diethoxy-2-methylbutane 4. I , I-Diethoxy-4-methyl-2-pentanol

    VI. Ketones (19) 1. Acetone 2. 2-Butanone 3. Hydroxyacetone (acetol) 4. 2.3-Butanedione (diacetyl) 5. 3-Hydroxy-2-butanone (acetoin) 6. 2-Cyclohexin-I-one 7. 4-Methyl-3-penten-2-one 8. 4-Methyl-2-pentanone 9. 2-Hexanone

    10. 2,3-Pentanedione 1 1. 3-Hydroxy-2-pentanone 12. 2-Hydroxy-3-methyl-2-cyclopenten-I-

    one (cyclotene) 13. 5-Methyl-2-hexanone 14. 2,3-Hexanedione 15. Acetophenone 16. 3-Octanone 17. 2,6-Dimethyl-4-heptanone 18. 2-Methyl-3-octanone 19. 3-Methyl-3-decen-2-one

    VII. Acids (24) 1. Formic acid 2. Acetic acid 3. Propionic acid 4. (E)-ZButenoic acid (crotonic acid) 5. 2-Methylpropanoic acid

    (iso-butyric acid)

    106 112 I 20 132 140



    118 160 160 190

    58 72 74 86 88 96 98

    100 100 100 102 112

    I I4 1 I4 120 128 142 I42 I68

    46 60 74 86 88

    5, 20 1 5 17 2


    3, 26

    2, 28 29 9 29

    5 I 2 2, 30 5, 14, 11 33 I 1 5 2, 30 31 2

    1 5 , 30 2 2 1 I 33

    19 5, 32, 35 2, 19 2 17, 5


    TABLE XLVII (Continued)

    Compound Molecular Molecular Reference

    weight formula numbera

    VII. Acids (24) (continued) 6. Butanoic acid (sec-butyric acid) 7. 2-Oxopropanoic acid (pyruvic acid) 8. 2-Hydroxypropanoic acid (lactic acid) 9. 2-Methyl-2-butenoic acid

    10. 2-Methylbutanoic acid 11. 3-Methylbutanoic acid 12. Pentanoic acid (n-valeric acid) 13. 2-Oxobutanoic acid (2-ketobutyric acid) 14. 4-Methylpentanoic acid 15. Hexanoic acid (caproic acid) 16. 4-Oxopentanoic acid (levulinic acid) 17. Butanedioic acid (succinic acid) 18. Benzoic acid 19. Phenylacetic acid 20. Octanoic acid (caprylic acid) 21. Dodecanoic acid (lauric acid) 22. Hexadecanoic acid (palmitic acid) 23. 9,12-Octadecadienoic acid

    24. 9-Octadecenoic acid (oleic acid) (linoleic acid)

    VIII. Phenols (17) 1. Phenol 2. 1,2-Benzenediol (pyrocatechol) 3. 4-Vinylphenol 4. 4-Ethylphenol 5. 2-Methoxyphenol (guaiacol) 6. 4-(2-Hydroxyethyl)phenol (tyrosol) 7. 4-Ethyl-l,3-benzenediol

    (4-ethylresorcinol) 8. 4-Hydroxybenzoic acid 9. 2-Methoxy-5-vinylphenol

    10. 4-Ethyl-2-methoxyphenol (4-ethylguaiacol)

    11. 2,6-Dimethoxyphenol 12. 3,4-Dihydroxybenzoic acid

    13. 3-(4-Hydroxyphenyl)propenoic acid (protocatechuic acid)

    @-coumaric acid), @-hydroxy- cinnamic acid)


    (vanillic acid)

    14. 4-Hydroxy-3-methoxyacetophenone

    15. 4-Hydroxy-3-methoxybenzoic acid

    138 C7H603 150 CgHlo2 152

    166 CgHlo3

    168 CsH804

    5 , 19 24 17 2 2 5 , 19 17, 2 34 17, 2 2, 19 17 17 5, 19 17, 2, 19 17 17 35, 36, 22 22


    2 2 33 5 , 12 5 37 1

    12 2 5 , 35

    5 19


    2, 19

    35. 38



    TABLE XLVII (Continued)

    Molecular Molecular Reference Compound weight formula numbera

    VIE Phenols (17) (continued) 16. 3-(4-Hydroxy-3-methoxyphenyl)

    propenoic acid (ferulic acid) 17. 4-Hydroxy-3,5-dimethoxybenzoic acid

    (syringic acid) IX. Furans (16)

    1. Furan 2. 2-Methylfuran 3. 2-Furfural 4. Furfuryl alcohol 5. Tetrahydrofurfuryl alcohol 6. 1 -(2-Furyl)- 1 -ethanone

    (2-fury1 methyl ketone) 7. 5-Methyl-2-furfural 8. 1-(2-Tetrahydrofuryl)-l-ethanone

    (2-tetrahydrofuryl methyl ketone) 9. 1-(2-Furyl)-l-propanone

    (ethyl 2-fury1 ketone) 10. 2-furfuryl acetate 11. 1-(3-Hydroxy-2-furyl)-l-ethanone

    12. 5-Hydroxymethyl-2-furfural 13. 1 -(2,5-Dimethyl-3-furyyl)- 1 -ethanone 14. Ethyl 2-furoate 15. 3-Phenylfuran 16. 2-Pcopenyl 2-furoate


    X. Lactones (4) 1. 4-Butanolide 2. 2-Penten-4-olide [5-methyl-2(5H)-

    3. 2-Methyl-4-butanolide 4. 4-Pentanolide

    1. 3-Methyl-2(5H)-furanone 2. 2-Methyl-3-tetrahydfuranone 3. 4-Hydroxy-5-methyl-3(2H)-furanone 4. 4-Hydroxy-2,5-dimethyl-3(2H)-

    5. 4,5-Dihydro-5-(l-hydroxyethyl)-2(3H)

    6. 4-Hydroxy-2(or 5)-ethyl-5-(or 2)-

    furanone] (P-angelicalactone)

    XI. Furanones (6)




    1 94


    68 82 96 98

    102 110

    110 114


    126 126

    126 138 140 144 152

    86 98

    100 100

    98 100 114 128





    4 4 5 , 14, 40 5, 41 2 8, 5

    2 2


    5 2

    42 1 1 1 2

    2 2

    2 5

    33 5 43 2




    TABLE XLVII (Continued)

    Compound Molecular Molecular Reference weight formula numbera

    XII. Pyrones ( 5 ) I . 3-Hydroxy-2-methyl-4H-pyran-4-one

    2. S-Hydroxy-2-rnethyl-4H-pyran-4-one 3. 3-Methoxy-2-methyl-4H-pyran-4-one 4. 3,5-Dihydroxy-2-methyl-4H-pyran-4-

    5. 3,5-Dihydroxy-6-methy1-2,3-dihydro-



    4H-pyran-4-one XIII. Pyrazines (27)

    1. Pyrazine 2. 2-Methylpyrazine 3. 2.3-Dimethylpyrazine 4. 2,5-Dimethylpyrazine 5. 2.6-Dimethylpyrazine 6. 2-Ethylpyrazine 7. SH-Cyclopenta[ blpyrazine 8. 2-Methyl-6-vinylpyrazine 9. 6,7-Dihydro-SH-cyclopenta[b]pyrazine

    10. 2,3,5-TrimethyIpyrazine 1 1. 2-Ethyl-5-methylpyrazine 12. 2-Ethyl-6-methylpyrazine 13. 2(0r 3)-Methyl-5H-cyclopenta[b]-

    14. 6-Methyl-5H-cyclopenta[b]pyrazine 15. 7-Methyl-5H-cyclopenta[b]pyrazine 16. Pymlo[ I ,2-a]-3-methylpyrazine 17. 2-Methyl-6,7-dihydro-SH-cyclo-

    penta[ b]-pyrazine 18. Tetramethylpyrazine 19. 3-Ethyl-2,5-dimethylpyrazine 20. 2,3-Diethylpyrazine 21. 2,6-Diethylpyrazine 22. 2(or 3),6(or 7)-Dimethyl-SH-cyclo-

    penta[ blpyrazine 23. Pymlo[ 1,2-a]-3,4-dimethylpyrazine 24. 2-Ethyl-6,7-dihydro-5H-cyclopenta[b]-

    25. 2-Ethyl-3,5,6-trimethylpyrazine 26. 2.6-Diethyl-3-methylpyrazine 27. 2,3,5-TrimethyI-6,7-dihydro-SH-cyclo-



    penta[ blpyrazine


    126 140 142


    80 94 I08 108 108 108 118 120 120 122 I22 122 132

    132 132 132 134

    I36 136 I36 I36 146

    I46 148

    150 I50 162

    5, 13

    2 2 2


    28 5 , 28 5 , 28 28 5 , 28 28 28 2 28 28 28 5, 28 28

    28 28 33 28

    5. 28 5. 28 28 28 28

    33 28

    28 28 28



    TABLE XLVII (Conrinued)

    Compound Molecular

    weight Molecular formula

    Reference numbera

    XIV. Pyridines (7) 1. Pyridine 79 2. 3-Methylpyndine 93

    5. 2-Pyridylmethanol 109 6. 3-Methoxypyridine 109

    3. 2,6-Dimethylpyridine 107 4. 2-Ethylpyridine 107

    7. Ethyl 3-pyridinecarboxylate 151 (ethyl nicotinate)

    XV. Miscellaneous nitrogen-containing compounds (8) 1. 1-Methyl-2-pyrrolidinone 99 2. I - (2-~lyl)- l -e thanone 109 3. Benzoxazole 119 4. 1-(5-Methyl-2-pyrrolyI)-I ethanone 123 5 . 1.5-Dimethyl-2-pyridone I23 6. 2-Methylbenzoxazole I33 7. Ethyl 2-pymlidone-5-carboxylate 157 8. Ethyl 2-(acetylamino)-4-methyl-penta- 201

    noate (N-acetylleucine ethyl ester) XVI. Sulfur-containing compounds (16)

    1. Hydrogen sulfide 34 2. Methanethiol 48 3. Dimethyl sulfide 62 4. Ethanethiol 62

    6. Thiophene 84 5. 2-Ropene- 1 -thiol 74

    7. Dimethyl disulfide 94 8. 4-Methyl-l,3-oxathiolane 104 9. 3-Methylthiopropanal (methional) 104

    10. 3-(Methylthio)-l-propanal (methionol) I 0 6 I 1. Phenylmethanethiol 124 12. Dimethyl trisulfide 126 13. 3,4-Dimethyl-2,5dihydmthiophen-2- 128

    one 14. 2-Ethyl-6-methyl- 1,3-oxathiane 146 15. 3-(Methylthio)propyl acetate 148 16. I , I-Bis(methylthio)-2-methylpropane 150

    I . 2-Ethoxythimle 129 2. 2-Butoxythiazole 157 3. N-Acetyl-IH-benzothiazol I79 4. 2-(Methy1thio)benzothiazole 181

    1. Borneo1 154

    XVII. Thiazoles (4)

    XWI. Terpenes (3)

    28 28 28 28 1 28 28

    1 8, 5, 28 28 1 1 2, 28 33 33

    46 29 4 24 17 33 2 33 46 5. 16, 47 1 1 33

    1 1 24

    1 1 33 1



    TABLE XLVII (Continued)

    Molecular Molecular Reference Compound weight formula numberu

    XVIII. Terpenes (3) (continued) 2. 4-Methyl-2-(2-methyl- I -propenyl)- 154 CloH180 1

    tetrahydropyran (cis-rose oxide) 3. Bornyl acetate 1% C12H2002 5

    2. P-Methoxystyrene 134 CgHlo0 1

    XIX. Miscellaneous compounds (3) 1. I ,rl-Dioxane 88 C4Hs02 33

    3. I ,5-Dimethoxynaphthalene 188 C12H1202 1

    a References: 1: Nunomura er 01. (1979); 2: Nunomura er 01. (1980); 3: Shoji (1936); 4: Sasaki and Nunomura (1979); 5: Nunomura et al. (l976a); 6: Morimoto and Murakami (1966); 7: Yamada and Goan (1969); 8: Goto (1973); 9 Sasaki and Nunomura (1981); 10: Taira (1926a); 11: Tomiyasu (1927); 12: Asao er al. (1967); 13: Kihara (1940); 1 4 Akabori (1936); 15: Taira (1926b); 16: Akabori and Kaneko (1936); 17: Ishizu (1969); 18: Yokotsuka (1951a); 19: Yokotsuka (1953b); 20: Yokot- suka (1951~); 21: Fukai (1929); 22: Yokotsuka (1951b); 23: Yamada (1928); 24: Yokotsuka (1953~); 25: Nakajima and Takei (1949); 26: Yokotsuka (1954); 27: Ikeda and Kawaguchi (1922); 28: Nunomura er 01. (1978); 29: Yokotsuka (1950); 3 0 Asao and Yokotsuka (1961b); 31: Asao and Yokotsuka (1963); 32: Matsumoto (1921); 33: Nunomura and Sasaki (1981); 34: Yokotsuka and Asao (1961); 35: Yokotsuka (1953a); 36: Yokotsuka (1953d); 37: Yukawa (1917); 38: Yokotsuka (1956); 3 9 Asao and Yokotsuka (1958a); 40: Shoji and Onuki (1932); 41: Morimoto and Murakami (1967); 42: Yokotsuka (1949); 43: Nunomura er al. (1979); 44: Nunomura er al. (l976b); 45: Kosuge er al. (1971); 4 6 Nunomura and Sasaki (1982); 47: Morimoto and Murakami (1966); 48: Ishizu ( I 963).


    The author wishes to express his sincere appreciation to his colleagues, Dr. M. Sasaki, Mr. N. Nunomura, Dr. Y. Asao, Dr. A. Okuhara, Dr. H. Hashimoto, Mr. T. Iwasa, Mr. T. Sakasai, Dr. F. Noda, Dr. S. Ishi, Dr. T. Kikuchi, Dr. S. Sugiyama, Mr. K. Oshita, Mr. A. Yasuda, and Dr. K. Hayashi for their valuable contributions and cooperation in the preparation of this work. I am also grateful to Dr. K. Sakaguchi, Professor Emeritus of Tokyo University, and Dr. T. Obara, Professor Emeritus of Tokyo University of Education, for their kind guidance and encouragement. Finally, I wish to express my hearty thanks to Chancellor Emeritus Dr. E. M. Mrak, Professor F. S. Stewart, and Professor B. S. Schweigert of the University of California at Davis, and to Dr. C. 0. Chichester, Vice President of the Nutrition Foundation, Inc., New York, for their assistance in publishing this review.


    Abe, T., Saito, Y., Fujita, E., and Kataoka, J. 1975. Studies on shyu koji (I). NSI and material balance of shoyu-koji making. J . Japan Soy Sauce Res. I m . 1, 195-200.


    Aiba, T. 1982. Modernized heat treatment of wheat kernels used for brewing. Proc. 14th Symp.

    Aibara, K., and Miyaki, K. 1965. Qualitative and quantitative analysis of aflatoxin. Proc. J . Agric.

    Aishima, T. 1979a. Evaluation and discrimination of soy sauce by computer analysis of volatile

    Aishima, T. 1979b. Classification of soy sauce on principal components in GC profiles. Agric. Biol.

    Aishima, T. 1979c. Objective evaluation of soy sauce by statistical analysis of GC profiles. Agric.

    Aishima, T., and Nobuhara, A. 1976. Evaluation of soy sauce flavor by stepwise multiple regression

    Aishima, T., and Nobuhara, A. 1977. Statistical analysis for soy sauce flavor and sensory evalua-

    Ajinomoto Co., and Shimazaki, H. 1970. Japanese patant announcement. Sho45-30184. Akabori, S. 1936. On the flavor components of shoyu. I. J . Chem. SOC. Japan 57, 828-831. Akabori, S. , and Kaneko, T. 1936. On the flavor components of shoyu. 11. Isolation and synthesis of

    a sulfur containing flavor compound from shoyu. J . Chem. SOC. Japn 57, 832-836. Akabori, S . , and Kaneko, T. 1937. Flavor components of shoyu. 111. On the fhnethylmercapto-

    propionaldehyde. J. Chem. SOC. Japan 58, 236-237. Akao, T., and Okamoto, Y. 1983. Cultivation of microorganisms in an air-solid fluidized bed. Inr.

    Conf. Fluid., 4th Eng. Found. Conf. SOC. Chem. Eng. Japan, Kashikojima, Mieken June 2 . Akao, T., Sakasai, T., Sakai, M., Takano, A., and Yokotsuka, T. 1972. A new plant for continuous

    production of koji in the shoyu industry. Proc. First Pac. Chem. Eng. Congr., Ist, Kyoto, Oct.

    Akao, T., Yasuda, A., Watanabe, H., and Yuasa, T. 1980. Method of explosive puffing of cereal powder. Patent announcement, Japan, Sho55-42814. Kikkoman Corp.

    Akao, T., Nagata, S . , Osaki, M., and Okamoto, Y. 1982. Production method of liquid seasoning. Japanese patent announcement, Sho60-6184.

    Aonuma, T., Yasuda, A,, Yuasa, T., Arai, A., Mogi, K., and Yokotsuka, T. 1970. Preparing method of shoyu and miso. Japanese patent 794915. Kikkoman Shoyu Co. Ltd.

    Asao, Y., and Yokotsuka, T. 1958a. Studies on flavorous substances in soy (16). Flavorous sub- stances in raw soy. J . Agric. Chem. SOC. Japan 32, 617-622.

    Asao, Y., and Yokotsuka, T. 1958b. Studies on flavorous substances in soy (17). Guaiacyl com- pounds in soy (1). J . Agric. Chem. SOC. Japan 32, 622-628.

    Asao, Y., and Yokotsuka, T. 1961a. Studies on flavorous substances in shoyu (18). Isolation of C7H1205 and its structure. J . Agric. Chem. SOC. Japan 35, 831-837.

    Asao, Y., and Yokotsuka, T. 1961b. Studies on flavorous substances in shoyu (20). a-Diketon compounds (2). J . Agric. Chem. SOC. Japan 35, 1211-1218.

    Asao, Y., and Yokotsuka, T. 1963. Studies on flavorous substances in shoyu (23). a-Diketon compounds (2). J . Agric. Chem. SOC. Japan 37, 563-568.

    Asao, Y., and Yokotsuka, T. 1977. Flavorous components of shoyu. I n Chemical Components in Fermented Foods, pp. 367-381. Brew. Soc. Japan, Takinogawa 2-6, Kitaku, Tokyo.

    Asao, Y., Sakasai, T., and Yokotsuka, T. 1967. Studies on flavorous substances in shoyu (26). Flavorous compounds produced by yeast fermentation (2). J. Agric. Chem. SOC. Japan 41,434- 441.

    Asao, Y., Sakasai, T., and Yokotsuka, T. 1969. The role of yeasts,for the production of shoyu flavor. J. Ferment. Technol. 47, 318-326.

    Baker, J. C., Parker, H. K., and Fortmann, K. L. 1953. Flavor of bread. Cereal Chem. 30,22-30.

    Brew. SOC. Japan p. 20.

    Chem. SOC. Japan p. 86.

    profiles. Agric. Biol. Chem. 43, 1171-1718.

    Chem. 43, 1905-1910.

    Biol. Chem. 43, 1935-1943.

    analysis of gaschromatographic profiles. Agric. Biol. Chem. 40, 2159-2167.

    tion. Agric. Chem. SOC. Japan 51, 65-74.

    10-14 p. 127.


    Blockman, H., und Franck, B. 1955. Bouillon-Geruch der Protein-hydrolyzate. Angew. Chem. 67,

    Bo, T. A. 1982. Origin of chiang and chian-yu and their production technology ( I and 2). Brew SOC.

    Bo, T. A. 1984. Origin of tao-shi and its production technology ( 1 and 2). Brew Soc. Japan 79,221-

    Brand, J. 1894. Ber. Dtsch. Chem. Ges. 27, 816. Buchanan, R. E., and Gibbons, N. E. 1974. Bergeys Mannual of Determinative Bacteriology, p.

    515. Williams & Wilkins, Baltimore. Bureau of Foods, Japan. 1976. Standard Production and Circulation of shoyu (H. Ebine, E.

    Kimura, K. Fukuzaki, T. Furuta, H. Maeda, and T. Yokotsuka, eds.), p. 97. Burton, H. S., McWeeny, D. J., and Biltcliffe. 1963. Non-enzymatic browning: The role of unsatu-

    rated carbonyl compounds as an inhibitor of browning. J . Sci. Food Agric. 14, 91 1-920. Buttery, R. G., Seifert, R. M., Lundin, R. E., Guadagni, D. G., and Ling, L. C. 1969. Charac-

    terization of an important aroma component of bell peppers. Chem. Ind. 490. Diemar, W., and Hala, H. 1959. Contribution to the formation of 2-methyl-3-hydroxypyrone (Malt-

    01). 2. Lebensm. Unters. Forsch. 110, 161-168. Dunn, G., Neuboid, G. T., and Spring, F. S. 1949. Synthesis of Flavacol, a metabolic product of

    Aspergillus jlavus. J . Chem. SOC. 2586-2587. Dutcher, J. D. 1957. Aspergillic acid, an antibiotic substance produced by Aspergillus jlavus. 111.

    The structure of hydroxyaspergillic acid. J . Biol. Chem. 232, 785-795. Ebine, H. 1968. Application of rhizopus for shoyu manufacturing. Season. Sci. 15, 10-18. Ebine, H., Kimura, E., Fukuzaki, K., Furuta, T., Maeda, H., and Yokotsuka, T. 1976. Standard

    Production and Circulation of Shoyu, p. 73. Bureau of Food, Japan. Eguchi, U. 1977. HTST cooking of soybeans utilizing NK-cooker. J. Brew. SOC. Japan 72, 250-

    255. Fujimoto, H., Aiba, T., and Goan, M. 1980. Studies on lactic acid bacteria in shoyu mash. Part 3.

    Effect of temperature on the growth and fermentation of lactic acid bacteria. J. Japan Soy Sauce Res. Inst. 6, 5-9.

    Fujita, M. 1961. Flavoring method of shoyu and resembling seasonings. Japanese patent announce- ment. Sho35- 14485, Sho36- 14295, Sho36- 14296, Sho36- 14297.

    Fujita, A., and Kishi, K. 1975. Characteristic of Koji Mugi (granular substances for soy sauce use). J. Japan Soy Sauce Res. Inst. 1, 28-30.

