advances in food and nutrition research volume 42

ADVANCES IN

FOOD AND NUTRITION RESEARCH

VOLUME 42

ADVISORY BOARD

DOUGLAS ARCHER Gainesville, Florida

JESSE F. GREGORY I11 Gainesville, Florida

SUSAN K. HARLANDER Minneapolis, Minnesota

DARYL B. LUND New Brunswick, New Jersey

BARBARA 0. SCHNEEMAN Davis, California

SERIES EDITORS

GEORGE F. STEWART (1948-1982)

EMIL M. MRAK ( 1948- 1987)

C. 0. CHICHESTER (1 959-1988)

BERNARD S. SCHWEIGERT (1984-1988)

JOHN E. KINSELLA (1989-1993)

STEVE L. TAYLOR (1995- )

ADVANCES IN

FOOD AND NUTRITION RESEARCH

VOLUME 42

Edited by

STEVE L. TAYLOR Department of Food Science and Technology

University of Nebraska Lincoln, Nebraska

ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto

This book is printed on acid-free paper. @

Copyright 0 1998 by ACADEMIC PRESS

All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any

means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher.

The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 0 1923). for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1998 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 1043-4526/98 $25.00

Academic Press a division of Harcoiwt Bruce & Cornpuny 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com

Academic Press Limited 24-28 Oval Road, London NW1 7DX, UK http://www.hbuk.co.uk/ap/

International Standard Book Number: 0- 12-01 6442-6

PRINTED IN THE UNITED STATES OF AMERICA 98 9 9 0 0 01 02 0 3 Q W 9 8 7 6 5 4 3 2 I

CONTENTS

CONTRIBUTORS TO VOLUME 42 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

The Role of Flavoring Substances in Food Allergy and Intolerance

Steve L. Taylor and Erin Stafford Dormedy

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

111. Types and Uses of Flavoring Substances in Foods . . . . . 13 IV. Review of Reported Allergic Reactions to Food

Flavoring Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 V. Appropriate Diagnostic Tests for Investigation of

Sensitivity to Food Flavoring Substances . . . . . . . . . . . . . 31 VI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

11. Food Allergies and Intolerance ...................... 2

The Use of Amino Acid Sequence Alignments to Assess Potential Allergenicity of Proteins Used in Genetically Modified Foods

Steven M. Gendel

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 11. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

111. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 IV. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

V

vi CONTENTS

Sequence Databases For Assessing the Potential Allergenicity of Proteins Used in Transgenic Foods

Steven M . Gendel

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 I1 . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

111 . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 IV . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Design of Emulsification Peptides

David Sheehan. Kathleen Carey. and Siobhan O’Sullivan

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 I1 . Secondary Structure of Peptides ..................... 95

I11 . Modeling of Peptide Structures ...................... 99 IV . Synthesis of Designed Peptides ...................... 112 V . Testing of Peptide Emulsification Properties . . . . . . . . . . 115

VI . Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

X-Ray Diffraction of Food Polysaccharides

Rengaswami Chandrasekaran

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

I11 . Molecular Shapes and Interactions . . . . . . . . . . . . . . . . . . . 146 IV . Mixed Polysaccharides ............................. 200 V . Morphology to Macroscopic Properties . . . . . . . . . . . . . . . 202

V1 . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

I1 . Basic Principles of Solving Three-Dimensional Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

Cellular Signal Transduction of Sweetener-Induced Taste

Michael Naim. Benjamin J . Striem. and Michael Tal

I . Introduction ...................................... 211 I1 . Recognition Stage At the Taste-Receptor Cell . . . . . . . . 214

CONTENTS vii

I11 . IV .

V . VI .

VII .

I . I1 .

I11 . IV . V .

VI . VII .

VIII . IX .

Components of the Downstream Transduction Pathway 219 Involvement of Gustducin/Transducin in Sweet-Taste Transduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Amiloride-Sensitive Sweet-Taste Transduction . . . . . . . . 229 The Hypothesis of Receptor-Independent Activation of Sweet Taste By Amphipathic Nonsugar Sweeteners . . . . 230 Summary and Research Needs ...................... 233 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

Antioxidant Activity of the Labiatae

Susan L . Cuppett and Clifford A . Hall. I11

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Evolution of Labiatae as Antioxidant Sources . . . . . . . . . 247 Plant Tissue Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 Labiatae Essential Oils as Antioxidants . . . . . . . . . . . . . . 249

Isolation and Identification of Rosemary Compounds . . 254 Rosemary Extracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Compound Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Rosemary Synergism(s) and Heat Stabilities . . . . . . . . . . 258 Health Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

This Page Intentionally Left Blank

CONTRIBUTORS TO VOLUME 42

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

Kathleen Care , Department of Biochemistry, University College, Cork, Ireland (933

Rengaswami Chandrasekaran, Whistler Center for Carbohydrate Research, Purdue University, West Lafayette, Indiana 47907 (131)

Susan L. Cuppett, Department of Food Science and Technology, University of Nebraska-Lincoln, Lincoln, Nebraska 68583 (245)

Erin Stafford Dormedy, Department of Food Science and Technol- ogy, University of Nebraska-Lincoln, Lincoln, Nebraska 68583 ( 1 )

Steven M. Gendel, Biotechnology Studies Branch, Food and Drug Adminis- tration, National Center for Food Safety and Technology, Summit-Argo, Illinois 60501 (45; 63)

Clifford A. Hall 111, Department of Food and Nutrition, North Dakota State University, Fargo, North Dakota 58105 (245)

Michael Naim, Institute of Biochemistry, Food Science and Nutrition, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Rehovot 76-100, Israel (211)

Siobhan O’Sullivan, TEAGASC Dairy Products Research Centre, Moore- park, Fermoy, Co., Cork, Ireland (93)

David Sheehan, Department of Biochemistry, University College, Cork, Ireland (93)

Benjamin J. Striem, Institute of Biochemistry, Food Science and Nutrition, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Rehovot 76-100, Israel (21 1)

Michael Tal, Sigma Israel Chemicals Ltd., Park Rabin, Rehovat 76-100, Israel (21 1)

Steve L. Taylor, Department of Food Science and Technology, University of Nebraska- Lincoln, Lincoln, Nebraska 68583 ( 1 )

ix

ADVANCES IN FOOD AND NUTRITION RESEARCH. VOL. 42

THE ROLE OF FLAVORING SUBSTANCES IN FOOD ALLERGY AND INTOLERANCE'

STEVE L. TAYLOR AND ERIN STAFFORD DORMEDY

Department of Food Science and Technology University of Nebraska-Lincoln

Lincoln. Nebraska 68583

I. Introduction 11. Food Allergies and Intolerance

A. Definitions and Perceptions B. True Allergic Reactions C. Intolerances

111. Types and Uses of Flavoring Substances in Foods A. Nature and Composition of Flavoring Substances B. Manufacturing of Flavoring Substances C. Protein Content of Flavoring Substances D. Usage Levels of Flavoring Substances E. Labeling of Flavoring Substances

IV. Review of Reported Allergic Reactions to Food Flavoring Substances A. B. C. Occupational Sensitivities Appropriate Diagnostic Tests for Investigation of Sensitivity to Food Flavoring Substances A. Challenge Tests B. Skin Tests C. Patch Tests

References

Published Examples of Allergic Reactions to Flavoring Substances Likelihood of Allergic or Intolerance Reactions of Flavoring Substances

V.

VI. Conclusions

I . INTRODUCTION

Flavorings are concentrated preparations, with or without flavor adjuvants, used in foods to impact flavor. Flavorings are not intended to be

' This article was reviewed and accepted by the Scientific Advisory Board to the Allergy and Immunology Institute of the International Life Sciences Institute.

1

1043-4526/98 $25.00 Copyright 0 1998 by Academic Press.

All rights of reproduction in any form reserved.

2 STEVE L. TAYLOR AND ERIN STAFFORD DORMEDY

consumed as such, but rather used at low levels in finished consumer foods to enhance or improve quality. Flavoring substances are either chemically defined or natural products, the primary function of which is to impart flavor.

Food allergies and intolerances are an increasingly important concern to consumers and food manufacturers alike. Flavoring substances are rarely implicated as causative factors in food allergies and intolerances. Since thousands of different flavoring substances are used in foods, usually at very low levels, the likelihood of allergies or intolerances triggered by these substances is quite small.

However, some cases of adverse reactions to foods remain idiopathic or unexplained. In some such cases, the major ingredients listed on the incriminated product’s ingredient listing have been clinically tested and eliminated as possible causative agents. The physician is then left to consider the possible role of minor ingredients, such as flavoring substances. This review provides a perspective of the likelihood that flavoring substances could be involved in allergy. The review also provides a diagnostic approach to evaluating the role of flavoring substances in adverse reactions.

II. FOOD ALLERGIES AND INTOLERANCE

A. DEFINITIONS AND PERCEPTIONS

Food allergies and intolerances are adverse reactions that affect some, but not all, individuals in the population. These individualistic adverse reactions to foods can occur through a variety of different mechanisms. Food allergies involve abnormal immunological reactions in which the person’s immune system overreacts to ingestion of ordinarily harmless substances, usually naturally occurring proteins in foods (1-4). In contrast, food intolerances do not involve the immune system. Nonimmu- nological food intolerances include non-IgE-mediated histamine release (anaphylactoid reactions), metabolic food disorders, and a host of food idiosyncrasies (5). Although food allergies and intolerances are manifested in only a minor proportion of the population, the public views such illnesses as a major health concern and fails to distinguish between the different types of illnesses that fall within this general category. As many as 10-20% of all consumers believe that they have food allergies (6), although only 1-2% of the population are truly afflicted with these conditions.

FLAVORING IN FOOD ALLERGY AND INTOLERANCE 3

B. TRUE ALLERGIC REACTIONS

Two types of allergic reactions to foods can be distinguished (Table I), reflecting differences in the immune mechanisms involved in the course of the reaction. Allergic reactions may involve either an antibody-mediated or a cell-mediated response. Antibody-mediated responses are associated with IgE antibodies and are designated immediate allergic reactions because symptoms are experienced within minutes or hours following exposure of the sensitive individual to the allergenic food. The IgE-mediated mechanism is involved in other types of environmental allergies such as reactions to pollens, mold spores, animal danders, insect venoms, and drugs. Cell- mediated reactions are mediated by T lymphocytes and are designated delayed-type allergic reactions denoting the delay in the onset of symptoms

TABLE I CLASSIFICATION OF ALLERGIC REACTIONS OF FOODS

Initiation Descriptive name time Mechanism Typical manifestations

IgE-mediated hypersensitivity

2-30 min Antigen cross-links IgE Systemic anaphylaxis bound t o mast cells and basophils with release of Localized anaphylaxis vasoactive mediators Respiratory (histamine and many asthma, wheezing, others) rhinitis, bronchospasm

(anaphylactic shock)

Cutaneous dermatatis o r eczema (rash), urticaria (hives), angioedema, pruritis

vomiting, nausea, diarrhea, abdominal cramps

hypotension. palatal itching, oral swelling, include tongue and

Gastrointestinal

Other

larnyx Cell-mediated 24-72 hr Sensitized T lymphocytes Localized tissue damage

hypersensitivity release cytokines that (especially in the activate macrophages, intestinal tract with which mediate direct food) cellular damage


of 24-72 hr following exposure to the allergenic food. Cell-mediated reactions can also occur with agents from sources other than foods, such as medications and cosmetics. The term “food allergy” should be applied only to immunologically based adverse reactions invoked by food, but this convention is not always followed by consumers or even some physicians.

1. IgE-Mediated Allergies

The primary mechanism involved in food allergies is the antibody- mediated allergic reaction involving the formation of the IgE class of antibody. The mechanism of IgE-mediated food allergies has been thoroughly reviewed (3,7). Briefly, in IgE antibody-mediated allergic reactions (Figure l), the offending food protein (allergen) triggers production by the B lymphocytes of specific IgE antibodies. IgE binds in a highly specific fashion to certain high-affinity IgE receptors on the membranes of blood basophils and tissue mast cells. Upon a second exposure to the same allergen, the allergen cross-links two membrane-bound IgE molecules, initiating the

FIG. 1. Mechanism of an IgE-mediated allergic reaction.


release of a large number of different mediators, including histamine, pros- taglandins, and leukotrienes (3).

Mediators released from mast cells and basophils interact with a variety of tissue receptors, causing numerous symptoms (Table I), primarily involving the gastrointestinal tract, skin, and respiratory system (2). Not all of these symptoms are experienced by any one allergic individual. Most affected individuals experience only mild symptoms, although severity can increase with frequent exposure. The severity of the symptoms varies depending on the sensitivity of the allergic individual, the amount of the offending food ingested, and the length of time since the last previous exposure. On rare occasions, life threatening anaphylactic shock responses occur upon exposure to the offending food (8,9). Anaphylactic shock is the most common cause of death in the rare fatalities associated with true food allergies (9). However, the most common manifestation of IgE-mediated food allergies is the so-called “oral allergy syndrome,” in which symptoms such as swelling, itching, and hives are confined to the oropharyngeal area. These reactions are typically associated with fresh fruits and vegetables.

The prevalence of IgE-mediated food allergies is not precisely known in the total population. Studies have demonstrated that the prevalence in young infants ranges from 4 to 8% (10). However, food allergies, especially to certain foods including cow’s milk and eggs, are commonly outgrown (11,12). While the prevalence of IgE-mediated food allergies in adults is not precisely known, it has been estimated to affect <1% (2,12). Any food or food ingredient that contains protein has the potential to cause allergic reactions in some individuals. However, the protein constituents of only a few foods are responsible for the majority of food allergies (Table 11) (3). Perhaps 90% of allergic reactions in the United States are caused by the foods on this relatively short list (12). Foods often implicated by consumers as causes of food allergies such as chocolate, strawberries, and citrus fruits do not give positive results in double-blind, placebo-controlled food chal-

TABLE I1 COMMON ALLERGENIC FOODS

Cereals containing gluten (wheat, rye, barley, oats, spelt or their hybridized strains and products of these)

Crustacea and products of these (shrimp, crab, lobster) Eggs and egg products Fish and fish products Milk and milk products (lactose included) Peanuts, soybeans, and products of these Tree nuts and nut products


lenges (13). However, the prevalence of allergic sensitivities to specific foods varies from one country to another depending on the frequency with which the food is eaten in that country and the age at its introduction into the diet. For example, peanuts are a much more frequent cause of food allergies in the United States than in most other countries because U.S. citizens eat peanuts more often and introduce peanut butter into the diets of children at an early age. For similar reasons, the Japanese seem to experience more soybean and rice allergies than other cultures, and Scandi- navians have a very high prevalence of codfish allergy.

The major mode of treatment for all food allergies and intolerances is the specific avoidance diet. The construction of avoidance diets and the difficulties faced by consumers with such diets have been extensively reviewed (14,1.5). The effectiveness of avoidance diets is dependent on the patient’s ability to identify the offending food or the ingredients derived from that food. Therefore, ingredient labeling is critical for effective avoidance. Food labels, especially certain labeling terms, can sometimes be con- fusing to patients, who need to be educated in the proper use of label information. Careful scrutiny of the ingredient lists on food labels will usually reveal the presence of the offending food.

Trace levels of the offending food can elicit adverse reactions. The safe level of the offending food that can be tolerated by individuals with true food allergies is unknown, but likely to be very low and variable. Allergic reactions have occurred from the inadvertent ingestion of milligram levels of the offending food (16). Foods may also become contaminated with trace amounts of other foods through a wide variety of means. Errors in processing and preparation of foods can result in the contamination of foods with allergenic foods. Examples of processing errors include use of shared equipment or facilities, use of work (i.e., material from one batch blended with material from another batch), formulation errors, and packaging or labeling mistakes. Food preparation practices, including use of shared utensils or cooking equipment, use of the same frying oil for more than one type of food, and creative formulation (e.g., peanut butter in chili) can also lead to the unsuspected presence of an allergenic food, especially in food service situations. Furthermore, restaurant foods are not labeled, so allergic individuals must inquire about the presence of specific foods in specific menu items.

When considering the allergenicity of a food ingredient, the presence of protein is important. If the protein is removed or rendered nonallergenic by processing, the resultant food product is safe for allergic individuals to consume. However, food allergens tend to be quite stable to processing treatments, especially heat treatments (17-20). For example, a study by Nordlee et al. (17) showed that the extracts of most processed peanut


products retained their ability to bind specific IgE from peanut-allergic sera, indicating that peanut allergens are highly heat-stable. One exception is the allergens present in certain fresh fruits and vegetables, which are quite heat-labile. The allergens in fresh fruits and vegetables are also sensitive to digestion. Hence, these proteins tend to elicit primarily oral symptoms, the oral allergy syndrome. Proteolysis has been suggested as a more effective way to destroy food allergens but most of these proteins also tend to be resistant to proteolysis (20-22). While casein hydrolysates are frequently used in hypoallergenic infant formulas, allergic reactions have been associated in a few cases even to this highly hydrolyzed protein (23). Soy products, such as hydrolyzed soy protein and soy sauce, which are subjected to considerable proteolysis during processing, remain unsafe for soy-allergic consumers in many cases because allergenic fractions of protein may remain in the final consumer product (20). Allergens can be effectively removed from some food ingredients. Herian et al. (20) determined that some highly refined soy products such as soy oil, which does not normally contain soy protein, may be safe for consumption by soy-allergic individuals. Double- blind challenge tests to determine the allergenicity of peanut, soybean, and sunflower oils indicate that the highly refined oils did not cause allergic reactions in sensitive individuals (24-26). Allergenic activity was lost because the protein fraction is typically removed from the oil during the refining process.

Virtually all food allergens are proteins. However, foods, even commonly allergenic foods, contain enormous numbers of proteins, and only a few of these proteins are food allergens. The nature of the IgE-mediated reaction, especially the need for cross-linking, conveys certain structural requirements on food allergens, including the presence of multiple IgE-binding sites. Although most allergens are proteins having a molecular weight between 10 and 60 kDa (27), attempts to identify some common chemical property of these proteins, for example, one that conveys allergenicity, have failed. It appears that allergenicity is a consequence of a complex series of interactions involving not only the allergen but also the dose, the sensitizing route, sometimes an adjuvant, and most importantly, the genetic constitution and immunological status of the consumer. The nature of the proteins involved in most food allergens has not been identified and characterized. The nature of the well-characterized food allergens has been reviewed elsewhere (12).

Only a few contaminants or food additives are implicated in true food allergies, and the overall impact of allergic reactions to these categories of food-borne chemicals is quite small in comparison to that of naturally occurring substances. One (perhaps the only) food additive associated with true food allergies is papain (28) . This proteolytic enzyme from papaya, a


component of meat tenderizers and not a flavoring, is a known although not very common allergen. Food additives and ingredients derived from allergenic foods such as wheat starch may contain residues of the protein fraction of the food and thus would be allergenic. Penicillin, a veterinary antibiotic sometimes found in the milk or meat of treated animals, is one of the few nonprotein food contaminants associated with true, IgE-mediated food allergies (29,30). In this case, penicillin acts as a hapten by binding to proteins and eliciting the formation of IgE antibodies specific to the penicillin-protein complex. A hapten is a low-molecular-weight compound that is not allergenic by itself, but when coupled with a carrier protein, can elicit immune responses. Where naturally occurring proteins exist as components of food additives or inadvertent, unlabeled contaminants of other foods, a risk exists to those consumers with food allergies. It is unlikely that flavoring substances could be associated with true IgE-mediated food allergies unless the flavoring substances contain proteins, or can act as significant haptens. There are no nonproteinaceous flavoring substances associated with true IgE-mediated food allergies.

2. Cell-Mediated Allergies

Cell-mediated allergies or delayed hypersensitivities occur without the direct involvement of antibodies. Cell-mediated reactions develop when allergens activate sensitized T lymphocytes. Activation of T lymphocytes by allergen on appropriate antigen-presenting cells results in the secretion of various lymphokines (cytokines produced by activated lymphocytes). At the cellular level, the overall effect of these lymphokines is to attract macrophages and other cells into the area. Due to the local environment, these cells are activated, promoting increased phagocytic activity and increased concentration of lytic enzymes for more effective killing of foreign cells. As lytic enzymes leak out of the activated phagocytic cells into the surrounding tissue, localized tissue destruction can ensue. These reactions typically take 48-72 hr to develop, the time required for initial T lymphocyte activation and lymphokine secretion to mediate accumulation of additional cells and the subsequent release of their lytic enzymes (31).

Evidence for the involvement of cell-mediated immune reactions in food allergies is sparse but reasonably compelling (32-37). No estimates of the prevalence of cell-mediated food allergies have been made. The symptoms of cell-mediated food allergies are likely to be more localized than those of systemic IgE-mediated reactions. Thus, life-threatening responses, such as anaphylactic shock, are not known to occur in cell-mediated allergic reactions.