    Fujita, A., Watanabe, Y., Kishi, K., and Ogawa, G. 1977. The effect of granular distribution of starchy materials on their Koji qualities and soy sauce. J. Japan Soy Sauce Res. Inst. 3, 200- 204.

    Fukai, T. 1929. The relation between soy sauce fermentation and fats and oil. 1. On the ethylesters of fatty acids and free acids in soy sauce oil. J . Agric. Chem. SOC. Japan 5 , 458-470.

    Fukushima, D., and Mogi, M. 1955. Treating method of soybean protein. Japanese parenr 236368, 237805.

    Fukushima, D . , and Mogi, M . 1957. Treating method of soybean protein. Japanese patent 248103. Fukuzaki. K . 1972. Naturally fermented shoyu. Shokuno Kagaku 7 , 47-59. Furuta, T., and Ohara, K. 1954. The effect of iron salts on the color tone and intensity of shoyu.

    Furuya, T., Ishige, M., Uchida, K., and Yoshino, H. 1983. Koji mold breeding by protoplast fusion

    Gianturco, M. A., and Friedel, P. 1963. The synthesis of some cyclic diketones isolated from coffee.


    Japan 77, 365-371, (7), 439-445.

    223, (6). 395-402.

    Season. Sci. 2, 6-12.

    for soy sauce production ( I ) . J . Agric. Chem. SOC. Japan 57, 1-8.

    Tetrahedron 19, 2039-2049.


    Gianturco, M. A., Giammarino, A. S., and Pitcher, R. G . 1963. The structures of five cyclic diketones isolated from coffee. Tetrahedron 19, 205 1-2059.

    Gianturco, M. A,, Friedel, P., and Giammarino, A. S. 1964. The volatile constituents of coffee. 111. The structures of two heterocyclic compounds and the synthesis of tetrahydrofuranones. Tetrahedron 20, 1763- 1772.

    Goan, M. 1969. Glutamic acid content in shoyu. Season. Sci. 5 , 1-5. Goto, S. 1973. Constituents and composition of aroma concentrate from soy sauce (I). On neutral

    Guadagni, D. G., Buttery. R. G., and Okano. 1%3. Odour threshold of some organic compounds

    Haga, H. 1%8. Three-day Koji cultivation and cold mash making in shoyu brewing. J . Brew. SOC.

    Haga, H., Sasaki, S., Nakamura, K., and Umeda, 1. 1967. Studies on the new Koji manufacturing

    Haga, H., Fujii, T., Moriguchi, S., Umeda, I., and Shimizu, K. 1970. The report of test brewing

    Hanaoka, Y. 1976. Microbial quality change of shoyu and its prevention. J . Japan Soy Sauce Res.

    Hashiba, H. 1971. Studies on browning compound in shoyu (soy sauce). J . Agric. Chem. Soc. Japun

    Hashiba, H. 1973a. Studies on melanoidin from soy sauce. J. Agric. Chem. SOC. Japan 47, 727- 732.

    Hashiba, H. 1973b. Non-enzymatic browning of soy sauce. Use of ion exchange resins to identify types of compounds involved in oxidative browning. Agric. Biol. Chem. 37, 871-877.

    Hashiba, H. 1973c. Non-enzymatic browning of soy sauce decolorized by ultrafiltration. J. Agric. Chem. SOC. Japan 47, 289-291.

    Hashiba, H. 1974. Effect of aging on the oxidative browning of sugar-amino acid model systems. Agric. Biol. Chem. 38, 551-555.

    Hashiba, H. 1975. A glucose-diglycine condensation product participating in oxygen-dependent browning. J. Agric. Food Chem. 23, 539-542.

    Hashiba, H. 1976. Participation of Amadori rearrangement products and carbonyl compounds in oxygendependent browning of soy sauce. J. Agric. Food. Chem. 24, 70-73.

    Hashiba, H. 1978. Isolation and identification of Amadori compounds from soy sauce. Agric. Biol. Chem. 42, 763-768.

    Hashiba, H. 1981. Oxidative browning of soy sauce. Part 111. Effect of organic acids, peptides and penlose. Part IV. Oxidative browning of Amadori compounds. Part V. Isolation and identifica- tion of Amadori compounds from soy sauce. J . Japan Soy Sauce Res. Inst. 7, 19-23, 116-120,

    Hashiba, H., and Abe, K. 1984. Oxidative browning and its control with the emphasis on shoyu. Proc. Annu. Meet. Agric. Chem. Soc. Japan p. 708.

    Hashiba, H., Koshiyama, I., Sakaguchi, K., and Iguchi, N. 1970. Effects of iron and some trace metal ions in shoyu on the increase of color of shoyu. J . Agric. Chem. SOC. Japan 44,3 12-3 16.

    Hashiba, H., Koshiyama, I., and Fukushima, D. 1977. Oxidative browning of Amadon compounds from amino acids and peptides. In Protein Crosslinking-B. Nutritional and Medical Conse- quences (M. Friedman. ed.), pp. 419-448. Plenum, New York.

    fraction. Nurr. Foods 26, 135-138.

    associated with food flavors. 1. Sci. Food Agric. 14, 761-765.

    Japan 63, 931-936.

    methods (3). Shoyu brewing by low-temperature mashing. Sason. Sci. 14, 1-11.

    using American wheat. Season. Sci. 17, 1-29.

    Inst. 2, 221-226.

    45, 29-35.

    12 1- 124.

    Hashimoto, Y. 1960. Kanro shoyu. Brew. Id. (special issue of Food Id . ) 204-206. Hashimoto, H., and Yokotsuka, T. 1971. Studies on sediment of shoyu (5 ) . Composition of sedi-

    ments fonned by heating of raw shoyu and isolation of precursors of sediment. J . Ferment. Technol. 49, 642-649.


    Hashimoto, H., and Yokotsuka, T. 1972. Studies on sediment of shoyu (6). Sediment forming activity in fermented raw shoyu. J . Ferment. Technol. 50, 257-263.

    Hashimoto, H., and Yokotsuka. T. 1973. The significant effect of thermostable acid-protease on sediment formation of shoyu. J. Ferment. Technol. 51, 661-669.

    Hashimoto, H., and Yokotsuka, T. 1974. Mechanisms of sediment formation during heating of raw shoyu. J. Ferment. Technol. 52, 328-334.

    Hashimoto, H., and Yokotsuka, T. 1976. Mechanism of sediment formation during heating of raw shoyu. J. Brew. Soc. Japan 71,496-499.

    Hashimoto, H., Yoshida, H., and Yokotsuka, T. 1970. Removal of iron from soy sauce. J. Ferment. Technol. 48, 50-55.

    Harada, Y., and Kawaguchi, S. 1968. Dealing method of soybeans for soy sauce brewing. Applica- tion for Japanese patent. Sho47-26708. Nihonmamten Shoyu Co. Ltd.

    Harada, Y., and Sotoyama, N. 1966. Application of cellulase in brewing. (IV). Isolation of soybean protein with cellulase and a cell separating enzyme. J. Ferment. Technol. 44, 835-841.

    Hayashi, K., and Kawase, K. 1970. New synthetic method of maltol and its application as flavor enhancer. New Food Indu. 12, 20-23.

    Hayashi, K., Terada, M., Mogi, K., and Yokotsuka, T. 1%8a. Japanese patent 673114. Hayashi, K., Terada, M., Mogi, K., and Yokotsuka, T. 1%8b. Japanese patent 687804. Hayashi, K., Terada, M., Mizunuma, T., and Yokotsuka, T. 1979. Retarding effect of acetic acid on

    growth of contaminated bacteria during shoyu Koji making process. J . Food Sci. 44,359-362. Hayashi, K., Terada, M., and Mizunuma, T. 1981. Some differences in characteristics of soy sauce

    fermentation between species of Koji-molds, Aspergillus sojae and Aspergillus oryzae. J. Japan Soy Sauce Res. Inst. 7, 166-172.

    Hesseltine, C. W., Shotwell, 0. L., Ellis, J. J., and Stublefield, R. D. 1966. Aflatoxin formation by Aspergillus jlavus. Bacteriol. Rev. 30, 795-805.

    Hodge, J. E. 1953. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1,

    Hodge, 1. E., Fisher, B. E., and Nelson, E. C. 1963. Dicarbonyls, reductones, and heterocyclics produced by the reactions of reducing sugars with secondary amine salts. Am. Soc. Brew. Chem. Proc. 84-92.

    Hodge, J. E., Mills, F. D., and Fisher, B. E. 1972. Compounds of browned flavor derived from sugar-amine reactions. Cereal Sci. Today 17, 34-38.

    Hugli, T. E., and Mooore, S. 1972. Determination of the tryptophan content of proteins by ion exchange chromatography of alkaline hydrolyzates. J. Biol. Chem. 247, 2828-2834.

    Iguchi, N., and Nasuno, S. 1978. Shoyu brewing and microbiology. J. Ferment. Technol. 56,645- 655.

    Iijima, K., Dejima, M., Tsuji. A., Watanabe, T., and Kagami, T. 1973. A new method for the cooking of soybean meal by a rotary NK cooker. Season. Sci. 20, 13-19.

    Ikeda, K., and Kawaguchi, M. 1922. J. Chem. Soc. Japan 43,950. Imai, S., and Suzuki, K. 1%7. Studies on the soy sauce Koji processed under the automatic control

    (3). The properties of the soy sauce Koji processed under the middle temperature in the early stage. Season. Sci. 14, 22-32.

    Imai, S., Suzuki, K., and Kaneko, M. 1969. The components of soy sauce mash under various temperature in the early stage. Season. Sci. 16, 1-5.

    Inamori, K., Miyauchi, K., and Uchida, K. 1977. Studies on environmental factors involved in fermentation of shoyu mash (I). J. Japan Soy Sauce Res. Inst. 3, 300-313.

    Inamori, K., Miyauchi, K., Uchida, K., and Yoshino, H. 1984. Interaction between Pediococcus halophylus and Saccharomyces rouxii. J. Agric. Chem. SOC. Japan 58, 771-777.

    Ishigama, Y., Ishikawa, H., Ueda, R., and Taketsuru, H. 1%5. Studies on the changes in microflora



    and its metabolisms during Koji making process in soy sauce brewing. (I). The changes of various components with some microorganisms. J . Ferment. Technol. 43, 165- 173.

    Ishigami, Y., Ishikawa, H., Taketsuru, H., and Ueda, R. 1967. Studies on the changes in microflora and its metabolism during the process of making koji in soy sauce brewing. J . Ferment. Technol. 45, 1003-1008.

    Ishii, S., Kikuchi, T., Ogami, T., and Yokotsuka, T. 1972. Effect of degrading enzyme of plant tissue on s h o p production. J . Agric. Chem. SOC. Japan 46, 340-354.

    Ishizu, H. 1963. Gas chromatography of chemical components in soy sauce (11). Variation of odour on heating. Rep. Bukogawa Woman College 11, S-113-120.

    Ishizu, H. 1969. Flavor components of shoyu. Chori Kagaku 2, 156-163. Ito, Y. 1972. Flavor of maltol and its related compounds. Koryo 102, 11-18. Iwasa, T., Fujii, S., Kakinuma, T., and Yokotsuka, T. 1972a. Glutaminase produced by Cryprococ-

    cus albidus ATCC 20293. Purification and properties of enzyme. Annu. Meer. Agric. Chem. SOC. Japan. Sendai, April 2 .

    Iwasa, T., Fujii, S., and Yokotsuka, T. 1972b. Microbial glutaminases and their properties. Inr. Ferment. Symp. 4rh, Kyoro. Mar. 19-25,

    Jose, H., and Sugimori, T. 1973. Studies on Terracoccus sp., the lactic acid bacteria relating to soy sauce brewing. Season. Sci. 20, 19-21.

    Jose, H., Ysukada, Y., and Sugimori, T. 1976. Fermentation of lactic acid in soy sauce. J . Japan Soy Sauce Res. Insr. 2, 172-176.

    Kamata, H., and Sakurai, Y. 1964. Color formation of soy sauce. Season. Sci. 11, 21-26. Kanbe, C., and Uchida, K. 1984. Selection of shoyu-lactobacilli having strong reducing power and

    their application for the stabilization of shoyu color. Proc. 16rh Symp. Brew. SOC. Japan pp. 37-40.

    Kaneko, Y. 1965. New pterin produced by Aspergillus. Agric. Biol. Chem. 29, 965-966. Karrer, P., Salomon, H., Schopp, K., Schlitter, E., and Fritzshe, H. 1934. Helv. Chim. Acra 17,

    1010. Katagiri, J., Abe, T., Fujita, E., and Saito, Y. 1976. Studies on shoyu koji (11). Peptide and amino

    acid composition of shop-koji making. J . Japan Soy Sauce Res. Ins?. 2 , 76-81, Kato, H. 1956. Studies on browning reactions between sugars and amino compounds. Part I. pH

    Dependency of browning interactions between various kinds of reducing sugar and amino compound. Bull. Agric. Chem. SOC. Japan 20, 273-278.

    Kato, H. 1958. Studies on browning reactions between sugars and amino acids. Part Ill. Identifica- tion of red pigments and furfural produced from aromatic amine-N-xylosides. Bull. Agric. Chem. SOC. Japan 22, 85-91.

    Kato, H. 1959. Studies on browning reactions between sugars and amino acids. Part IV. Identifica- tion of 5-hydroxymethylfurfural formed from aromatic amine-N-hexosides and its role in the browning reaction. Bull. Agric. Chem. SOC. Japan 23, 551-554.

    Kato, H. 1960. Studies on browning reactions between sugars and amino acids. Part V. Isolation and characterization of new carbonyl compounds, 3-deoxy-osones formed from N-glycosides and their significance for browning reactions. Bull. Agric. Chem. SOC. Japan 24, 1-12.

    Kato, H., and Sakurai, Y. 1962. Studies on the mechanism of browning of soybean products. Part 11. The role of 3-deoxy-osones, hexose, and pentose in the color increase of pasteurized shoyu. J . Agric. Chem. Soc. Japan 36, 131-137.

    Kato, H., and Sakurai, Y. 1963. Studies on the mechanism of browning of soybean products. Part 111. Meaning of 3-deoxy-osones and furfurals in the browning of pasteurized shoyu. J . A@. Chem. Soc. Japan 37, 423-425.

    Kato, H., Yamada, Y., Isaka, K., and Sakurai, Y. 1961. Studies on the mechanism of browning of soybeans products. Part I. Isolation and identification of 3-deoxy-~-glucosone for shoyu and miso. J . Agric. Chem. SOC. Japan 35, 412-415.


    Kawamori, T. 1940. Brewing method of shoyu from the view point of food problem. J . Brew. 16,

    Kawano, Y. 1938. The ingredient change of soybeans during cooking. J . Brew. 16, 755-767. Kihara, K. 1940. Studies on the flavor of shoyu. 1-111. J. I n d . Chem. 43, 876-883. Kihara, K. 1983. Generation, separation and quantity of malt01 precursor in the growing soybean and

    the relation with water content (1 and 2). J. Japan Soy Sauce Res. Inst. 9, 49-58. Kikuchi, T. 1975. Effect of the acidic polysaccharide in soy sauce on viscosity of soy sauce (Note).

    J. Agric. Chem. Soc. Japan 49, 231-232. Kikuchi, T. 1976. Changes of soybean cell wall polysaccharide during soy sauce fermentation. J .

    Agric. Chem. Soc. Japan 50, 273-277. Kikuchi, T. 1977. Behavior and role of soybean polysaccharides on soy sauce fermentation. J . Japan

    Soy Sauce Res. Inst. 3, 154-159. Kikuchi, T., Sugimoto, H., and Yokotsuka, T. 1976. Polysaccharides in pressed cake and their

    effects on difficulty in press filtration of fermented soy sauce mash. J. Agric. Chem. Soc. Japan

    Kinoshita, R., Ishiko, T., Sugiyama, S., Seto, T., Igarashi, S., and Goetz, I. E. 1968. Mycotoxin in

    Kodama, S. 1922. Chemistry of aldehydes (4). On some new aldehydes in shoyu. J. Chem. Soc.

    Koizumi, K., and Takahashi, K. 1974. The control of fermentation of shoyu mash. Season. Sci. 21,

    Komatsu, Y. 1968. Changes of a few kinds of enzyme in soy sauce brewing (1). Changes of constituent and enzyme activity in soy sauce fermentation by the low temperature mashing. Season. Sci. 15, 10-20.

    Kosuge, T., Zenda, H., Tsuji, K., and Yamamoto, T. 1971. Studies on flavor components of foodstuffs. Part 1 . Distribution of tetramethylpyrazine in fermented foodstuffs. Agric. Biol. Chem. 33,693-696.

    Kurata, H . , Tanabe, H., Kanota, K., Udagawa, S., and Ichinoe, M. 1968. Studies on the population of toxic fungi in foodstuffs. IV. Aflatoxin-producing fungi isolated from foodstuffs in Japan. J .

    Kurata, H., Udagawa, S., Ichinoe, M., Natori, S., Tanabe, H., and Saito, M. 1969. Studies on the population of toxic fungi in foodstuffs. XI. Aflatoxin in foodstuffs produced in the districts of a high rate of gastrocancer. J . Food H y g . 10, 2-3.

    Kurono, K., and Fukai, T. 1928. Studies on the flavorous components of shoyu. J. Agric. Chem. SOC. Japan 4, 361-383.

    Kurono, K., and Katsume, H. 1927. Chemical composition of shoyu pigments. J. Agric. Chem. Soc. Japan 3, 594-613.

    Kuroshima, E., Oyama, Y., Matsuo, T., and Sugimori, T. 1969. Biosynthesis and degradation of glutamic acid in microorganisms relating to the soy sauce brewing. (111). Some factors affecting the glutamic acid and its related substances formation in soy sauce brewing. J. Ferment. Technol. 47, 693-700.


    50, 279-286.

    fermented foods. Cancer Res. 28, 2296-231 1 .

    Japan 43, 956-981.


    Food H y g . 9, 29-34.

    Kurtzman, C. 1983. Inr. Mycol . Conf., 3rd, Tokyo, Sept. p. 153 (Abstr.). Kusumoto, E., Inamori, K., Uchida, K., and Yoshino, H. 1977. Studies on microorganisms isolated

    Li, T. 1970. The present condition and prospect for future of shoyu industry in Korea. J . Brew. Soc.

    Liardon, R., and Ledermann, S. 1980. Volatile components of fermented soya hydrolyzate. 11.

    Liardon, R., and Philippossian, G. 1978. Volatile components of fermented soya hydrolyzate. I.

    in fermentation of soy sauce mash ( I ) . J. Japan Soy Sauce Res. Inst. 3, 283-289.

    Japan 67, 106-108.

    Composition of the basic fraction. Z. Lebensm. Unters. Forsch. 170, 208-213.

    Identification of some lactones. A . Lebensm. Unrers. Forsch. 167, 180-185.


    Lin, J. Y., Wang, H. I., and Yen, Y. C. 1978. Mutagenicity of soybean sauce. Consumer Toxicol.

    Lucian0 Re, Mauer, B., and Ohloff, G. 1973. Ein einfacher Zugang zu 4-hydroxy-2,5,-di- methyl-3(W)-furanon (Furaneol), einem Aromabestandteil von Ananas und Erdbeere. Helv. Chim. Acra 56, 1882-1894.

    17, 329-33 1.

    MacDonald, 1. C. 1961. Biosynthesis of aspergillic acid. 1. Biol. Chem. 236, 512-514. MacDonald, J. C. 1%2. Biosynthesis of hydroxyaspergillic acid. J. Biol. Chem. 240, 1977-1981. MacDonald, J. C. 1965. Biosynthesis of pulcherriminec acid. Biochem. J . 96, 533-538. MacDonald, J. C. 1%7. Aspergillic acid and related compounds. In Antibiotics 2, Biosynthesis

    (D. Gottlieb and P. D. Shaw, eds.), pp. 43-51. Springer-Verlag Berlin and New York. MacDonald, W. C.. and Dueck, J. W. 1976. Long term effect of shoyu on the gastric mucosa of the

    rat. J . Natl. Cancer Inst. 56, 1143-1 147. MacDonald, J. C., Micetich, R. G., and Haskins, R. H. 1964. Antibiotic activity of neoaspergillic

    acid. Can. J. Microbiol. 10, 90-92. Machi, M. 1966. Comparative studies of soy mashes in natural and warmed brewing. Season. Sci.

    Manabe, M., Matsuura, S., and Nakano, M. 1968. Studies on the fluorescent compounds in fer- mented foods. Part 1. Chloroform-soluble fluorescent compounds produced by koji molds. Nippon Shokuhin Kogyo Gakkaishi 15, 7-12.

    Manabe, M., Shinshi, E., Tanaka. K., Goto, T., and Matsuura, S. 1985. Analysis of cyclopiazonic acid by HPLC. Proc. Annu. Meet. Jpn. Soc. Food Sci. Technol. p. 50.

    Masuda, Y., Mori, K., and Kuratsune, M. 1%5. Annu. Meet. Jpn. Cancer Assoc., Fukuoka p. 50. Matsumoto, K. 1921. Studies on the organic acids in the constituents of shoyu. Part 1. Rep. Narl.

    Matsumoto, J. et al. 1977. Production method of processed cereal powder used for brewing. Public

    Micetich, R. G., and MacDonald, J. C. 1965. Biosynthesis of neoaspergillic acid and neohydrox-

    Mitsui, M., and Kusaba, E. 1957. Studies on shoyu color. (11). J. Brew. Assoc. Japan 52,816-822. Miyauchi, K., Inamori, K., and Uchida, K. 1981. Studies on environmental factors involved in

    fermentation of shoyu mash (11). Effect of aeration on growth of shoyu yeasts ( I ) . J . Japan Soy Sauce Res. Inst. 7, 13-18; Effect of aeration on growth of shoyu yeasts (2). J . Japan Soy Sauce

    Miyazaki, K., and Honkawa, Y. 1964. Studies on various temperatures of soy Koji-making. Season. Sci. 11, 21-25.

    Mogi, Keitaro. and Mogi, K. 1968. An observational method of yeast flora in soy sauce fermenta- tion. J. Agric. Chem. Soc. Japan 42,466-470.

    Mori, S . 1979. Quality evaluation of shoyu and multivariate analysis. J . Brew. Soc. Japan 74,526- 531.