Celiac disease is probably the best characterized example of a cell- mediated food reaction. There is increasing evidence to support an immunologic basis of gluten sensitivity in this condition (38). However, even in this case, the mechanism has not been proven. Celiac disease, also known as celiac sprue or gluten-sensitive enteropathy, is an inherited malabsorption disease that occurs in certain individuals following the ingestion of the naturally occurring protein fraction of wheat, rye, barley, triticale, and perhaps oats (39-41). Following ingestion, the absorptive epithelial cells in the small intestine become damaged. This results in a decreased number of epithelial cells that are critical for digestion and absorption. The mucosal enzymes necessary for digestion and absorption are also altered in the damaged cells. The absorptive cells are thus functionally compromised. This process results in severe malabsorption characterized by diarrhea, bloating, weight loss, anemia, bone pain, chronic fatigue, weakness, muscle cramps, and, in children, failure to gain weight and growth retardation (42).

The treatment of celiac disease typically involves the total avoidance of wheat, rye, barley, triticale, oats, and all products produced from these grains (40). As with other food allergies, the symptoms of celiac disease are thought to be triggered by ingestion of rather small quantities of these grains. A safe tolerance level cannot yet be established, so complete avoidance is still advised. However, adherence to these strict avoidance diets can be quite difficult. Flavorings are not involved in celiac disease except in the rare circumstance where the flavoring contains proteins from wheat, rye, barley, oats, or triticale.

3. Contact Sensitivities

Contact sensitivity is another common form of cell-mediated allergy. Many contact sensitivity reactions are mediated by T lymphocytes or Lang- erhans cells in the skin. Most of the causative substances are small molecules that can act as haptens and complex with skin or perhaps mouth mucosal proteins. This complex is internalized by antigen presenting cells in the mucosa, causing activation of sensitized lymphocytes or Langerhans cells and an ensuing cell-mediated reaction as detailed above. Contact sensitization reactions are most common with consumer products that are in frequent and/or prolonged contact with the mucosa. Several flavoring substances are known to cause contact sensitivity reactions, but primarily in the skin only from cosmetic uses. Such reactions are not relevant to most food uses where mucosal exposure is short. In products such as chewing gum or hard candy, there is prolonged and frequent exposure of the flavoring substances to the mucosal cells of the oral cavity. Contact sensitivity reactions in the mouth should be distinguished from the oral allergy syndrome, which


involves IgE antibodies and exposure to proteins without the need for prolonged exposure.

C. INTOLERANCES

In contrast to true food allergies, many of the individualistic adverse reactions to food do not involve the immune system. Food intolerances, like food allergies, affect a limited number of individuals. Food intolerances occur through a number of nonimmunologic mechanisms. The three major classes that we discuss are (1) non-IgE-mediated histamine release (referred to as anaphylactoid reactions, to be distinguished from systemic anaphylaxis), (2) metabolic food disorders, and (3) idiosyncratic reactions. Intoler- ances can be distinguished from true food allergies both by the lack of immune system involvement and by the degree of tolerance for the offending food or food ingredient. With food intolerances, the affected individual can typically tolerate ingestion of a much greater amount of the offending substance than is the case with true food allergies.

Most of the nonimmunological food intolerances are caused by food- borne substances other than proteins. These substances may be naturally present in the food or from constituents added to foods. However, the situation is complicated by the fact that the evidence supporting the existence of some of these illnesses is far from complete. Food additives appear to be a focus of concern for unexplained maladies, but the possible role of naturally occurring substances in these conditions has received considerably less attention.

1. Non-IgE-Mediated Histamine Release (Anaphylactoid Reactions)

Non-IgE-mediated histamine release reactions are the result of substances in food that cause mast cells and basophils to release spontaneously histamine and other mediators of allergic reactions. However, unlike food allergies, there appears to be no involvement with IgE or other immuno- globulins, and prior exposure is not a prerequisite (2).

Occasionally, histamine poisoning, also known as scombroid fish poisoning, is included as an example of a non-IgE-mediated release reaction (43) or as a separate category of nonimmunological food intolerance (44). However, histamine poisoning is actually a food-borne intoxication associated with the ingestion of foods containing unusually high levels of histamine, which is formed during bacterial spoilage (45). Everyone is susceptible to histamine poisoning, so it does not truly fit into this review of food allergies and sensitivities.


In non-IgE-mediated histamine release reactions, it is presumed that some substance in the implicated food destabilizes the mast cell membranes causing a spontaneous release of the histamine and other mediators. Actu- ally, none of these histamine-releasing substances has ever been isolated or identified in foods, although this mechanism is well established with certain drugs. Therefore, the only evidence for the existence of anaphylactoid reactions from foods is highly circumstantial.

The most often cited example is strawberry “allergy.” Strawberries are known to cause adverse reactions (frequently urticaria) in some individuals. Yet, strawberries contain little protein, no strawberry allergen has ever been identified, and no evidence has been obtained for the existence of strawberry-specific IgE. Since the symptoms of strawberry “allergy” are reminiscent of a true food allergy, in vivo release of histamine and other mediators seems a plausible mechanism. However, it must be emphasized that this mechanism has not been proven in strawberry “allergy” or any other food sensitivity.

There is no evidence that flavoring substances are involved in non-IgE- mediated histamine release reactions. This is likely due to low use levels of flavoring substances since significant dosage levels are likely required for this type of reaction.

2. Metabolic Food Disorders

Metabolic food disorders are adverse reactions to a food or food additive that occur through some effect of the substance on the metabolism of the individual (5) . Metabolic food disorders involve genetically determined enzyme deficiencies that either affect the host’s ability to metabolize a food component (e.g., lactose intolerance) or enhance the sensitivity of the host to some food-borne chemical via an altered metabolic pattern (e.g., favism). Lactose intolerance and favism are by far the most common metabolic food disorders. Lactose intolerance is caused by a deficiency of the enzyme p- galactosidase in the intestine, which results in an inability to digest and absorb lactose. Favism is caused by a deficiency of the enzyme glucose- 6-phosphate dehydrogenase in erythrocytes, which results in hemolytic anemia upon ingestion of fava beans, which contain several otherwise innoc- uous hemolytic agents. Metabolic food disorders have a very defined tolerance level (most lactose-intolerant individuals can safely ingest 5 g or less of lactose, a level that exceeds by orders of magnitude the levels of use as flavor ingredients), although the degree of tolerance may be individually variable. Flavors do not contain the substances that cause these disorders and have not been linked to metabolic food disorders.


3. Idiosyncratic Reactions

Idiosyncratic reaction is the general term used to describe a variety of food intolerances thought to have nonimmunological, but unknown, mechanisms (2). Many of the reported adverse reactions to food additives are placed in this category owing to the lack of understanding of their modes of action. Although consumers may wrongly refer to idiosyncrasies as food allergies, physicians should explain this aspect to patients. Food idiosyncrasies fall into three categories: (1) illnesses whose association with specific foods or food ingredients is well documented, (2) illnesses whose association with specific foods or food ingredients have not been firmly established or, in some cases, remain controversial, and (3) illnesses widely believed by consumers to be associated with specific foods or food ingredients despite considerable evidence to the contrary.

While some illnesses such as the involvement of sulfites in asthma are well documented (46,47), other associations such as chocolate in migraine headaches (48-50), food coloring agents and sugar in hyperkinesis (48-50), butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) in hives (53,54), and monosodium glutamate in headache, facial flush, and chest pain (55,56) are unproven.

In the vast majority of food idiosyncrasies, the role of specific foods or food ingredients remains to be established. Although tests have been done to prove the existence of some reactions, for the most part, carefully controlled challenge studies have not been conducted to examine the role of foods or food ingredients in these situations. The existence of many of these illnesses is based primarily on unconfirmed, anecdotal reports, or poorly designed clinical trials that leave the cause and effect relationship in doubt. Some of these illnesses may exist, but others may be psychosomatic or attributable to causes other than foods. The role of psychological disorders in perceived reactions to foods has been the subject of several notable studies (56-61). In some cases, including food-induced migraines and monosodium glutamate (MSG)-induced headache and flushing, the symptoms are so subjective that the confirmation of the responses is difficult. In other cases, such as the alleged roles of tartrazine (FD&C Yellow No. 5 ) in asthma and BHT and BHA in chronic urticaria (54), the results of several challenge studies are not in agreement. The design of the studies, especially the withholding of medications for the chronically ill patients involved in the studies, is a particularly critical factor.

Sulfite-induced asthma is the best example of a well-established idiosyncratic illness of food-borne origin. Sulfites have been widely used in


foods for a variety of purposes, and sulfite residues are present in a variety of foods at levels ranging from a few ppm to >lo00 ppm in dried fruits (62). Sulfites can also occur naturally in foods, especially fermented foods, but the residues of naturally occurring sulfites are usually low (62). The pathogenesis of sulfite-induced asthma is not understood, although several mechanisms have been hypothesized including IgE- mediated reactions, hyperreactivity to inhaled SOz, and sulfite oxidase deficiency (54,62,63). In controlled challenges with capsules and/or acidic beverages, the threshold level of sulfite ranges from 3 to 130 mg of SO2 equivalents (62). Thus, sulfite-sensitive asthmatics must avoid highly sulfited foods (64).

Flavoring substances, with the exception of MSG, have not been implicated in idiosyncratic reactions. MSG is a widely used food ingredient and is best characterized as a flavor enhancer. MSG and various protein hydrolysates, which contain glutamate, are used widely as ingredients of flavor formulations. However, as noted later in this review, the role of MSG in food-borne illness remains controversial and unproven.

Ill. TYPES AND USES OF FLAVORING SUBSTANCES IN FOODS

Many aspects of flavoring substances must be examined when assessing the allergenic potential of these substances, including their chemistry, the presence of protein, the manufacturing of the substance, and the typical usage level.

Flavors are composed of many materials, chemicals, although most are present in extremely low amounts. In the United States, flavors are composed of almost 2000 substances of which over 400 are well-characterized products of natural origin containing numerous individual components, while approximately 1600 are structurally defined substances. Flavors can be classified by physical appearance (liquid, solid, or paste) as summarized in Table 111 (65). The physical characteristics of the flavor must be compatible with the product in which it is to be used. Flavors can also be classified as either simple or compounded (65). Simple flavors are those consisting of a single ingredient or a single substance diluted in an appropriate neutral carrier. Compounded flavors are blends of several substances. Flavors can also be classified as natural or artificial, a common distinction on the package label. However, artificial flavors can, and often do, contain natural ingredients. Flavoring substances meet the specifications for identity and purity of the Food Chemicals Codex, established by a committee of experts for the National Academy of Science.


TABLE I11 PHYSICAL CLASSIFICATION OF FLAVOR INGREDIENTS

Solids

Crystals Powders

Freeze dried Spray dried Dried extracts Plated

Encapsulated flavors

Liquids

Essential oils Tincture Folded Infusions Recitified Distillates Terpeneless Spirits Sesquiterpeneless Soluble essences

Oleoresins Emulsions Absolutes Fractions and isolates Fluid extracts Concentrated juices Compounded oils Single-strength juices Alcoholates

Pastes

Soft extracts Resins

Natural Prepared

Resinoids Concretes Emulsions (creams)

A. NATURE AND COMPOSITION OF FLAVORING SUBSTANCES

The flavor characteristics of a food are derived mainly from its volatile components. Volatile substances, while often intense flavors, are seldom major components of food (carbohydrates, proteins, and fats), although they may be considered to be derived from these major components (66) . Flavors may be naturally occurring in foods or generated from precursors during the cooking or processing of foods. Flavors can be produced by thermal reactions between naturally occurring compounds in foods, such as the creation of meat flavor by the thermal reaction of certain amino acids and sugars. Flavors may also be generated by enzymatic reactions or enzymatic modification, as is cheese flavor, or by microbial fermentation, as is butter flavor. The most important classes of flavors in various foodstuffs and the mode of their formation from precursors are summarized in Table IV. In general, these and all other flavoring substances are low-molecular- weight substances.

The classes of flavoring substances listed in Table IV are unlikely to elicit allergic reactions except when they act as haptens. These reactions are usually manifested as contact sensitivity reactions. Antibodies are elicited for the hapten and the carrier protein. Hapten-protein complexes can elicit either antibody-mediated or cell-mediated immune responses. For example, cinnamic aldehyde may act as a hapten through its reactivity with protein. Therefore cinnamic aldehyde can produce contact allergy and urticaria under some conditions (67). However, digestive processes


TABLE IV MAJOR CLASSES OF FLAVORING SUBSTANCES IN VARIOUS FOODS A N D THEIR MODE OF

FORMATION FROM PRECURSORS

Main mode Food Precursors of formation Classes of flavoring substances

Fruit

Vegetables. spices, herbs

Wine, beer

Dairy products

Meat

Fish

Coffee, cocoa

Bread

Roasted nuts

Sugars

Fats

Sugars

Fats Amino acids,

Amino acids,

Amino acids Sugars Amino acids Fats Amino acids Amino acids,

sugars

sugars

sugars

Amino acids Fats Amino acids,

sugars

Amino acids,

Amino acids, sugars

sugars

Natural formation

Natural formation

Natural formation

Natural formation Natural formation

Enzymes

Heat Fermentation Fermentation Enzymes, heat Fermentation Heat

Enzymes, heat Enzymes Heat

Heat

Heat

Terpenes, terpenoids, acids, furans, pyrans, esters

Aldehydes, ketones, esters, lactones

Terpenes, terpenoids, isoprenoids, phenols

Lactones, aldehydes Pyrazines

Sulfur compounds, aldehydes

Sulfur compounds, cyanides Esters, acetals Lactones, amines Ketones, lactones, aldehydes Acids, amines Acids, aldehydes, ketones, furans,

pyrazines, phenols, sulfur compounds

Amines, sulfur compounds Aldehydes, ketones Heterocyclic amines, aldehydes,

ketones, phenols, sulfur compounds

Pyrroles, pyrazines, pyridines

Pyrazines, aldehydes, ketones, sulfur compounds

commonly destroy the hapten-protein complexes. Thus, low-molecular- weight chemicals in foods are unlikely to be effective haptens in systemic reactions.

Process flavors are created by the combination of reducing sugars with amine compounds at elevated temperatures. In the case of process flavors, proteins or peptides may be used as the nitrogen source materials. Thus, the process flavors could contain protein residues and potentially be allergenic (see next section).


B. MANUFACTURING OF FLAVORING SUBSTANCES

There are many general methods for preparation of flavoring substances (Table V). Each method can be modified or combined with other methods. When manufacturing a flavoring substance, many processing methods are utilized, including chemical synthesis, distillation, expression, solvent extraction, concentration, microencapsulation, vacuum drying, and spray drying.

Distillation is a process requiring heat to volatilize the components at elevated temperature. Each component boils at a different temperature, depending on molecular weight, functional groups present, molecular composition, structure, etc. The resulting distillate leaves the protein behind in the source material. Therefore the distillate is not likely allergenic.

Expression is a process by which oils and juices are removed from plant sources by applying external pressure. If the source of extraction contains allergens, the extract may also contain allergens.

Solvent extraction is a procedure in which solvents are used to extract selectively desirable ingredients. Extracted natural flavor substances generally consist of rather dilute solutions in which the solvent may or may not contribute to the overall flavor strength. Proteins, if any, would be present at very low levels.

The process of concentration results in an increase in organoleptic princi- pals resulting from the partial elimination of the inert solvent.

Microencapsulation consists of forming capsules that envelope the flavor substances when added gum solidifies. Microencapsulation will not remove or change any allergen if present.

Vacuum drying of liquid extracts takes place under vacuum with removal of solvents as the main objective. The moisture content of vacuum-dried products is approximately 3 to 4%.

TABLE V GENERAL MANUFACTURING METHODS FOR

FLAVORING SUBSTANCE PREPARATION

Distillation Expression Concentration Crystallization Lyophilization Vacuum drying Spray drying Microencapsulation


Spray drying is an operation in which the product is mixed with a carrier and sprayed into a stream of hot air. If allergens are present in the flavor source, drying will not remove them or render them nonallergenic. In fact, the drying process would concentrate any allergens present in the liquid extract. However, it must be emphasized that the source materials for production of distilled, concentrated, extracted, encapsulated, and dried flavors seldom contain allergenic components.

Process flavors are generated by enzymatic or thermal treatment, or combined treatments, of various materials. Enzymes, especially lipases and proteases, are added to food materials and incubated at appropriate temperatures to create such products as hydrolyzed proteins and enzyme- modified cheeses. Alternatively, fermentations with bacterial or yeast cultures can be used to create similar products. The thermal processes used to create reaction flavors typically involve extensive nonenzymatic brow- ning, the reaction of reducing sugars with amino acids, to create unique flavors. The nature of precursor materials used in the creation of process flavors dictates the type of flavor that will evolve. For example, the most prominent methods in which the various classes of sulfur-containing flavor substances are formed from their natural precursors are summarized in Table VI (66). As noted below, process flavors are often prepared from proteinaceous sources, for example, meat, and do contain protein residues. But, because extensive conditions (typically 100°C for 15 min) and other reactions occur in the processing of process flavors, the allergenicity of the protein may be diminished or destroyed. Some food allergens are known to be quite heat-stable. Therefore, process flavors should be assumed to be allergenic if the source protein is a known allergen unless testing proves otherwise.

C. PROTEIN CONTENT OF FLAVORING SUBSTANCES

Nonproteinaceous substances make up the majority of flavors. For most flavoring substances, protein residues are not allowed in Food Chemicals Codex specifications. Most flavoring substances contain little, if any, protein. Even fewer of these proteins are derived from commonly allergenic foods. Yet, virtually all allergens are proteins, so the protein content of flavors is a key indicator of its possible allergenicity. Table VII lists the common protein sources that are used in the production of flavors, especially process flavors. Several of these sources, including eggs, peanuts, milk, sea foods, soybeans, tree nuts, and wheat, are considered to be common allergenic foods. However, some of the substances derived from such sources (e.g., refined oils) are free of protein residues. The protein content of reaction flavors is highly variable depending on the source materials used in the


TABLE VI SULFUR-CONTAINING FLAVOR COMPOUNDS FORMED FROM NATURAL PRECURSORS

Class of sulfur compound Precursor Formation Food

~

Hydrogen Sulfide

Thiols

Monosulfides

Disulfides

Trisulfides Thiosulfinates

Isothiocyanates Thiocyanates Methional (derivatives)

Thiazoles

Thiophenes

Thiolanes, thianes, thiepanes

Cysteine Cysteinekystine Thiamine Methionine Methionine Thiamine Methionine Methionine S-methylmet hionhe Dimethyl-b-

Disulfide Sulfide Thiosulfinates Disulfide S-AI kylcysteine

sulfoxide Glucosinolate Isothiocyanate Met hionine Methionine Cysteine (?) Cysteinekystine Thiamine Disulfide Cysteinekystine Thiamine Furanones + H2S Peptides (?)

propiothetin

Peptides Hydrogen sulfide

Enzymes Heat Heat Enzymes Heat Heat Enzymes Heat Heat Bioformation

Seafood, yeast Meat, eggs Meat Yeast Meat, fish Meat Yeast, milk Vegetables, meat, seafood Vegetables Seafood

Heat Vegetables Heat, oxidation Meat, vegetables Heat Allium vegetables Heat, oxidation AIlium vegetables Enzymes ANium vegetables

Enzymes Enzymes, heat Enzymes Heat Bioformation Heat Heat Heat Heat Heat Heat Bioformation

Cruciferae vegetables Vegetables Yeast, soy sauce Vegetables, meat, milk Tomato Meat, coffee, nuts, bread Meat Onion Meat, coffee, nuts, bread Meat, milk Meat Asparagus

Enzymes, heat Mushroom Heat, oxidation Potato, meat

formulation, the protein content of these substances, processing conditions (especially in the case of reaction flavors), and extraction conditions, where applicable. Identifying these proteins on flavor labels is an appropriate means of informing the food processor of their presence. The use of such flavors would determine if subsequent identification of the protein on the consumer product label is warranted.


TABLE VII PROTEINACEOUS SOURCE MATERIALS USED AS INGREDIENTS IN

FLAVOR MANUFACTURING

Protein source Substances used in flavor manufacturing

Milk and milk products

Peanut and peanut products

Seafood

Seeds

Tree nuts

Egg and egg products

Meat and meat products

Dried egg albumen, egg yolk solids, egg yolk extract, and frozen egg products.

In meat, the characteristic flavor compounds are formed in thermal reactions from sugars and amino acids. For sulfur-containing volatiles, the paramount source of sulfur is cysteine, either in the free form or as a part of the meat proteins.

Butter, butter oil, buttermilk solids, nonfat dry milk, enzyme modified milk powders, casein (dry and hydrolyzed), whey, sour cream, sour cream solids, cheese (including enzyme- modified cheeses), butter starter distillates (as defined by 21 CFR), lipolyzed butter oil, and butter powders. Note that butter esters, although the name implies dairy, are often mistaken to be a butter product. These esters tend to be those flavor chemicals known to be present in butter.

Peanuts extracts, peanut distillates, and peanut oils, which may or may not be refined.

For the most part, seafood ingredients include extracts of cod, shrimp, crab, lobster, oyster, scallop, crayfish, and molluscs. Some of these ingredients are available in powdered form.

These ingredients tend to be the oils derived from the seed. In particular, both refined and unrefined oils of sesame and sunflower as well as the partially hydrogenated oils of these seeds.