    Mori, S., Nunomura, N., and Sasaki, M. 1982. Correlation coefficients between the flavor compo- nents and the overall flavor of shoyu. Proc. Annu. Meer. Agric. Chem. Soc. Japan p. 36.

    Mori, S., Nunomura, N., and Sasaki, M. 1983. The effects of 4-ethyl-2-methoxy phenol and 3-(methylthio)-l-propanol to the overall flavor of shoyu. Proc. Annu. Meet. Agric. Chem. Soc. Japan p. 236.

    Mori, S., Nunomura. N., and Sasaki, M. 1984. The characteristics of sensory flavor preference of marketing shoyu and their relations to the volatile flavor components. Proc. Annu. Meet. Agric. Chem. Soc. Japan p. 129.

    Moriguchi, S., and Ishikawa, H. 1-a. Difference of color of Usukuchi shoyu made from whole and defatted soybean. Part 2. Anti-oxidation chemical component in whole soybeans. J. Fer- ment Technol. 38, 413-422.

    13, 1-16.

    Brew. Res. Ins?. 85, 1-33.

    opening of patent application. Sho52-34988. Nisshin Milling Co., Ltd.

    yaspergillic acid. J. Biol. Chem. 240, 1692-1695.

    Imt. 7, 54-58.


    Moriguchi, S., and Ishikawa, H. 1960b. Difference of color of Usukuchi shoyu made from whole and defatted soybean. Part 1. Formation and intensity of color. J . Ferment. Technol. 38, 271- 274.

    Moriguchi, S., and Nishiyama, T. 1960. Roasting degree of the wheat used for Usukuchi shoyu. J . Ferment. Technol. 38, 274-276.

    Moriguchi, S., and Kawaguchi, T. 1961. Studies on the difference of the shoyu made from raw and defatted soybeans. V. The rise and fall of microorganisms during Koji cultivations. Season. Sci.

    Moriguchi, S., and Ohara, H. 1961. Studies on the effect of sulfite applied to the treatment of raw material for lightcolored shoyu (1). Determination of the kind and amount of sulfite for use, and small-scale brewing test therewith. J . Ferment. Technol. 39, 283-288.

    Moriguchi, S. , Kawaguchi, T., Ishigami, Y., and Ueda, R. 1961. Studies on the difference of the shoyu made from raw and defatted soybeans. IV. The composition of organic acids and other components in shoyu moromi produced in every season. Season. Sci. 9, 34-38.

    Morimoto, S. , and Murakami, M. 1966. Studies on a quick ripening process for soy sauce (V). A new quick ripening process based on alcohol fermentation (VI). A qualitative analysis of soy sauce flavor components using gas-liquid chromatography. J . Ferment. Technol. 44,467-475, 476-482.

    Morimoto, S . , and Murakami, M. 1967. Studies on fermentation products from aldehyde by micro- organisms (I). The fermentative production of furfuryl alcohol from furfurol by yeasts. Part 1. J . Ferment. Technol. 45, 442-45 1.

    9, 39-41.

    Motai, H. 1976. Browning of shoyu. Nippon Shokuhin Kogyo Gakkaishi 23, 372-384. Motai, H., and Inoue, S . 1974a. Conversion of color components of melanoidin produced from the

    glicine-xylose system. Agric. Biol. Chem. 38, 233-239. Motai, H., and Inoue, S . 1974b. Oxidative browning in color of shoyu. Studies on color of shoyu.

    Part 2. J . Agric. Chem. SOC. Japan 48, 329-336. Motai, H., Inoue, S . , and Nishizawa, Y. 1972. Studies on color of shoyu. Part 1. Characteristics of

    shoyu color, and separation of color components. J . Agric. Chem. Soc. Japan 46, 631-637. Motai. H., Inoue, S . , and Hanaoka, Y. 1975. The contribution of soybean and wheat on the color

    formation of shoyu. Nippon Shokuhin Kogyo Gakkaishi 22, 494-500. Motai, H., Hayashi, K., Ishiyama, T., and Sonehara, T. 1983. Role of the main enzyme proteins of

    koji-mold Aspergillus in sedimentation of coagula in soy sauce during pasteurization. J . Agric. Chem. SOC. Japan 57, 27-36.

    Mu~akami, H. 1971. Classification of Koji molds. J . SOC. Brew. Japan 66, 658-662, 759-762,

    Murakami, H. 1973. Classification of Koji molds and its meaning. Season. Sci. 20, 2-14. Murakami, H., Takase, S. , and Ishii, T. 1967. Non-productivity of aflatoxin by Japanese industrial

    strains of Aspergillus. I. Production of fluorescent substances in agar slant and shaking cultures. J. Gen. Appl. Microbiol. 13, 323-334.

    Murakami, H., Takase, S . , and Kuwabara, K. 1968. Non-productivity of aflatoxin by Japanese industrial strains Aspergillus. 11. Production of fluorescent substances in rice koji, and their identification by absorption spectrum. J . Gen. Appl. Microbiol. 14, 97- 110.

    Nagahara, W., Onishi.X., and Nasuno, S. 1984. On the mutagenicity of shoyu. Meet. Food Hyg. Soc. Japan, 48rh, Oct. 18. Fukui.

    Nagahori, T., Motai, H., and Okuhara, A. 1980a. On the substances in shoyu to suppress the nitrosamine of dimethylamine. Nufr. Foods 33, 151-160.

    Nagahori, T., Motai, H., and Okuhara, A. 1980b. On the substances in shoyu to suppress the nitrosation of dimethylamine. Eiyo-ro Shokuryo 33, 15 1-160.

    Nagase, I., Omae, N., and hose, K. 1971. On the effect by an addition of fermentative micro- organism in the brewing of soy sauce. Season. Sci. 18, 6-20.

    859-863. 866-969, 1042-1045, 1150-1153.


    Nakadai, T. 1977. Estimation of composition of various enzymes produced by shoyu Koji culture of Aspergillus sojae. J . Japan Soy Sauce Res. Insr . 3, 99-104.

    Nakadai, T., and Aishima, T. 1983. Statistical estimation of key-enzymes determining the viscosity and filtration rate of soy mash. J . Agric. Chem. SOC. Japan 57, 307-3 I 1.

    Nakajima, M., and Takei, S. 1949. Studies on the flavorous components of shoyu. 11. Isolation and its chemical structure of shoyu-aldehyde from shoyu. J. Chem. Soc. Japan 70, 47-49.

    Nakamura, S . , and Shiro, T. 1960. Studies on growth inhibition of Hiochi-bacteria, specific saprophytes of sake. IV. Hydroxyaspergillic acid as a growth inhibitant against Hiochibacteria. Bull. Agric. Chem. Soc. Japan 23,418-427.

    Nakamura, S., and Shiro, T. 1%1. Studies on growth inhibition of Hiochi bacteria, specific saprophytes of sake. VI. Mutaaspergillic acid as a new growth inhibitant against Hiochi- bacteria. Agric. Biol. Chem. 25, 573-579.

    Nakayama, S., Takeda, R., and Sotoyama, N. 1965. Application of cellulase in brewing (111). Decomposition of soybean seedcoat and cotyledon using cellulase with a cell separating en- zyme. J. Ferment. Technol. 43, 648-652.

    Nasuno, S., and Nakadai, T. 1971. Formation of glutamic acid from defatted soybean by Aspergillus oryzae 460. J . Ferment. Technol. 49, 544-551.

    Nasuno, S . , and Nakadai, T. 1977. Proteases produced by Koji-molds. J. Japan Soy SauceRes. Inst.

    Nasuno, S . , and Ohara, T. 1971. Comparison of alkaline proteinase from hyperproductive mutants and parent strain of Aspergiffus sojae. Agric. Biol. Chem. 35, 836-843.

    Nasuno, S . , and Ohara, T. 1972a. Improvement of Koji-mold for soy sauce production by induced mutation (11). Hyperproduction of proteinese and some hydrolytic enzymes by mutants of Aspergillus sojae. Season. Sci. 19, 32-39.

    Nasuno, S . , and Ohara, T. 1972b. Improvement of Koji-mold for soy sauce production by induced mutation (111). Comparison of alkaline proteinase from hyperproductive mutants and parent strain of A s p e r g i f h sojae. Season. sci . 19, 41-47.

    Nasuno, S . , Ohara, T., and Iguchi, N. 1971. Successive isolation of protease hyperproductive mutants of Aspergillus sojae. Agric. Biol. Chem. 35, 291-294.

    Nasuno, S. , Ohara, T., and Iguchi, N. 1972. Improvement of Koji-mold for soy sauce production by induced mutation (I). Successive isolation of proteinase hyperproductive mutants of Aspergillus sojae. Season. Sci. 19, 27-31.

    Nesbitt, B. F., OKelly, J., Sargeant, K., and Sheridan, A. 1962. Aspergiffusflavus and turkey X disease. Narure (London) 195, 1062.

    Nishikawa, E. 1933. Biochemistry of molds. 11. One metabolic compound by Aspergiffus melleus (I). J . Agric. Chem. Soc. Japan 9, 772-774.

    Noda, F., Sakasai, T., and Yokotsuka, T. 1975. Studies on fermentation of soy mash (3). Effect of ether extract of soy sauce on Saccharomyces rouxii. J . Agric. Chem. SOC. Japan 22,530-535.

    Noda, F., Sakasai, T., and Yokotsuka, T. 1976a. Studies on fermentation of soysauce mash (4). Identification of soysauce components which have an inhibitory effects on growth and an enhancing effect on fermentation activities of soysauce yeasts. Nippon Shokuhin K O ~ Y O Gak- kaishi 23, 53-58.

    Noda, F., Sakasai. T., and Yokotsuka, T. 1976b. Studies on fermentation of soysauce mash ( 5 ) . On the antibiotic activities of vanillin, guaiacol and isomers of vanillin. Nippon Shokuhin K O ~ Y O Gakkaishi 23, 59-66.

    Noda, F., Sakasai, T., and Yokotsuka, T. 1976c. Studies on fermentation of soysauce mash (6). On the enhancing effect of o-vanillin, vanillin and guaiacol on fermentation activity of S . rouxii. Nippon Shokuhin K O ~ Y O Gakkaishi 23, 67-73.

    Noda, Y., and Nakano, M. 1979. Studies on the control of the fermentation of soy sauce mash. 1. Determination of volatile phenols in soy sauce. J . Japan Soy Sauce Res. Insr. 5 , 299-302.

    3, 309-313.


    Nunomura, N., Sasaki, M., Asao, Y., and Yokotsuka, T. 1976a. Identification of volatile com- pounds in shoyu by gas-chromatography-mass spectrometry. Agric. Biol. Chem. 40,485-490.

    Nunomura, N., Sasaki, M., Asao, Y., and Yokotsuka, T. 1976b. Isolation and identification of 4- hydroxy-2(or S)-ethyl-S(or 2)-methyL3(2H)-furanone, as a flavor component in shoyu. Agric. Biol. Chem. 40, 491-495.

    Nunomura, N., and Sasaki, M. 1981. Studies on flavor components of shoyu (X). Higher boiling flavor components. Proc. Annu. Meet. Agric. Chem. SOC. Japan p. 104.

    Nunomura, N., and Sasaki, M. 1982. Studies on flavor components of s h o p (XI). Qualitative analysis of sulfur containing flavor constituents of shoyu. Annu. Meet. Agric. Chem. SOC. Japan Tokyo, April.

    Nunomura, N., Sasaki, M., Asao, Y., and Yokotsuka, T. 1977a. Simple discrimination between genuine fermented and chemical shoyu on the basis of volatiles. J . Agric. Chem. SOC. Japan 51,

    Nunomura, N., Sasaki, M., Asao, Y., and Yokotsuka, Y. 1977b. On the flavor components of

    Nunomura, N., Sasaki, M., Asao, Y., and Yokotsuka, T. 1978. Shoyu (soy sauce) volatile flavor

    Nunomura, N., Sasaki, M., and Yokotsuka, T. 1979. Isolation of 4-hydroxy-5-methyl-3(W)-

    Nunomura, N., Sasaki, M., and Yokotsuka, T. 1980. Shoyu (soy sauce) flavor components: Acidic

    Ochiai, M., Suwa, Y., Wakabayashi, K., and Sugimura, T. 1982. Annu. Meet. Cancer Assoc.

    Ogasawara, T. 1963. Odor of amino acids. J. Brew. SOC. Japan 58, 521-525. Ohara, H., Moriguchi, S., and Nishiyama, T. 1959. The effect of the temperature and moisture upon

    the cultivation of shoyu Koji. (1),(2). Season. Sci. 7, 25-29, 30-42. Ohloff, G. 1978. Importance of minor components in flavors and fragrance. Perfum. Flavor. 3, 11-

    22. Oka, S., and Nagata, K. 1974a. Isolation and characterization of neutral peptides in soy sauce.

    Agric. Biol. Chem. 38, 1185-1 194. Oka, S., and Nagata, K. 1974b. Isolation and characterization of acidic peptides in soy sauce. Agric.

    Biol. Chem. 38, 1195-1202. Okuhara, A,, and Saito, N. 1970. Color of soy sauce. 111. Effect of reducing reagents on brewing of

    soy sauce. IV. Decolorization and depression of browning by reduction. J. Ferment. Technol.

    Okuhara, A, , and Yokotsuka, T. 1958. Studies on the analytical method of shoyu (I). Analytical method of glycerol in shoyu. J . Agric. Chem. SOC. Japan 32, 138-143.

    Okuhara, A., and Yokotsuka, T. 1961. Studies on flavorous substances in soy (21). Chemical changes of extractable organic acids in the course of heating of soy. J. Agric. Chem. SOC. Japan 35, 447-454.

    Okuhara, A., and Yokotsuka, T. 1962. Studies on the analytical method of shoyu (111). Analytical method of total polyol in shoyu. J . Agric. Chem. SOC. Japan 36, 320-326.

    Okuhara, A., and Yokotsuka, T. 1963. Studies on the flavorous substances in shoyu (22). Dif- ferences of shoyu made from whole and defatted soybean. J . Agric. Chem. SOC. Japan 37,255- 261.

    Okuhara, A., and Yokotsuka, T. 1977. Shoyu, Japanese fermented soy sauce. Presented by Yokot- suka at SCOGS Hearing, Hilton Hotel, Bethesda, Maryland, July 26.

    Okuhara, A., Nakajima, T., Tanaka, T., Saito, N., and Yokotsuka, T. 1969. Color of soy sauce. 11. Stability of color. J. Ferment. Technol. 47, 57-68.

    Okuhara, A., Tanaka, T., Saito, N., and Yokotsuka, T. 1970. Color of soy sauce. V. Application of

    709-7 1 1.

    shoyu. The 1 Ith symposium on taste and flavor. Tokyo University. No. 26-27.

    components: basic fraction. Agric. Biol. Chem. 42, 2123-2128.

    furanone, a flavor component in shoyu. Agric. Biol. Chem. 43, 1361-1363.

    fractions and the characteristic flavor component. Agric. Biol. Chem. 44, 339-351.


    48, 177-189, 190-201.


    multiple correlation analysis to analysis of browning mechanism. J. Fermenr. Technol. 48,

    Okuhara, A,, Saito, N., and Yokotsuka, T. 1971. Color of soy sauce. VI. The effect of peptides on

    Okuhara, A., Saito, N., and Yokotsuka, T. 1972. Color of soy sauce. VII. Oxidative browning. J.

    Okuhara, A., Saeki, M., and Sasaki, T. 1975. Browning of soy sauce. J. Jupan Soy Sauce. Res. Ins?.

    Omata, S., and Ueno, T. 1953a. On the colour of shoyu. Part 1. Colour of s h o p and its charac- teristics for the active charcoal. J . Agric. Chem. SOC. Japan 27, 570-575.

    Omata, S., and Ueno, T. 1953b. On the colour of shoyu. Part 2. Factors causing the colour change during storage. J . Agric. Chem. Soc. Japan 27, 575-580.

    Omata, S. , Ueno, T., and Nakagawa, Y. 1955a. On the colour of soy sauce. Part 3. On the substances increasing colour intensity of soy sauce. J. Agric. Chem. SOC. Japan 29, 165- 168.

    Omata, S., Ueno, T.. and Nakagawa, Y. 1955b. On the colour of soy sauce. Part 4. On the effect of furfural on the colour of soy sauce. J . Agric. Chem. SOC. Japan 29, 215-219.

    Omata, S., Ueno, T., and Nakagawa, Y. 1955~. On the colour of soy sauce. Part 5 . Effect of reductones on the colour of soy sauce. J. Agric. Chem. Soc. Japan 29, 251-256.

    Omata, S., Ueno, T., and Nakagawa, Y. 1955d. On the color of soy sauce. Part 6. On the soy sauce pigment (I). J. Agric. Chem. Soc. Japan 29, 256-260.

    Onishi, T. 1970a. Changes of aroma and color during process of soy sauce. Season. Sci. 17, 1-5. Onishi, T. 1970b. Studies on color formation by heating of soy sauce. Season. Sci. 17, 47-53. Onishi, T. 1971. Changes of aroma and color during process of soy sauce (2). Preserving heated soy

    Onishi, T. 1972. Changes of aroma and color during process of soy sauce (3). Effect of the added

    Onishi, T. 1975. Pastuerization of shoyu in the middle class producers. J. Brew. SOC. Japan 70,

    Onishi, T. 1976. Refining and pasteurization of shoyu. J. Japan Soy Sauce Res. Insr. 2, 269-273. Ono, T., and Harada, H. 1981. Reduction method of raw materials used for brewing. Patent

    announcement. Sho56-42562. Ajinomoto Co., Ltd. Oshita, K., Nakashima, Y., Sugiyama, S., and Takahashi, R. 1977. Safety evaluation of Shoyu.

    Oral presentation by T. Yokotsuka, on Shoyu. SCOGS (Select Committee of GRAS Sub- stances) Hearing, Hilton Hotel, Bethesda, Maryland, July 26.

    Patton, S. 1950. The formation of maltol in certain carbohydrate-glycine systems. J . Biol. Chem.

    Raper, K. B., and Fennel, D. 1. 1965. The Genus Aspergillus, p. 36. Williams & Wilkins, Baltimore.

    Rodin, J. 0.. Himel, C. M., Silverstein, M., k p e r , R. W., and Gortner, W. A. 1965. Volatile flavor and aroma components of pineapple. 1. Isolation and tentative identification of 2.5- dimethyl-4-hydroxy-3(W)-furanone. J. Food Sci. 30, 280.

    Sahai, B. I., and Shaw, H. 1961. A Salkowski-positive derivative of ferulic acid from wheat leaves. Can. J. Bor. 39, 1553-1555.

    Sakaguchi, G., Oishi, I., and Shinagawa, K. 1975. Behavior of Staphylococci and Clostridium borulinum inoculated in Shoyu. Dept. of Agriculture, Osakafuritsu University. Unpublished.

    Sakaguchi, K., and Yamada, K. 1944. Morphology and classification of koji molds. J. Agric. Chem. SOC. Japan 20, 65-73, 141-154.

    Sakaguchi, Kenji. 1958. Studies on the activities of bacteria in soy sauce brewing. 111. Taxonomic


    browning. J. Fermenr. Technol. 49, 272-287.

    Ferment. Technol. 50, 264-272.

    1, 185-189.

    sauce by covering polyethylene film. Season. Sci. 18, 46-49.

    ethyl alcohol on the quality of soy sauce. Season. Sci. 19, 30-35.

    47 I -475.

    184, 131-134.


    studies on Pediococcus soyae nov. sp., a soy sauce lactic acid bacteria. Bull. Agric. Chem. SOC. Japan 22, 353-362.

    Sakaguchi, Kinichiro. 1981. Searching for the route of shoyu. S E W 398, 252-266. Sakasai, T., Noda, F., and Yokotsuka, T. 1975a. Studies on fermentation of soy sauce mash (I).

    Effect of components in soy sauce on growth of microorganisms. Nippon Shokuhin Kogyo Gakkaishi 22, 474-480.

    Sakasai, T., Noda, F., and Yokotsuka, T. 1975b. Studies on fermentation of soy sauce mash (11). Effect of components in soy sauce and physical environmental factors on growth of lactic acid bacteria and yeasts in soy sauce mash. Nippon Shokuhin Kogyo Gakkaishi 22, 481-487.

    Sakurai, H., and Okuhara, A. 1977. Rapid determination of glycerin in soy sauce by gaschromatog- raphy. J. Japan Soy Sauce Res. Insr. 3, 290-294.

    Salo, P., Nykanen, L., and Suomalanien, H. 1972. Odor thresholds and relative intensities of volatile aroma components in an artificial beverage imitating whisky. J . FoodSci. 37,394-398.

    Sargeant, K., Sheridan, A., OKelly, J., and Camaghan, R. B. A. 1961. Toxicity associated with certain samples of groundnuts. Nature (London) 192, 1092-1097.

    Sasaki, M. 1975. Characteristics of various kinds of shoyu from the viewpoints of flavorous compo- nents. Proc. Symp. Brew., 7th p. 30-33.

    Sasaki, M., and Nunomura, N. 1978. Minor components of shoyu flavor. Chem. Id. 29, 73-83. Sasaki, M., and Nunomura, N. 1979. Studies on flavor components of shoyu (VII). Lower boiling

    fractions. Proc. Annu. Meet. J . Agric. Chem. Soc. Japan p. 31. Sasaki, M., and Nunomura, N. 1981. Flavor components of topnote of shoyu. J . Chem. SOC. Jupan

    (3, 736-745. Sasaki, M., Kikuchi, T., Asao, Y., and Yokotsuka, T. 1967. Studies on the compounds produced by

    molds. 11. Fluorescent compounds produced by Japanese industrial molds (2). J . Agric. Chem. SOC. Japan 41, 154-158.

    Sasaki, M., Asao, Y., and Yokotsuka, T. 1968a. Studies on the compounds produced by molds. 111. Fluorescent compounds produced by Japanese industrial molds. J . Agric. Chem. Soc. Japan 42,

    Sasaki, M., Asao, Y., and Yokotsuka, T. 1968b. Studies on the compounds produced by molds. V. Isolation of nonfluorescent pyrazine compounds. J. Agric. Chem. SOC. Japan 42, 351-355.

    Sasaki, M., Asao, Y., and Yokotsuka, T. 1968~. Studies on the compounds produced by molds. V. Isolation of nonfluorescent pyrazine compounds (2). J. Agric. Chem. SOC. Japan 42, 35 I - 355.