As with seeds, these ingredients tend to be the oil and partially hydrogenated oil. The oil is often used in concentrations up to 0.5% to eliminate dusting from ingredients such as autolyzed yeast extract. Also included in this category is soy sauce and soya flours.

nut, walnut, hazelnut [filbert], cashew, almond, pinenut, pistachio, pecan, macadamia) include the extracts and distillates.

Soybean or soybean-containing products

As with peanuts, the tree nut products (Brazil

(conrinues)


TABLE VII (Continued)

Protein source Substances used in flavor manufacturing

Wheat and wheat-containing products, and other grains

Barley, bran, cereals, gluten, and oats; starch ingredients used in flavor production tend to be modified cornstarch, maltodextrin, and barley malt flour as carriers. Various enzymes are used to prepare hydrolyzed products including lipolyzed butter and enzyme-modified cheeses. The enzymes include proteases, lipases. and peptidases. Various degrees of hydrolysis of various sources of plant proteins fall under this category. Included are corn gluten, wheat gluten, wheat protein, and soy protein. Yeast extracts prepared from bakers yeast, brewer’s yeast, Torula yeast, and other yeast extracts as well as autolyzed yeast extracts.

Enzymes

Hydrolyzed plant proteins

Yeast extracts

D. USAGE LEVELS OF FLAVORING SUBSTANCES

When considering the allergenic potential of flavoring substances, usage levels should also be taken into consideration. Because flavoring substances are used at low levels, ranging from parts per million (ppm) to perhaps 1% of the total product, very small amounts of many flavoring substances are present in finished food products. The wide variety of uses for flavoring substances makes it impossible to establish firm rules defining their usage levels, but many are self-limiting at high levels. In some cases, the use of a substance to impart a dominant note may be at a level more than 1000 times that of the same substance used as a trace ingredient (69). It is reasonable to conclude that in actual manufacturing practice, a level of 5 times the average maximum use will include nearly all normal applications of the ingredient (69). Table VIII provides examples of the average maximum use levels of some ingredients reported by food category. As noted earlier, an enormous variety of flavoring substances is allowed for use in foods. Thus, exposure to any single flavoring substance would be quite limited, especially given the low levels of typical use. Thus, allergic sensitization to most food flavors is rather unlikely due to the chemical composition and low level of use.

E. LABELING OF FLAVORING SUBSTANCES

Information required to be on the label of U.S. food products is itemized in Section 403 of the Food, Drug and Cosmetic Act. Among the required


disclosures is the need to list ingredients used in fabricated food, in descending order of predominance. This requirement was by no means meant to be rigidly comprehensive. From the start, Congress recognized that reasonable limits to any such required disclosure are essential to preserving an efficient and competitive food industry and the concomitant benefit of an abundant food supply. The intent of Congress is clear: the labeling provisions of the FDC Act do not reflect a consumer right-to-know orientation. Even on issues as sensitive as food ingredients, Congress has specifically exempted the identification of certain ingredients and provided the agency authority to, in its sound judgment, exempt further ingredients from disclosure requirements.

The purpose of Section 403(1)(2) of the FDC Act is to ensure that the consumer obtains “reasonable information regarding the composition of the food he buys.” This section of the Act generally excludes spices, flavors, and certain colors from its identification requirements. Congress also recognized that trade secret information regarding formulas falls outside the scope of “reasonable” information needed to inform the consumer.

Section 201(n) recognizes that certain foods or food ingredients may present risks for some consumers that can and should be averted through affirmative labeling. Under this authority, the FDA has required declara- tions identifying the presence of ingredients such as sulfiting agents (21 CFR 101.100(a)(4)), FD&C Yellow No. 5 (21 CFR 74.705(d)(2)), and other ingredients that the agency has concluded possess the potential to trigger adverse reactions in people sensitive to such ingredients.

The Code of Federal Regulations has outlined terms describing the physical or technical functional effects for which direct human food ingredients may be added to foods (Table IX). In Europe and many other countries, if any compounded ingredient (an ingredient composed of two or more ingredients) contains an ingredient that is less than 25% of the compounded ingredient, it need not be listed separately (70).

In some cases, when a flavoring substance is used for a function other than imparting flavor, the substance may be used at higher levels. If a known allergenic protein is added to the flavor in a larger quantity for such dual functional purposes, such as using whey as a flavor carrier, the likelihood of an allergic reaction is increased.

IV. REVIEW OF REPORTED ALLERGIC REACTIONS TO FOOD FLAVORING SUBSTANCES

Allergens are typically proteins of natural origin. As previously noted, flavors and flavoring substances rarely contain proteins. Most frequently,

w h)

TABLE VIII AVERAGE MAXIMUM USE LEVELS REPORTED BY FOOD CATEGORY (IN

Frozen Gelatins Meat and Nonalcoholic Alcoholic dairy Baked and meat Condiments,

Substance beverages beverages desserts Candy goods puddings products relishes

Anethole 42.5 723 53.5 531 495 52.8

Anise oil 31.3 570 61.3 681 182 46.3 Balsam of Peru 6.04 6.99 9.46 15.33 15.1 8.88 Carvone 41 145 497 226 115 90 Cassia 1947 2 1940 140 14.006 4000

Cassia bark oil 48.6 228 123 375 469 99.8

Cinnamon 57 616 155 12,967 1429

Ciannamic acid 175 Cinnamaldehyde 68.7 Cinnamyl alcohol 8.06 Clove stem oil" 52.6 Ethyl vanillin 29.7

712 263 356 384 290 499 78 550 367 109

600 129 414 40.9 125 8.97 10 21.3 17.9 14.8

10 26.6 89.6 92.9 39.9

10

27.1

0.1 402 1

237

1570

59.3

525 3.9

180

1198

264

214

Eugenol 2.18 1 3.79 14.9 Fennel. sweet. oil 55.4 234 70 67.9 d-Limonene 199 812 437 512 &Mentholh 16.1 15.9 63.5 591 Phenyl salicylate 1.21 1.9 2.07 3.25 Peppermint oil 47.4 282 175 1040

Propylene glycol 1239 5885 1401

21.3 2.75 102 84 61.9 185

500 452 79.1 53.2 70 4.86 1.8

327 145 17.1

2441 685 534

Spearmint oil

Vanillin

136

97.4

154 130 560 1318 95.4

47 55.2 408 186 117 2.72

~~

' The data used were from clove stem oil (vs clove bud or leaf oil) as it appeared to have the highest The data used were from &menthol (vs I-form) as the average maximum use levels reported are

N w


TABLE IX TERMS DESCRIBING THE PHYSICAL OR TECHNICAL FUNCTIONAL EFFECTS FOR WHICH

DIRECT HUMAN FOOD INGREDIENTS MAY BE ADDED TO FOODS

Anticaking agents Antimicrobial agents Antioxidants Colors and coloring

adjuncts Curing and pickling

agents Dough strengtheners Drying agents Emulsifiers and emulsifier

salts Enzymes Firming agents Flavor

Flavoring substances and adjuvants Flour texturizing agents Formulation aids Free-flow agents Fumigants Humectants Leavening agents Lubricants and release agents Nonnutritive sweeteners Nutrient supplements Nutritive sweeteners Oxidizing and reducing agents pH control agents

Processing aids Propellants, aerating

agents, gases Separation/filtration aids Sequesterants Solvents and vehicles Stabilizers and

thickeners Surface-active agents Surface-finishing agents Synergists Texturizers Tracers

flavor formulations are mixtures of many small-molecular-weight chemicals. Thus, it is not surprising that flavors have rarely been implicated in allergic reactions. Yet, several published examples of oral allergic reactions to flavors exist, and these are reviewed here.

A. PUBLISHED EXAMPLES OF ALLERGIC REACTIONS TO FLAVORING SUBSTANCES

Most of the few published cases of allergic reactions to flavors have been reactions to flavoring substances found in toothpaste, dental sup- plies, tobacco products, hard candy, and chewing gum (Table X). These cases do not reflect the situation that exists for most flavored food products.

These reactions are contact sensitivity reactions resulting from the flavor acting as a hapten and binding to mucosal proteins in the mouth, thus eliciting a cell-mediated immune response. These contact sensitivity reactions require repeated exposure to comparatively higher concentrations of the offending substance and prolonged contact with the affected tissue. The reactions are localized and are likely delayed hypersensitivity reactions.

Balsam of Peru is the most notable of the many flavoring substances in its ability to induce such reactions. The major components of Balsam of


TABLE X DOCUMENTED CASES OF ORAL SENSITIVITIES TO FLAVORS OR FLAVORING SUBSTANCES

Substance Patient reacted to substance found in Reference

Anethole found in oil of anise. star anise, and fennel

Anise oil Balsam of Peru Car v o n e

main constituent of spearmint oil; also found in caraway oil

compounds: cinnamonkassia oil, cinnamic aldehyde, cinnamic acid, cinnamic alcohol

Cinnamon (cassia) and cinnamic

Clove oil Eugenol

integral part of oil of carnation, oil of bay, and other essential oils

Menthol Peppermint oil

free menthol (45%). limonene, pinene

carvone (50%), limonene, pinene Spearmint oil

toothpaste

denture cream dentures, toothpaste toothpaste

toothpaste, chewing and bubble gum, mouthwash tablets, breath spray, hard candy, antacid, dental cleaner

dental cleaner oral postsurgical dressing, temporary

cavity preparation, dental cement, impression paste

toothpaste, cigarettes, hard candy toothpaste

toothpaste

71

72 73 71

74-82

80 83, 84

83-88 71-86

72-86

Peru are listed in Table XI (89,90). Several components of Balsam of Peru (cinnamic aldehyde in particular) are notorious contact sensitizing agents (74). Most reported reactions to Balsam of Peru occur from its use in cosmetics (90) and over-the-counter (OTC) drugs where frequent and prolonged exposure enhances the likelihood of contact sensitization.

Thus, reactions in the oral cavity associated with the use of Balsam of Peru in toothpaste and dental materials are not surprising. Similar com-

TABLE XI COMPONENTS OF BALSAM OF PERU

Cinnamic aldehyde 2% Cinnamic alcohol 2% Cinnamic acid 5% Methyl cinnamate 0.5% Eugenol 5% Vanillin 10%


ments could be made about the other flavor substances involved in oral contact sensitivities noted in Table X. Balsam of Peru is also allowed for use as a food flavoring substance, although it is not commonly used. Foods containing Balsam of Peru include products containing citrus peel (marma- lade, baked goods, juice), products flavored with essences (baked goods, candy, chewing gum), ice cream, some nonalcoholic beverages, and products made with cinnamon, cloves, or vanilla. Contact sensitivity reactions to foods containing Balsam of Peru have been documented (91-93). Typical usage levels are listed in Table VIII, with the highest concentrations used in chewing gum and hard candy. Sensitization to Balsam of Peru almost certainly occurs through dermal exposure. The individuals described in earlier studies of sensitivities to Balsam of Peru in foods (91-93) were likely sensitized through exposure to Balsam of Peru in cosmetics and OTC drugs, although they do react to Balsam of Peru in ingested foods thereafter in some cases (91-93). A cook who was sensitized to cinnamon was observed by Fisher (94). The patient’s chronic hand eczema cleared upon avoidance of contact with cinnamon, but he suffered from a subsequent flare of dermatitis and from urticaria whenever he drank vermouth, which contains cinna- mates. Another individual was observed who had marked positive reactions to Balsam of Peru, oil of cinnamon, and cinnamic aldehyde, and suffered from generalized urticaria whenever he drank soft drinks containing these substances (94). The mechanism of these adverse reactions to Balsam of Peru is not known but several of the active components, for example, cinnamic aldehyde, are known to act as haptens. Thus, these are likely cell- mediated, delayed hypersensitivity reactions.

IgE-mediated allergic reactions resulting from other flavoring substances in typical food products are even more rare, and in such cases the flavoring contained well-known allergens from milk or peanuts. While the principal examples involve milk, similar situations could arise with the use of proteins from other prominent allergenic foods (soy, wheat, eggs, peanuts, tree nuts, etc.) found in Table 11. In a report by Gern et al. (95), four patients had reactions after eating foods that were labeled “nondairy” or “pareve” (containing no milk products), either a beef hot dog (two cases) or a bologna product (two cases). Traces of milk protein were found in these products. On each occasion, a parent had carefully read the product label in order to avoid exposing the child to foods containing milk or milk proteins. Each of the patients had subsequently eaten different brands of similar foods without experiencing adverse reactions. The manufacturer of the hot dog and bologna stated that the “natural flavorings” component of the product formulation had recently been changed from yeast autolysate to hydrolyzed sodium caseinate. Since hydrolyzed sodium caseinate was considered a “natural flavoring” by this manufacturer, it was determined that it was


unnecessary to specifically list the flavoring ingredient on the label. How- ever, enough milk allergen existed in the product to elicit allergic reactions in milk-allergic individuals. Under current FDA regulations, hydrolyzed sodium caseinate must be labeled on such consumer products.

St. Vincent and Watson (96) documented the cases of two patients who developed allergic reactions after ingesting an unsuspected source of cow’s milk protein found in dill pickle-flavored potato chips. The package label listed “dill pickle seasoning,” but not milk or milk products as an ingredient. Information from one manufacturer showed that the dill pickle seasoning contained whey as an ingredient. Another manufacturer who makes dill pickle-flavored chips did not list milk as an ingredient, but includes lactose and “spices.” The lactose was derived from cow’s milk, and may contain residues of milk proteins (96).

In 1993, the Food Allergy Network and Kellogg’s Company (97) jointly alerted allergic individuals that the new Kellogg’s Rice Krispies Treats Cereal contained natural and artificial flavorings that may contain milk proteins, in an effort to prevent milk-allergic individuals from inadvertently consuming a product that contained milk proteins but was not labeled as such. The ingredient statement originally reflected only the presence of natural and artificial flavorings.

In 1996, a similar episode occurred involving a soup product sold by Fantastic Foods, Inc. (98). The flavoring in the soup contained peanut flour and elicited an allergic reaction. The ingredient statement revealed only the presence of natural flavors.

Persons allergic to milk and peanuts, including those who read product labels carefully, may consequently continue to be exposed to quantities of milk and peanut proteins sufficient to cause generalized reactions. Physi- cians should therefore consider this possibility in patients with milk and peanut sensitivity who have adverse reactions of unclear origin.

Some of the most commonly used flavoring substances are spices. The most common spice allergy is mustard; the allergens from mustard have been identified (99-102). Other spices that have been known to cause allergic reactions are allspice (103), anise (101,104,105), caraway (106), cardamon (107). clove (103), coriander (105-107), cumin (104), curry (103- 105), dill (100). fennel (104-log), ginger (103-105), mace (109), nutmeg (104), paprika (103), parsley (1 lo), pepper-cayenne (106), pepper-red (104), and pepper-white (103-105).

Difficulties in distinguishing between irritant and allergic reactions from spices are well noted. Skin tests from some spices are notorious for false- positive responses. Even in modest amounts, putting cinnamon on the skin can cause a wheal and flare in a nonallergic individual due to irritation caused by the spice. Thus, while some of the reported adverse reactions


to spices are probably true IgE-mediated reactions, others may be irritant reactions with false-positive skin tests. While allergic reactions to spices have been reported more frequently than allergic reactions to other flavoring substances, such sensitivities are still considered rather rare.

B. LIKELIHOOD OF ALLERGIC OR INTOLERANCE REACTIONS OF FLAVORING SUBSTANCES

I . Allergic Reactions

IgE-mediated allergic reactions to flavor substances are unlikely for several reasons: (1) virtually all allergens are proteins (3,4) and proteins are infrequently used in flavors, (2) when proteins are used as a flavor or flavoring substance, the proteins used are rarely allergenic proteins, (3) in some cases, if an allergenic protein is used, the manufacturing process will remove the protein fractions through refinement or distillation, and (4) if an allergenic protein is used and the protein survives the manufacturing process, in most cases it will be at such a low level that it would not elicit a reaction from an allergic individual. If the source of the flavoring substance is not allergenic, then the flavoring substance will not be. If the source is allergenic then it must be determined if any protein survived the manufacturing of the flavor, and if it remains allergenic. As noted above, few flavor substances contain allergenic proteins.

Delayed hypersensitivity or cell-mediated allergies are also unlikely to occur with flavor substances except in a few special circumstances. Cell- mediated reactions involving proteins, such as celiac disease, are highly unlikely due to the reasons noted above.

It also must be determined if the ingredient can act as a hapten. Nonpro- teinaceous, low-molecular-weight flavoring substances can sometimes act as haptens. Haptens will more frequently elicit cell-mediated reactions than IgE-mediated sensitivities. Cinnamic aldehyde is an outstanding example. However, to cause a cell-mediated response, the flavoring ingredient (1) must be reactive with proteins; (2) must be present in comparatively large amounts in the food product; (3) as the protein-hapten complex, must survive digestion or react in the oral cavity; and (4) must have a long residence time in the oral cavity for localized oral reactions. Contact sensitivity, a form of cell-mediated reaction, has occurred in the oral cavity with highly reactive flavoring substances in toothpaste, chewing gum, or hard candy. These cases are limited to the few flavoring substances that can act as haptens, bind to proteins, and elicit immune responses in the oral mucosa. However, such reactions have not been documented with other food applications of these same flavoring ingredients. Substances


used as flavoring substances that are associated with contact sensitization are listed in Table XII. However, the contact sensitization reactions are primarily associated with cosmetic use and not with the use of flavors in foods. Such reactions in food would be more likely to occur in prolonged exposure situations as with cosmetics, toothpaste, dental materials, and perhaps chewing gum and hard candy. Contact sensitivities could also con- ceivably occur from handling of flavor preparations in occupational settings. The low concentrations of flavoring substances in food, the diversity of flavoring substances used in foods, and the comparatively short contact time with specific tissues argues against contact sensitivity as a mechanism for adverse rections to food flavor substances, except in a few unique circumstances.

Contact dermatitis to flavors was first described in 1993 by Hutchinson (lll), who mentioned a skin eruption caused by vanilla. Later case reports identified flavors in toothpastes, dental preparations, and cosmetics as causes of allergic contact dermatitis in cheilitis, gingivostomatitis, or hand eczema. Some of the allergic reactions described in Table X, especially those involving toothpaste and various dental and denture products, are most likely contact sensitivity reactions.

TABLE XI1 SOME COMMON FLAVORING SUBSTANCES REPORTEDLY ELICITING

MUCOSAL CONTACT SENSITIZATION REACTIONS

Substance Reference

Anise oil (anethole) Balsam of Peru Carvone Cinnamic acid Cinnamic alcohol Cinnamic aldehyde Cinnamon (cassia) Cinnamon (cassia) oil Ethyl vanillin Eugenol (oil of clove) Fennel oil (anethole) d-Limonene Menthol Peppermint oil (menthol) Phenyl salicylate Spearmint oil (carvone) Vanillin

71, 72, 112

71 89, 125 125, 126, 134 71. 74, 77, 79-81, 89, 125, 126, 135, 136 74, 78, 79, 82, 118 75, 76, 80, 81, 129, 130, 136 129, 130 76, 80, 84, 125, 134 71, 72 129, 137 71, 82, 85-88, 138, 130 71, 86 140, 141 71. 86 88, 90, 125, 131

73.76, 79. 89. 90.92, 113-133


2. Intolerances

Other types of food intolerances are unlikely to occur with flavoring substances. Known metabolic food disorders do not involve flavoring substances. In addition, such reactions display high thresholds, and the low concentrations of flavoring substances would not likely exceed these thresholds, The identities of food-borne substances involved in non-IgE-mediated histamine release reactions remain unknown, but no evidence exists for an association with food additives, including flavoring substances. The role of flavoring substances in idiosyncratic reactions is also speculative since the association with food remains to be established with most of these reactions. In these cases as well, high threshold responses have been noted with the well-documented idiosyncratic reactions, such as sulfite-induced asthma, and the low concentrations of flavoring substances would not likely exceed such thresholds. Sulfites are not widely used in flavoring formulations.

Adverse reactions to flavor enhancers, such as monosodium glutamate, and related flavor enhancers, such as hydrolyzed vegetable protein and yeast autolysate, would be classified as idiosyncratic reactions. Flavor enhancers may be defined as compounds exhibiting little or negligible odor and taste, but they magnify, usually manyfold, the characteristic flavor of certain food substrates when used in small or even trace amounts. The flavor enhancers used in food are primarily MSG, yeast autolysate, hydrolyzed proteins made from soy, wheat, corn, gelatin, casein, and other sources, and several nucleotides. The involvement of MSG in these idiosyncratic reactions remains to be proven. While numerous adverse reactions have been attributed to MSG, especially in the lay media, a recent and extensive review of MSG conducted by a group of independent scientists under the auspices of the Federation of American Societies for Experimental Biology (FASEB) helped put these safety concerns into perspective and reaffirmed the Food and Drug Administration’s belief that MSG and related substances are safe ingredients for most people when eaten at customary levels (142). Some evidence exists to suggest certain people may develop short-term reactions (flushing, etc.) when they consume large doses (3 g or more) of MSG (142). No evidence was found linking the “MSG Symptom Complex” to consumption of low levels of MSG (142). There may be a small subgroup of people with severe asthma who respond to large doses of MSG (3 g or more) (145). However, the mechanism is unknown, and the role of MSG in asthma has not been confirmed (55,56,143). While a complete discussion of MSG and its possible role in idiosyncratic reactions is beyond the scope of this review, the level of MSG in the diet originating from flavor formulations would be insufficient (far below 3 g) to elicit the adverse reactions noted in the FASEB report.