    Sasaki, M., Kaneko, Y., Oshita, K., Takamatsu, H., Asao, Y., and Yokotsuka, T. 1970. Studies on compounds produced by molds. VII. Isolation of isocoumarin compounds. Agric. Biol. Chem.

    Sasaki, M., Takamatsu, H., Oshita, K., Kaneko, U., and Yokotsuka, T. 1974. Isolation of Lumi- chrome from the culture filtrate of Aspergillus oniki 7784 (Note). J. Agric. Chem. SOC. Japan

    Sasaki, M., Ohshita, K., and Yokotsuka, T. 1975. Studies on the compounds produced by molds. IX. Studies on the fluorescent compounds produced by Aspergillusflavus Link ex Fr. (WF-3-8) and Aspergillusflavus Link ex Fr. (68-SA-R-7). J . Agric. Chem. SOC. Japan 49, 553-555.

    Sasaki, Y., and Yoshida, T. 1966a. Microbial studies of soya mashes in Hokkaido (I). A taxonomic study and distribution of yeasts. J . Ferment. Technol. 44, 61-71.

    Sasaki, Y., and Yoshida, T. 1966b. Microbial studies of soya mashes in Hokkaido (11). Influence of salt on the growth and fermentation of soya mash yeasts. J . Ferment. Technol. 44, 158-165.

    Sasaki, Y., Yoshida, T., Fukuyama, T., and Nishizawa, I. 1964. Microorganisms in soy-mash of Hokkaido. I. On the yeasts. Season. Sci. 11, 26-32.

    Sasaki, M., Nunomura, N., and Mori, S. 1984a. Comparison between Japanese fermented soysauces


    34, 1296-1300.

    48, 569-57 I .


    and other Asian products regarding volatile flavor constituents. Proc. Annu. Meet. Agric. Chem. SOC. Japan p. 129.

    Sasaki, M., Nunomura, N., and Mori, S. 1984b. Comparison of volatile flavor compounds produced by shoyu yeasts. Proc. Annu. Meet Agric. Chem. Soc. Japan p. 130.

    Secliff, J. A., and Mower, H. F. 1977. Mutagens in decomposition products of carbohydrate. Fed. Proc., Fed. Hum. SOC. Exp. Biol. 36, 304.

    Sekine, H. 1972. Neutral protease I and I1 of Aspergillus sojae. Isolation in homogenous form, and some enzymatic properties. Agric. Biol. Chem. 36, 198-206, 207-216.

    Sekine, H. 1976. Neutral protease I and I1 of Aspergillus sojae. Action on various substrates. Agric. Biol. Chem. 40, 703-709.

    Shibamoto, T. 1983. Possible mutagenic constituents in nitrite-treated soy sauce. Food Chem.

    Shibata, Y., and Nakanishi, K. 1967. Effect of C/N ratio and cultural temperature on the composi- tion of protease system, and mycelial weight in asbestos Koji. Season. Sci. 14, 16-21.

    Shibuya, Y. 1969. Practical Shoyu Brewing, pp. 69-70. Chijin Shokan, Ushigome, Tokyo. Shikata, H., Omori, K., Okuno, T., and Moriguchi, S. 1971a. Color of Usukuchi soy sauce. I.

    Browning reaction by reducing sugars and amino acids. Season. Sci. 18, 28-39. Shikata, H., Omori, K., Okuno, T., and Moriguchi, S. 1971b. Color of Usukuchi soy sauce. 11.

    Browning reaction of soy sauce mash. Season. Sci. 18, 149-162. Shikata, H., Yasui, T., Ishigami, U., and Omori, K. 1978. Studies on the glutaminase of shoyu Koji

    (Part 1). J. Japan Soy Sauce Res. Inst. 4, 48-52. Shimizu, Y., Matsuto, S., Mizunuma, Y., and Okada, 1. 1970. Studies on the flavor of roasted

    barley (Mugi-cha). VI. Separation and identification of 5-hydroxymaltol, maltol, 5-meth- ylcyclopent-2-en-2-01- 1-one and other compounds. Agric. Biol. Chem. 34, 845-853.

    Shinshi, E., Manabe, M., Tanaka, K., Goto, T., and Matsuura, S. 1985. Degradation of cyclopia- zonic acid added to a fermenting salty shoyu mash. Proc. Annu. Meet. Jpn. SOC. Food Sci. Technol. p. 50.

    TOX~CO~. 21, 745-747.

    Shoji, K. 1936. Flavor of shoyu. 111. Sci. Pap. Phys. Chem. Res. 15, 223-229. Shoji, K., and Onuki, M. 1932. Studies on fermentation products by polarography. IV. On the soy

    Stahl, W. H. 1973. Odor and Taste Threshold Value Data. American Society for Testing and

    Sulser, H., Depizzol, J., and Buchi, W. 1967. A probable flavoring pineapple in vegetable-protein

    Suzuki, T., Shibuya, N., Watanabe, Y., and Tazaki, R. 1972. The control of soy sauce fermentation

    Tagami, H., and Sugawara, S. 1977. Proteinase composition of various Koji molds. J . Japan Soy

    Taira, T. 1926a. Studies on flavorous substances in shoyu (I). J . Brew. SOC. Japan 21, 31-37. Taira, T. 1926b. Distributions of p,y,-butylenglycol in Japanese products. J . Agric. Chem. Soc.

    Japan 2, 734-741. Takamatsu, H., Iwasa, T., and Yokotsuka, T. 1980. Production method of liquid seasoning. Ja-

    panese patent No. 1,120,428, Kikkoman Shoyu Co., Ltd. Tanaka, T., Saito, N., and Yokotsuka, T. 1969a. Relation between chemical factors and preference

    of soy sauce. Season. Sci. 16, 21-26. Tanaka, T., Saito, N., Nakajima, T., and Yokotsuka, T. 1969b. Studies on sensory evaluation of soy

    sauce (VI). Interrelation and harmony of flavor. J . Ferment. Technol. 47, 137-145. Tanaka, T., Saito, N., and Yokotsuka, T. 1970. Studies on sensory evaluation of soy sauce (XI).

    Relation between odor patterns and chemical factors ( 2 ) . J . Ferment. Technol. 48, 56-62.

    flavor. Sci. Pap. h i . Phys. Chem. Res. 11, 277-283.


    hydrolyzates. J. Food Sci. 32, 61 1-615.

    (2). The additional effect of yeast. Season. Sci. 19, 30-35.

    Sauce Res. Inst. 3, 1 11-1 15.


    Tateno, M., and Umeda, I. 1955. Cooking method of soybeans and soybean cake as raw material for

    Tazaki, R., and Ouye, K. 1962. Action of cellulase on soybean oil meal. J. Ferment. Technol. 40,

    Tazaki, R., and Watanabe, Y. 1966. The relationship of enzyme activities and temperatures during shoyu Koji processes. Season. Sci. 13, 1-9.

    Tazaki, R., Watanabe, Y ., Kumaki, T., and Kozaki, M. 1969. The relationship of enzyme activities and temperatures during shoyu Koji processes (3). Microflora of shoyu moromi with different Koji processes and shikomi process. Season. Sci. 16, 20-27.

    Terada, M., Hayashi, K., and Mizunuma, T. 1980. Distinction between Aspergillus oryzae and Aspergillus sojae by the productivity of some hydrolytic enzymes. J . Japan Soy Sauce Res. Inst.

    Terada, M., Hayashi, K., and Mizunuma, T. 1981. Enzyme productivity and identification of species of Koji-molds for soy sauce distributed in Japan. J . Japan Soy Sauce Res. Insr. 7 , 158- 165.

    Terasawa, M., Kadowaki, K., Fujimoto, H., and Goan, M. 1979. Studies on lactic acid bacteria in shoyu mash. Part 2. Citric acid metabolism by lactic acid bacteria in shoyu mash. J . Japan Soy Sauce Res. Inst. 5 , 15-20.

    Tomiyasu, Y. 1927. Distribution of acetylbutylcarbinol and 2,3-butylenglycol in fermented foods. J . Agric. Chem. SOC. Japan 3, 758-764.

    Uchida, K. 1978. The preferential fermentation of pentoses by Pediococcus hulophylus X-160, lactobacilli in shoyu mash. Proc. Annu. Meet. Agric. Chem. SOC. Japan p. 458.

    Ushijima, S . 1984. Breeding of Aspergillus sojae by protoplast fusion. Proc. Symp. Brew. SOC. Japan pp. 16-19.

    Ushijima, S., and Nakadai, T. 1983. Breeding of Aspergillus sojae, a shoyu koji mold, by protoplast fusion. Proc. SOC. Ferment. Technol. Japan p. 260.

    Uchiyarna, N., and Matsurnura, K. 1974. Treatment method of raw materials used for koji. Japanese patent announcement. Sho46-39074. Yamasa Co., Ltd.

    Udo, S . 1931. Studies on the taste of soy sauce. J. Agric. Chem. SOC. Japan 7 , 852-858. Ueda, R., and Tanitaka, S. 1977. Quantity changes of organic acids and amino acids in the course of

    Ueda, R., Nagai, S . , Kanbara, K., and Moriguchi, S. 1958. Studies on the organic acids in soy

    Ueno. H. 1963. Report of Kikkoman Shoyu Co., Ltd. Vol. 5 . Ujiie, F., and Yokoyama, K. 1956. The bactericidal nature of fermented shoyu for the intestinal

    pathogenic bacteria. J. Japan Food Hyg. Assoc. 6 , 1- 1 I . Umeda, I. 1967. Manufacturing method of Shoyu (3). J . Brew. Soc. Japan 62, 467-471. Umeda, I., and Saito, N. 1956. On the standard color of shoyu. Season. Sci. 4, 1-20. Umeda, I., Ishizuka, Z., and Saito, N. 1953. Bioassay for soy sauce. Season. Sci. 1, 29-40. Wakabayashi, K., Ochiai, M., Nagao, M., and Sugimura, T. 1983a. Proc. Annu. Meet. Environ.

    Wakabayashi, K., Ochiai, M., Saito, H., Tsuda, M., Sawa, Y., Nagao, M., and Sugimura, T.

    Wan, Hwa L., and Fuang, S. F. 1981. History of Chinese fermented foods. USDA Misc. Publ. FL-

    Watanabe, Y., Ishii, M., and Tazaki, R. 1970. The control of soy sauce fermentation (1). The

    Webb, A. D., and Schultz, H. W. 1969. Symposium on Foods: The Chemistry and Physiology of

    brewing. Japanese patent 204,858.


    6 , 75-81.

    heating of shoyu. J . Japan Soy Sauce Res. Inst. 3, 19.

    sauce. 1-11, J . Ferment. Technol. 37, 94-104.

    Mutagen. Soc. Japan p. 82.

    1983b. Proc. Narl. Acud. Sci. U.S.A. 80, 2912-2916.

    MS-333. NRRL, Peoria, Illinois.

    addition of yeast and Lactobacteliaceae. Season. Sci. 17, 35-40.

    Flavors, p. 217. AVI, Westport, Connecticut.


    Weiss, U., Strelitz, F., Flon, H., and Asheshov, I. N. 1958. Antibiotic compounds with action against bacterial viruses; Neohydroxyaspergillic acid. Arch. Biochem. Biophys. 74, 150- 157.

    White, E. C., and Hill, J. H. 1943. Studies on antibacterial products formed by molds. I. Aspergillic acid, a product of a strain of Aspergillusjlavus. J. Bacreriol. 45, 433-442.

    Wieland, Th., und Wiegandt, H. 1955. Bouillon Geruch bei der Threonin-Spaltung. Angew. Chem.

    Yamada, S . 1928. The formation of aldehydes in fermented foods. 11. J . Agric. Chem. Soc. Japan 4,

    Yamada, S . 1929. The formation of aldehydes in fermented foods. V. Fulfural-like compounds. J . Agric. Chem. Soc. Japan 5 , 223-234.

    Yamada, K., and Goan, M. 1%9. Studies on soy sauce flavour by gas-chromatography (Part 1). Research of flavour components by head gas analysis. Season. Sci. 16, 1-5.

    Yamaguchi, S. 1954. Brewing method of shoyu. Japanese Patent 211529. Kikkoman Shoyu. Co. Ltd.

    Yamamoto, K. 1957. Studies on Koji (3). Effects of cultural temperature on the production of mold protease. Bull. Agric. Chem. Soc. Japan 21, 319-324.

    Yamamoto, Y., Moriya, T., and Tsujihara, T. 1978. Studies on sterilize ability in seasoning foods. Part I . Sterilize ability of composition in soy sauce (No. I ) . J . Japan Soy Sauce Res. Inst. 4, 101-104.

    Yamanaka, T., and Akao, T. 1981. Apparatus for continuous cooking of cereal materials. Patent announcement. Sho56-12128. Kikkoman Cop.

    Yasuda, A., Mogi, K., and Yokotsuka, T. 1973a. Studies on the cooking method of proteinous materials for soy sauce brewing. Part 1. High temperature and short time cooking method. Season. Sci. 20, 20-24.

    Yasuda, A.. Arai, A., Tsukada, N., Mogi, K., Aonuma, T., and Yokotsuka, T. 1973b. Studies on the cooking method of proteinous materials for soy sauce brewing. Part 2. Brewing test of soy sauce. Season. Sci. 20, 25-27.

    67, 399-400.


    Yen, G. C., and Lee, T. C. 1984. Annu. Meet. I F T . Yokotsuka, T. 1949. Studies on flavorous substances in soy (1). Isolation of flavorous substances

    from soycake, chiefly on the flavors of lower boiling point. J . Agric. Chem. Soc. Japan 23, 200-205.

    Yokotsuka, T. 1950. Studies on flavorous substances in soy (2). (3). Isolation of Cl&z03 as heat flavor of shoyu. J . Agric. Chem. Soc. Japan 24, 355-358, 358-362.

    Yokotsuka, T. 1951a. Studies on flavorous substances in soy (4). Flavors resulting from heating. 3. Color reactions of acetals. J . Agric. Chem. Soc. Japan 24, 402-406.

    Yokotsuka, T. 1951b. Studies on flavornus substances in soy (5). Flavors resulting from heating. 4. Decomposition of flavors by alkali or light. 1 . Agric. Chem. Soc. Japan 24, 407-409.

    Yokotsuka. T. 1951~. Studies on flavorous substances in soy (6). Flavors with higher boiling point. Ethylpalmitate and ethyllinolate. J . Agric. Chem. Soc. Japan 25, 206-210.

    Yokotsuka, T. 1953a. Studies on flavorous substances in soy (9). Isolation of phenolic flavorous substances. J. Agric. Chem. Soc. Japan 27, 276-281.

    Yokotsuka, T. 1953b. Studies on flavorous substances in soy (10). Precursors of phenolic flavorous substances. J. Agric. Chem. Soc. Japan 27, 334-339.

    Yokotsuka, T. 19%. Studies on flavorous substances in soy (1 I ) . Flavors of lower boiling fractions (2). 1. Agric. Chem. Soc. Japan 27, 359-363.

    Yokotsuka, T. 1953d. Studies on flavorous substances in soy (12). Flavorous substances in raw soy. J . Agric. Chem. Soc. Japan 27, 549-553.

    Yokotsuka, T. 1954. Studies on flavorous substances in soy (13). On the yeast-static substances in brewed soy and their increase in the heated soy. J . Agric. Chem. Soc. Japan 28, 114-118.


    Yokotsuka, T. 1960. Aroma and flavor of Japanese soy sauce. Adv. Food Res. 10, 75-134. Yokotsuka, T. 1972. Brewing aptitude of raw materials of shoyu. Proc. 4th Symp. Brew. Assoc.,

    Yokotsuka, T. 1975. The flavor of shoyu. Koryo 112, 57-71. Yokotsuka, T. 1977. Oral presentation. Shoyu. SCOGS (Select Committee of GRAS Substances)

    Hearing. Hilton Hotel, Bethesda, Maryland, July 26. Yokotsuka, T., and Asao, Y. 1961. Studies on flavorous substances in shoyu (19). a-Diketon

    compounds (1). J . Agric. Chem. SOC. Japan 35, 837-845. Yokotsuka, T., and Asao, Y. 1969. Production method of shoyu. Japanese patent No. 1,042,917. Yokotsuka, T., and Takimoto, K. 1956. Studies on flavorous substances in soy (14). Flavorous

    substances in heated soy ( I ) . J . Agric. Chem. SOC. Japan 30, 66-71. Yokotsuka, T., and Takimoto, K. 1958. Studies on flavorous substances in soy (15). Flavorous

    substances in heated soy (2). J. Agric. Chem. SOC. Japan 32, 23-26. Yokotsuka, T., Hayashi, K., Sakasai, T., and Mogi, K. 1965. Production method of shoyu-Koji.

    Japanese patent 539215. Kikkoman Shoyu Co., Ltd. Yokotsuka, T., Mogi, K., Fukushima, D., and Yasuda, A. 1966a. Dealing methodof proteinous raw

    materials for brewing. Japanese patent 929910. Kikkoman Shoyu Co., Ltd. Yokotsuka, T., Sasaki, M., Kikuchi, T., Asao, Y ., and Nobuhara, A. 1966b. Production of fluores-

    cent compounds other than aflatoxins by Japanese industrial molds. In Biochemistry of Some Foodborne Microbial Toxins. (R. Mateles and G. N. Wogan, eds.), pp. 131-152. MlT Press, Cambridge, MA.

    Yokotsuka, T., Sakasai, T., and Asao, Y. 1967a. Studies on flavorous substances in shoyu (25). Flavorous components produced by yeast fermentation (I). J . Agric. Chem. SOC. Japan 41,428.

    Yokotsuka, T., Sakasai, T., and Asao, Y. 1967b. Studies on flavorous substances in shoyu (27). The production of 4-ethylguaiacol during shoyu fermentation and its role for shoyu flavor. J . Agric. Chem. SOC. Japan 41, 442.

    Yokotsuka, T., Sasaki, M., Kikuchi, T., Asao, Y., and Nobuhara, A. 1967~. Studies on the compounds produced by molds. I. Fluorescent compounds produced by Japanese industrial molds. J. Agric. Chem. SOC. Japan 41, 32-38.

    Yokotsuka, T., Asao, Y., Sasaki, M., and Ohshita, K. 1968a. Fyrazine compounds produced by molds. Proc. U.S. Japan Conf., 1st. Oct. 7-10, Honolulu p. 133 (UJNR Joint Panel on Toxic Microorganisms and U.S. Dept. of the Interior).

    Yokotsuka, T., Asao, Y., and Sasaki, M. 1968b. Studies on the compounds produced by molds. IV. Isolation of nonfluorescent pyrazine compounds. J . Agric. Chem. SOC. Japan 42, 346-350.

    Yokotsuka, T., Kikuchi, T., and Sasaki, M. 1968~. Aflatoxin-G like compounds with green fluores- cence produced by Japanese industrial molds. J . Agric. Chem. SOC. Japan 42, 581-585.

    Yokotsuka, T., Iwasa, T., and Fujii, S. 1%8d. Degrading method of protein, Japanese patent 827414. Roduction method of seasoning. Japanese patent 667036; and production method of protein degradate. Japanese patent 682848.

    Yokotsuka, T., Oshita, K., Kikuchi, T., Sasaki, M., and Yokotsuka, T. 1969. Studies on the compounds produced by molds. VI. Aspergillic acid, Koji acid, P-nitropropionic acid, and oxalic acid in solid-Koji, J . Agric. Chem. SOC. Japan 43, 189-196.

    Yokotsuka, T., Iwasa, T., and Fujii, S. 1970. Degrading method of protein. Japanese patent 753376. Yokotsuka, T., Iwasa, T., Fujii, S., and Kakinuma, T. 1972. The role of glutaminase in shoyu

    Yokotsuka, T., Sasaki, M., Nunomura, N., and Asao, Y. 1980. Flavor of shoyu (1.2). J . Brew. Soc.

    Yoshida, T. 1979. Studies on the yeasts in shoyu mash produced in Hokkaido Island. Proc. 11th

    Japan pp. 31-33.

    brewing. Annu. Meet. Agric. Chem. SOC. Japan, Sendai, Apr. 1 .

    Japan 75, 516-622, (9), 717-728.

    Sympo. Brew. Sept. 13 p . 62.

  • This Page Intentionally Left Blank




    * Ralston Purina Company, St. Louis, Missouri 63164

    fEnergy and Protein Nutrition Laboratory, Beltsville Human Nutrition Research Center,

    Agricultural Research Service, U S . Department of Agriculture,

    Beltsville, Maryland 20705 $Agricultural Nutritional Consultants, Inc.,

    Cedar Rapids, Iowa 52406 Departments of Medicine, and Community and Family Medicine,

    Georgetown University School of Medicine, Washington, D.C. 20007

    I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. The Energy-Protein Interaction

    111. Food Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Conventional Sources of Protein Foods

    A. Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Land Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Marine Animals

    V. Reflections on Foods ......................... VI. New Protein Foods B

    A. B. C. Nutritious Beverage Model D. Properties of Plant Protein Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. C. Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Government Regulations E. Comment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cereal-Legume Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Animal Flesh Model

    Nutritional Equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemistry and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . . . . . . VII.

    332 332 334 335 335 339 349 352 354 354 355 351 358 360 360 316 316 311 318

    33 1 Copyright Q 1986 by Academic Press, Inc.

    All rights of repduction in any form reserved.


    VIII. Intervention to Improve Energy and Protein Nutrition . . . . . . . . . . . . . . . . . 378 A. Short Term.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378 B. Long Term ................................................ 379 References .................................................... 38 1


    A treatise entitled New Protein Foods has recently been completed (Alt- schul, 1974, 1976; Altschul and Wilcke, 1978, 1981, 1985). Aside from dealing with new protein foods, the treatise reviewed new ways of producing and market- ing classical protein foods. The Editors undertook to review the treatise and to look for any fundamental generalizations that might arise. They invited C. E. Bodwell and D. T. Hopkins to join them in this review and to provide additional summary information on oilseed proteins.

    The purpose of the work was to describe the role of science and technology in providing greater flexibility for utilizing protein food resources. More flexibility in providing protein foods means greater capability for increasing world food supply. The two nutrients that make the greatest demands on photosynthetic capacity are energy and protein; failure to deal adequately with both of them as a unit reduces total food supply.


    There was a time when any emphasis on protein foods was taken to mean that the world food problem was considered to be primarily a protein problem. In opposition, many others insisted that there was no evidence of a protein problem: It was total food supply, total energy supply, that was the problem. And there were two camps: those who insisted that the major effort to increase food supply should be on total food supply, and the other that the effort must be concentrated on protein supply. Neither is correct. Protein and energy are interrelated and inseparable in individual human and animal nutrition (Fordyce and Christakis, 1981; Jansen, 1981; Wilcke and Hopkins, 1981); they are equally inseparable in considering agricultural resources. The food problem, wherever it exists, is a joint protein-energy problem, just as severe malnutrition in children is a pro- tein-calorie problem. Surely other essential nutrients are required, but in rela- tively small quantities for which there are alternate sources (Brin, 1976). It is the protein-energy axis that is limiting.

    Wilcke and Hopkins (1981) concluded that, given the diet prevalent in most of the world, i.e., one that includes cereal grains of some type and pulses (legumes)


    or fish or other types of animal protein, when a mixed diet of this type provides sufficient calories, there is also sufficient protein.

    Protein and energy interact in other ways. Protein crops and energy crops compete for the same land (Byerly, 1978; Stonaker, 1978; Cunha, 1978; Phil- lips, 1981). Protein foods (e.g., legumes and livestock) are more expensive to produce than energy foods (e.g., cereals and tubers). (See Reid and White, 1978, on the energy cost of food production by animals; Harada and Saito, 1978, on the relative efficiency of solar energy utilization of energy and protein crops.) There- fore, only lands unsuitable for production of energy foods can be devoted ex- clusively to protein foods, e.g., livestock. Or only countries that have excess capacity to produce energy foods can afford to devote substantial acreage to protein foods. Some countries that have the climate and the soil to produce soybeans do not do so because that acreage would compete with land for corn, rice, or wheat production. Rapeseed or sunflower seed become more attractive in areas with shorter growing seasons or less rainfall.

    Even though it is possible to categorize foods as energy foods or protein foods on the basis of the percentage of calories furnished as protein, taste and cultural preference often ovemde nutrient composition (Pyke, 1978; Wilcke and Hop- kins, 1981). Animal foods are given high status because they are readily con- sumed. As income rises, the demand for animal foods increases. Countries that have adequate energy food supplies compete with those lacking sufficient energy foods, such as corn: the latter group to feed humans, the former to feed livestock. So, in addition to the stress of increasing population (Bean, 1978), there is the stress of rising income, wherever that occurs.

    Food changes that follow rising income can cause new nutritional and medical problems because of the unbalanced nutrient composition that results (Altschul and Schertz, 1981). Animal products then provide a greater proportion of the energy intake: The protein and fat content increases at the expense of carbohy- drate intake (Wilcke and Hopkins, 1981). Excessive fat consumption is consid- ered a risk factor for coronary heart disease and other chronic diseases. In most industrialized countries, the pressure from medical and public health groups is to reduce fat as a percentage of the calories, and this would reduce the proportion of dietary animal protein (Fordyce and Christakis, 1981).

    Attention to the protein partner of this axis still means that we affirm the relationship between energy and protein, but consider that the problem of total food supply is made easier by increasing the versatility of interchange of protein foods, and we concentrate on that aspect of the problem. If soy protein can be eaten directly by humans, then there is less stress on animal production to provide protein of high quality (Horan, 1974; Smith, 1976; Horan and Wolff, 1976; Bodwell and Hopkins, 1985; Kolar et al. , 1985; Campbell et al., 1985). If peanut protein can be mixed with milk, the need for milk protein is reduced


    (Rhee, 1985). If there are better ways of feeding poultry and swine with a greater proportion of energy coming from nutritional material unsuitable for human consumption, then more energy and protein becomes available for humans. Low- protein energy foods such as cassava can be better utilized in a complete food program if cheaper protein foods are made concomitantly available. The thesis is that anything that makes it easier to provide protein makes it easier to adjust to limited food availability by providing more alternative ways of increasing total food supply (Altschul , 1974).


    Byerly (1978) divided the nations of the world into five categories based on average daily per capita consumption of food energy (Table I). He then listed the



    Food energy Fat (glcaplday) Protein (g/cap/day) (kcal/cap/day)

    Population ~ Meat, milk, Meat, milk, Category b (millions) Cereal Total and eggs Total and eggs Fish Total

    1 (less than 2100 kcal) 1768 1320 2000 12 29 6 2 52 2 (2100-2450 kcal) 468 1210 2250 11 42 10 4 62

    20 54 21 2 68 3 (2450-2800 kcal) 348 1330 2635 4 (2800-3150 kcal) 278 1 1 6 0 2965 31 85 30 3 81 5 (more than 3150 kcal) 748 980 3225 76 141 51 3 94 Total or average 3610 1224 2421 21 62 19 3 61

    From Byerly (1978). Countries in each category: (1) Afghanistan, Algeria, Angola, Bolivia, Burma, People's Republic of China,

    Dominican Republic, Ecuador, El Salvador, Ethiopia, Ghana, Guatemala, Guinea, Guyana, Haiti, India, Indonesia, Iran, Iraq, Laos, Mauritania, Nepal, North Vietnam, Philippines, Rwanda, Saudi Arabia, Somalia, Sudan, Tanzania, Upper Volta, Yemen, Yemen Democratic Republic, Zaire. (2) Albania, Bangladesh, Barbados, Benin, Burundi, Cameroon, Central African Republic, Chad, Colombia, Congo, Gabon, Gambia, Honduras, Hong Kong, Ivory Coast, Jamaica, Jordan, Kenya, Khmer, Lebanon, Liberia, Madagascar, Malawi, Malaysia, Mali, Mauritius, Morocco, Mozambique, Nicaragua, Niger, Nigeria, Pakistan, Panama, Peru, Senegal, Sierra Leone, South Korea, Singapore, South Vietnam, SriLanka, Surinam, Thailand, Togo, Trinidad, Tunisia, Uganda, Venezuela, Zambia. (3) Chile, Cuba, Egypt, Japan, Libya, Mexico, Mongolia, North Korea, Paraguay, Rhodesia, South Africa, Spain, Syria, Taiwan, Turkey, Uruguay. (4) Brazil, Bulgaria, Czechoslovakia, East Germany, Finland, Greece, Israel, Italy, Norway, Poland, Portugal, Romania, Sweden. (5) Argentina, Australia, Austria, Belgium, Canada, Denmark, France, Hungary, Ireland, Netherlands, New Zealand. Soviet Union, Switzerland, United Kingdom, United States, West Germany, Yugoslavia.


    amounts and types of protein and energy foods produced and consumed for each category. This same classification was utilized by Wilcke and Altschul (1978) to depict per capita meat production as it relates to total energy consumption. Wilcke and Hopkins (1981) continued this analysis of countries by categories by providing information on protein and fat intake in these countries for two peri- ods: 1966-1968 and 1975-1977. Some countries changed category in the inter- val by showing an increase in average calorie intake, many remained in the same category, some showed a decrease in total energy availability, and some even a decrease in total protein availability.

    The categories cited by Byerly (1978) were based on statistics available for 1970. At that time 1.7 billion people or close to 50% of the world population were in category 1 (average energy intake was less than 2100 kcal/capita/day). By 1980, a number of countries moved in position from category 1 to category 2 (2100 to 2450 kcal/capita/day). These included the Dominican Republic, Equa- dor, Iraq, Sudan, Burma, China, Indonesia, the Phillipines, Algeria, and Rwan- da. By 1980,20 countries, 20% of the world population, remained in category 1, still a very substantial number.

    Surely many individuals in countries of category 1 do not receive sufficient food energy, and a substantial number in countries of category 2 are in the same position. The total food problem diminishes markedly or disappears in countries of categories 3-5. But even in countries in the highest energy consumption category there may be, from time to time, pockets of poverty and hunger, depending on social and economic conditions.

    The number of people considered as suffering from inadequacy of protein intake would depend on interpretation of experimental data as related to protein needs (Wilcke and Hopkins, 1981). Similarly, a decision on the number of people suffering from low total food energy intake would depend on what goes in to determining energy requirements: Is it simply achievement of energy balance or is it the attainment of energy balance at a desired level of quality of life? We make no quantitative judgments, but state that the energy-protein axis for too many is inadequate and that the food problem remains a most serious world problem that could worsen in the face of strife and economic and social insta- bility.


    A. PLANT

    1 . Cereals, Legumes, and Oilseeds

    Cereals are the major source of calories and protein for most of the world. Where animal products are scarce or less plentiful, cereals may account for 60%


    of the daily protein intake (Mitsuda and Yasumoto, 1974). It is no wonder that efforts continue to be made to improve the protein content or quality of cereals, particularly for the more vulnerable. Neither breeding for better protein quality (e.g., high-lysine corn) nor for higher protein content (higher protein rice) has been successful as practical solutions because the yield of the improved varieties, with few exceptions, is lower than for their conventional counterparts (Altschul and Schertz, 1981; Nielsen, 1985). The great achievement has been to increase the yield of the cereals and to transfer the technology, in part, to less indus- trialized countries (Bean, 1978).

    Legumes are the plant protein counterpoint to the cereals. Together, they provide for balanced protein nutrition. But there are problems; the major one is low yield that discourages major increase in the cultivation of legumes. The great strides in increasing yields of cereals have not been matched with legumes (Bressani and Elias, 1974). In competition with wheat and rice, production of legumes has actually decreased in India (Bean, 1978). Another classic problem of legume foods is their content of inhibitors that restrict the quantities that can be eaten. Additional processing in the factory or home is required to reduce inhibitor levels. Much more is now known about the chemistry and biochemistry of these compounds; perhaps the most is known about those in the soybean (Nielson, 1985; Kinsella et af., 1985; Bodwell and Hopkins, 1985).

    The oilseeds are a special category of food source, since by proper processing the oil is removed, the protein concentrated, and interfering materials removed or inactivated. About 37 million metric tons of protein were forecasted to be har- vested as oilseed in 1981-1982 (Bodwell and Hopkins, 1985). Nearly 70% of this predicted production was from soybeans, with cottonseed, sunflower, rapeseed, and peanuts furnishing most of the remaining protein at levels of 10,5, 5 , and 5%, respectively.

    At the present time most of the oilseed protein is fed to animals; a relatively small proportion enters the human food system. However, the potential contribu- tion of oilseed proteins to human nutrition is very promising; undoubtedly, developing technology will make increased human consumption possible. The domination of soybeans over other oilseeds is striking and suggests that future technological breakthroughs will most likely originate from research with soy- beans. A review of the status of the major oilseeds follows.

    2. Soybeans

    Nearly two-thirds of the worlds soybeans are grown in the United States (Table 11), while Brazil and the Peoples Republic of China are estimated to be responsible for 15 and 10% of the world production, respectively. Soy products are entering the food system in increasing amounts in the industrialized countries



    Production, million metric tons

    Oilseed Producer Oilseed Oilseed protein

    soybean united states Brazil Peoples Republic of China Argentina Paraquay Other Total

    Cottonseed Soviet Union Peoples Republic of China United States India Palcistan Other Total

    Peanut (in shell) India Peoples Republic of China United States Senegal Sudan Brazil South Africa Other Total

    Sunflower Seed Soviet Union Argentina United States Romania Bulgaria Other Total

    Rapeseed India Peoples Republic of China Canada Poland France Other Total

    54.4 12.8 9.3 4.0 0.6 5. I


    4.9 5.9 5.8 2.8 1.5 7.0


    6.6 3.8 1.8 0.8 0.8 0.3 0.1 4.6


    4.6 1.8 2.1 0.8 0.5 4.4


    2.5 4.1 I .8 0.5 I .O 2.6


    16.49 3.88 2.80 1.21 0.19 I .56


    0.64 0.77 0.76 0.36 0.19 0.91 3.63

    0.68 0.40 0. I9 0.08 0.09 0.03 0.02 0.47 1 .%

    0.64 0.25 0.29 0.11 0.07 0.61 I .97

    0.41 0.67 0.30 0.08 0.16 0.43 2.05

    Data calculated from USDA Foreign Agricultud Service (1982).


    throughout the world, but the actual amounts consumed by humans is impossible to ascertain. According to one estimate, 945,000 tons of soy foods were pro- duced in the United States in 198 1 from 1.3 million tons of soybeans (Leviton, 1984). However, this estimate included the production of 600,000 tons of soy flour, grits, and extruded soy flour, a large portion of which was undoubtedly incorporated into pet foods.

    The International Nutritional Anemia Consultative Group (Bothwell et al. , 1982) estimated that the average consumption of soy products in the United States is about 5 lb/person. This would indicate a total consumption of about 550,000 tons. This figure would seem reasonable if one-third of the soy flour, grits, and textured vegetable protein entered human diets.

    The consumption of soy foods outside of the United States is even more difficult to estimate. The Peoples Republic of China is estimated to produce about 9 million metric tons of soybeans which represents nearly 3 million metric tons of protein. A significant portion of this protein is likely to be consumed by people, but accurate figures are impossible to obtain. In Japan (Dronne, 1981; Fauconneau, 1983) it was estimated that over 580,000 tons of defatted soybean meal equivalents (DSME) were consumed by humans in 1977; this included over 100,OOO tons (DSME) of tofu or tofu-like products, 385,600 tons (DSME) of fermented products, 9500 tons (DSME) of soy isolates and concentrates, and 7500 tons (DSME) of texturized soy products.

    Soy protein is consumed in a number of different forms. It is estimated that in the United States, isolated soy protein is used in the production of 20% of commercial infant formulas.Three million pounds of defatted soy flour are in- cluded in PL 480 programs (Bothwell et al., 1982). Soy milk is currently being produced in Japan at a rate of 30,000 tons annually and is predicted to reach 300,000 tons in the next 3 to 4 years (USDA Foreign Agricultural Service, 1982). In the United States in recent years, the consumption of traditional orien- tal-type soy products has grown to a total production of about 26,000 tons of tofu and tofu-like products and of about 148,000 tons of soy milk and soy milk products. Thus, it appears that high production of soybeans in the world coupled with developing technology for the use of soy products, both in traditional foods and as ingredients in new foods, will lead to an increasingly higher impact of soybean protein in human nutrition.

    3 . Cottonseed

    Cottonseed is the second largest source of oilseed protein in the world, with the leading producers being the Peoples Republic of China, the Soviet Union, the United States, India, and Pakistan. However, on a practical basis, little of this protein goes directly into human consumption. The only significant con-


    sumption of this protein in the United States is in the form of partially roasted, glandless cottonseed kernels as a confectionary. Cottonseed flour has been used in limited amounts in the weaning food Incaparina in Guatemala. However, other than for such specific limited applications, cottonseed protein has little impact on the nutritional quality of human diets.

    4 . Peanuts

    Peanuts are produced in the largest amounts in India, the Peoples Republic of China, and the United States. Although the total production of peanut flour, concentrates, or isolates is small compared to soybeans, a significant amount of peanuts is eaten as foods in the United States either as whole peanuts or as peanut butter. According to the estimates of the U.S. Department of Agriculture (USDA, 1981), the consumption of peanuts in peanut butter, salted peanuts, candy, and other uses in 1979 was 638,000 short tons, equivalent to 163,000 short tons of peanut protein. Data on the consumption of peanut products in other countries are not available.

    5. Other Oilseeds

    Sunflower seed is an important source of oilseed protein in the Soviet Union, the United States, and Argentina. However, the bulk of the world production is crushed for oil, and the residual meal is fed predominantly to animals. A small amount of dehulled sunflower seeds is eaten as snack-type foods and as a topping in confectionary-type products. It has been estimated that in 1982 -30 million Ib of protein from sunflower seed were used in food (Adams, 1982).

    Rapeseed, an important oilseed in the Peoples Republic of China, Canada, Northern Europe, and India, is most exclusively a source of oil, and the resultant meal is primarily in manufactured animal feeds.


    1. Production: General

    The numbers of livestock and wild animals in relation to feed supplies and the competition between man and animal for the energy and protein resources were discussed by Eyerly (1978). Surely there is a competition, but he emphasized the role of animals in recycling plant nutrients and as scavengers and suggested that present world resources are enough to support the present human, livestock, and wildlife population. These resources must be managed through skillful applica- tion of technology to provide for twice the human population and the proportion-


    ate increase in animal population as might be expected early in the twenty-first century.

    Is it possible to define the upper limits of livestock production? Phillips (1981) considered land and water resources. The total land resources per capita are expected to be reduced in the year 2000 to one-half of that available in 1959. Even assuming that the land under cultivation and tree crops could be doubled by 2050, this would still provide less than one-half as much land per capita as was available in 1959. About seven times as much water is required to support human life on a modest mixture of animal and bread compared to bread alone (2500 gal/person/day compared to 300).

    Phillips (1981) divided the world land resources into 10 categories, described in Table 111. He saw possibilities for increases in animal production in exten- sive grazing areas, mixed farming areas, intensively tilled areas, and highly intensive livestock enterprises. In the year 2000, in his view, the protein gap probably will have widened rather than narrowed, both because the supply of animal protein will not have kept pace witb the demand, and because much of the increase will have gone to meet the increased demands in the industrialized countries.

    Under conditions of production in the United States, the efficiency of protein production (grams per megacalorie of digested energy) is in the order: milk, 12.8; broiler, 11.9; eggs, 10.1; pork, 6.1; and beef, 2.1 (Reid and White, 1978). When food-energy supply is scarce, swine have the most precarious position as food producers. But beef cattle, despite the low output of protein per unit of digestible energy, can be produced under an all-forage system with very little fossil energy subsidization. Hence, the future role of beef cattle and other rumi- nants as food producers is assured.

    Much depends on the utilization of existing technology and on new improved technologies. Cunha (1978) considered animals as a good reservoir of food and as environmental improvers. There are many opportunities for improving animal production practices, and this information and practice is transferable to most climates, including the tropics.

    Cunha set production efficiency goals for the year 2000 (Table IV). Many of these goals have already been attained by individual producers, but to broaden the numbers who can attain these goals requires a reversal of trends in expendi- tures for agricultural research, an increase in research effort.

    It comes down to the role of government policy and its effects on the condition of the livestock sector (Schuh, 1981). Because agricultural production processes involve a biological factor, they differ from a typical production process in the nonagricultural sector. With livestock such as cattle, the production process is especially complex, and this can be a source of much mischief and confusion in devising a rational policy.

    Policies toward research and extension vary enormously throughout the world,













    Arctic and high mountain areas Examples: Arctic and Antarctic land masses; northern portions of Canada, Greenland, Siberia, and Alaska; northernmost portions of Norway, Sweden, and Finland; above the treeline in mountains Desert areas Examples: Sahara; deserts of central and western Australia; the Gobi; much of the Arabian Peninsula; portions of southwestern United States and adjacent Mexico Tropical rain forest areas Examples: Amazon and Congo basins; hot, humid tropics Semiarid areas Examples: Drier grazing areas in southwestern United States; northern and southern edges of the Sahara; eastern Africa; many parts of the Near East Extensive grazing areas Examples: Perimeters of Australia; parts of New Zealand; south and central Africa up to the sub-Sahara; parts of eastern Europe; much of central Asia; parts of Mexico and Central Amer- ica; Argentina, Uruguay, and southern Brazil; areas around Gulf of Mexico from Texas to Florida Pasture grazing areas Examples: Intensively managed man-made pastures; parts of New Zealand; Gulf Coast of the United States; tropical and subtropical islands; portions of the pampas in Argentina that are plowed and reseeded at intervals of about 10 years Extensive grazing linked with fattening areas Examples: United States and Canada where cattle and sheep are grown in ranges and then moved to feedlots; movement of feeder cattle from Ireland to England for fattening Mixed farming areas Examples: Farms found throughout most of the agricultural areas of the United States, Canada, western Europe, the British Isles, Australia, New Zealand, Latin America, South Africa, eastern Europe Intensively tilled areas Examples: Areas of great population density; Bangladesh, eastern portions of the Peoples Republic of China, India, Indonesia, Pakistan, Japan, portions of Egypt, and the rice-producing areas of Cambodia, Laos, Vietnam, Thailand, the Philippines, and Burma; terraced hillsides of some portions of Italy Highly intensive livestock enterprises Examples: Maintenance of many animals in limited space; beef cattle feedlots in western United States; large-scale broiler production in the United States; large-scale cattle feeding and pig-raising enterprises in Romania

    0 According to Phillips (1981).

    as do national agricultural policies, as shown by Schuh (1981) for selected agricultural areas. He concluded that government policy plays an important and growing role in how the worlds resources are utilized; consequences of policy are quite diverse and complex.

    A specific example of the interaction of government policy with production and marketing trends was given by Graham and Whitted (1978) for milk supplies



    Goal for year 2000

    Animal Production Average Average

    in 1974 of all producers of top producers

    Beef cattle Calf crop (a) Weaning weight (Ib) Daily gain, 500 Ib to market (Ib) Feed per 100 Ib gain; 500 Ib to market (Ib)

    Pigs weaned per litter Daily gain, birth to market (Ib) Feed per 100 Ib gain; birth to market (Ib) Weaning age (weeks) Market age (months)

    Lamb crop (%) Feed per 100 Ib gain (Ib)

    Broilers feed per 100 Ib gain (Ib) Turkeys, feed per 100 lb gain (Ib) Eggs per hen

    Milk per cow (Ib)

    Foal crop ('3%)




    Dairy cattle


    80 400

    900 2.5

    7.4 1.2

    350 6 6

    130 55OC

    200c 360c 220c



    1506 500

    675 3.25

    1 I b 1.6

    275 2 4.5

    3006 4006

    1 7 0 ~ 25OC 260=



    2006 700

    4 575

    156 2

    210 Day 1


    450b -

    - - -



    a According to Cunha (1978). bThe estimated increased numbers of young with beef, swine, and sheep are based on the

    assumption that the problems associated with multiple births will have been solved. Kottman and Geyer (1973) estimates.

    in the United States; They suggested that no other segment of the food industry has been more extensively regulated than the dairy industry. Free flow of milk and milk products in commerce has been impeded; surpluses have become com- mon. But the intensive regulation has developed standards of the highest order for milk and milk products.

    2. Production: Species

    When grains are low cost, productivity of ruminants can generally be in- creased by feeding grains, thereby increasing production of milk, meat, and/or


    work per individual animal. When grains are costly, the ruminant can convert otherwise nonusable plant energy and protein to produce valuable animal pro- teins at low cost. This flexible, noncompetitive symbiosis with the food and power needs of man accounts for the great numbers of buffalo, cattle, sheep, and goats found in nearly all tropical countries irrespective of population density or economic level. India, with its high population density, has one of the worlds highest densities of ruminants-living symbiotically-not competitively (Stonaker, 1978).