The FASEB report indicates that no evidence exists for allergic reactions to such ingredients as hydrolyzed proteins and yeast autolysates (142). Fully hydrolyzed proteins would be mixtures of amino acids, including glutamate, and would be unlikely to elicit adverse reactions in the levels used in flavoring formation (see the preceding MSG discussion). Thus, reactions from MSG and fully hydrolyzed proteins (1) have not been proven, (2) remain highly controversial, and (3) would not be true allergic reactions in any case but fall under the category of idiosyncratic reactions (56,142,143). Furthermore, allergic reactions to yeast autolysates would not be expected except on rare occasions, because few individuals have IgE directed toward yeast protein. However, the partially hydrolyzed proteins from commonly allergenic sources such as soy, wheat, and casein could induce true IgE-mediated allergies if the allergenic epitopes remained intact after hydrolysis. In the United States, hydrolyzed proteins must be listed and declared by source when used as flavor enhancers or for other technological functions, including flavoring.

Thus, both scientific logic and experience dictate against a major role for flavoring substances in allergic and other sensitivity reactions, except in those cases when a large quantity of a known allergen is added to the flavor formulation.

C. OCCUPATIONAL SENSITIVITIES

Occupational sensitivities to flavor substances are unlikely to occur. Fur- thermore, they would not provide clues to adverse reactions in consumers exposed to much lower amounts of these substances. If occupational sensitivities due to flavor substances existed, it would be found in flavor manufacturing sites where the workers would be exposed to higher levels, rather than in food manufacturing. Hand eczema due to contact allergy to spices was found in growers and handlers of spices and people who have frequent contact with spices such as bakers, confectionerykandy makers, chemists, cooks, and homemakers (74,103,129,131,134,144). Garlic, onion, lettuce, endive, parsley, and carrots were reported to be responsible for most hand contact dermatitis in homemakers and catering workers (145). Very few, if any, occupational sensitivities due to inhalation of flavoring substances have been reported. Garlic and onion sensitivity has been documented only at high levels found in dehydration facilities (145-147).

V. APPROPRIATE DIAGNOSTIC TESTS FOR INVESTIGATION OF SENSITIVITY TO FOOD FLAVORING SUBSTANCES

Allergists frequently encounter patients who provide histories of recent adverse reactions to a specific food product. If the historical account is


convincing and if the patient has a known food allergy, the physician should suspect that the patient has inadvertently ingested that food even if a source of that food is not clearly identifiable on the package label. Cross- contamination can occur in the food and food service industries from the use of shared equipment or inadequate clean-up procedures. For example, Yunginger et al. (148) described reactions to peanut residues remaining in sunflowers seed butter from the use of shared equipment. A determination can be made regarding the presence of that particular allergenic food in the incriminated product using the RAST inhibition assay or some similar test (149). If this patient does not have a known food allergy or if the known allergenic food is not detected in the incriminated product, the physician must first confirm a cause and effect association between the incriminated food product and the alleged adverse reaction using challenge procedures as documented below. If a positive response is observed in the challenge test, then the search for the causative ingredient in the food product must commence. Since the ingredients are listed in descending order of prevalence on the label, the obvious strategy would involve testing these ingredients in their descending order of prevalence. Skin testing could be conducted initially if an IgE-mediated reaction was suspected. However, a challenge test is again the best means of establishing a cause and effect relationship.

Since flavors are grouped and usually appear at or near the end of the ingredient list on a food label, they are the least prevalent ingredients in most foods. However, the physician occasionally encounters negative results from evaluation of the more prevalent ingredients, which results in a desire to evaluate the flavoring substances in the product. An approach for evaluating the likelihood of the involvement of flavoring substances in allergic reactions is provided in Table XIII. The assessment strategy detailed in Table XI11 presumes that alternative explanations as noted above have been considered and eliminated. Since flavor formulations are proprietary and are quite complex in most foods, physicians should seek information from the food manufacturers and flavor suppliers in such assessments. Flavor formulations are confidential information that is the “life-blood’’ of the flavor industry. However, there are certain conditions under which the flavor industry must be willing to disclose enough formula information for a physician to treat a patient. Normally verbal communication between the physician and the manufacturer will suffice. If the patient has known food allergies, celiac disease, or contact sensitivity, then the physician may be able to make a probable diagnosis, without need for further testing, if sources of known offending substances are found in the flavor formulation. Flavor manufacturers should be able to provide information on the protein content and origin of flavors. Skin tests, patch tests, and double-blind,


TABLE XI11 APPROACH TO THE DIAGNOSIS OF ALLERGIC REACTIONS TO FOOD FLAVORS

I. Systemic reactions (anaphylactic shock, generalized urticaria, dermatitis, gastrointestinal symptoms. etc.) A. Does the patient have a history of known IgE-mediated food allergy?

If yes, continue and answer parts a-d. If no, go to No. 2. 1. Does the food contain any dual-functional flavors?“

a. What is the flavor source material? b. Was it derived from an allergenic source? c. Does it contain proteins from that source? d. Does the source match with the sensitivity of the patient?

If yes, the diagnosis is confirmed. If no. go to No. 2. 2. Was the flavor formulation derived in whole or in part from an allergenic source?”

If yes, continue and answer parts a-c. If no, the likelihood of involvement of the flavors in the allergic reaction is remote. a. What is that source? b. Does it contain proteins? c. Does the source match with the sensitivity of the patient?

If yes, the diagnosis is confirmed. If no, do skin tests to assess allergenicity to other known allergenic sources in the formulation. Consider DBPCFC to confirm positive skin test. Positive skin tests confirm the diagnosis.

B. If the patient has a history of celiac disease, answer the following: 1. Does the food contain any dual functional flavors?“

If yes, continue and answer a and b. If no, go to No. 2. a. Was the ingredient derived from wheat, rye, barley, triticale, or oats? b. Does it contain proteins from that source?

If yes, the diagnosis is confirmed. If no, go to No. 2. 2. Was the flavor derived in whole or in part from wheat, rye, barley, triticale, or

oats?” a. Does it contain proteins?

If yes. the diagnosis is confirmed. If no, consider other sources in the patient’s diet for this delayed type of hypersensitivity.

11. Localized oral reactions A. If the patient has a history of known IgE-mediated food allergy, repeat parts 1 and

2 from Part I,A. B. If the patient has a history of contact sensitivity to cosmetics, answer the following:

1. Does the flavor formulation contain any known contact sensitizing substances? (see Table XII)”

2. Does the flavor formulation contain any substances that are highly reactive with proteins?h If yes, continue and answer part a and consider patch tests of those substances to confirm diagnosis. If no, the involvement of the flavors in the allergic reaction cannot be confirmed. a. Is this substance present in high proportion in the flavor formulation?

‘’ The food manufacturer should be able to answer this question in most cases. Dual functional

” The food manufacturer can only answer this question after consultation with flavor sup- flavors are ingredients used for more than one purpose.

plier(s).


placebo-controlled challenge tests can be considered to suggest and confirm possible diagnoses of the role of flavors in allergic reactions, but are unlikely to be necessary except in very unusual circumstances. The preferred approaches to challenge testing and skin testing for the causative agent are provided below.

A. CHALLENGE TESTS

The double-blind, placebo-controlled food challenge (DBPCFC) is considered the “gold standard” for the diagnosis of adverse reactions to foods (150). This procedure should have similar status in the evaluation of the adverse reactions to food ingredients including flavoring substances. The best use of the DBPCFC is to confirm or refute a cause and effect association between the adverse reaction and the specific food product. Any challenge tests should be conducted using the oral route of exposure since the alleged adverse reaction also involved that route of exposure. If the incriminated food is available, the double-blind challenge should be done with the product. If the actual incriminated food is not available, an identical product purchased from the marketplace is the next best choice. The physician may wish to conduct a skin test using a freshly prepared extract of the whole food before conducting the challenge test. A positive response in the skin test may indicate the presence of an allergenic food in the product, whether acknowledged on the label or not. If the patient has a known food allergy, the physician may want to search for the presence of that allergenic food in the product before conducting a direct challenge test. If the patient does not have a known food allergy, a positive skin test may indicate the existence of an IgE-mediated allergy to one of the ingredients in the specific food product. With a positive skin test, the physician should approach the challenge test with some caution since patients with positive skin tests are more likely to react in the challenge trial. Extra precautions would need to be exercised in the challenge test especially if severe reactions are thought possible. Challenge studies will have some serious limitations for the evaluation of adverse reactions to flavors and flavoring substances. It will be especially difficult to identify the specific flavoring substance responsible for any adverse reaction. It is, in general, only necessary to determine the protein content and source. Flavoring formulations are also difficult to disguise so adequate placebo controls are difficult to obtain. Finally, flavors are typically used at rather low concentrations, so the amount used in the DBPCFC should be similar to the level encountered in food applications. Since flavor formulations are difficult to acquire and use in such challenge tests, their use is not advocated in the approach outlined in Table XIII.


Where they are used, the evaluation of adverse reactions to the flavoring formulation rather than the individual flavoring substances would suffice for diagnostic purposes.

B. SKINTESTS

The skin pick test is probably the most frequently used test for the diagnosis of IgE-mediated allergies including food allergies. The skin prick test is typically conducted by the prick/puncture method with diluted (1 : 10 or 1 : 20) extracts of the specific food or food ingredient. A positive skin test is indicated by the development of a wheal at the puncture site that is 3 mm or larger in diameter than a diluent control wheal measured within 10-30 min of treatment. The proper application of the skin prick test in the evaluation of food allergies has been described by Bock et al. (151).

Skin tests may be quite useful for the detection of allergenic foods in a food product incriminated in a specific incident. If the patient is known to have food allergies, then a search should be made for sources of those specific foods in the product. If the patient does not have a history of food allergies, a positive skin test may indicate the existence of such an allergy to some component of the incriminated food product.

However, the use of skin tests in the evaluation of adverse reactions to food flavoring substances or formulations is limited. Skin tests identify the presence of IgE-binding allergens, which are typically proteins of natural origin. As discussed earlier, flavors and flavoring substances will not often contain such ingredients. Most frequently, flavors will be mixtures of small- molecular-weight chemicals that are not likely allergens. If the patient has a known food allergy, a skin test with the product’s flavoring formulation could be used to confirm the presence of an allergenic protein, probably from that source in the formulation. If other sources of proteins are identified in the suspect flavor through the assessment approach outlined in Table XIII, skin tests could be used to assess the possible existence of allergic sensitivity to those foods. However, in any other situation, the use of skin tests must be approached with care because of the likelihood of false- positive reactions. Many of the chemicals used in flavors may be irritating to the skin even in modest amounts. Therefore, the injection of small amounts of an undiluted flavor may cause an irritational response, which can be easily read as a false positive. Even with dilution, this possibility will exist. Control skin tests must be conducted on nonsensitive individuals to identify such false-positive responses. The ideal diluent for flavor or flavoring substance is buffered physiological saline. Histamine is used as a positive control substance for skin testing.


C. PATCH TESTS

Patch tests are used in the diagnosis of contact sensitivity reactions. A patch test involves applying a specific substance to the skin with the inten- tion of producing a small area of allergic contact dermatitis. The patch test is generally placed on the skin for 48-72 hr and then observed for the appearance of localized dermatitis. The test material is applied to a small area of the skin and covered by gauze held in place with tape. The concentration of the test material is critical, but, in the case of flavors, the concentration should not exceed the level found in the suspect food by more than severalfold. False-positive reactions can occur with patch testing especially with irritant substances. Since some flavoring substances likely have irritant properties, precaution must be taken to avoid misinterpretation of the results. The use of relevant concentrations is one reasonable precaution. The comparison to results obtained with normal, control individuals should also help to identify false positives. Patch tests could be extremely useful in the diagnosis of contact sensitivity reactions to flavors found in food products such as chewing gum and hard candy where the exposure to the flavor might be sufficiently long and frequent to induce such reactions.

VI. CONCLUSIONS

Flavoring substance in food are highly unlikely to elicit adverse reactions among consumers. The extremely low amounts of individual flavoring substances used and the wide variety of such flavoring substances in foods argues against any significant role for these substances in the adverse reactions of foods. The nonprotein nature of most flavoring substances implies that these substances will not likely be involved in IgE-mediated allergic reactions or celiac disease. While some of these chemicals could possibly act as haptens to initiate an IgE-mediated response, this mechanism is unlikely given the low usage levels of flavoring substances and the effect of digestion on protein-hapten complexes. However, contact sensitivity reactions may occur in some situations where prolonged oral exposure occurs to a flavoring substance, for example, chewing gum or hard candy. Other forms of food intolerances are unlikely to be triggered by flavoring substances because affected consumers typically have a finite tolerance for the offending food with these illnesses. The small amounts of flavors used in foods would argue against the involvement of flavoring substances in such food intolerances.

Flavors do occasionally contain food proteins in small quantities. If these proteins are obtained from known allergenic foods such as peanuts or


cows’ milk, listed in Table 11, then these particular flavors might trigger IgE-mediated allergic reactions in allergic consumers. However, the very low quantities of proteins from the flavors in the final product would in most instances likely result in a mild adverse reaction by comparison to other types of inadvertent exposures to the same protein, except in the cases when higher levels are added for dual functionality as mentioned earlier.

The term “food allergy” should only be applied to immunologically based adverse reactions invoked by food. In cases where the patient has a known food allergy, the presence of that allergenic food should be sought in the incriminated food product before any consideration is given to possible adverse reactions to flavoring formulations. The role of flavoring substances in any adverse reaction should be confirmed by DBPCFC using appropriate dosages of the flavoring substance(s) and effective blinding approaches with the objective assessment of symptoms wherever possible.

ACKNOWLEDGMENTS

The author acknowledges, with appreciation, the critical review of this paper by the following members of the Council of Scientific Advisors of the International Life Sciences Institute’s Allergy and Immunology Institute: John A. Anderson, Professor Bengt BjorkstCn, Jean Bous- quet, Sheldon C. Cohen, David J. Hill, Dean D. Metcalfe, Lanny J. Rosenwasser, Stephen I. Wasserman.

REFERENCES

I. Lehrer, S B. 1993. Clin. Rev. Allergy 11, 155. 2. Taylor, S. L. 1985. Food allergies. Food Technol. 39, 98. 3. Taylor, S. L., Nordlee, J. A,, and Rupnow, J. H. 1989. Food allergies and sensitivities.

In “Food Toxicology: A Perspective on the Relative Risks’’ (S. L. Taylor and R. A. Scanlan, eds.), p. 225. Dekker, New York.

4. Aas, K. 1983. The critical approach to food allergy. Ann. Allergy 51, 256. 5. Anderson. J. A. 1986. The establishment of common language concerning adverse reac-

tions to foods and additives. J . Allergy Clin. Immunol. 78, 140. 6. Sloane, A. E. 1986. A perspective on popular perceptions of adverse reaction to foods.

J. Allergy Clin. Imniunol. 78, 127. 7. Barrett, K. E., and Metcalfe, D. D. 1988. Immunologic mechanisms in food allergy. In

“Food Allergy: A Practical Approach to Diagnosis and Management” (L. T. Chiara- monte, A. T. Schneider, and F. Lifshitz, eds.), pp. 23-43. Dekker, New York.

8. Sampson. H. A,, Mendelson, L., and Rosen. J. P. 1991. Fatal and near-fatal anaphylactic reactions to foods in children and adolescents. N . Engl. J. Med. 327, 280-284.

9. Yunginger. J. W., Sweeney, K. G., Sturner, W. Q., Giannandrea, L. A,, Teigland, J. D., Bray, M., Benson, P. A,, York, J. A,, Biedrzycki, L., Squillace, D. L., and Helm, R. M. 1988. Fatal food-induced anaphylaxis. J . Am. Med. Assoc. 260, 1450.


10. Bock, S. A., and Atkins, F. M. 1990. Patterns of food hypersensitivity during sixteen

11. James, J. M., and Sampson, H. A. 1992. Immunologic changes associated with the develop-

12. Burks, A. W., and Sampson, H. A. 1993. Food allergies in children. Curr. Probl. Pediatr.

13. Sampson, H. A. 1988. IgE-mediated food intolerance. J. Allergy Clin. Immunol. 81,495. 14. Taylor, S. L., Bush, R. K., and Busse, W. W. 1986. Avoidance diets-How selective

should we be? N. Engl. Allergy Proc. 7,527. 15. Taylor, S. L., Bush, R. K., and Busse, W. W. 1987. Avoidance diets-How selective

should they be? In “Food Allergy” (R. K. Chandra, ed.), p. 253. Nutrition Research Education Foundation, St. John’s Newfoundland.

16. Yman, I. M., Eriksson, A,, Everitt, G., Yman, L., and Karlsson, T. 1994. Analysis of food proteins for verification of contamination or mislabeling. Food Agric. Immunol. 6, 167.

17. Nordlee, J. A., Taylor, S. L., Jones, R. T., and Yunginger, J. W. 1981. Allergenicity of various peanut products as determined by RAST inhibition. J . Allergy Clin. Zmmunol. 68, 376.

18. Barnett, D., Baldo, B. A., and Howden, M. E. H. 1983. Multiplicity of allergens in peanuts. J . Allergy Clin. Immunol. 72, 61.

19. Barnett, D., and Howden, M. E. H. 1986. Partial characterization of an allergenic glyco- protein from peanut (Arachis hypogaea L). Biochim. Biophys. Acta 882, 97.

20. Herian, A. M., Taylor, S. L., and Bush, R. K. 1993. Allergenic reactivity of various soybean products as determined by RAST inhibition. J. Food Sci. 58, 385.

21. Haddad, Z . H., Kalra, V., and Verma, S. 1979. IgE antibodies to peptic and peptic- tryptic digests of betalactoglobulin: Significance in food hypersensitivity. Ann. Allergy 42, 369.

22. Burks, A. W., Williams, L. W., Tresher, W., Brooks, J. R., and Sampson, H. A. 1991. Identification of soy protein allergens in patients with atopic dermatitis and positive soy challenges, determination of change in allergenicity after heating or enzymatic digestion. Adv. Exp. Med. Biol. 289, 295.

23. Kelso. J. M., and Sampson, H. A. 1993. Protein-mediated gastroenteropathy to casein hydrolysate formulas. J. Allergy Clin. Immunol. 91, 50.

24. Taylor, S. L., Busse, W. W., Sachs, M. I., Parker, J. L., and Yunginger, J. W. 1981. Peanut oil is not allergenic to peanut-sensitive individuals. 1. Allergy Clin. Zmmunol. 68, 373.

25. Bush, R. K., Taylor, S. L., Nordlee, J. A,, and Busse, W. W. 1985. Soybean oil is not allergenic to soybean-sensitive individuals. J. Allergy Clin. Zmmunol. 76, 242.

26. Halsey, A. B., Martin, M. E., Ruff, M. E., Jacobs, F. 0.. and Jacobs, R. L. 1986. Sunflower oil is not allergenic to sunflower seed-sensitive patients. J . Allergy Clin. Immunol. 74,235.

27. Lemanske, R. F., and Taylor, S. L. 1987. Standardized extracts of foods. Clin. Rev. Allergy 5, 23.

28. Mansfield, L. E., and Bowers, C. H. 1983. Systematic reaction to papain in a nonoccupa- tional setting. J . Allergy Clin. Immunol. 71, 371.

29. Minikin, W. P.. and Lynch, P. J. 1969. Allergic reactions to penicillin in milk J. Am. Med. Assoc. 209, 1089.

30. Dewdney, J. M., and Edwards, R. G. 1984. Penicillin hypersensitivity-Is milk a significant hazard?: A review. J . Royal Soc. Med. 77, 866.

31. Sampson, H. A. 1990. Immunologic mechanisms in adverse reactions to foods. Zmmunol. Allergy Clin. North Am. 11, 701.

years of double-blind, placebo controlled food challenges. J . Pediatr. 117, 561.

ment of tolerance in children with cow milk allergy. J. Pediatr. 121, 377.

23,230-252.


32. Stern, M.. and Walker, W. A. 1985. Food allergy and intolerance. Pedintr. Clin. North Am. 32,471.

33. Phillips, A. D.. Rice, S. J., France, N. E., and Walker-Smith, J. A. 1979. Small intestinal intraepithelial lymphocyte levels in cow’s milk protein intolerance. Gut 20, 509.

34. Stern, M., Dietrich, R., and Muller, J. 1982. Small intestinal mucosa in coeliac disease and cow’s milk protien intolerance: Morphometric and immunofluorescent studies. Eur. J . Pediatr. 139, 101.

35. Ashkenazi. A., Levin, S., Idar, D., Or., A,, Rosenberg, I., and Handzel, 2. T. 1980. In vitro cell-mediated immunologic assay for cow’s milk allergy. Pediatrics 65, 399.