    Stonaker (1978) emphasized that there is no cheaper way of improving rumi- nant production than by genetic improvement. He provided illustrations for animal production in the tropics and concluded that the effects of increases in production of 1-2% per year through better directed selection coupled with greater utilization of heterosis would be multiplied by the great number of ani- mals that harvest that area. Perhaps another 15-20% improvement could be expected through planned crossbreeding.

    Within the world of ruminants, small ruminants have a special place (Fitz- hugh, 1981). The majority of the worlds small ruminants (sheep, goats, cam- elids, and countless wild ruminants) are providing food and fiber in those coun- tries where poor nutrition and low income are most prevalent. Small ruminants have several advantages over cattle for small holders. Financial investment per head is small, as is risk. Continuity, more so than quantity, of animal protein is critical in the diet of children and young women in small farm communities. Near the equator, managed mating of goats can ensure a continuous supply of milk from females freshening periodically throughout the year; few 5-ha farms can support sufficient cows to provide this same continuous supply of milk.

    Even though nonruminants depend heavily on grain production and their num- bers would decrease as grain prices increase, the nonruminants are of immense value as scavengers. As such they could be improved by the same breeding plans available to commercial producers.

    Originally the pig was raised as a means of utilizing food wastes. Commercial pork production in the United States and Europe developed under conditions where cereal grains and other high-energy feeds were available at low cost. What about the pros and cons of the pig as a food producer (Pond, 1981)? Pond listed the improvements achieved in efficiency of production and the opportunities available for further improvement. He emphasized that new technology will be needed to ensure the continued growth and viability of the swine industry. Such new developments must include both changes in the pig and the way the pig is raised. Their large variability in size, appearance, growth rate, efficiency of feed utilization, and other production traits provides the opportunity for genetic selec- tion to meet any particular need. At the moment, agricultural technology must provide the feeds appropriate to meeting the pigs needs, and some of this is


    being done. New high-lysine varieties of maize and barley show promise of saving in protein supplementation of swine diets. In less industrialized countries, local products not suitable for human consumption must form the basis for feeding programs. Later on, industrial, nonagricultural products can reduce the stress on competitive feed sources and ease conditions for swine feeding.

    Milligan and Wilcke (1981) continued the analysis begun by Byerly (Table I) of dividing countries into categories based on their per capita food energy avail- ability and applied it to poultry products. Per capita availability of poultry meat followed general food energy availability: Category 1 countries had the lowest and category 5 the highest. Availability of poultry meat products rose in the period from 1972 to 1977 for countries of all groups, but most of all for those in group 3 (2450-2800 kcal/capita/day). The same relationship held for egg pro- duction; in this instance, the largest increase in the above-mentioned 5-year period was for the countries with the lowest food energy availability (category 1:


    without these amino acids. Methionine is largely utilized in poultry feeds and lysine in swine feeding. Methionine is produced by chemical synthesis from industrial chemicals; lysine is produced by fermentation with cane molasses as the substrate. Hence the price of lysine depends in part on the price of this by- product of sugar manufacture.

    Beigler (1976) reviewed the history of costs of synthetic amino acids and predicted that the time would come when the amino acid requirements of an adult human could be met at the cost of $0.11 (US.) per day. Harada and Saito (1978) were not quite so optimistic. They thought that threonine and tryptophan will also become needed as the pressure increases to devote more land to cultivation of rice and corn, which are more efficient in utilization of solar energy, at the expense of soybeans.

    Instead of dealing with isolated amino acids, one could consider the entire microorganism as a noncompetitive source of nutrients, and consumption of microorganisms antedates pure amino acids. In recent years, the concept of producing protein from petroleum in the form of microorganisms grown on that substrate attracted wide interest. Waldroup (198 1) concluded that the future of microorganisms grown on petroleum substrates is uncertain. The problem of increased nucleic acid ingestion with the accompanying risk of raised blood uric acid levels will prevent inclusion of high levels of microbial proteins in human feeding. But, after reviewing the status of proteins from yeast, bacteria, algae, and fungi, he foresaw increased application of microorganisms of all sorts, grown on other than hydrocarbon substrates, in feeding of animals. Physical forms and their level in the diet play almost as important a role in acceptance by the animal as does nutritive value.

    How sophisticated must microbial production systems be? Waldroup (198 1) stated that care must be taken to ensure that production of undesirable or less desirable species does not take place at the expense of organisms of known quality. It is well documented that the nutritive quality of the various microbial products is influenced by processing techniques; this may limit extensive utiliza- tion of microbial proteins grown in less sophisticated systems. Much progress has been made in developing algae growing systems based on sewage by-prod- ucts and effluent. The possibility of accumulation of heavy metals in such sys- tems remains a consideration that requires more study.

    In 1974 the amount of cellulosic waste in the form of straw alone was 1.8 billion tons/year worldwide. It is estimated that all cellulosic wastes, worldwide, exceed 100 billion tons/year. These constitute a great potential source of energy for ruminants to take the place of grains fit for human consumption. Even nonruminants such as the horse and rabbit are actually pseudoruminants in terms of digestion (Ward, 1981).

    It is possible to influence the degree of utilization of cellulose by controlling


    the state of maturity of the plant, by physical state, and by chemical and biolog- ical treatment. Indeed, it is possible to improve digestibility of cellulosic wastes by several treatments, and this lends credence to the notion that utilization of cellulosic wastes will increase in ruminant feeds. Cellulose utilization is limited in swine.

    Where animals are managed under intensive systems, the waste can be re- covered and recycled into animal feeds. Fontenot (1981) estimated that 50 mil- lion tons (dry matter) of livestock and poultry waste is collectable annually in the United States. Such wastes constitute a disposal problem, yet have nutritional value. Satisfactory performance has been obtained from feeding wastes to farm animals without affecting the taste of meat and milk. He considered the potential hazards, such as pathogens, pesticides, medicinal drug residues, and accumula- tion of heavy metals, and found that these do not appear to pose serious prob- lems.

    Production efficiency could be increased by judicious feeding of animal wastes; the kind and amount fed would depend on the class of animal involved. Only a limited level of animal waste would be fed to lactating dairy cows and fattening cattle; the level could be much higher in dry cows or other low- producing animals.

    4. Marketing

    The underlying trend dominating food production and consumption is the greater and greater reliance on fossil fuel energy. Whatever newness is inherent in animal production is associated with intensive systems which make great demands on fossil fuel consumption. Brockmann (1978) compared corn produc- tion in the United States and Mexico. Production of corn in the United States requires only 22 man-hours/ha, yielding 5080 kg, while in parts of Mexico, where corn is raised mainly by hand labor, 1144 man-hours are required to produce 1944 kg in the same area. An expenditure of 0.37 kcal of fossil fuel is required for every kcal of corn produced in the United States; in the Mexican example, 0.0078 kcal are required, one-fiftieth as much.

    The same holds for the entire food chain. In the United States, one-third of the total food systems consumption of energy goes for production, another third for processing that includes transportation and packaging, and the rest is for re- frigeration, cooking, and waste disposal in the home and institutions. The meat- packing industry accounts for 0.17% of the energy budget in the United States and 10% of the energy consumed by the food industry. Slaughter of beef requires 415-830 kcal/kg and of pork 830-6650 kcal/kg. Processing of sausage requires 775- 1050 kcallkg and canning, 2500 kcallkg (Brockmann, 1978).

    These changes in dependence on fossil fuels have taken place so slowly and


    gradually as not to appear revolutionary, and yet they are. They emphasize the great and enlarging gulf between the industrialized and nonindustrialized coun- tries. Both Deatherage (1974) and Brockmann (1978) looked at emerging new technologies available for industrialized countries and projected further improve- ments in the marketing of meat products. Both of them were more reserved about how much can be done to improve marketing of meat in less industrialized countries given the constraints of inadequate refrigeration and inadequate in- frastructure. They made suggestions, but, as Brockmann (1978) put it, large installations for the preservation of meat would require elements of a social revolution. Deatherage (1974) cited control of foot-and-mouth disease as a factor that could promote greater trade in meat products to the advantage of less industrialized countries.

    Much the same marketing story obtains for poultry and eggs. Brant (1974) diagramed the complex industrial and financial arrangements that were devel- oped to support the intensive production of these foods. Farm flocks, while not completely gone, account for only 10% of the production of poultry and eggs in the United States. The improved systems for production and marketing account for the uninterrupted rise in consumption of poultry products in most of the industrialized world. In the United States, another significant rapid change, in addition to the general increase in poultry consumption, was the increase in the amount going into institutional markets: In the decade from 1960 to 1970 the increase was from 400 million Ib to 2 billion. Another major change in poultry marketing was the increase in the quantities that were further processed- halved, quartered, cut into pieces, deboned raw, cooked, canned, or given other forms of additional processing. These new products have opened new home and institutional markets for poultry meat.

    Progress in intensive production and marketing together with a proliferation of new products will continue in the short run as the new products become more attractive and competitive in cost. Sooner or later, intensive poultry production will need to face up to the problems of increasing grain costs and increasing competition for grain by humans. But alternate feeds may reduce the economic pressure.

    Brant (1974) was less willing to assess the future of poultry production in less industrialized areas. The system is not there; all support functions need to be in place. Milligan and Wilcke (1981) provided examples of less sophisticated ver- sions of a complete system suitable for more primitive circumstances. Nev- ertheless, such systems were incapable of supporting the kind of growth in the poultry industry seen in the industrialized countries.

    Profound changes took place in milk production and marketing in the United States in the two decades of 1950-1970 (Weisberg, 1976). While milk produc- tion remained constant, there were fewer milk cows, and more fluid milk was


    handled in plants and not on the farm. The man-hours of labor per unit of milk production were drastically reduced as were the number of milk plants. Home delivery routes were cut by 75%, fluid milk sales remained at the same level, milk fat for butter production declined, whereas milk fat for production of cheese and nonfat dry milk solids increased. There is an increased and lasting interest in low-fat dairy products and in fermented products such as yogurt.

    Both Weisberg (1976) and Graham and Whitted (1978) saw numerous oppor- tunities to design new dairy food products to fit the desires of consumers for more convenience products and to meet nutritional requirements. One of the major advantages to dairy products is their value as a source of calcium, consid- ered deficient in modem diets, especially among women. Dairy products alone in their many forms or mixed with other protein sources such as soy protein will find increasing new markets. But the dairy industry has suffered from inadequate expenditure for research on new products, and this will affect the rate of ap- pearance of new developments.

    As previously noted for the meat industries, industrialization is required for maintenance of a viable dairy industry. Many of the required components such as trained labor, transport, good dairy animals, adequate veterinary service, and dairy farmers are lacking in less industrialized countries. Real progress will require many years for the development of the above-mentioned components to a reasonable level in such countries.

    Marketing programs are oriented to the needs of the consumer. One type of consumer of vastly increased importance is the institution. Livingston (1976) documented the great growth of the food service industry in the United States. He estimated that restaurant and bar sales will range between 77 and 81 billion dollars in 1985. This will certainly influence the food industry and its products. There will be a premium on convenience foods that allow the serving of a full range of menu items without need of a skilled cook or baker to ensure customer acceptance. All traditional and novel methods of processing protein foods are represented, including freezing, refrigeration, dehydration, and freeze-drying. However, frozen foods dominate the scene.

    Some of the trends that might be expected are more multiunit operations with strong centralized direction and standardization in facility design, menu plan- ning, procurement, preparation, service, and management controls. While fran- chised food operations can be expected to increase, there will remain a demand for quality restaurants that prepare food specialities and provide services that the fast-food operator cannot duplicate. Consumers will demand continuation of elements that have characterized food service in the past: predictability of menu, consistent quality of food, modest prices, rapid service and/or self-service, lim- ited choice, snack-type minimeals, and ethnic specialties. The practice of sepa- rating food preparation from food service will become more widespread. This


    will be the place that pioneers in the introduction of food analogs. This is where (in the School Lunch Program) textured vegetable protein was introduced into the food system. However, institutional facilities are not limited to restaurants and bars. There are the armed services, school feeding programs, and the hospi- tals. All these together will affect the marketing of protein food products. And it is in institutions where the most efficient and least costly use of existing nutrients can be organized and fed in situations of food scarcity.


    1. Production

    Van Cleve (1978) reviewed the various efforts to estimate the maximum potential fish catch from the Oceans. About 100 million metric tons annually is a reasonable approximation. The world catch of marine and diadromous fish in 1972 was 50 million tons; this was 77% of the total fish catch that includes freshwater fish, crustaceans, and mollusks.

    The fish catch grew steadily after World War 11. The early period represented steady growth for Japan and the Soviet Union who were building distant water fleets. They were joined by distant water fleets of other industrialized countries. Although the world catch of fish has been growing rapidly since 1948, most of the increase has been landed by the fleets of industrialized countries (aside from Peru) and much by their distant water fleets.

    A study of the growth of the fish catch shows clearly that the oceans stocks of fish have not been managed properly. The total catch of fish has been unstable since 1968. Overfishing is expensive and devastating. A study of maximum and current yields of selected species made around 1970 shows that the maximum yield of 10 species was 24 million tons; the current yield of these same species at that time was 5.5 million tons. It was estimated that in 1976 the world suffered a loss of 5 million tons of fish as the cost of overfishing.

    Heretofore, the worlds fishing fleet has followed a policy of boom and bust: New stocks of fish and unexploited areas were always available somewhere as a reserve. This policy is now coming to an end. The future development of fish- eries and the continued production of fish from natural stocks will depend on what is known as rational use. This will require rational national management of coastal waters and international agreements for management of fishing on the high seas. If rational programs can be put into place, Van Cleve suggested that we can look forward to a maximum sustained production from the oceans. This may reach from 90 to 115 million tons, larger than what can be expected under the current practice of overfishing, which is more comparable to the aboriginal system of shifting agriculture.


    Aquaculture or fish farming has a role in providing protein foods; such farm- ing was described by Love11 et al. (1978). Production can depend exclusively on natural aquatic foods; the natural foods may be supplemented with artificial feed or the fish can be farmed intensively under artificial conditions. The second category, supplementation with artificial feed, is of the greatest interest: Yields are much higher than without supplementation, and there is the opportunity to find uses for low-cost feeds and to recycle waste products.

    The energy requirement for metabolism is less in fish than in warm-blooded animals because fish do not have to maintain constant body temperature, and they exert relatively little energy to maintain position in the water. Hence, the amount of protein synthesized per unit of metabolizable energy is higher than for poultry and livestock.

    Fish farming utilizes resources unused in other agriculture. Swamp or poorly drained land is suitable for some types. Any available aquatic environment where a portion can be confined or controlled is a potential resource for fish culture. The food materials needed to supplement the aquatic environment could be low cost, such as chicken manure or ground trash fish mixed with rice by-products, or they may be more complete foods designed to meet the requirements of the species produced. More efficient utilization of food resources can be brought about by polyculture-different species of fish with divergent feeding charac- teristics stocked in the same environment. Species of fish successfully farmed include carp, perch, channel catfish, trout, eels, and tilapies. Fish farming is suitable both for industrialized and for less industrialized countries, and it is adaptable to large and small operations. However, expansion has been most rapid in countries with the highest level of capital, technology, research, and extension.

    Invertebrate seafoods also are a source of protein, although their cost generally is too high for most low-income groups. Idyll (1978) reviewed progress and prospects in farming mollusks and crustaceans.

    Production of oysters in the United States is one-third of what it was at the turn of the century. Noteworthy is the drastic reduction in the productivity of the Chesapeake Bay-down from about 120 million tons of meats annually in 1880 to less than 25 million tons in 1960. Far better yields are obtained when the oysters are grown by vertical suspension on ropes as in Japan instead of bottom culture as in the United States. The yields in the United States could be increased 2-fold by increasing the efficiency of the present culture and 10-fold by changing to vertical culture.

    Clams have been cultured in a simple way in Asia for centuries; there is now sufficient knowledge to support more sophisticated culture. There is great de- mand for expansion in Asia of the low-intensity culture presently practiced. The technology already exists for hatchery production and profitable rearing, al-


    though improvements are required to increase survival of the young. Commercial culture in the United States is marginally feasible.

    Mussels can be produced cheaply because of the relatively simple culture required, the lack of need for supplemental food, and the high yields per unit area, and they are relatively free from deseases and parasites. Despite these advantages, culture of mussels has lagged behind that of oysters. This is related to demand, especially in North America. But demand is spreading. Mussel culture may become one of the most significant and productive of all forms of estuarine aquaculture.

    Shrimp culture is locally significant in some parts of Asia, mostly as a side activity to the pond culture of estuarine fish. Intensive culture of shrimp by Japanese methods, involving nearly complete control of the life cycle of the animals, has been accomplished with limited economic success, with the pos- sibility of future development.

    Lobster culture techniques are well advanced, but the slow growth, aggressive behavior, and cannibalism of northern lobsters have prevented these advances from being translated into commercial culture. A small quantity of crabs is raised in Asia, incidental to fish and prawn culture, but this activity is insignificant. The technology for more sophisticated culture is not far advanced; commercial opera- tions for this group cannot be expected in the near future.

    Idyll (1978) concluded that aquatic farming offers great promise, with yields per unit area rivaling or sometimes greatly exceeding agricultural land. The development of large-scale commercial culture of aquatic invertebrates has been delayed longer than many observers expected. Rising costs, increasing pollution, and mounting competition for the water areas will cause further delays. But increasing demand will make it likely that commercial culture will be established in time.

    2 . Marketing

    Pigott (1976) described the type of processing practiced on the world catch of aquatic animals in 1970. Most fish are sold fresh, but all fish sold fresh can also be sold frozen. Many are canned and a few are dried, salted, or smoked. When properly packaged and processed, frozen fishery products maintain the quality of a seafood product closest to the fresh condition. The original development of the frozen fish industry was in industrialized countries that had the basic processing, storage, transport, and marketing sophistication to handle the frozen items. With the development of small package freezing units that can be operated eco- nomically, frozen foods have come within the reach of most countries. But the retail price of frozen seafoods is beyond the pocketbooks of many people, and this situation will persist.


    Pigott commented that it is ironic, with much of the world placing an emphasis on improving technology and sophistication of food preservation techniques, that the other and larger part of the world actually needs the implementation of preservation processes known to primitive man. Hence, one portion of the world will consume seafood products fresh or will freeze them for future consumption, and the other will improve its protein supply through curing processes that stabilize the flesh through chemical action (e.g., salting) or by removing suffi- cient water to prevent bacterial growth during ambient temperature storage.

    Unlike the agricultural areas of the world, only a portion of the potential of the sea for harvesting natural populations and for fish farming is being utilized. Aside from the reduced harvest mentioned earlier, the majority of landed weight of fish is wasted or underutilized. Forty million metric tons of fish products are discarded or processed into cheap animal feed. The technology for doing better is cvailable. The increased emphasis on controlling pollution could well force the seafood industry toward a total utilization concept whereby most of the cur- rent waste material must be viewed as secondary raw material. Pigott (1976) asserted that the technology for supporting major expansion of the seafood indus- try is available. The removal of financial and political barriers could enhance the world protein supply from fisheries.


    As we look back over this summary, certain observations recur. And, because they were made independently by the various authors, they reinforce each other and have the effect of building up to great, albeit simple, generalizations. Animal foods are accepted, sought for, and respected now as they were from the begin- nings of history. In this respect there is nothing new about their status. But the way that they are produced and marketed, the opportunities for more efficient production, and the challenges facing continued production by present methods are new, and therein lie the generalizations.

    The production of animals and their marketing is highly industrialized and, in fact, is no different than production and marketing of nonfood industrial prod- ucts. The gulf between the industrialized and nonindustrialized countries has widened in this respect as in all other aspects. No longer can nonindustrialized countries hope to approach the efficiency of animal food production of the industrialized countries simply by setting out to do so with the best of intentions. Access to fossil fuel is required, trained professionals, trained workers, and an infrastructure of education and industry must be in place, as well as the kinds of operations that go under the rubric of sophistication. Less industrialized coun- tries are a world apart; some hints are given here and there on what they can do to


    improve their status, e.g., simplified poultry production, small ruminants, and aquaculture, but no one pretends that such efforts at developing appropriate technology can do more than be appropriate to the ability, but not to the need.

    Intensive animal production is a highly industrialized undertaking, even on the water. Marketing of perishable products requires freezing and refrigeration facil- ities, or other industrialized ways of preservation, and superb distribution facili- ties. Otherwise the alternatives are the millenia-old ways of drying and salting. But with this quantum leap in production and marketing of animal products there is now the refrain echoed by author after author: What will happen when grain prices reach the point at which intensive animal production is jeopardized? All authors agree that industrialized countries are moving in that direction, that competition from humans for grains now fed to animals is increasing, that adjustments will be required-not all at once or perhaps never to the extent that some might fear-and everyone is thinking about alternatives. This is the second generalization.

    The first reaction is to concentrate more research on better breeds and better management to reduce the grain costs of production. There is room for progress, as pointed by Cunha (Table IV) in his proposed goals for the year 2000.

    As was stated earlier, the ruminant stands out as the animal with the greatest flexibility. No one doubts the continued dominance of the ruminant as a food producer. Even though in some stages under intensive production in fattening pens the ruminant was treated almost as a nonruminant in terms of grain cost of food production, there can be some backing off without serious damage to efficiency. Actually, the fat content of meat could be decreased thereby; most medical authorities would consider this a good thing.

    Not so with swine and poultry. They lend themselves well to intensive produc- tion. Consumption of poultry meat has made great gains in this century prin- cipally because industrialized production has provided low-cost animal protein. Even some nonindustrialized countries have increased their poultry production. Commercial pig production too has achieved a high efficiency of production, but depends heavily on availability of low-cost grain. Hence, commercial pig pro- duction is the most vulnerable to increased grain costs. Much depends on changes that can be induced in the pig and the manner in which it is raised, particularly as this affects ability to utilize industrial, nonagricultural sources of energy as partial replacements for grain. Both types of animals will need to be adjusted to being fed some level of nonagricultural products, but probably will not do as well as the ruminant, programmed from the very beginning to do so.