36. Minor, J. D., Tolber, S. G., and Frick, 0. L. 1980. Leukocyte inhibition factor in delayed- onset food allergy. J . Allergy Clin. Imnzunol. 66, 314.

37. Fallstrom, S. P., Lindholm, L., and Ahlstedt, S. 1983. Cow’s milk protein intolerance in children is connected with impaired lymphoblastic responses to mitogens. Int. Arch. Allergy Appl. Immunol. 70, 205.

38. Strober, W. 1986. Gluten-sensitive enteropathy: A nonallergic immune hypersensitivity of the gastrointestinal tract. J. Allergy Clin. Immunol. 78, 202.

39. Kasarda, D. D. 1978. The relationship of wheat protein to celiac disease. Cereal Foods World 23, 240.

40. Hartsook. E. I . 1989. Celiac sprue: Sensitivity to gliadin. Cereal Foods World 29, 157. 41. O’Mahoney, S., and Ferguson, A. 1991. Gluten-sensitive enteropathy (celiac disease).

In “Food Allergy-Adverse Reactions to Foods and Food Additives” (D. D. Metcalfe, H. A. Sampson, and R. A. Simon, eds.), p. 186. Blackwell Scientific Publications, Boston.

42. Skerritt, J. H., Devery, J.. and Hill, A. S. 1990. Gluten intolerance: Chemistry, celiac- toxicity, and detection of prolamins in foods. Cereal Food World 35, 638.

43. American Academy of Allergy and Immunology and National Institute of Allergy and Infectious Diseases. 1984. “Adverse Reactions to Foods,” NIH Publication 84-2442. US. Dept. of Health and Human Services, Washington, D.C.

44. Taylor, S. L. 1987. Allergic and sensitivity reactions to food components. In “Nutritional Toxicology” (J. N. Hathcock, ed.), Vol. 11, p. 173. Academic Press, Orlando, FL.

45. Taylor, S. L., Guthertz, L. S., Leatherwood, M., Tillman, F., and Lieber, E. R. 1978. Histamine production by food borne bacterial species. J . Food Safety 1, 173.

46. Stevenson, D. D., and Simon, R. A. 1981. Sensitivity to ingested metabisulfites in asthmatic subjects. J . Allergy Clin. Immimol. 68, 26.

47. Bush, R. K.. Taylor, S. L., Holden, K., Nordlee, J. A, and Busse, W. W. 1986. The prevalence of sensitivity to sulfiting agents in asthmatics. Am J. Med. 81,816.

48. Monro, J., Rostoff, J., Carini, C., and Zilkha, K. 1980. Food allergy in migraine. Study of dietary exclusion and RAST. Lancer ii, 1.

49. Egger, J.. Carter, C. M., Wilson, J., Turner, M. W., and Soothill, J. F. 1983. Is migraine food allergy? A double-blind controlled trial of oligoantigenic diet treatment. Lancet ii, 865.

SO. Perkin, J. E., and Hartje, J. 1983. Diet and migraine: A review of the literature. J. Am. Diet. Assoc. 83, 459.

51. Stare, F. J.. Whelan E. M., and Sheridan, M. 1980. Diet and hyperactivity: Is there a relationship Pedintrics 66, 521.

52. Harper, A. E., and Cans, D. A. 1986. Diet and behavior-An assessment of reports of aggressive, antisocial behavior from consumption of sugar. Food Technol. 40, 142.

53. Juhlin, L., Michaelson, and Zetterstrom, 0.1972. Urticaria and asthma induced by Food- and-drug additives in patients with aspirin hypersensitivity. J . Allergy Clin. Immunol. 50, 92.

54. Simon, R. A. 1986. Adverse reactions to food additives. N. Engl. Reg. Allergy Proc. 7,533.


55. Kennedy, R. A. 1979. Placebo-controlled studies of human reaction to oral monosodium L-glutamate. In “Glutamic Acid: Advances in Biochemistry and Physiology” (L. J. Filer, Jr., S., Garattini, M. R. Kare, W. A.Reynolds, and R. J. Wurtman, eds.), p. 363. Raven Press, New York.

56. Tarasoff, L., and Kelly, M. F., 1993. Monosodium L-glutamate: A double-blind study and review. Food Chem. Toxicol. 31, 1019-1035.

57. King, D. S. 1989. Psychological and behavioral effects of food and chemical exposure in sensitive individuals. Nutr. Health 3, 137.

58. Rippere, V. 1984. Some varieties of food tolerance in psychiatric patients. An overview. Nutr. Health 3, 125.

59. Rix, K. J. B., Pearson, D. J., and Bentley, S. J. 1984. A psychiatric study of patients with supposed food allergy. Br. J . Psych. 145, 121.

60. Pearson, D. J., Rix, K. J. B., and Bentley, S. J . 1983. Food allergy: How much in the mind? Lancet i, 1259.

61. Selner, J. C., and Staudenmayer, H. 1986. The relationship of the environment and food to allergic and psychiatric illness. In “Psychobiological Aspects of Allergic Disorders” (S. H. Young, J. M. Rubin, and H. R. Daman, eds.), p. 102. Praeger Publ., Westport, CT.

62. Taylor, S. L., Higley, N. A,, and Bush, R. K. 1986. Sulfites in foods: Uses, analytical methods, residues, fate, exposure assessment, metabolism, toxicity, and hypersensitivity. Adv. Food Res. 30, 1.

63. Delohery, J., Simmul, R., Castle, W. D., and Allen, D. 1984. The relationship of inhaled sulfur dioxide reactivity to ingested metabisulfite sensitivity in patients with asthma. Am. Rev. Resp. Dis. 130, 1027.

64. Taylor, S. L., Bush, R. K., Holden, K., Nordlee, J. A., and Busse, W. W. 1986. Food challenges in sulfite-sensitive asthmatics. J . Allergy Clin. Zmmunol. 77, 159.

65. Burdock, G. A. (Ed.). 1994. “Fenaroli’s Handbook of Flavor Ingredients,” 3rd ed. CRC Press, Boca Raton, FL.

66. Schutte, L. 1975. Presursors of sulfur-containing flavor compounds. In “Fenaroli’s Hand- book of Flavor Ingredients” (T. E. Furia and N. Bellanca, eds.), 2nd ed., Vol. 1, pp. 132-183. CRC Press, Boca Raton, FL.

67. Majeti, V. A,, and Suskind, R. R., Mechanism of cinnarnaldehyde sensitization. Contact Derm. 3, 16-18.

68. Flavor and Extract Manufacturers’ Association of the United States. (1975). “Results of Second FEMA Survey of Flavoring Ingredients Average Maximum Use Levels,’’ April.

69. Hall, R. L., and Oser, B. L. 1965. Recent progress in the consideration of flavoring ingredients under the food additives amendment Il l . GRAS substances. Food Techno/. 19, 253.

70. Directive 79/112/EEC as amended by 86/1971EEC, 89/395/EEC, 93/102/EEC, 94/54/EEC. 71. Anderson, K. E. 1978. Contact allergy to toothpaste flavors. Contact Derm 4, 195. 72. Loveman, A. B. 1938. Stomatitis venenata. Report of a case of sensitivity of the mucous

membranes and the skin to oil of anise. Arch. Dermatol. Syphilol. 37, 70-81. 73. Kaaber, S., Thulin, H., and Nielsen, E. Skin sensitivity to denture base materials in the

burning mouth syndrome. Contact Derm. 5, 90-96. 74. Fisher, A. A. 1975. Dermatitis due to cinnamon and cinnamic aldehyde. Curr. Contact

News 16,383-384. 75. Miller, J. 1941. Che s from sensitivity to oil of cinnamon present in bubble gum.

J . Am. Med. Associ. 116, 131-132. 76. Kirton, V., and Wilkinson, D. S. 1975. Sensitivity to cinnamic aldehyde in a toothpaste.

2. Further studies. Contact Derm. 1, 77-80.


77. Miller, R. L., Gould, A. R., and Berstein, M. L. 1992. Cinnamon-induced stomatitis venenata. Clinical and characteristic histopathologic features. Oral Surg. Oral Med. Pathol. 76, 708-716.

78. Drake, T. E.. and Maibach. H. I., 1976. Allergic contact dermatitis and stomatitis caused by a cinnamic aldehyde-flavored toothpaste. Arch Dermatof. 112, 202-203.

79. Magnusson, B., and Wilkinson, D. S. 1975. Cinnamic aldehyde in toothpaste. 1. Clinical aspects and patch tests. Contact Derm. 1, 70-76.

80. Silvers, S. H. 1939. Stomatitis and dermatitis venanata with purpura, resulting from oil of cloves and oil of cassia. Dent. Items Interest 61, 649-651.

81. Laubach, J . L., Malkinson. F. D., and Ringrose, E. J., 1953. Cheilitis caused by cinnamon (cassia) oil in toothpaste. JAMA 152, 404-405.

82. Millard. L. 1973. Acute contact sensitivity to a new toothpaste. J. Dent. I, 168-170. 83. Barkin, M. E.. Boyd, J. P., and Cohen, S. 1984. Acute allergic reaction to eugenol. Oral

Surg. 57, 441 -442. 84. Koch, G., Magnusson, B., Nobreus, N., Nyquist, G., and Soderholm, G. 1973. Contact

allergy to medicaments and materials used in dentistry. IV. Sensitizing effect of eugenol? Colophony in surgical dressing. Odont. Rev. 24, 109-1 14.

85. Papa, C. M., and Shelley, W. B. 1964. Menthol hypersensitivity. Diagnostic basophil response in a patient with chronic urticaria, flushing, and headaches. J . Am. Med. Assoc. 189, S46-548.

86. Subiza, J.. Subiza, J. L., Valdivieso, R., Escribno, P. M., Garcia, R., Jerez, M., and Subiza, E. 1992. Toothpaste flavor-induced asthma. J . Allergy Clin. Immunol. 90, 1004-1006.

87. McGowan. E. M. 1966. Menthol urticaria. Arch. Derm. 94, 62-63. 88. Camarasa. G., and Alomar, A. 1978. Menthol dermatitis from cigarettes. Contact Derm.

89. Forsbeck, M., and Skog, E. 1977. Immediate reactions to patch tests with balsam of

90. Hjorth, N. 1961. Eczematous allergy to balsams. Acra Dermatovenereol. 41, 1-216. 91. Veien, N. K., Hattel. T., Justesen. O., and Norholm, A. 1983. Oral challenge with balsam

of Peru in patients with eczema: A preliminary study. Contact Derm. 9, 75-76. 92. Veien, N. K., Hattel. T., Justesen, O., and Norholm, N. 1985. Oral challenge with balsam

of Peru. Contact Derm. 12, 104-107. 93. Veien, N. K., Hattel, T., Justesen, 0.. and Norholm, N. 1985. Reduction of intake of

balsams in patients sensitive to balsam of Peru. Contact Derm. 3, 270-273. 94. Fisher, A. A. 1966. Systemic eczematous “contact-type’’ dermatitis medicamentosa. Ann.

Allergy 24, 415. 95. Gem, J. E., Yang, E., Evrard, H. M., and Sampson, H. A. 1991. Allergic reactions to

milk-contaminated %on-dairy” products. N . Engl. J . Med. 324, 976-979. 96. St. Vincent, J. C. M., and Watson, W. T. A. 1994. Unsuspected source of cow’s milk

protein in food. J. Am. Med. Assoc. 93, 209. 97. Food Allergy Network and Kellogg’s Company. March 2. 1993. “Product Alert for

Kellogg’s@ Rice Krispies Treats Cereal.” 98. Food Allergy Network and Fantastic Foods. February 19, 1996. “Inadvertant Use of

Peanut Flour in Several Fantastic Foods Products.” 99. Gonzalez de la Pena, M. A., Menendez-Arias, L., Monsalve, R. I., and Rodriguez, R.

1991. Isolation and characterization of a major allergen from oriental mustard seeds, Brajl. Int. Arch. Appl. Immunol. 96, 263-270.

100. Kidd, J. M., Cohens, S. H., Sosman. A. J., and Fink, J . N. 1983. Food-dependent exercise induced anaphylaxis. J. Allergy Clin. Immunol. 71,407-41 1.

4, 169-170.

Peru. Contact Derm. 3,201-205.


101. Stricker, W. E., Anorve-Lopez, E., and Reed, G. E. 1986. Food skin testing in patients

102. Malet, A,, Valero, A., Lluch, M., Bescos, M., Amat, P., and Serra, E. 1993. Hypersensitiv-

103. Niinimaki, A. 1984. Delayed-type allergy to spices. Contact Derm. 11, 34-40. 104. Stager, J., Wuthrich, B., and Johansson, S. G. 0. 1991. Spice allergy in celery-sensitive

patients. Allergy 46, 475-478. 105. van Toorenenbergen, A. W., Huijskes-Heins, M. I. E., Leijinse, B., and Diegas, P. H.

1988. Imrnunoblot analysis of IgE-binding antigens in spices. Int. Arch. Allergy Appl. Imrnunol. 86, 117-120.

106. Niinimaki, A., Bjorksten. F. Puukka, M., Tolonen, K., and Hannuksela, M. 1989. Spice allergy: Results of skin prick tests and RAST with spice extracts. Allergy 44, 60-65.

107. Bock, S. A. 1993. Anaphylaxis to coriander: A sleuthing story. J . Allergy Clin. Imrnunol.

108. Ortolani, C., Ispano, M., Pastorella, E. A., Ansaloni, R., and Magri, G. C. 1989. Cornpari- son of results of skin prick tests (with fresh foods and commercial food extracts) and RAST in 100 patients with oral allergy syndrome. J . Allergy Clin. Imrnunol. 83,683-690.

109. van Toorenenbergen, A. W., and Diegas, P. H. 1987. Demonstration of spice-specific IgE in patients with suspected food allergies. J . Allergy Clin. Immunol. 79, 108-113.

110. Hebling, A., Lopez, M., Schwartz, H. J., and Lehrer, S. B. 1993. Reactivity of carrot specific IgE antibodies with celery, apiaceous spices, and birch pollen. Ann. Allergy

with “idiopathic anaphylaxis.” J . Allergy Clin. Immunol. 77, 516-519.

ity to mustard seed. Allergy 48, 62-63.

91, 1232-1233.

70,495-499. 111. Hutchinson, J. 1993. An eruption caused by vanilla. Arch. Surg. 4, 49-50. 112. Rudzki, E., and Grzywa, Z. 1976. Sensitizing and irritating properties of star anise oil.

Contact Derm. 2,305-308. 113. Bedello, P. G., Goitre, M., and Cane, D. 1982. Contact dermatitis and flare from food

flavoring agents. Contact Derm. 8, 143-144. 114. Bruckner-Tuderman, L., Konig, A., and Schnyder, U. W. 1992. Patch test results of

the Dermatology Clinic Zurich in 1989: Personal computer-aided statistical evaluation. Dermatology 184,29-33.

115. Bruynzeel, D. P., van den Hoogenband, H. M., and Koedijk, F. 1984. Purpuric vasculitis- like eruption in a patient sensitive to balsam of Peru. Contact Derrn. 11, 207-209.

116. Cronin, E. 1992. Contact dermatitis. XVII. Reactions to contact allergens given orally or systemically. Br. J. Derm. 86, 104-107.

117. Dooms-Goossens, A., Dubelloy, R., and Degreef, H. 1990. Contact and systemic contact- type dermatitis to spices. Dermatol. Clin. 8, 89-93.

118. Etain, C. 1972. Reactions to contact allergens given orally or systemically. Br. J. Derm.

119. Hausen, B. M., Evers, P., Stuwe, H. T., Konig, W. A., and Wollenweber, E. 1992. Propolis allergy. IV. Studies with further sensitizers from propolis and constituents common to propolis, poplar buds and balsam of Peru. Contact Derm. 26,34-44.

120. Hjorth, N. 1982. Skin reactions to balsams and perfumes. Clin. Exp. Dermatol. 7, 1-9. 121. Liden, S. , Beckman, L., Cedergren, B., Goransson, K., and Nyquist, H. 1978. HLA

antigens in allergic contact dermatitis. Acta Dermatovenereol. (Stockholm) 79, 53-55. 122. Liden, S., Beckman, L., Cedergren, B., Groth, 0.. Goransson, K., and Wahlby, L. 1981.

Lack of association between allergic contact dermatitis and HLA antigens of the A and B series. Acta Dermatovenereol. (Stockholm) 61, 155-156.

123. Marzulli, F. N., and Maibach, H. I. 1976. Contact allergy: Predictive testing in man. Contact Derm. 2,l-17

86, 104-107.


124. Meneghini, C . L., and Angelini, G. 1982. Contact dermatitis from pyrolnitrin. Contact

125. Mitchell, J. C. 1978. Patch testing with some components of balsam of Peru. Contact

126. Nethercott, J. R. et al. 1991. Patch testing with a routine screening tray in North America.

127. Opdyke. D. L. J. 1974. Monographs on raw materials. Food Cosmet. Toxicol. 12,953-954. 128. Rudner, E. J., Clendenning, Epstein, E., Fisher, A. A., Jillson, 0. F., Jordan, W. P.,

Kanof, N.. Larsen. W., Maibach, H., Mitchell, J. C.. O’Quinn, S. E., Schorr, W. F., and Sulzberger, M. B. 1975. The frequency of contact sensitivity in North America 1972-1974. Contact Derm. 1, 277-280.

129. Rudzki, E., and Grzywa, Z. 1976. Immediate reactions to balsam of Peru, cassia oil and ethyl vanillin. Contact Derm. 2, 360-361.

130. Rudzki, E., and Grzywa, Z. 1978. Immediate reactions to balsam of Peru, cassia oil, and ethyl vanillin. Contact Derm. 4, 374.

131. Spencer, L. V.. and Fowler, J. F. 1988. “Thin mint” cookie dermatitis. Contact Derm.

132. Temesvari, E., Soos, G.. Podanyi, B., Kovacs, I., and Nemeth. I. 1978. Contact urticaria

133. Young, F. 1987. Sensitivity to propolis. Contact Derrn. 16,49-50. 134. Malten, K. E. 1980. Four bakers showing positive patch-tests to a number of fragrance

materials, which can’ also be used as flavors. Acta Dermatovenereol. 85, 117-121. 135. Kirton, V. 1978. Contact urticaria and cinnamic aldehyde. Contact Derm. 4, 374. 136. Rietschel, R. L. 1978. Contact urticaria from synthetic cassia oil and sorbic acid limited

to the face. Contact Derm. 4, 347-349. 137. Karlberg, A. T., Shao, L. P., Nilsson, U., Gafvert., E., and Nilsson, J. L. G. 1994.

Hydroperoxides in oxidized d-limonene identified as potent contact allergens. Arch. Dermatol. Res. 286, 97-103.

Derm. 8, 55.

Derm. 4, 391-392.

1987 through 1989. IV. Occupation and response. Am. J . Contact Derm. 2, 247-254.

18, 185-186.

provoked by balsam of Peru. Contact Derm. 4, 65-68.

138. Chrisman, B. B. 1978. Menthol and dermatitis. Arch. Dermatol. 114, 286. 139. Highstein, B.. and Zeligman, I. 1951. Nonthrombocytopenic purpura caused by mentho-

lated cigarettes. J. Am. Med. Assoc. 146, 816. 140. Calnan, C. D.. Cronin, E., and Rycroft, R. J. G. 1981. Allergy to phenyl salicylate. Contact

Derm. 7 , 208-211. 141. Sonnex, T. S., and Rycroft, R. J. G. 1986. Dermatitis from phenyl salicylate in safety

spectacle frames. Contact Derm. 14, 268-270. 142. Federation of American Societies for Experimental Biology (FASEB). 1995. “Analysis

of Adverse Reactions to Monosodium Glutamate (MSG).” Prepared for Food and Drug Administration, Washington, D.C.

143. Schwartstein, R. M. 1992. Pulmonary reactions to monosodium glutamate. Pediatr. Al- lergy Immunol. 3, 228-232.

144. Lehrer, S. B., and O’Neil, C. 13. 1992. Occupational reactions in the food industry. Food Technol. 46, 153-155.

145. Niinimaki, A. 1987. Scratch-chamber tests in food handler dermatitis. Contact Derm.

146. Lybarger, J. A.. Gallagher, J . S., Pulver, D. W., Litwin, A,, Brooks, S., and Bernstein, I. L. 1982. Occupational asthma induced by inhalation and ingestion of garlic. J. Allergy Clin. Immunol. 69, 448-454.

147. Falleroni, K. A. E., Zeiss, C. R., and Levitz, D. 1981. Occupational asthma secondary to inhalation of garlic dust. J . Allergy Clin. Immunol. 68, 156-160.

16, 11-20.


148. Yunginger, J. W., Gauerke, M. B., Jones, R. T., Dahlberg, M. J. E., and Ackerman, S. J. 1983. Use of radioimmunoassay to determine the nature, quantity, and source of allergenic contamination of sunflower butter. J. Food Prot. 46, 625-628.