    These concerns lead into the third generalization: the greater emphasis on recycling of waste products of agriculture and industry. Scavenging was the original rationale for poultry and swine, and these today together with fanned fish perform this useful service for most of the less industrialized societies. The


    question is whether these and the ruminant can convert what were originally home and backyard operations into industrialized recycling operations, which would solve two problems at once: the need to reduce dependence for feed energy on grains that can be eaten directly by humans, and the need to control pollution from waste of agricultural and industrial operations. Recycling will need to become a major concern for agriculture, particularly intensive animal production.

    The treatise does not deal with one major issue: losses from spoilage, losses owing to products being eaten by pests. Here too, industrialization carries with it better means of protection, but much can be done even under primitive condi- tions. The entire subject merits a treatise all of its own.

    Aside from these three major generalizations, other points might be men- tioned. The productivity of the oceans not only depends, as does agricultural productivity, on practice of more sophisticated operations, but on international cooperation to prevent overfishing. This problem is not unlike that of equitable distribution of fresh water from the great and small rivers to support land agriculture.

    Although milk and animal meats are prized as foods, there are medical con- cerns, not altogether settled, that may encourage shifts in consumption toward reduction in their intake, particularly in intake of animal fat. How significant this trend may be in its effect on consumption of animal products in industrialized countries remains to be seen.

    Are there any new animal protein foods? We commented on the process, how about the products? There are the many new forms: Brant (1974) described further processed poultry products; Weisberg (1976), new dairy products; and Pigott (1976), new products from fish. But they are no more than extensions of the classic products, more convenient to store and eat, providing more flexibility in composition, such as low-fat dairy foods.



    Cereals are the major source of protein worldwide. Yet when the protein content of a diet must be adjusted to a higher level to meet the special needs of infants and children, pregnant and nursing women, and those recovering from illness or injury, this is done classically by mixing cereals with other protein foods. Typically these are the legumes, containing 20% or more protein, and the animal protein foods-dairy, egg, fish, and meat products. Otherwise this can be


    done by improving the protein quality or content of cereals or by introducing protein concentrates produced from oilseeds. These latter two might be consid- ered new concepts, although the soybean, the major practical source of protein concentrates, has a long and ancient history as a food in the Orient.

    Altschul (1976) and Altschul and Schertz (1981) introduced the concept of protein food models as applied to new protein foods from plant sources. The presently known ones can be classified into one or another of three models: cereal-legume, nutritious beverage, and animal flesh.


    1 . Amino Acid Enrichment of Cereals

    The forerunner of amino acid enrichment of cereals is the successful public health outcome of the fortification of certain staple foods such as cereals with vitamins and minerals (Brin, 1976). If cereals can be carriers for vitamins and minerals, why not add deficient amino acids? Actually, the protein value of wheat is greatly improved by the addition of lysine, so that now its quality approaches that of a cereal-legume mixture. Lysine need not be added phys- ically; it can be bred into the grain, as reported by Forman and Hornstein (1976) and Bliss (1985). However, Bliss reported that such endosperm mutants of maize, sorghum, and barley have been a practical disappointment because these types usually have lower yields and less desirable agronomic traits than the normal counterparts. In either case, the theoretical predictions of higher protein quality have been proven experimentally (Jansen 1974, 1981). Amino acids can be added as powder to wheat flour, as preformed rice grains admixed with normal rice (Mitsuda and Yasumoto, 1974), or as pseudo corn kernels containing soy protein as well as amino acids added to normal corn before grinding (Forman and Hornstein, 1976). The cost of adding the amino acids for the cases of wheat, corn, and rice were estimated: Addition of lysine, vitamins, and minerals would add 5.5% to the cost of wheat, 2.5% to the cost of corn, and 7-10% to the cost of rice (Forman and Hornstein, 1976; Mitsuda and Yasumoto, 1974).

    These concepts were tested in large-scale pilot tests in three countries: Tunisia, wheat; Thailand, rice; and Guatemala, corn (Forman and Hornstein, 1976; Alt- schul and Schertz, 1981). In all instances the nutrients were delivered as ex- pected and generally accepted by the population. It is indeed possible to fortify cereals with added nutrients. The pilot tests were set up to measure possible nutritional advantages of such fortification in children. There were some, but they were insufficient to warrant continued interest in this procedure from a general public health viewpoint (Altschul and Schertz, 198 1).


    2 . Protein Enrichment of Cereals

    The classic example is the practice, popular worldwide in its many forms, of mixing cereals and legumes: rice and pulses, corn and beans, rice and soybean products (Bressani and Elias, 1974). Jansen (1974) and Bodwell and Hopkins (1985) provided examples of the variety of opportunities to complement and supplement the protein in plant sources with proteins from other plant products or with animal proteins.

    There is another classic example: the fortification of bread (Hulse, 1974). Eggs (challa) and milk represent excellent protein sources compatible with bread, but in less industrialized regions supplies of these protein sources are inadequate and prices too high to permit widespread application. Grain legumes and oilseed proteins offer probably the best protein alternative for such regions, but more research is required to develop novel methods of bread baking so that these protein-rich materials can be combined with wheat and other cereal flours to form cheap, nutritious bread. Synthetic amino acids might be of benefit were they to be manufactured at low cost within the countries themselves, or more nutritious cereals could be bred to be included in breadmaking.

    Mixtures containing cereal and oilseed, or milk protein, or both have been formulated and tested extensively as weaning foods in controlled, small-scale child-feeding trials. Such mixtures could be considered as new food products. They are designed to have the same nutritional impact as milk and are attractive where milk supplies are limited. They cost less than milk. Incaparina (Altschul and Schertz, 1981) containing cottonseed flour developed at INCAP in Guatema- la was one of the earliest of these weaning foods; corn-soy-milk (CSM) (Smith, 1976; Altschul and Schertz, 1981) and other blended foods have been and are being distributed widely in bilateral or international feeding programs.

    The adequate nutritional value of these mixtures, as determined in controlled small-scale trials, has been well documented. Yet, when these products were evaluated after years of being distributed in large-scale feeding programs in terms of improved growth, health, or better prognosis for the target children, it was not possible to provide evidence of efficacy in this regard. Even so, none of those who examined the data was willing to conclude that the programs were worthless (Altschul and Schertz, 1981). We have here yet another example of an intervention which theoretically should improve nutrition of children not sup- ported by detailed analysis. It is becoming difficult in all societies to prove the efficacy of almost any kind of an intervention in large-scale studies.

    As was pointed out, the weaning mixtures cost less than milk. But they cost more than the grains themselves,and they look and are eaten like grain prod- ucts.


    3. Commentary on the Cereal-Legume Model

    The examples of this model have the improved nutritive value expected of them. Large quantities have been distributed free by United Nations and national agencies. They have been accepted by children; even though no clear proof of efficacy was shown, they represent, especially the weaning foods, a viable way of dealing with problems of hunger and malnutrition among children. Efforts have been made to commercialize these products. Incaparina especially was set up as a profit-making, marketing venture, but neither this one nor others suc- ceeded on their own. In some instances, commercial production is supported by government subsidy. Altschul and Schertz (1981) suggested that these kinds of products will not become viable commercial ventures without government sup- port because, although their nutritional quality is better, this is not readily evi- dent to the consumer who only sees a product that looks and is eaten like a cereal, but costs more.

    Only the animal feeder who can tell by performance, which can be measured, the differences that can be induced by improved nutritional quality will be willing to pay more for such anticipated performance, regardless of how the mixture looks. Poultry and swine feeders have regularly fed mixed feeds not unlike the blended foods; amino acid supplementation is commonplace in their programs (Harada and Saito, 1978).


    Milk is the prototype of a nutritious beverage and is the one that most nutri- tionists would prefer for delivery of good nutrition. But the knowledge of nutri- tional requirements and the capability of food sciences have broadened the concept to many other examples. All kinds of balanced liquid diets utilizing milk solids or casein or soy protein as the protein source are available for a variety of feeding situations, either to lose or to gain weight. Vitasoy, a soy-based drink, is a commercial success in Hong Kong (Altschul and Schertz, 1981). Infant for- mulas have been available for most of this century (Altschul and Schertz, 1981). Formulas containing isolated soy protein as the sole protein ingredient are avail- able for infants who cannot tolerate milk products (Kolar et al . , 1985). Elemental diets are available for oral and intravenous feeding of those patients who cannot obtain their required energy and nutrients by the normal route (Beigler, 1976).

    A most interesting development of this model is the partial replacement of milk protein by either soy or peanut protein (Altschul and Schertz, 1981; Rhee, 1985). Some products of this type are commercially available in India (peanut protein) and China (soy protein).


    Moving in cost in either direction from milk we have a set of products more expensive-elemental formulas, infant formulas, special sources of balanced nutrition-which have a special market for those who want or need specialized nutrition and will pay for it. Then there are the less expensive drinks such as those in which less expensive vegetable protein replaces part of the milk protein. Their nutritional value should be adequate for most purposes. In this instance they look like milk, are drunk like milk, serve a similar purpose as milk, and are cheaper. This model surely could become the basis for viable commercial enter- prises. But such enterprises require a degree of industrialization for production, distribution, and marketing. They require more sophisticated treatment of the protein materials, to be discussed in the next section, and they require particular attention to proper flavoring (Mason and Katz, 1974). But pockets of indus- trialization may be adequate for local application of this idea and as a beginning.


    I. The Concept

    Meats (and animal flesh, generally) have characteristic flavors and textures (Mason and Katz, 1976; Deatherage, 1974) that appeal and make for appetizing foods. This is one of the reasons that meat products have enjoyed high status from the beginning of recorded history (Pyke, 1978). Meat products contribute important nutrients (Brant, 1974), but they are more expensive than plant prod- ucts. Ever since protein concentrates from seed storage proteins became avail- able (in this century), it has been tempting to reproduce the qualities of meat products and replace products of animal origin, in part or completely, by analogs of plant origin. There have been failures in the attempt to create meat analogs, but also successes. What follows is the story of the failures and successes, and the portents for the future.

    The key invention needed to create analogs was the ability to produce texture from high-protein products (Horan, 1974). This has been done first by spinning dissolved protein through spinnerettes (the textile model). This method did pro- vide excellent texture, but was too expensive and was replaced by the cheaper and more versatile method of thermoplastic extrusion (the plastics model) de- scribed by Horan (1974) and Smith (1976). Other ways of producing texture are described by Horan (1974) and Horan and Wolff (1976). Japanese approaches to creating texture are described by Watanabe et al. (1974).

    2 . Raw Materials

    Removal of lipids from an oilseed, the conventional process for producing edible vegetable oils, leaves a high-protein residue that can be fed directly to


    ruminants. After further purification, it can be fed to monogastric animals and, in some instances, can be converted into food for humans. The major sources include soybean (Horan, 1974; Horan and Wolff, 1976; Nielson, 1985; Kinsella et al., 1985; Kolar et al., 1985; Campbell et al., 1985), cottonseed (Table 11: Bodwell and Hopkins, 1985), peanuts (Rhee, 1985), sunflower seed (Lusas, 1985), rapeseed (Ohlson, 1985), and sesame (Kinsella and Mohite, 1985). For practical purposes most development of new textured food products on the meat or animal flesh model has been with soybean proteins.

    We distinguish three general types of soy protein raw materials: soy flour (-50% protein), soy protein concentrate (-70% protein), and isolated soy pro- tein (more than 90% protein).

    Soy flour production and applications were described by Horan (1974), Horan and Wolff (1976), and Smith (1976). Smith included a description of textured products made directly from full-fat soy. Most of the products described by Smith can go into cereal products, but the principle of texturizing that he de- scribed is the operation of choice for producing meatlike texture. Food products made with soy flour have the advantage of being cheaper, but mixing with meat causes flavor problems and creates physiological problems of flatulence from the oligosaccharides that are not hydrolyzed by the gut enzymes and reach the bowel where they are fermented; this limits the amount that can be added to meat products.

    One of the first applications of mixing textured vegetable proteins with meat was in ground meat patties. The U.S. Department of Agriculture school lunch regulations permit substitution of 30% of the product in meat patties with recon- stituted textured vegetable protein (Campbell et al., 1985). Originally the tex- tured protein was primarily based on soy flour. These products are now being replaced more and more with textured products from soy concentrate and isolate.

    Isolated soy protein was an early soy protein product needed first for industrial applications, but later as the ingredient for spinning to produce structured prod- ucts for food application (Kolar et al., 1985). Now it is sold for incorporation into meat, poultry, and seafood products and is the protein source in certain infant formulas and bakery products. An enormous technology for its incorpora- tion in foods has been developed. Its nutritional properties have been extensively studied, and its composition is well defined. The early history and utilization patterns were described by Horan (1974) and Horan and Wolff (1976). A thor- ough description of the isolated protein and its applications is provided by Kolar er al. (1985). Soy protein concentrate also was first produced in the late 1950s for industrial applications, but food applications began soon thereafter.

    Horan (1974) showed how all three types of protein products from the soybean relate to each other in the manufacturing process: Removal of oil from the seed produces a meal or flour; removal of the soluble sugars from the meal produces a


    concentrate; removal of the protein and its reprecipitation produces an isolate. Campbell et al. (1985) provide a detailed description of the properties and utilization in food products of the concentrate. They describe how concentrates are texturized and show how functional properties in food uses can be tailored to satisfy particular needs. Here, too, there is information on nutritional value and on composition. Most of the current production is used in the meat industry, and there is some application as milk replacers for infant animals.

    The present role of soybean protein products in the meat model is in mixtures with meat. With the exception of some specialty products, such as bacon analogs and meatlike foods for vegetarians, the technology has not developed to the point where complete meat analogs can be marketed successfully.

    The chapters in the treatise describing soy protein in food products were written by scientists employed in the soy industry. Each group sought to put their own product in the best light, and there is a degree of competitiveness in the presentations. However, the presentations are accurate and fair; the advantages and disadvantages of each type of product emerge. But beyond the individual stories arises a picture of a sophisticated, well-developed industry supported by research and development experience. This is truly a first-class food protein industry well able to take its place in the human protein arsenal, and its products surely demand serious consideration.



    When a new product replaces all or part of a well-known and accepted food and begins to make a significant contribution to the diet, it becomes necessary to be sure that nutritional status does not suffer or, better yet, that flexibility for improved nutrition is enhanced. Bressani and Elias (1974) and Jansen (1981) provided the background for considering the nutritional value of vegetable pro- teins as a source of protein. Bodwell and Hopkins (1985) discussed in detail the questions that arise in considering nutritional equivalence of vegetable proteins compared to animal protein. As shown in Table V, soy and rapeseed protein have the potential to provide, per gram of protein, all of the essential amino acids at 99% or more of the specified levels. Less than 80% of the specified level of lysine can be provided by peanut, safflower, sesame, and sunflower protein, of isoleucine by cottonseed protein, and of threonine by safflower protein.

    Excluding infants, both soy and rapeseed protein would be adequate for all age groups as a sole source of protein if all their constituent amino acids were completely available. Cottonseed protein would not be adequate as a sole source




    Children Adults

    No. No. Protein source of reports Mean Range of reports Mean Range

    Oilseeds Cottonseed Peanuts Sesame flour Soy flour Soy isolate Sunflower seed (flour)

    Other sources Meat, poultry, fish Milk, casein or lactalbumin Egg or egg albumen Corn Whole wheat Wheat, white flour Rice, polished

    1 1 I 3 3 -

    - 9 3 2

    3 3


    87 92 88 87 93 -

    - 89 93 72

    89 87


    - - -

    84-88 92-95 -

    - 86-93 89-97 62-82 -

    85-93 85-89

    3 4

    5 5 1

    10 12 19 4 6 2 4


    90 94

    86 99 90

    94 95 97 93 87 96 89


    70-98 91-98 -

    75-92 93- 107 -

    90-99 90- 100 92- 106 92-95 80-93 96-97 82-91

    a Adapted from summary by Hopkins (1981).

    of protein for children 2-5 years of age, but would be adequate for persons 10- 12 years of age and older. Peanut, safflower, sesame, and sunflower, as a sole source of dietary protein, could provide adequate amounts of essential amino acids for adults, but not for those 10-12 years old or younger.

    In general, as shown in Table VI, well-processed products from any of the oilseeds can be anticipated to have true nitrogen digestibilities of 90% or higher.

    The picture that emerges is that a satisfactory protein level and quality can be achieved by feeding vegetable proteins, some alone, but most in mixtures with other animal or vegetable proteins, or supplemented with certain amino acids. Most other nutrients are provided by the vegetable protein mixtures; those not available can be provided as supplements, e.g., vitamin B,*.

    Some trace minerals such as iron and zinc are not as well absorbed in vege- tarian diets. A review of experimental data bearing on this subject is given in Tables VII and VIII. In the meat model where soy protein constitutes 30% of the protein mixed in ground beef, no problems of trace metal deficiencies were observed. There is concern, however, that foods in the cereal-legume model fed in countries where an iron-poor diet prevails, iron and perhaps other trace miner- al supplementation may be required.


    Another problem area is the antinutritional factors in vegetable protein prepa- rations. In the soybean, these principally are trypsin inhibitors and hemag- glutinins (Nielsen, 1985; Bodwell and Hopkins, 1985). Their level is controlled to a presumably innocuous level by present methods of processing, particularly in the concentrate and isolate.

    Peanuts and sesame seed have the advantage of a long history as human foods. This advantage turns out to be a disadvantage, since only relatively small per- . centages of these seeds are available for oil extraction to produce high-protein flours. Rhee (1985) cites the economic position of peanut flour as the major drawback to it being considered seriously as a source of concentrated plant protein. Kinsella and Mohite (1985) cite the need to develop a commercially viable indehiscent variety (to reduce losses due to shattering of the seed pod before harvest) so that sesame may be grown in larger quantities as a source of oil and concentrated protein.

    CottQnseed contains gossypol which is toxic and must be reduced to very low levels before protein concentrates and isolates can be produced from the meal. The meal itself can be made suitable for human consumption by appropriate processing. Indeed, it was the protein concentrate component of Incaparina. However, the production of protein concentrates or isolates requires a purer starting material (Bodwell and Hopkins, 1985).

    Rapeseed has two compositional problems that interfere with human consump- tion of its products (Ohlsson, 1985; Bodwell and Hopkins, 1985). The major deleterious factor in the lipids of normal rapeseed is erucic acid, not considered suitable for human consumption. For example, normal rapeseed oil cannot be used in human foods in the United States. The meal contains glucosinolates, antinutritional factors for nonruminants. Hence the meal is now fed to ruminants. The principal cause for optimism is the success in breeding out both of these antinutritional factors. The most successful of all is the Canola type created in Canada and accounting for close to 95% of the production there. There is good reason to expect that its oil and meal will find greater utilization among humans and nonruminants. This type of meal could be an excellent starting material for some of the processes for production of protein concentrate described by Ohlson (1985). As Ohlson points out, rapeseed has the highest protein content of an oilseed grown in cooler climates and has great attraction for countries that cannot grow soybeans because of the climate.

    Sunflower seed also has great potential, particularly in cooler climates. It is the fourth major oilseed, yet it is not now a commercial source of concentrated protein. The major economic reason for production of sunflower seed is the income from the oil; markets for the oil determine the amount of seed grown. Sunflower seed also serves directly as a human food, hence demonstrating that no serious nutritional problems stand in the way of its being a suitable source of


    concentrated protein. However, the seed contains chlorogenic acid which causes green and brown discoloration of food products at alkaline pH and an irreversible green color in protein isolates produced by the tradtional method of alkaline extraction followed by acid precipitation. Lusas (1985) reviewed the status of sunflower seed protein as a human food and the prospects for further develop- ment.

    Some evidence suggests that the source of dietary protein could affect blood lipids independently of other dietary factors such as fat and cholesterol intake. Some clinicians have even gone so far as to shift hypercholesterolemic patients to vegetable protein diets from their prevailing predominantly animal protein diet. A summary of research experience in this important subject area is given in Table IX .


    Much of the early technology developed before adequate knowledge of the chemistry and structure of seed proteins was available. It was an empirical technology based on the known generalized properties of proteins in food prod- ucts. Many of the examples cited by Horan (1974), Watanabe et al. (1974), Smith (1976), and Horan and Wolff (1976) fall into this category, and it is amazing that so much could be done.

    In the meantime, progress was being made in understanding the seed storage proteins as species of proteins, i.e., their composition and various structures (Nielson, 1985; Kinsella et al . , 1985) and their chemistry and modifications by chemical and physical means (Feeney and Whitaker, 1985). Most of the progress is reported on the soy proteins, but Rhee (1985) and Kinsella and Mohite (1985) show progress in knowledge of peanut and sesame proteins, respectively.

    The chapters in the treatise by Campbell et al. (1985) and Kolar et al. (1985) represent a transition. Many of the applications are still based on empirical knowledge, but more of the explanations depend on the newer knowledge of specific protein structures, promising that the next generations of technology will be less empirical and derive more from specific knowledge of the chemistry of the oilseed proteins.


    As with technology, so too with genetics. Breeding for improved phenotypes depends on better knowledge of the gene structures involved. It is now possible to breed for better protein quality-be it amino acid composition, reduction in level of antinutritional components, or better functional properties wherein one protein component is emphasized as compared to the others. Locations on the gene that control for these protein structures are becoming better understood.


    It is now possible to consider homologies in protein structure common to species of seeds and evolutionary patterns in these structures (Dieckert and Dieckert, 1985). These authors conclude that reserve proteins fit the expectation of a protein that can function satisfactorily with a minimum of structural specifi- cations. Diversity seems to be the result of genetic accidents that leave the biosynthetic processes intact. If this is true, it should be possible to replace residual structural genes for the reserve proteins with strange genes to satisfy the needs of man and the plant.

    Bliss (1985) notes that quantitative variation in seed protein content is quan- titiatively inherited, but environmental effects are substantial. Alterations in the amounts of specific proteins have been associated with the effects of one or a few major genes and are highly heritable.

    Beachy and Fraley (1985) use a model gene system-one encoding a subunit of P-conglycinin, the 7 S seed storage protein of soybean-to discuss the poten- tial problem that might arise in transferring a gene from one plant to another. They then review the current methodologies for gene transfer and the potential for genetic engineering of soybeans. Success in gene transfers to new plants with derivatives of the Ti-plasmid of Agribucterium turnefaciens indicates the poten- tial of the new methods. However, transfer of techniques to soybeans which were successful with tobacco, carrot, petunia, and sunflower will demand research on problems unique to the soybean.