149. Nordlee, J. A., and Taylor, S. L. 1995. Immunological analysis of food allergens and other food proteins. Food Technol. 49(2), 129-132.

150. Bock, S. A., Sampson, H. A., Atkins, F. M., Zeiger, R. S., Lehrer, S., Sachs, M., Bush, R. K., and Metcalfe, D. D. 1988. Double-blind, placebo-controlled food challenges (DBPCFC) as an office procedure: A manual. J . Allergy Clin. Immunol. 82, 986-997.

151. Bock, S. A,, Buckley, J., Holst, A., and May, C. D. 1977. Proper use of skin tests with food extracts in diagnosis of hypersensitivity to food in children. Clin. Allergy 7,375-383.

ADVANCES IN FOOD A N D NUTRITION RESEAKCH, VOL. 42

THE USE OF AMINO ACID SEQUENCE ALIGNMENTS TO ASSESS POTENTIAL ALLERGENCITY OF PROTEINS USED

IN GENETICALLY MODIFIED FOODS

STEVEN M. GENDEL

Biotechnology Studies Branch Food and Drug Administration

National Center for Food Safety and Technology Summit-Argo, Illinois 60501

I. Introduction 11. Methods

111. Results A. B. C. D. E. Extending Local Alignments

IV. Discussion References

Global Alignment-GAP with Control Sequences Local Alignment-FASTA with Control Sequences Local Alignment-FASTA with Transgene Proteins Local Alignment-BLASTA with Transgene Proteins

I. INTRODUCTION

Food allergies occur in approximately 5% of children and 2% of adults (Sampson, 1992; Hefle 1996). Although allergic reactions to foods can range from mild to life threatening, sensitive individuals experience extreme reactions when exposed to small amounts of allergen (Sampson, 1992). The only reliable way to deal with food allergy is to avoid the offending food; therefore, it is important that allergic individuals be aware of the content of all foods consumed. An allergic individual must avoid both whole foods that cause reactions (for example milk) and mixtures that contain components of the allergenic food (such as casein).

The production of transgenic foods raises two major concerns regarding allergenicity: the transfer of allergenic proteins to new hosts, and the poten-

45

46 STEVEN M. GENDEL

tial for proteins from organisms that have not previously been part of the food supply to become allergens (FDA, 1992; Fuchs and Astwood, 1996). If a transferred protein is from a “commonly allergenic” donor, it may be possible to obtain some measure of the allergenicity of the protein in a new host by testing with sera from allergic individuals. This has been done in one case in which a Brazil nut protein was transferred to soybean and retained allergenicity (Nordlee et af., 1996). However, if a transferred protein is derived from a “less commonly allergenic” source, this approach to safety assessment is not practical because of the difficulty in obtaining a sufficient number of allergic sera. Similarly, there are no direct biochemical tests for potential allergenicity that can be used to assess new proteins in the food supply.

Food allergens, and allergens in general, are a diverse group of proteins. Food allergen proteins have been described as generally being between 10 and 70 kDa, highly expressed, possibly glycosolated, and resistant to degradation (Hefle, 1996). However, there are no data to show that any of these properties are necessary or sufficient to cause sensitization or an allergic reaction in a previously sensitized individual.

Several publications, including a recent report from the International Food Biotechnology Council, have suggested that the potential allergenicity of a transferred protein can be assessed by examining a set of physiochemi- cal properties (including stability to digestion, prevalence, and stability to processing) and by comparing the sequence of the protein to those of known allergens. (Astwood and Fuchs, 1996; Fuchs and Astwood, 1996; Metcalfe et al., 1996). The sequence-based component of an allergenicity assessment can be carried out by aligning a query sequence with each member of a database of allergen sequences. A negative result, the failure to find significant sequence similarity between the query sequence and any known allergen, can be considered an indication of low probability of potential allergenicity. Such comparisons have, in fact, been used in the safety assessment process for several transgenic foods, although little specific information has been published on how these comparisons were performed (Astwood and Fuchs, 1996; Fuchs and Astwood, 1996).

Sequence alignments of this type can be carried out using programs that implement a number of different algorithms (Gribskov and Devereux, 1991). Most of these algorithms were developed primarily to detect evolutionary or functional relationships. However, allergenicity assessment involves the detection of short regions of structural similarity that are not evolutionarily or functionally related. Therefore, some assumptions that are built into the alignment programs may not be relevant in this context. In addition, these programs often have multiple user-definable input parameters that affect their functioning. The data sets used for alignment also

ALLERGAN SEQUENCE ALIGNMENTS 47

affect the results obtained and the ease with which significant results can be recognized.

To determine the best method for utilizing sequence information in assessing the potential allergenicity of proteins used in new food varieties, I compared the results obtained by using different sequence alignment strategies with several test sequences and two allergen sequence databases. The test sequences included both synthetic control sequences and sequences for proteins that are currently being used in transgenic foods. The results of these tests showed that local alignment algorithms are more appropriate for use in this context than global allignment algorithms, use of the proper scoring matrix is necessary to reliably locate significant matches, and the lack of reliable criteria for defining an allergenic epitope makes it difficult to assess the biological significance of the matches that are identified.

I I . METHODS

All of the sequence analysis programs used were part of Version 8 of the GCG package (Genetics Computer Group, Inc., Madison, WI) running on an AXP 2100 computer (Digital Equipment Corp., Maynard, MA) under a VMS 6.1 operating system. The individual programs and parameters used are described in detail under Results.

Construction of the two allergen sequence databases has been described (Gendel, chapter 3 of this volume). Briefly, accessions for food and nonfood allergen sequences were identified in three large reference databases. These sequences were compared both within and between the reference databases to identify a complete set of accessions that includes all available allergen sequence variants. Because it is not known whether common sequence properties are involved in the allergenicity of food and nonfood allergens, each group of sequences was treated as a separate database. The overall composition of the allergen databases is described in Table I. The food allergen sequence database does not include wheat gluten proteins because it is not clear whether food allergies and gluten-associated enteropathies share a common etiology (O’Mahony and Ferguson, 1991; Metcalfe, 1992).

All known food allergens are proteins (Taylor, 1992; Hefle, 1996). There- fore, amino acid sequence comparisons should be used for assessing potential allergenicity. Direct comparison of amino acid sequences avoids three problems with nucleic sequence comparison that could obscure significant matches. First, because the genetic code is degenerate, proteins with identical amino acid sequences can have significantly different coding sequences. Second, because all known food allergen sequences originate from eukaryotes, the genomic sequences may contain introns. Although it may be

48 STEVEN M. GENDEL

TABLE I

DATABASES (GENDEL, CHAFTER 3 OF THIS VOLUME)

SUMMARY OF THE CONTENTS OF THE ALLERGEN

Food allergen database Unique sequences GenPept accessions SwissProt accessions PIR accessions Species Proteins

Unique sequences GenPept accessions SwissProt accessions PIR accessions Species Proteins

Nonfood allergen database

138 89 53 90 15 44

218 118 105 162 65

142

possible in many cases to identify and use only the coding regions of the nucleic acid sequences, this can be much more complex than simply using the translated amino acid sequence. Third, some allergen sequences have been obtained from cDNA while others represent genomic clones. Again, this means that the possible presence of introns needs to be considered when making comparisons at the nucleic acid level.

Construction of three positive control sequences was carried out as follows. The sequence of a known food allergen, the 113-amino-acid cod parvalbumin protein known as allergen M or Gad cl (SwissProt accession A94236), was randomized by using the program SHUFFLE. This produces a random sequence with the same amino acid composition as the original sequence. The 10 amino acids numbered 51-60 in the original sequence were used to replace amino acids 11-20,51-60, or 101-110 in the shuffled sequence. This produced sequences with regions located near the N- terminus (control sequence Cl). the middle (control sequence C2), or the C-terminus (control sequence C3) that are identical to part of a known food allergen.

A set of transgene test sequences, proteins currently being used in transgenic plants, were identified from a variety of sources, including direct searching of database annotation, regulatory documents from both the Food and Drug Administration and the U.S. Department of Agriculture, and literature sources (Table 11). Sequence testing was carried out using the accessions listed in Table 11; the actual transgenic plants may express

TABLE I1 ACCESSIONS OF GENES USED TO CONSTRUCT TRANSGENIC FOOD PLANTS

Target Gene Original source organism organism(s) GenPept accession SwissProt

ACC deaminase Bacillus toxin CrylA(b)

Bacillus toxin CrylA(c) Bacillus toxin Cry3A Bacillus toxin

Neomycin phosphotransferase I1 (NW)

Nitrilase (BXN) Phosphinothricin

Phosphinothricin

Thioesterase

ZYMV coat protein MWV coat protein

acetyltransferase 1

acetyltransferase 2

Pseudomonas 6G5 Bacillus thuringiensis kurstaku

Bacillus thuringiensis kurstaki Bacillus thuringiensis kurstaki Bacillus thuringiensis

tenebrionis Escherichia coli

berliner

Klebsiella pneumoniae Streptomyces

viridochromogenes Streptomyces hygroscopicus

Umbellularia californica

Zucchini yellow mosaic virus Watermelon mosaic virus

(California bay)

Tomato Tomato

Cotton Potato Patato

Tomato, cotton, potato, etc.

Cotton Corn

Corn

Brassica

Squash Squash

M80882 A09398”

X54159 X70979 M30503

V00618

503196 M22827

X 17220

M94159

M35095’ D0053Sd

‘ 1. Fuchs and Astwood (1996); 2. FDA (1994); 3. Klee et al. (1991); 4. Noteborn and Kuiper (1995); (1995); 7. USDA Petition 94-357-01; 8. USDA Petition 94-257-01; 9. Voelker et al. (1992); 10. USDA 12. USDA Petition 92-204-01.

” A09398, from a from Ciba-Ceigy patent, is 100% identical at the nucleotide level to M15271 and (the accessions translated to create PO6527 in SwissProt and JD0002 in PIR. The amino acid sequences in the 100% identical. ‘ The coat protein is made from a polyprotein that is processed proteolytically. The actual coat protein

The coat protein is made from a polyprotein that is processed proteolytically. The actual coat protein

50 STEVEN M. GENDEL

proteins with slightly different amino acid sequences if the gene used origi- nated from a different strain or if it was modified during construction of the transgenic plant.

Note. To avoid any implication of evolutionary or functional relationship between the sequences involved in this work, all sequence comparisons are referred to in terms of sequence identity or similarity, rather than homology. Identical sequences have the same amino acid sequence over the region involved, while similar sequences have some amino acid mismatch. Se- quence homology is a consequence of evolutionary divergence from a common ancestor.

Ill. RESULTS

A. GLOBAL ALIGNMENT-GAP WITH CONTROL SEQUENCES

In general, sequence alignment algorithms can be divided into global algorithms that optimize alignments across the entire length of the sequences involved and local algorithms that attempt to optimize alignments only with regions of high similarity (Gribskov and Devereux, 1991). Global alignment algorithms are of greatest utility when the sequences involved are related. Allergenicity assessment involves sequence alignments between proteins that are not evolutionarily related. Therefore, it is likely that local alignment will be more useful. To confirm this, the three control sequences were each aligned with the unmodified Gad c l sequence by using the GCG implementation of the Needleman and Wunsch algorithm in the program GAP (Needle- man and Wunsch, 1970). The program was able to find the correct 10-amino- acid match only for the control sequence that contained the sequence identity in the middle, that is, in the original position (Fig. 1). Because there is no reason to expect regions of potential allergenic cross-reactivity to be located in the same region of different allergens, global alignment algorithms appear unsuitable for assessing potential allergenicity.

B. LOCAL ALIGNMENT-FASTA WITH CONTROL SEQUENCES

The most widely used local alignment program is FASTA, developed by Pearson and Lipman (Pearson and Lipman, 1988; Pearson, 1990). FASTA has several user-definable parameters that can be altered depending on the nature of the search being conducted and that significantly affect the matches obtained. A series of FASTA searches was carried out using the

ALLERGAN SEQUENCE ALIGNMENTS

A. Global Alignment of Control Sequence C1 with Gad cl.

1 . . . AFLDIKER- ...- KAKEGGEKWSFKGFGADFDAGGAE 44

I I l l I l l I I I I 1 AFKGILSNADIKAAEAACFKEGSFDE . . . . . . . . . . . DGNAKVGLDAFS 39

45 ADSEDDDLFKDGADKDKLCAALEEFL.ALDIGLFTFSWKEDTKDGF1 93

I I I I I I I I I I / 40 ADELKKLFKIAPEDXEGFIE EDELKLFLIAFAADLRALT..DAETKAFLK 87

94 DDLAKALGKVFASIAELEAI . . . . . . . . 113 I I I t

88 AGDSDGDGKI..GVDEFGALVDKWGAKG 113

B. Global Alignment of Control Sequence C2 with Gad cl.

1 AFLDIKERKAVIFLAANGGKKAKEGGEKWSFKGFGADFDAGGAEADSEDD 50

I I I I I I I 1 1 AFKGILSNADIKAAEAACFKEGSFDEDGFYAKVGLDAFSADELKKLFKIA 50

51 PEDXEGFIEE KLCAAEALEEFL.ALDIGLFTFSVYKEDTKDGF1DDLAK.A 99

I I I I I I I I I I I I I I I I I 51 D g D K E O F I E E . . . . . DELKLFLIAFAADLRALT..DAETKAFLKAGDSDG 93

100 LGKVFASIAELEAI . . . . . . . . 113 I I I I

94 DGKI..GVDEFGALVDKWGAKG 113

C. Global Alignment of Control Sequence C3 with Gad cl.

1 AFLDIKERKAVIFLAANGGKKAKEGGEKWSFKGFGADFDAGGAEADSEDD 50

I I I l l I I I I I I I 1 . . . . . . . . . AFKGILSNADI KAAEAA... CFK..EGSFDEDGFYAKVGLD 36

51 DLFKDGADKDKLCAAEALEEFL . . . ALDIGLFTFSVYKEDTKDGFIDDLA 97 I I I I I I I I I I

37 AFSADELKKLFKIAPEDKEGFm DELKLFLIAFAADLRALTDAETKAFL 86

98 KALDBDREGFIEEEAI . . . . . . . . . . . 113 I l l 1 I I

87 KAGDSDGDGKIGVDEFGALVDKWGAKG 113

FIG. 1. Alignment of the three control sequences with the original Gad cl sequence using the Needleman and Wunsch (1970) global alignment algorithm. The 10-amino-acid match present in each control sequence is underlined. The control sequence is the top sequence in each case: the Gad cl sequence is on the bottom.

52 STEVEN M. GENDEL

control sequences and allergen databases to determine the optimal parameters for allergenicity assessment.

To confirm that FASTA could locate short regions of sequence identity regardless of location, each of the control sequences described above was aligned with the unmodified Gad cl sequence by using FASTA (Fig. 2). FASTA located the correct alignment in all three cases.

FASTA can compare a query sequence to all members of a sequence database (Pearson, 1990). Therefore, there are two possible approaches to using FASTA to assess potential allergenicity. First, the query sequence could be aligned with all sequences in a large reference database, such as GenPept, and the resulting “hits” examined to determine if any allergen sequences are present. Alternatively, the query sequence could be aligned with the allergen sequence database, and those sequences with the most extensive similarities could be examined in detail. The first approach seems inherently undesirable because no single large database contains all the relevant allergen sequences. Therefore, it would be necessary to carry out multiple searches and analyses to ensure complete testing. Also, in most cases, searches of large reference databases produce a number of high scoring alignments with sequences that are evolutionarily related to the

A. Local Alignment of Control Sequence C1 with Gad cl.

1 AFLDIKERWEDKEQFIEEKAKEGGEKWSFKGFGAD

I I I I I I I I I I I I I l l 21 EGSFDEDGFYAKVGLDAFSADELKKLFKI-----DELKLFLIAFAAD

B. Local Alignment of Control Sequence C2 with Gad cl

21 KAKEGGEKWSFKGFGADFDAGGAEADSEDDDEDKEQFI~EKLCAAEALEEF~DIGLFT

I I I I I I I I I I 21 EGSFDEDGFYAKVGLDAFSAELKKLFKI~EDKEQFIEEDEL~FLIAFAADLRALTDA

C. Local Alignment of Control Sequence C3 with Gad cl.

63 CAAEALEEFLALDIGLFTFSVYKEDTKDGFIDDLAKALDEDKEGFIEREAI

I I I I I I I I I I I I 21 E G S F D E D O F Y A K V G L D A F S A D E L K K L F K 1 A e B I ) K E Q F I E

FIG. 2. Alignment of the three control sequences with the original Gad cl sequence using the FASTA local alignment algorithm (Pearson, 1990). The 10-amino-acid match present in each control sequence is underlined. The control sequence is the top sequence in each case; the Gad cl sequence is on the bottom. Compare these alignments to those shown in Fig. 1.


query sequence. Sequences containing regions of short, but immunologically significant. sequence similarity produce alignments with low scores. Depending on the gene involved and the parameters used to determine how many matches are displayed, immunologically significant alignments may be lost. Direct searching of allergen databases ensures complete testing and, because the databases involved are relatively small, all alignments can be examined.

Each of the three control sequences was aligned with the complete PIR database, using FASTA with the default parameters to demonstrate the difficulty in using large reference databases for allergenicity assessment. Because the control sequences were derived from a randomized sequence, matches to evolutionarily related sequences were eliminated. Despite this simplification, the correct match was not found in the top 50 scores for two of the control sequences and was the 30th highest scoring sequence for the third. In contrast, when the control sequences were aligned with the food allergen database using FASTA and the default parameters, the correct sequence produced the highest score in one case and was among the top 5 scores for the other two control sequences.

FASTA, as well as all other commonly used sequence alignment programs, uses a scoring system that allows for evolutionarily conservative substitutions. The score for each alignment is determined by using a scoring matrix that assigns a numerical value to each possible pair of amino acid matches. Although several different scoring matrices exist, they all assign positive scores to some amino acid substitutions on the basis of observed evolutionary patterns (Gribskov and Devereux, 1991). As a result, an alignment of two proteins that contains a number of “evolutionarily conservative” substitutions may produce a higher score than an alignment of the same proteins that contains a shorter region of sequence identity. The alignment programs only reports the highest scoring alignment between the two sequences. Because allergenic cross-reactivity likely involves short regions of high sequence similarity, the use of an evolutionary scoring matrix may produce alignments that miss significant matches. Therefore, an alternative scoring matrix (an identity matrix) that assigned a positive score only to exact amino acid matches was constructed. When FASTA was run using this identity matrix instead of the default evolutionary matrix with the three control sequences and the food allergen database, the correct match produced the highest score in all three cases.

C. LOCAL ALIGNMENT-FASTA WITH TRANSGENE PROTEINS

The efficacy of a FASTA database alignment for assessing potential allergenicity was further tested with the 12 transgene sequences listed in

54 STEVEN M. GENDEL

Table 11. These proteins were chosen because they (or closely related proteins) have been used to produce transgenic crop plants that are being used in commerce (Astwood and Fuchs, 1996; Fuchs and Astwood, 1996). Each test sequence was aligned with each entry in both the food and nonfood allergen databases by using both the evolutionary and identity scoring matrices. In each case, the entire set of matches was scanned to locate the longest contiguous region of sequence identity (Table 111). Longer stretches of sequence identity were found by using the identity matrix than by using the evolutionary matrix in more than half of the database alignments. Figure 3 is an example of the different alignments produced between the same two proteins by using the different matrices. In this case, the alignment produced using the evolutionary matrix missed a possibly

TABLE I11 THE LONGEST CONTIGUOUS REGION OF SEQUENCE IDENTITY FOUND BY FASTA

ALIGNMENT OF EACH TRANSGENE SEQUENCE; WORDSIZE = 2

Scoring matrix

Gene Database Ident. Evol.

ACC deaminase Food 5 5 Nonfood 5 5

Bacillus toxin CrylA(b) Food 6 6 Nonfood 7 7

Bacillus toxin CrylA(c) Food 6 6 Nonfood 5 4

Bacillus toxin Cry3A Food 5 3 Nonfood 5 4

Bacillus toxin Food 5 5 Nonfood 5 4

NPT Food I 4 Nonfood 5 5

Nitralase Food 5 4 Nonfood I 5

Nonfood 5 4

Nonfood 5 5 Thioesterase Food 5 4

Nonfood 5 5

Nonfood 5 4

Nonfood 5 5

Phosphinothricin acetyltransferase 1 Food 5 4

Phosphinothricin acetyltransferase 2 Food 5 4

WMV coat protein Food 4 4

ZYMV coat protein Food 6 5


A. FASTA Alignment using the Evolutionary Matrix

210 220 230 240 250 260

NPT CGRLGVADRYQDIALATRDIAEELGGEWADRF-LVLYGIAAPDSQRIAFYRLLDEFF

Lectin P A K S T V G R V L H S T Q V R L W E K S T N R L T N F Q A Q F S F V I K S P N P D S Q I P K : : : I : : : : : : : : : : : : : : : : : I : I : : : : ::: I l l :

50 60 70 80 90 100

B. FASTA Alignment using the Identity Matrix

220 230 240 2 50 260

NPT DRYQDIALATRDIAEELGGEWADRFLVLYGIAAPDSQRIAFYRLLDEFF

Lectin STNRLTNFQAQFSFVIKSPNIGADGIAFFIAAPDSQIP~SAGGTLGLFDPQTAQNPSA I I I I I I I

70 80 90 100 110 120

FIG. 3. An example of different FASTA alignments produced using the evolutionary and identity scoring matrices. In both cases, the top sequence is part of NPT, the bottom sequence is part of a peanut lectin protein. Lines indicate exact amino acid matches; dots indicate conservative substitutions, as defined by the evolutionary scoring matrix. The seven amino acids that are aligned using the identity matrix are underlined.

significant exact match. This confirms the previous conclusion that use of the identity matrix is more appropriate for allergenicity assessment.

The distribution of lengths of sequence identity shown in Table I11 suggests that exact matches of five or fewer amino acids are likely to occur frequently by chance. This was confirmed independently by using the GCG WORDSEARCH implementation of the Wilbur and Lipman algorithm (data not shown) (Wilbur and Lipman, 1983). Matches of seven amino acids or greater were sufficiently rare (approximately the top 10% of scores) to suggest that they should be evaluated for biological relevance (see below).

One critical user-definable parameter in FASTA is wordsize. The wordsize defines the window size used in the initial steps of the sequence alignment. The default value for proteins, which was used above, is 2, meaning that only those sites containing at least two adjacent matching amino acids are used for initiating possible alignments. The complete set of database alignments was also run using a wordsize of 1 with the identity scoring matrix (data not shown). Use of the smaller wordsize did not locate any longer regions of sequence identity, and in some cases decreased the size of the longest region found.

Although it is reasonable to assume that immunologically cross-reactive sequences will have a high degree of sequence identity, it is possible that

56 STEVEN M. GENDEL

some amino acid mismatches might be tolerated. Therefore, the FASTA results used to construct Table I11 were also examined to identify regions containing 270% or 280% sequence identity over 210 amino acids (Table IV). Eight of the test sequences matched at least one allergen at the 70% identity level, while two sequences produced matches at the 80% level. In most cases, as expected, the regions of sequence similarity identified by these criteria were the same as, or only slightly larger than, the longest regions of contiguous identity located previously. It should be noted that the two proteins that had matches at the 80% level are related, and the high similarity occurs in a region that has the same sequence in both.

D. LOCAL ALIGNMENT-BLAST WITH TRANSGENE PROTEINS

The other major local alignment tool that is used for database searching is the BLAST implementation of the Altschul et al. (1990) algorithm. Al- though BLAST is usually used with large databases, it is possible to construct local BLAST databases. Both the food allergen and nonfood allergen databases were converted to the BLAST format, and the test sequences in Table I1 were aligned with each using two scoring matrices, as was done with FASTA. The results of these alignments (Table V) were similar to

TABLE IV PRESENCE OF AT LEAST ONE HIGH HOMOLOGY MATCH AT THE INDICATED LEVEL OF

IDENTITY OVER 2 I 0 AMINO ACIDS AFTER FASTA ALIGNMENT OF THE TRANSGENE

SEQUENCES WITH EACH ALLERGEN DATABASE; WORDSIZE = 2

270% identity 280% identity

Gene Food Nonfood Food Nonfood

ACC deaminase Bacillus toxin CrylA(b) Bacillus toxin CrylA(c) Bacillus toxin Cry3A Bacillus toxin NPT Nitralase Phosphinothricin acetyltransferase 1 Phosphinothricin acetyltransferase 2 Thioesterase WMV coat protein ZYMV coat protein

ALLERG AN SEQUENCE ALIGNMENTS 57

TABLE V

ALIGNMENT OF EACH TRANSGENE, SEQUENCE*

THE LONGEST CONTIGUOUS REGION OF SEQUENCE IDENTITY FOUND BY BLAST

Scoring matrix

Gene Database Ident. Evol.

ACC deaminase

Bacillus toxin CrylA(b)

Bacillus toxin CrylA(c)

Bacillus toxin Cry3A

Bacillus toxin

NPT

Nitralase

Phosphinothricin acetyltransferase 1

Phosphinothricin acetyltransferase 2

Thioesterase

WMV coat protein

ZYMV coat protein

Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood Food Nonfood

5 5 6 7 6 5 5 5 5 5 7 5 5 7 5 5 5 5 5 5 5 5 6 5

5 5 6 7 6 4 4 5 5 5 7 5 5 7 4 5 4 4 5 5 5 4 5 5

‘ I Compare with Table 111.

those obtained using FASTA. BLAST did not locate any regions of sequence identity 2 6 amino acids in length that were not found by FASTA.

E. EXTENDING LOCAL ALIGNMENTS

Although it is likely that immunological cross-reactivity requires extensive sequence similarity, absolute identity may not be necessary (for example, see Elsayed et af., 1982). Therefore, the regions of the alignments described above with highest degree of sequence identity were examined to determine if additional sequence similarity was present. This was done using FASTA to realign these regions containing the sequence identities employing either the evolutionary scoring matrix or a biochemical scoring

58 STEVEN M. GENDEL

matrix. The biochemical scoring matrix divides the amino acids into six classes based on biochemical characteristics (i.e., hydrophilic acid amino acids, hydrophilic basic amino acids, etc). (GCG Program Manual, Version 8). Alignment of members of the same class is scored as self-match; alignment of members of different classes is scored as a mismatch. The realignment was confined to a region of 15 to 20 amino acids in each case to preserve the previously located identities.

The relevant regions of each of the alignments containing 7 amino acid exact matches (Table 111) and all of the alignments containing 270% identity over 210 amino acids (Table IV) were realigned and examined for similarity. For two of the three sequence pairs that included 7 amino acid exact matches, the realignment did not indicate that significant additional similarity was present in these regions. In the third case, use of the evolutionary matrix indicated that the region contained a stretch of 17 amino acids with 82% similarity, but the matches involved were not similar biochemically.

Realignment did not indicate that additional similarity was present in 9 of the 11 regions with 270% identity over 210 amino acids. However, in two cases, significant additional similarity was present (Fig. 4). First, the

A C r y 3 A QLPPETTDEPLEKGYS

I I : : I I : I I I D-lactoglobin VRTPEVDDEALEKFDK

B C w l A (b) FLLSEFVPGAGFVLGLVDIIW

Vitellogenin I I I I I I : I : I I : :

EWFYEFVPGAAFMLGFSERMD

C C r y l A (b) FLLSEFVPGAGFVLGLVDIIW

Vitellogenin I I I I I I : I : I I : :

EWFYEFVPGAAFMLGFSERMD

FIG. 4. Realignment of sequences with regions of high similarity using the evolutionary and biochemical similarity scoring matrices. (A) Bacillus toxin gene Cry3A aligned with p- lactoglobin using the similarity matrix. (B) Bacillus toxin gene CrylA(b) aligned with vitellogenin using the evolutionary matrix. (C) Bacillus toxin gene CrylA(b) aligned with vitellogenin using the similarity matrix. The lines indicate exact matches; the dots indicate similar amino acids.


initial alignment between Cry 3A and P-lactoglobin located subsequences in which 7 of 10 amino acids matched exactly. Realignment with both the evolutionary and biochemical matrices indicated that the intercalary amino acids were similar, meaning that the alignment was 100% similar over 10 amino acids (Fig. 4). Second, the initial alignment between CrylA(b) and vitellogenin located subsequences in which 9 to 11 amino acids were identical (82% similarity). Realignment indicated that these regions contained stretches of 11 biochemically similar and 12 evolutionarily similar amino acids (100% similarity over 11 or 12 amino acids) (Fig. 4). (CrylA(c) has the same sequence as CrylA(b) in the region involved, and therefore produced the same alignments, but this was not considered an independent alignment because the proteins are closely related).

IV. DISCUSSION

In the absence of reliable biochemical indicators, assessing the potential allergenicity of new proteins in the food supply is difficult. Sequence analysis may contribute to this analysis by determining whether the test protein shares any significant sequence features with known allergens. Three conditions must be met for this analysis to be meaningful. First, the database of allergen sequences used must be extensive enough to encompass all relevant sequence features. Second, the sequence analysis must be carried out with the proper algorithms and scoring parameters. Third, the criteria that define significant matches must be clear and technically valid.

The allergen sequence databases used here have beed described (Gendel, chapter 3 of this volume). Although they are the most extensive allergen databases available, it is clear that a large number of allergens, particularly food allergens, have not been identified or sequenced. Interestingly, a number of important sequence motifs (such as nucleic acid or cofactor binding regions) have been identified from much smaller data sets (Bairoch et al., 1996). The absence of a common motif or sequence pattern for food allergens suggests that there is no single common pattern or that secondary structural patterns that are not easily recognized from the primary structure are critical. In the absence of an appropriate motif or pattern, allergenicity assessment should be carried out by aligning each test protein to all members of the allergen databases.

These results show that allergenicity assessment should be carried out using local alignment algorithms such as those implemented in FASTA or BLAST. Because BLAST requires compilation of special format databases and index files, FASTA is easier to use with small data sets, particularly if those data sets are subject to change. The optimal strategy for locating

60 STEVEN M. GENDEL

sequence matches with a high degree of sequence identity is to carry out an initial alignment between a test sequence and each database with FASTA using a wordsize of 2 and an identity scoring matrix. All regions showing a relatively high degree of sequence identity should then be realigned using either an evolutionary or biochemical similarity scoring matrix. The results of the second set of alignments can then be evaluated for biological relevance.

Because little is known about the etiology of food allergy, the criteria that define significant matches for potential allergens are not clear. It is difficult to determine how much sequence similarity between distantly related proteins is significant in this context when some closely related proteins that share extensive homology are not allergenicity cross-reactive. The only published recommendations assert that an exact match of at least eight amino acids is necessary to raise concerns about potential allergenicity (Astwood and Fuchs, 1996; Metcalfe et al., 1996). These results show that exact matches of eight or more amino acids occur infrequently by chance, and therefore any such occurrence may be of biological significance. How- ever, it is not clear whether this is the minimum level of significant sequence similarity for potential allergenicity.

The work of Elsayed et al. (1982, 1991), Miller et al. (1996), Walsh and Howden (1991) suggests that shorter sequences may determine recognition specificity as long as they occur within peptides that provide the proper (nonsequence specific) structural context. This is consistent with studies showing that the antigen binding site may be able to accommodate only six to eight amino acids (Berzofsky and Berkower, 1993). Further, although it is clear that some amino acid residues are critical for specific binding, some conservative substitutions may not affect allergenicity. Therefore, it may be prudent to treat sequence matches with a high degree of identity that occur within regions of similarity as significant even if the identity does not extend for eight or more amino acids. For example, the similarity between CrylA(b) and vitellogenin (Fig. 4) might be sufficient to warrant additional evaluation.

These results emphasize the need for further research to define minimal food allergen epitopes, determine the effects of amino acid substitutions on allergenicity, and clarify the possible role of so-called discontinuous epitopes in food allergy. The development of techniques for the rapid production of large numbers of synthetic peptides should greatly facilitate this research, especially in systems for which well-Characterized human sera can be readily obtained. These allergen databases are currently being used to determine if potential sequence motifs can be identified for food allergens.


ACKNOWLEDGMENT

This work was partially supported by Cooperative Agreement No. FD000431 between the FDA and the National Center for Food Safety and Technology.

REFERENCES

Altschul, S., Gish, W., Miller, W., Myers, E., and Lipman, D. 1990. Basic local alignment search tool. J . Mol. Biol. 215, 403-410.

Astwood, J., and Fuchs, R. 1996. Allergenicity of foods derived from transgenic plants. Monogr. Allergy 32, 105-120.

Bairoch, A., Bucher, P., and Hofman, K. 1996. The PROSITE database: Its status in 1995. NiicI. Acids Rex 24, 189-196.

Berzofsky, J., and Berkower, I. 1993. Immunogenicity and antigen structure. I n “Fundamental Immunology” (E. Paul, ed.) 3rd ed. Raven Press, New York.

Elsayed, S., Apold, J., Holen, E., Vik, H., Florvagg, E.. and Dybendal, T. 1991. The structural requirements of epitopes with IgE binding capacity demonstrated by three major allergens from fish, egg, and tree pollen. Scand. J. Clin. Lab. Invest. 51 (Suppl. 204). 17-31.

Elsayed, S., Sornes, S., Aplod, J., Vik, H., and Florvagg, E. 1982. The immunological reactivity of the three homologous repetitive tetrapeptides in the region 41-64 of allergen M from cod. Scand. J. Immunol. 16,77-82.

FDA. 1992. Statement of policy: Food derived from new plant varities. Fed. Reg. 57,22,984- 23,005.

FDA. 1994. Secondary direct food additives permitted in food for human consumption: Food additives permitted in feed and drinking water of animals; aminoglycoside 3’- phosphotransferase 11. Fed. Reg. 59, 26,700-26,711.

Fuchs, R.. and Astwood. J. 1996. Allergenicity assessment of food derived from genetically modified plants. Food Technol. 50(2), 83-88.

Fuchs. R., Rogan, G., Kech, P., Love, S., and Lavrik, P. 1995. Safety evaluation of Colorado potato beetle-protected potatoes. In “Application of the Principles of Substantial Equiva- lence to the Safety Evaluation of Foods or Food Components Derived by Modern Biotechnology.” World Health Organization, Geneva.

Gendel, S. 1998. Sequence Databases for Assessing the Potential Allergenicity of Proteins Used in Transgenic Foods. I n “Advances in Food and Nutrition Research” (Steve L. Taylor, ed.), vol. 42, pp. 63-92. Academic Press, San Diego.

Gribskov, M., and Devereux, J. 1991. “Sequence Analysis Primer.” Stockton Press, New York. Hefle, S. 1996. The chemistry and biology of food allergens. Food Technol. 50(3). 86-92. Klee, H., Hayford, M., Kretamek, K., Barry, G., and Kishore, G. 1991. Control of ethylene

synthesis by expression of a bacterial enzyme in transgenic tomato plants. Plant Cell

Metcalfe. D. 1992. The nature and mechanisms of food allergies and related diseases. Food Technol. 46(5) , 136-1 39.

Metcalfe, D.. Astwood, J.. Townsend. R., Sampson, H.. Taylor, S., and Fuch, R. 1996. Assess- ment of the allergenic potential of foods derived from genetically engineered corp plants. Crit. Rev. Food Sci. Nutr. 36(S), S165-Sl86.

Miller, K., Bradley, A,, Fitzgerald, R., Maggi E., Morgan, M., and Wal, J. 1996. Chemical composition and structure of food constituents: Defining allergenic potential. Monogr. Allergy 32, 100-104.

3, 1187-1193.

62 STEVEN M. GENDEL

Needleman, S., and Wunsch, C. 1970. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J. Mol. Bid. 48,443-453.

Nordlee, J., Taylor, S., Townsend, J., Thomas, L., and Bush, R. 1996. Identification of a Brazil- nut allergen in transgenic soybeans. N. Engl. J. Med. 334, 688-692.

Noteborn, H., and Kuiper, H. 1995. Safety evaluation of transgenic tomatoes expressing Bt Endotoxin. In “Application of the Principles of Substantial Equivalence to the Safety Evaluation of Foods or Food Components Derived by Modern Biotechnology.” World Health Organization, Geneva.

O’Mahony, S., and Ferguson, A. 1991. Gluten-sensitive enteropathy (celiac disease). In “Food Allergy: Adverse Reactions to Food and Food Additives” (D. Metcalfe, H. Sampson, and R. Simon, eds.), pp. 186-198. Blackwell Publisher, Boston.

Pearson, W. 1990. Rapid and sensitive sequence comparision with FASTF and FASTA. Meth. Enzymol. 83, 63-98.

Pearson, W., and Lipman, D. 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85, 2444-2448.

Quemada, H. 1995. Food safety evaluation of a transgenic squash. I n “Proceeding of the OECD Workshop on Food Safety Evaluation.” Organisation for Economic Cooperation and Development, Paris.

Redenbaugh, K., Lindemann, J., and Malyj, L. 1995. Application of the principles of substantial equivalence in the safety evaluation of FLAVR SAVR tomato, BXN cotton and oil- modified rapeseed. I n “Application of the Principles of Substantial Equivalence to the Safety Evaluation of Foods or Food Components Derived by Modern Biotechnology.” World Health Organization, Geneva.

Sampson, H. 1992. Food hypersensitivity: Manifestations, diagnosis, and natural history. Food Technol. 46(5), 141-144.

Taylor, S. 1992. Chemistry and detection of food allergens. Food Tecnol. 46(5), 146-152. Voelker, T., Worrell, A., Anderson, L., Bleibaum. J., Fan, C., Hawkins, D., Radke, S., and

Davies, H. 1992. Fatty acid biosynthesis redirected to medium chains in transgenic oilseed Plants. Science 257,12-14.

Walsh, B., and Howden, M. 1991. Epitope mapping of allergens for rapid localization of continuous allergenic determinants. Meth. Enzymol203, 301-311.

Wilbur, W., and Lipman, D. 1983. Rapid similarity searches of nucleic acid and protein data banks. Proc. Natl. Acad. Sci. USA 80,726-730.


SEQUENCE DATABASES FOR ASSESSING THE POTENTIAL ALLERGENCITY OF PROTEINS USED IN

TRANSGENIC FOODS

STEVEN M. GENDEL

Biotechnology Studies Branch Food and Drug Administration

National Center for Food Safety and Technology Summit-Argo, Illinois 60501

I . Introduction 11. Methods

111. Results IV. Discussion

References

I. INTRODUCTION

The development of transgenic food plants has progressed to the point that a significant number of these plants are in agricultural use, and many more will be introduced in the next several years. In 1992, the FDA issued a Statement of Policy to clarify the regulatory status of foods derived from new plant varieties, including transgenic plants (FDA, 1992). Part of this Statement of Policy was a Guidance to Industry outlining scientific considerations for evaluating the safety and nutritional aspects of foods from new plant varieties. One of the major issues discussed in the Guidance was allergenicity .

Food allergies occur in approximately 2-10% of the population (Sampson and Metcalfe, 1991; Chandra et al., 1995). Allergic reactions to foods can range from mild itching and redness to lethal anaphylactic shock; sensitive individuals can experience severe reactions when exposed to extremely small amounts of an allergen. Because the only reliable way to deal with food allergy is to avoid the offending food, it is important that allergic

63

64 STEVEN M. GENDEL

individuals be aware of the composition of all foods consumed (Sampson, 1992; Hingley, 1993). An allergic individual must avoid both whole foods that cause reactions (for example milk) and composite foods that contain components of the allergenic food (such as casein).

The production of transgenic foods raises two major concerns regarding allergenicity (FDA, 1992). The first is the possibility that an allergenic protein could be transferred to a host where the sensitive consumer does not expect it. Coupled with this is the question of whether the protein will remain allergenic in the new host. If the transferred protein is derived from a commonly allergenic donor, it may be possible to obtain some measure of the allergenicity of the protein in the new host by testing with sera from allergic individuals. Recently, this type of testing was carried out for a transgenic soybean containing a protein derived from Brazil nut (Nordlee et al., 1996). The transferred protein retained immunologic activity when tested with sera from Brazil nut-allergic individuals. Although there is no standard protocol for conducting such testing, this example provides a model for developing procedures for assessing the allergenicity of proteins derived from allergenic foods when allergic sera are available. However, if the transferred protein is derived from an allergenic food for which well-characterized sera are not readily available, this approach to safety assessment may be very difficult to carry out.

The second concern is the possibility that a protein not previously part of the food supply will become an allergen (FDA, 1992; Fuchs and Astwood, 1996). As with proteins derived from allergenic foods for which sera are not available, there are no appropriate immunologic tests for potential allergenicity that can be used in this case.

Food allergens, and allergens in general, are a diverse group of proteins. Food allergen proteins are often described as being between 10 and 70 kDa, highly expressed, possibly glycosolated, and resistant to degradation (Taylor, 1992; Fuchs and Astwood, 1996; Hefle, 1996). However, there are no data to show that any of these properties are either necessary or sufficient to cause either sensitization or an allergic reaction in a previously sensitized individual. Therefore, in the absence of reactive human sera, the assessment of potential allergenicity for transferred proteins requires consideration of a number of properties, including the original source of the protein, stability to digestion, stability to processing and/or cooking (which may not be relevant for all allergens, such as the heat-labile allergens associated with oral allergy syndrome), level of expression in the host, and similarity to known allergens.

Because the amino acid sequences of a number of protein allergens have been determined, it has been suggested that sequence comparison can be used as a tool for assessing potential allergenicity (Astwood and Fuchs,

ALLERGAN SEQUENCE DATABASES 65

1996; Fuchs and Astwood, 1996; Metcalfe et af., 1996). For example, a recent publication by the International Food Biotechnology Council suggests criteria for using sequence comparison as one component of an allergenicity assessment for foods derived from transgenic crops (Metcalfe et af., 1996). Although several papers report that such comparisons have, in fact, been used in the safety assessment process for transgenic foods, little information has been published on how these comparisons were performed, or on the allergen data sets used (Fuchs et af., 1995; Astwood and Fuchs, 1996; Fuchs and Astwood, 1996).