    HUB (1974) was concerned with laying down principles to govern the intro- duction of new protein foods, as might happen when products based on plant proteins reach the marketplace. He concluded that two principal concerns must underlie any regulatory policy governing new protein foods-the need to protect the public against unsafe or misrepresented food and the need to provide the food industry with an incentive to continue product development leading to improved new foods. Nutritional equivalence is one of the issues that must be part of regulatory concerns (Bodwell and Hopkins, 1985).

    Campbell et al. (1985) reviewed new regulations of the Food and Nutrition Service of the U.S. Department of Agriculture, which administers the national school lunch program. These have broadened somewhat the range of products, including soy protein concentrate and isolate, that may be included in this pro- gram. Labeling is important and regulations require that when plant protein substitutes for traditional protein foods are incorporated, the term vegetable protein product be included in the label. If the food resembles and substitutes for a traditional major protein food and is nutritionally inferior, the food must be labeled imitation.



    The only way to describe the accumulation of knowledge on the genetics, chemistry, and technology of oilseed storage proteins is to describe it as an explosion. Surely, primitive methodologies can and are still applied to produce foods for human consumption. But the great new knowledge now available exerts considerable pressure to raise the level of sophistication in dealing with oilseeds to make them more suitable to meet human needs, and more knowledge encourages more and better research.

    The oilseed protein food industry now finds itself with much in common with the most advanced animal production and marketing systems, Industrialized societies that were able to take advantage of new knowledge in land and marine animal production and marketing now can add concentrated proteins from plants as an additional resource. The advantage in capacity to meet food needs of the industrialized societies over the nonindustrialized societies now spills over into the utilization of plant proteins as well.


    New technologies affect how people eat by changing the relative price of some methods or their attractability relative to others. Thus cultural patterns change. Examples are margarine compared to butter, poultry meat consumption increas- ing rapidly in this century, the gradual introduction of proteins from plants into products on the animal flesh or meat model. New technologies make it possible to produce more food-energy and protein-to meet the needs of growing populations and their greater demand for animal products. In addition, new technologies affect how one nation or nations acting in concert can transfer food resources to other nations to cope with acute or chronic food shortages. The issues involved in intervention were discussed in a series of essays by Altschul (1974, 1976), Wilcke and Altschul (1978), Wilcke and Hopkins (1981), and Altschul and Schertz (1981).


    Food energy shipments to avert famine and hunger have classically consisted of grain and oil, and these remain the same. Shipments of protein-rich foods for vulnerable elements in a population were confined mostly to milk products. Now, weaning and child-feeding foods can be wholly of animal origin (nonfat dried milk), mixed animal and plant (e.g., CSM), and totally of plant origin


    (e.g., wheat-soy blend) (Altschul and Schertz, 1981). Animal-based foods are no longer essential. Hence, a new flexibility has been introduced in providing emergency shipments to cope with hunger and malnutrition.

    In less industrialized countries, there has been some modest success in for- mulating and distributing weaning foods that contain either some form of defat- ted soy flour or cottonseed flour. When attempting to incorporate oilseed pro- teins into more general diets in such countries, it is usually necessary, for reasons of cost, to use products of relatively low technology such as defatted flours and products textured by simplified extrusion. Use of an oilseed produced in the country where it is to be consumed is also desirable in order to avoid balance of payment problems. With these restrictions, it is difficult to produce foods con- taining oilseed protein that have the gustatory qualities of traditional foods. Oilseed products compete with staple foods such as grains and tubers for the consumers money, and the oilseed products usually cost more. This makes the distribution of oilseed ingredients and their incorporation into food extremely difficult to achieve.

    Probably the best way to distribute weaning foods which contain oilseed proteins is through government programs (Altschul and Schertz, 1981). In order to pursue such a program, it is necessary to obtain full backing of the govern- ment, and such a program probably must be tied into other programs, including improvement of hygiene and medical care. For continued government spon- sorship and support, demonstration of some measurable improvement of nutri- tional status, such as reduced child mortality, is essential.


    In general, the poorest countries have not been able to maintain a per capita protein supply (Wilcke and Hopkins, 1981). The population level in the poor countries is already such that primitive methods of protein production or of all food, for that matter, are inadequate to meet the needs of the vulnerable among them-the infants, children, pregnant and nursing women, and the ill. More- over, the rapidly expanding populations delay and postpone the assembly of the critical mass of knowledge and industrial capacity needed to provide the basis for increased and sustained new production, so that the opportunity to work out solutions diminishes or disappears.

    In the meantime, it has been possible to increase longevity (decrease infant mortality) by public health means (e.g., improved water, vaccination, and vi- tamin and iron fortification) that do not require the degree of industrialization required for optimum food and protein production. But such improvement in longevity is not proceeding rapidly enough, nor is it sufficiently sustained in itself to instigate effective means of control over rapid population growth.

    There are no easy solutions to this problem. The only moral alternative to


    chaos, famine, and social destruction or war is aggressive pursuit of acceptable measures to limit population growth. Wherever these can succeed, the chances improve for industrialization and for bringing into better balance the capacity to provide food for the large populations that are there to be fed.

    Given that the role of protein foods and foods in general is appreciated in its proper perspective, then some statements can be made about long-term prospects for protein food supply. Fortification of certain foods with vitamins and iron is possible and sufficiently cheap to be practical. The least costly way of increasing protein supply-supplementation of cereal grains with amino acids, principally lysine-adds 2-10% to the cost of the grain.

    As urbanization increases in less industrialized countries, high-technology oilseed ingredients or comparable ingredients from other plant sources may play a larger role in increasing nutritional well-being. If people enter the economy and have more money to spend, foods containing protein concentrates and isolates may help provide a more abundant supply of protein. Animal products, most notably meat and milk, will compete for the consumer's money, and traditional eating quality will be expected. Government support of nutritious products, such as emulsified meat containing both animal and vegetable protein, will be essential. Given the sometimes conservative nature of regulators and rulemakers around the world, an all too frequent occurrence is the prohibition of the use of high-quality oilseed protein in a milk or meat product when nutritional and economic condi- tions would suggest otherwise. In the future, more attention will be focused on writing specifications that would allow and encourage the utilization of plant protein products, while ensuring that safety and overall nutritional quality of the diet is at least maintained, if not enhanced.

    Foods on the cereal-legume model cost less than milk, but more than the grain itself. They are suitable as shipments from one country to another and may exist locally where the government is willing and able to subsidize their cost. They do not seem to be able to attract consumers to provide the basis for commercial sales.

    Protein foods on the milk model (mixtures of vegetable protein with milk) would seem to have real possibility, but requite pockets of advanced technology for their production and marketing. However, protein foods on the meat or animal flesh model (mixtures of textured plant protein and meat) are still con- fined to the most industrialized countries.

    The capacity of industrialized countries to provide protein foods continues to increase, be it in animal production, fish catch, or in the new plant protein products. At the present level of population, the industrialized countries have a better chance of solving problems affecting protein supply, be they higher grain costs, higher fossil fuel costs, or problems of pollution and disturbance of the ecology. They have the resources and their populations are relatively stable.



    Adams, I . (ed.) 1982. Sunflower. National Sunflower Association, Bismarck, North Dakota. Altschul, A. M. 1974. Protein food technologies and the politics of food: An overview. In New

    Protein Foods, Volume 1 , Technology, Part A (A. M. Altschul, ed.), pp. 1-39. Academic Press, New York.

    Altschul, A. M. 1976. Limits on technology. I n New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 280-304. Academic Press, New York.

    Altschul, A. M. (ed.) 1974. New Protein Foods, Volume I , Technology, Part A. Academic Press, New York.

    Altschul, A. M. (ed.) 1976. New Protein Foods, Volume 2, Technology, Part B. Academic Press, New York.

    Altschul, A. M., and Schertz, L. P. 1981. Protein food models. I n New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 335-364. Academic Press, New York.

    Altschul, A. M., and Wilcke, H. L. (eds.) 1978. New Protein Foods, Volume 3, Animal Protein Supplies, Part A, Academic Press, New York.

    Altschul, A. M., and Wilcke, H. L. (eds.) 1981. New Protein Foods, Volume 4, Animal Protein Supplies, Part 8. Academic Press, New York.

    Altschul, A. M., and Wilcke, H. L. (eds.) 1985. New Protein Foods, Volume 5 , Seed Storage Proteins. Academic Press, New York.

    Anderson, J. T., Grande, F., and Keys, A. 1971. Am. J. Clin. Nutr. 24, 524. Anjou, K., and Ohlson, R. 1978. Rapeseed Protein Concentrate. Karlshamns Research Laborato-

    ry, Karlshamn, Sweden. Beachy, R. N., and Fraley, R. T. 1985. Potentials for application of genetic engineering technology

    to soybeans. In New Protein Foods, Volume 5 , Seed Storage Proteins, (A. M. Altschul and H. L. Wilcke, eds.), pp. 89-105. Academic Press, New York.

    Bean, L. 1978. Food and people. I n New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 21-46. Academic Press, New York.

    Beigler, M. A. 1976. Complete synthetic foods. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 62-85. Academic Press, New York.

    Bliss, F. A. 1985. Relationships of genetic engineering to genetic technology and plant breeding. In New Protein Foods, Volume 5 , Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 65-87. Academic Press, New York.

    Bodwell, C. E., and Hopkins, D. T. 1985. Nutritional characteristics of oilseed proteins. In New Protein Foods, Volume 5 , Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 221-259. Academic Press, New York.

    Bodwell, C. E., and Marable, N. L. 1981. J . Am. Oil Chem. SOC. SS, 476. Bodwell, C. W., Judd, J. T., and Smith, J. C. 1981. 1980Annu. Rep. Am. Soybeon Assoc. Res. Bothwell, T. H., Clydesdale, F. M., Cook, J. R., Dallman, P. R., Hallberg, L., Van Campen, D.,

    and Wolf, W. J. 1982. The Effects of Cereals and Legumes on Iron Availability. The Nutrition Foundation, Washington, D.C.

    Brant, A. W. 1974. New approaches to marketing poultry products. I n New Protein Foods, Volume 1 , Technology, Part A (A. M. Altschul, ed.), pp. 337-366. Academic Press, New York.

    Bressani, R.. and Elias, L. G. 1974. Legume foods. In New Protein Foods, Volume I , Tech- nology, Part A (A. M. Altschul, ed.), pp. 231-297. Academic Press, New York.

    Brin, M. 1976. Nutrient intervention to improve nutritional status. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 222-238. Academic Press, New York.


    Brockmann, M. C. 1978. Technology of preservation of meat products. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 359-392. Academic Press, New York.

    Byerly, T. C. 1978. Competition between animals and man for agricultural resources. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 73-116. Academic Press, New York.

    Campbell, M. F., Kraut, C. W., Yackel, W. C., and Yang, H. S. 1985. Soy protein concentrate. In New Protein Foods, Volume 5 , Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 301-337. Academic Press, New York.

    Carroll, K. K., Geovannetti, P. N., Huff, W. M., Moarse, O., Roberts, D., and Wolfe, B. M. 1978. Am. J . Clin. Nutr. 34, 1312.

    Cook, J. D., Morck, T. A., and Lynch, S. R. 1981. Am. J . Clin. Nutr. 34, 2622. Cossack, Z. T., and h s a d , M. D. 1983. Nuir. Res. 3, 23. Cunha, T. J. 1978. Land animals: Opportunities for improved production. In New Protein Foods,

    Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 198-225. Academic Press, New York.

    Deatherage, F. E. 1974. New approaches to the marketing of red meats. In New Protein Foods, Volume 1, Technology, Part A (A. M. Altschul, ed.), pp. 298-336, Academic Press, New York.

    Descovich, G. C., Ceredi, C., Gaddi, A,, Benassi, G., Cattin, L., Fontana, G., Senin, U., Man- narino, E., Caruzzo, C., Bertelli, E., Fragiacomo, C., Noseda, G., Sirtori, M., and Sirtori, C. R. 1980. Lancet ii, 709.

    Descovich, G. C., Benassi, M. S . , Cappelli, M., Gaddi, A., Grossi, G., Piazzi, S . , Sangiorgi, Z., Marinino, G., and Lenzi, S. 1982. In Lipoproteins and Coronary Atherosclerosis (G. Noseda, C. Fragiocomo, R. Fumagalli, and R. Paoletti, eds.), p. 279. Elsevier, Amsterdam.

    Dieckert, J. W., and Dieckert, M. C. 1985. The chemistry and biology of the seed storage proteins. In New Protein Foods, Volume 5, Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 1-25. Academic Press, New York.

    Dronne, Y. 1981. Institute National de la Recherche Agronomique, Paris, France. FAO. 1970. Amino acid content of foods and biological data on proteins. F A 0 Nutrition Studies,

    No. 24, Food and Agriculture Organization, Rome. FAOWHOKJNU. 1985. Energy and protein requirements. Report of a joint FAO, WHO, UNU

    expert consultation. WHO Tech. Rept. Ser. 724. World Health Organization, Geneva. Fauconneau, G. 1983. In Plant Proteins for Human Food (C. E. Bodwell and L. Petite, eds.), pp.

    1-20. Nijhoff/Dr. W. Junk, The Hague. Feeney, R. E., and Whitaker, J. R. 1985. Chemical and enzymatic modification of plant proteins. In

    New Protein F o e s , Volume 5, Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 181-219. Academic Press, New York.

    Fitzhugh, H. A. 1981. Small ruminants as food producers. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 135-160. Academic Press, New York.

    Fontenot, J. P. 1981. Recycling of animal wastes by feeding. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 277-306. Academic Press, New York.

    Food and Nutrition Board. 1974. Improvement in protein nutriture. National Research Council, National Academy of Sciences, Washington, D.C.

    Food and Nutrition Board. 1980. Recommended dietary allowances (9th Ed. National Research Council, National Academy of Sciences, Washington, D.C.

    Fordyce, M. K., and Christakis, G. 1981. The nutritional and metabolic effect of variable protein


    intake in man. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 61-80. Academic Press, New York.

    Fonnan, M. J., and Hornstein, I. 1976. AID program for the introduction of new protein technolo- gies. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 239- 259. Academic Press, New York.

    Graham, D. M., and Whitted, S. F. 1978. Prospects for milk supplies in the United States. I n New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 145-173. Academic Press, New York.

    Hallberg, L., and Rossander, L. 1982. Am. J . Clin. Nurr. 36, 514. Harada, Y., and Saito, T. 1978. Amino acids as supplements in animal feeds. In New Protein

    Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 335-358. Academic Press, New York.

    Hodges, R. E., Krehl, W. A,, Stone, D. B., and Lopez, A. 1967. Am. J. Clin. Nurr. 20, 198. Holmes, W. L., Rubel, G. B., and Hood, S. S. 1980. Arheroschlerosis 36, 379. Hopkins, D. T. 1981. In Protein Quality in Humans: Assessment and in Virro Estimation (C. E.

    Bodwell, J. S . Adkins, and D. T. Hopkins, eds.). AVI Pub]., Westport, Connecticut. Horan, F. E. 1974. Meat analogs. I n New Protein Foods, Volume 1 , Technology, Part A (A. M.

    Altschul, ed.), pp. 367-414. Academic Press, New York. Horan, F. E., and Wolff, H. 1976. Meat analogs-a supplement. In New Protein Foods, Volume

    2, Technology, Part B (A. M. Altschul, ed.), pp. 260-279. Academic Press, New York. Hulse, J. H. 1974. The protein enrichment of bread and baked products. I n New Protein Foods,

    Volume I , Technology, Part A (A. M. Altschul, ed.), pp. 156-230. Academic Press, New York.

    Hutt, P. B. 1974. Regulatory aspects of new protein foods. In New Protein Foods, Volume I , Technology, Part A (A. M. Altschul, ed.), pp. 455-480. Academic Press, New York.

    Idyll, C. P. 1978. Progress and prospects in farming invertebrates. I n New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 293-334. Academic Press, New York.

    M a n , N., Murray, E., Janghorbani, M., and Young, V. R. 1983a. J. Nurr. 113, 2516. M a n , N., Murray, E., Janghorbani, M., Evans, W. J., and Young, V. R. 1983b. J. Nurr. 113,

    2524. Janghorbani, M., Istfan, N. W., Pagounes, J. 0.. Steinke, F. H., and Young, V. R. 1982. Am. J.

    Clin. Nurr. 36, 531. Jansen, G. R. 1974. The amino acid fortification of cereals. In New Protein Foods, Volume I ,

    Technology, Part A (A. M. Altschul, ed.), pp. 40-120. Academic Press, New York. Jansen, G. R. 1981. Amino acid fortification. In New Protein Foods, Volume 4, Animal Protein

    Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 161-205. Academic Press, New York.

    Kinsella, J. E., and Mohite, R. R. 1985. The physical characteristics and functional properties of sesame proteins. In New Protein Foods, Volume 5 , Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 435-456. Academic Press, New York.

    Kinsella, J. E.. Damodaran, S . , and Greman, B. 1985. Physicochemical and functional properties of oilseed proteins with emphasis on soy proteins. I n New Protein Foods, Volume 5, Seed Storage Proteins (M. Altschul and H. L. Wilcke, eds.), pp. 108-180. Academic Press, New York.

    Kolar, C. W., Richert, S . H., Decker, C. D., Steinke, F. H., VanderZanden, R. J. 1985. Isolated soy protein. In New Protein Foods, Volume 5 , Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 259-299. Academic Press, New York.

    Kottman, R. M., and Geyer, R. E. 1973. Feedstuffs 46, 12.


    Liebermeister, H., and Tolujipur, H. 1980. Senkung des Cholesterinspiegels durch Zusatz von Pectinen und Soja-protein zur Reduktiondiet. Drsch. Med. Wochenschr. 105, 333.

    Livingston, G. E. 1976. Mass and institutional feeding of protein foods. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 177-221. Academic Press, New York.

    Lovell, R. T., Smitherman, R. O., and Shell, E. W. 1978. Progress and prospects of fish farming. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 262-292. Academic Press, New York.

    Lusas, E. W. 1985. Sunflowerseed protein. In New Protein Foods, Volume 5, Seed Storage Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 393-433. Academic Press, New York.

    Mason, M. E., and Katz, I. 1976. Role of flavor in new protein technologies. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 122-176. Academic Press, New York.

    Milligan, J. L., and Wilcke, H. L. 1981. Issues in poultry production. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 105-134. Academic Press, New York.

    Mitsuda, H., and Yasumoto, K. 1974. The amino acid fortification of intact cereal grains. In New Protein Foods, Volume 1, Technology, Part A (A. M. Altschul, ed.), pp. 121-155. Academic Press, New York.

    Morck, T. A., Lynch, S. R., and Cook, J. D. 1981a. Am. J. Clin. Nutr. 34, 2630. Morck, T. A., Lynch, S. R., and Cook, J. D. (1981b). Am. J . Clin. Nurr. 36, 219. Moms, E. R., Bodwell, C. E., Miles, C. W., Mertz, W., Rather, E. S., and Canary, J. J. 1983.

    Fed. Proc., Fed. Am. Soc. Exp. Biol. 42, 530 (Abstr.). Nielsen, N. C. 1985. Structure of soy proteins. I n New Protein Foods, Volume 5, Seed Storage

    Proteins (A. M. Altschul and H. L. Wilcke, eds.), pp. 27-64. Academic Press, New York. Noseda, G., and Fragiacomo, C. 1980. In Diet and Drugs in Atherosclerosis (G. Noseda, G.

    Lewis, and R. Paoletti, eds.), pp. 61-65. Raven, New York. Ohlson, R. 1985. Rapeseed. In New Protein Foods, Volume 5, Seed Storage Proteins (A. M.

    Altschul and H. L. Wilcke, eds.), pp. 339-357. Academic Press, New York. Om, M. L., and Watt, B. K. 1957. Amino Acid Content of Foods. Home Economics Research

    Report No. 4, U.S. Department of Agriculture, Washington, D.C. Phillips, R. W. 1981. The upper limits of livestock production. In New Protein Foods, Volume 4,

    Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 5-31. Academ- ic Press, New York.

    Pigott, G. M. 1976. New approaches to marketing fish. In New Protein Foods, Volume 2, Technology, Part B (A. M. Altschul, ed.), pp. 1-37. Academic Press, New York.

    Pond, W. G. 1981. Developments in pork production. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.), pp. 81-104. Academic Press, New York.

    Pyke, M. 1978. The evolution of animal protein in the human diet. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 47-72. Academic Press, New York.

    Reid, J. T., and White, 0. D. 1978. Energy cost of food production by animals. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp! 117-144. Academic Press, New York.

    Rhee, K. C. 1985. Peanuts (groundnuts). In New Protein Foods, Volume 5, Seed Storage Pro- teins (A. M. Altschul and H. L. Wilcke, eds.), pp. 359-391. Academic Press, New York.

    Schuh, G. E. 1981. Government policy and the production of animal protein: An international perspective. In New Protein Foods, Volume 4, Animal Protein Supplies, Part B (A. M. Altschul and H. L. Wilcke, eds.) , pp. 32-60. Academic Press, New York. \


    Schwandt, P., Richter, W. 0.. and Weisweiler, P. 1981. Atherosclerosis 40, 371. Scrimshaw, N. S., and Young, V. R. 1979. In Soy Protein and Human Nutrition (H. L. Wilcke,

    Shorey, R. A., Bazan, B., Lo, G. S . , and Steinke, F. H. 1981. Am. J. Cfin. Nutr. 34, 1769. Sirtori, C. R., Agradi, E. Conti, F., Gatti, E., and Mantero, 0. 1977. Lancer i, 275. Sirtori, C. R., Gatti, E., Mantero, 0.. Conti, F., Agradi, E., Tremoli, E., Sirtori, M., Fratemgo,

    L., Tavazzi, L., and Kritchevsky, D. 1979. Am. J . Cfin. Nufr. 32, 1645. Smith, 0. B. 1976. Extrusion cooking. In New Protein Foods, Volume 2, Technology, Part B (A.

    M. Altschul, ed.), pp. 86-121. Academic Press, New York. Stekel, A. 1982. In The Effects of Cereals and Legumes on Iron Availability (T. H. Bothwell,

    S. M. Clydesdale, J. D. Cook, P. R. Dallman, L. Hallberg, D. Van Campen, and W. J. Wolf). A report of the International Nutritional Anemia Consultative Group (INACG, 1982). Nutrition Foundation, Washington, D . C .

    Stonaker, H. H. 1978. Increased animal production through breeding. In New Protein Foods, Volume 3, Animal Protein Supplies, Part A (A. M. Altschul and H. L. Wilcke, eds.), pp. 174-197. Academic Press, New Y