Sequence comparisons are frequently used to identify functional motifs or domains within proteins. Specific functional domains, such as protease digestion sites or DNA-binding sites, can be located by comparing a test sequence to a previously defined motif or consensus sequence. Unfortu- nately, too few allergenic epitopes have been identified to permit recognition of a common motif or consensus sequence (if one exists), particularly for food allergens. Consequently, the use of sequence information for assessing potential allergenicity requires that the sequence of each test protein be compared to the sequences of all known allergens. Because minor sequence variations might have major effects on allergenicity, the value of such comparisons depends on using the most complete set of allergen sequences possible.

Therefore, two databases of allergen sequences (food allergens and nonfood allergens) were constructed using information from three large reference protein sequence databases. Allergen sequences were identified in each of the reference databases and compared to homologous sequences in each reference database to identify equivalent sequences and allelic variants. This information was used to construct nonredundant allergen sequence databases that contain all currently available sequence variants for both food and nonfood allergens. In addition, beause of possible immunologic involvement in celiac disease, also known as gluten-associated enteropathy, a third database of wheat gluten protein sequences was also constructed. These databases are available for use in assessing the potential allergenicity of proteins introduced into transgenic foods.

I t . METHODS

All of the sequence analysis programs used were part of Version 8 of the GCG Wisconsin sequence analysis package (Genetics Computer Group, Inc., Madison, WI) running on a Digital Equipment Corp. (Maynard, MA) AXP 2100 computer under the Open VMS 6.1 operating system.

66 STEVEN M. GENDEL

All known food allergens are proteins. Therefore, amino acid sequence comparisons should be used for assessing potential allergenicity. The direct comparison of amino acid sequences avoids three problems that could occur with nucleic acid sequence comparisons. First, because the genetic code is degenerate, different nucleic acid coding sequences can specify proteins with identical amino acid sequences. Second, because all known food allergen proteins originate from eukaryotes, the genomic sequences that code for these proteins contain introns. Although it may be possible to identify and use only the coding regions of these sequences, this can be much more complex than simply using the translated amino acid sequence. Third, although most allergen sequences have been obtained by nucleic acid sequencing of cDNA or genomic clones, some have been obtained by direct amino acid sequencing. Therefore, the only way to access the complete set of allergen sequences is by using amino acid sequences.

The amino acid sequences for all proteins used in this study were obtained from the following reference databases: GenPept, release 94; Protein Identi- fication Resource (PIR), release 48; and SwissProt, release 33. The PIR and SwissProt databases were supplied by GCG; the GenPept database was obtained from the National Center for Biotechnology Information via FTP. The PIR is compiled by the National Biomedical Research Foundation (Washington, D.C.) (George et al., 1996), the SwissProt database by Amos Bairoch in collaboration with the European Molecular Biology Laboratory (Bairoch and Apweiler, 1996). Both contain amino acid sequences obtained by peptide sequencing and by translation of nucleic acid sequences as well as extensive annotation. The GenPept database is produced by translation of protein-coding sequences in the GenBank database, and all GenPept accession numbers match the corresponding GenBank accession (Benson et al., 1996). Not all coding sequences in GenBank are included in GenPept due to insufficient information in the annotation. In addition, GenPept does not include any separate annotation, so GenBank was used for all activities that required access to sequence annotation.

All sequences were identified and accessed by accession numbers. Acces- sion numbers were used rather than sequence names because related sequences that are listed independently in one release of a database may be merged into a single entry in subsequent releases. Although the original entry names may be altered or lost, all accession numbers are retained in the new entry.

Searches of database annotation were carried out using the GCG LOOKUP and STRINGSEARCH functions. Sequence comparisons were carried out using the BESTFIT implementation of the Smith and Waterman algorithm for pairwise alignments and the PILEUP implementation of the


Feng and Doolittle progressive alignment method for alignment of multiple sequences (Smith and Waterman, 1981; Feng and Doolittle, 1987).

Updated versions of the databases described here will be made available on-line at http://www.iit.edu/-sgendel.

Ill. RESULTS

The use of sequence comparisons for food safety assessment is justified only if the database of allergen sequences is as complete as possible. Key- word searching of the annotation in the reference databases did not adequately identify most food allergen sequences. Table I shows the number of accessions that were found in each database by using keyword searching. Most food allergens, and some nonfood allergens, have been sequenced because they are of nutritional, enzymatic, structural, or evolutionary interest. In many cases, the sequence annotation does not indicate that they are also allergens. For example, the keyword searches failed to find known allergens present in milk and eggs. Therefore, in addition to keyword searches, allergenic proteins were identified from several literature sources (Yucginger, 1991; Taylor, 1992; Matsuda and Nakamura, 1993; King et al., 1994; Bush and Hefle, 1996; Metcalfe et al., 1996).

It is important to note that not all allergenic proteins have been characterized with the same degree of precision. In some cases, such as Ara h l from peanuts or Gad cl from codfish, the allergenic proteins in a particular food have been studied in detail. In other cases, such as casein, it is not clear whether all components of a protein family are allergenic. Further, allergenic proteins differ in clinical significance, both in terms of the number of sensitive individuals and the severity of reaction. Because the etiology of food allergy is so poorly understood, all available accessions for proteins that have been identified as food allergens were included in the database (Yunginger, 1991; Taylor, 1992; Matsuda and Nakamura, 1993; King et al., 1994; Bush and Hefle, 1996; Metcalfe et al., 1996). Therefore, any use of

TABLE I NUMBER OF ALLERGEN ACCESSIONS FOUND IN EACH

REFERENCE DATABASE BY KEYWORD SEARCHING

Database Food allergens Nonfood allergens

GenBank 28 160 PIR 32 169 SwissProt 14 110

68 STEVEN M. GENDEL

these databases should include consideration of the clinical significance and the degree of characterization for each allergen.

Each reference database was searched to locate all accessions containing sequences for each allergen protein. All the accessions for each protein within each reference database were compared to determine whether any were redundant. Redundant sequences occur within a database for several reasons, including deposit of partial or preliminary sequences and sequencing of both cDNA and genomic clones of the same gene. Only accessions that represent unique sequences within each reference database were used to construct the allergen databases. Further, all sequences for each protein were compared between databases, and the allergen databases were constructed to show which accessions in each database contain identical sequences. In some cases, sequences for known food allergens (such as Pen al) were not included because these sequences had not yet been deposited in the reference databases, so no appropriate accession number was available. These sequences will be included in future updates as they become available.

The results of these searches and comparisons were used to construct two allergen databases, food allergen sequences (Table 11) and nonfood allergen sequences (Table 111). A third database of wheat gluten sequences was also constructed (Table IV). The wheat sequences were compiled separately because the relationship between gluten-sensitive enteropathy (celiac disease) and food allergy is not clear (O’Mahony and Ferguson, 1991; Metcalfe, 1992; Hefle, 1996). All three databases are available on- line (see Methods).

The overall content of the two allergen databases is summarized in Table V. The food allergen database contains 138 unique sequences and the nonfood allergen database contains 218 unique sequences. No single reference database contains more than about 60% of the unique food allergen sequences or 75% of the nonfood allergens. In addition, no combination of two of the reference databases contains all of the sequences in either allergen database. Therefore, a complete search of all allergen sequences requires the use of accessions from all three databases.

All accessions listed on the same line in all three databases have the same amino acid sequence; accessions for the same gene with differing sequences are listed on separate lines. For example, in Table 11, the GenPept accession 500922, the SwissProt accession P01014, and the PIR accession A01244 contain identical sequences for chicken ovalbumin and can be used interchangeably. However, SwissProt accession PO1012 does not exactly match any other ovalbumin sequence.

In some cases, it was necessary to combine two or more accessions from one reference database to completely match a single accession in another

TABLE I1 FOOD ALLERGEN SEQUENCES

Species Protein Allergen name GP accession SP accession

Animals Cod

Egg (chicken) Parvalbumin

Ovomucoid Ovalbumin Ovalbumin Ovalbumin Ovalbumin Ovalbumin

Ovalbumin Ovotransferrin Ovotransferrin Ovotransferrin Lysozyme Lysozyme Lysozyme Vi tellogenin Vitellogenin Vitellogenin Apovitellenin

BSA BSA

Milk (cow)

Gad c l

Gal d l Gal d2 Gal d2 Gal d2 Gal d2 Gal d2

Gal d2 Gal d3 Gal d3 Gal d3 Gal d4 Gal d4 Gal d4

500922 V00438 V00383

VOO385 + V00386+ V00387 V00382 YO0407 X02009

500885 M10640 X61002 K02113+ x00204 M18060 J00810

M73993

PO2622

PO1005 PO1014

PO1012 PO1013

PO2789 PO0698

P27042 PO2845

PO2659

PO2769

4 0 TABLE I1 (Continued)


Shrimp

Plants Apple

P-Lactoglobulin P-Lactoglobulin P-Lactoglobulin P-Lactoglobulin a-Lactalbumin a-Lactalbumin a-S1 Casein a-S1 Casein a - S 1 Casein a-S1 Casein a-S2 Casein /3 Casein P Casein

Casein K Casein K Casein K Casein K Casein

Tropomyosin

Profilin Profilin

Barley a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylaseltrypsin inhib.

X14712 248305 KO1086 M19088 J05 147 X06366 M33123 M38641 M38658 KO1084 M16644 M15132 M55158 M16645 M36641

PO2754

PO071 1

PO2662

PO2663

PO2666

PO2668 KO1085

Met e l UO8008

Ma1 d 1 X83672 Ma1 d 1 248969

Hor v 1 X63517 Hor v 1 Hor v 1 X69937 Hor v 1

P432 1 1

P16968 P28041

Brazil nut

Celery

Kidney bean

Mustard (leaf)

Mustard (white)

a-Amylaseltrypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylaseltrypsin inhib. a-Amylase/trypsin inhib. a-Amylaseltrypsin inhib. a-Amylase/trypsin inhib. a-Amylaseltrypsin inhib. a-Amylase/trypsin inhib.

2s Albumin 2s Albumin 2s Albumin 2s Albumin

PR Protein PR Protein

2s Albumin

Amylase inhibitor Amylase inhibitor

Amylase inhibitor Amylase inhibitor Amylase inhibitor Amylase inhibitor

Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1 Hor v 1

Ber e 1 Ber e 1 Ber e 1 Ber e 1

Api g 1

Bra j 1 L

Sin a 1 Sin a 1

Sin a 1 Sin a 1 Sin a 1 Sin a 1

X69938

X69939

X13443 M15207 X59264

X54490 X54491

248967

X61365 X61364

S54101

X91798 X91799

P32936 P34951

PI 1643 PO1086 P13691 P16969

PO4403

P25985 t P25986

P80215

P15322

=!

TABLE I1 (Continued) 4 N

Species Protein Allergen name G P accession SP accession

Papaya

Peanut

Rice

Soybean

Amylase inhibitor Amylase inhibitor Amylase inhibitor

Papain

Vicilin Vicilin Agglutinin Agglutinin Agglutinin Agglutinin Arachin Arachin

a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. a-Amylase/trypsin inhib.

Glycinin AlaBx Glycinin AlaBx Glycinin A2B1 a Glycinin A2B1 a

Sin a 1 Sin a 1 Sin a 1

Ara h 1 Ara h 1

RA 1 R A 2 RA 5 R A 5 b RA 14 RA 14b RA 14c RA 16 RA 17

X918o0 X91801 X91802

M15203

L34402 L38853

S42352 U22472 U22473

D11433 D11434 D11430

D11432

D11431

M36686 X02985 Yo0398 X02806

Po0784

P43238 P43237

P02872+

PO4149 P20780

Q01884 Q01885 Q01881

Q01882

Q01883

PO4776

PO4405

Glycinin A3B4 Glycinin A3B4 Glycinin A3B4 Glycinin A384 Glycinin A3B4

Glycinin A5A4B3 Glycinin Gy3 Glycinin Gy4 Glycinin AlaBlb Glycinin A7 Glycinin Glycinin A5A4B3 Glycinin A5A4B3 Glycinin Glycinin Glycinin P-Conglycinin a-subunit P-Conglycinin @subunit j3-Congl ycinin a'-subunit P-Conglycinin a-subunit 0-Conglycinin @'-subunit P-Conglycinin

p-Conglycinin a-subunit

P-subunit

M10962 M35671

X79467

X02626 X15123 X52863 X53404A+

X86970

X17698

S44893

M13759

M26128

PO4347

PO2858 P11828

P13916

P25974

P11827

-4 w

TABLE I1 (Continued )


Lectin Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Trypsin inhibitor Oil-body associated Oil-body associated

KO0821 X80039 x64447 X64448

S45035A S45035B S45092

Gly m 1 505560 Gly m 1

PO5046

PO1071

F'22895

a References: 1. Metcalfe et al. (1996) (used for accessions listed in Table 8.1 of the reference); al. (1994); 5. Matsuda and Nakamura (1993); 6. Bush and Hefle (1996); 10. Metcalfe et al. (1996) genes listed in Table 8.1 of the reference); 11. Used for accessions located by keyword searching other published reference at this time.

Notes: Coding? = Used for those genes in which one or more database entries indicate that the not match the sequence obtained by translation of the corresponding nucleic acid, or in cases where but are not reported in the sequence. See text for details.

TABLE 111 NONFOOD ALLERGEN SEQUENCES

Species GP

Protein Allergen name accession SP accession

Alder

Alternnria alternatn

Ant (jumper)

Aspergillus Mitogillin Mitogillin

Barley

Bee (honey bee) Phospholipase

Bent grass

Bermuda grass

Birch

Aln g 1

Aldehyde dehydrogenase Alt a 2 Ribosomal protein Alt a 6

Alt a 7 Ribosomal protein Alt a 2

Myr P 1 MYrP 1

Asp f 1 Asp f 1

Hor v 9

Api m 1 Api m 2 Api m 4

Agr a 1

Cyn d 1

Bet v 1

H yaluronidase Melittin

S50892

X78227 X78222 X78225 X84216

X70256

X56176 M83781

U06640

X16709 L10710 X02007

P38948

P42041 P42037 P42058 P49148

Q07932

PO4389

PO0630 Q08169 PO1501

4 m

TABLE I11 (Continued)

Species GP


Blue grass

Candida

Profilin Profilin

Bet v l a Bet v lb Bet v lc Bet v Id Bet v l e Bet v If Bet v lg Bet v l j Bet v lk Bet v 1L Bet v lm Bet v 2 Bet v 2 Bet v 3

Poa p 1 Poa p 1 Poa p 9 Poa p 9 Poa p 9 Pao p 9 Poa p 9

Alcohol dehydrogenase Cand a Alcohol dehydrogenase Cand a Alcohol dehydrogenase Cand a

X15877 P15494 X77200 P45431 X77265 P43176 X77266 P43177 X77267 P43178 X77268 P43179 X77269 P43180 X77271 P43183 X77272 P43184 X77273 P43185 X81972 P43186

M65179 P2.5816 X79267 P43187

M38342 P22284 M38343 P22285 M38344 P22286

X81694 P43067 U15924

Cat

Cladosporium Enolase Alcohol dehydrogenase Alcohol dehydrogenase Alcohol dehydrogenase Ribosomal Ribosomal HSP

Cockroach Protease

Cow (dander)

Lipocalin European hornet

European chestnut

Filarial worm

Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1 Fel d 1

Cla h 2 Cia h 3 Cla h 3 Cla h 3 Cla h 4 Cla h 4 Cla h ? Cla h 5

Bla g 2 Bla g 4

Ves c 5.01 Ves c 5.02

Cas s 1

M77341 X62478 M74952

M74953

X78226 X78228

X78223 X77253 X81860 X78224

U28863 U40767

L39834 L42867

U03103

P30440

P30438

P30439

P42040

P40108 P42039 P42038 P40918 P42059

P35781 P35782

4 4

4 00 TABLE I11 (Continued)

Species GP


Fire ant (S. invictu) red Phospholipase Sol i 2

Sol i 3 Sol i 4

Phospholipase Sol r 2 Sol r 3

Cor a 1-5 Cor a 1-6 Cora 1-11 Cora 1-11 Cor a 1-16

Fire ant (S. richteri) black

Hazel

Hornbeam tree

Hornet (D. urenuriu)

Hornet (D. nmculutu) Phospholipase Phospholipase Hyaluronidase

C a r b l Car b 1 Car b 1 Car b 1

Do1 a 5

Do1 m 1 Do1 m 1 Do1 m 2 Do1 m 5 Do1 m 5 Do1 m 5

X70999 X71000

X70997 X70998

X66932 X66918 X66933

M98859

X66869

L34548 503601 503602

P35775 P35778 P35777

P35776 P35779

P43216

P38949

P38950

Q05108

Q06478

P49371 P10736 P10737

Lilac

Maize

Meadow velvet

Syr v 1 Syr v 1 Syr v 1

Zea m 1

Hol L 1 Hol L 1 Hol L 1

Mite (Blomia)

Mite (D. furinae) Der f 1 Der f 1 Der f 2 Der f 2 Der f 2 Der f 2

Trypsin Der f 3 Chymotrypsin Der f 6

Der f mag Der f mag29

Der p 1 Mite (D. pteronyssinus)

L14271 S44171

227084

268893

U27479 U27702

X65196 D 10447 D10448 D10449

D13961 D17676

U11695

Q07154 P33050t

P43216

P16311

Q00855

P49275 P49276 P36973 P39674

PO8176 Der p 1 M24794 + Der p 1 Der p 1 Der p 2 P49278

~

.I W

co 0 TABLE 111 (Continued)

GP Protein Allergen name accession SP accession Species

Trypsin Der p 3 Trypsin Der p 3 Amylase Der p 4

Der p 5 Der p 5 Der p 5

Chymotrypsin Der p 6 Der p 7 Der p 15

U11719 P39675

P49274 P14004

X17699 S76340

P49277 U37044 P49273 S75286 P46419

Mite (D. microceras) P16312 Der m 1

Mite (Euroglyphus)

Mite (Lepidoglyphus) Proteinase Eur m 1 P25780

L e p d l L e p d l

A r t v 2

X83876 P80384+ X83875

Mugwort

Oak Que a 1

Olive tree Ole e 1 Ole e 1 Ole e 1 Ole e 1 Ole e 1 Ole e 1 Ole e 1

PI9963

Orchard grass

Pareiteria ( P . judaica)

Parietaria (P . ofticinalis)

Pea

Penicillium notatiim

Ragweed (A. arternisiifloria)

Ragweed (A. psilostachya)

Ole e 1 Ole e 1 Ole e 1 Ole e 1 Ole e 1

Dac g 2 Dac g 3

Par j 1 Par j 1 Par j 1

Par o 1

Amb a 1.1 Amb a 1.2 Amb a 1.2 Amb a 1.3 Amb a 1.3 Amb a 1.3 Amb a 1.4 Amb a 2 Amb a 2 Amb a 3 Amb a 5

Amb p 5

s45354

X77414

X85012

X85187

S77837

Ma0558 M80559

M62961 M80560

M80562 M80561

L24465

P43217

P27759 P27760

P27761

P28744 P27762

PO0304 PO2878

P43174

m

TABLE I11 (Continued ) 03 N

Species GP


Ragweed (A. trifidu)

Reed fescue

Roundworm (Ascuris)

Roundworm (Toxicuru)

Rye

Ryegrass

Amb p 5 A m b p 5 Amb p 5 Amb p 5

Amb t 5

Fes e l a Fes e l b

Asc 1 1 Asc s 1

Sec c

Lo1 p 1 Lo1 p 1 Lo1 p 1 Lo1 p 1 Lo1 p l b Lo1 p l b L o l p 2 Lo1 p 2a Lo1 p 2b Lo1 p 3 Lo1 p 4

L24466 L24467 P43175 L24468 L24469

S39336 P10414

LO3211

M57474

M57476+

M59163

X73363

P14964

P14947

P14948

Soybean

Sugi (Japanese cedar)

Sweet vernal grass

Timothy grass

Profilin

Tomato

Wasp (P. annularis)

L o l p 5 Lo1 p 5 Lo1 p 5 Lo1 p 9 Lo1 p 11

Gly m ciml

C r y j l a C r y j l b Cry j 2 Cry j 2

Ant o 1

Phl p 1 Phl p 1 Phl p 2 Phl p 5 Phl p 5 Phl p 5 Phl p 5a Phl p 6 Phl p 11 Phl p 32K Phl p 38K

Pol a 5

L13083

U03860

D26544 D26545 D29772 D37765

X78813 227090 X75925 227083

227082 X77583

X15855

M98857

P18632

P43212

P43213

P43214

P43215 P35079

P13447

Q05109

W W

TABLE I11 (Continued)

GP Protein Allergen name accession SP accession Species

Wasp (P. exclamans)

Wasp (P. fascatus)

Wheat

Yellow jacket (V. flavopilosa)

Yellow jacket (V. germanica)

Yellow jacket (v . rnaculifrons)

Pole e 5 P35759

Pol f 5 P35780

250867

P35783

Tre a 3

Ves f 5

Ves g 5 P35784

Phospholipase Ves rn 1 Ves m 5 P35760

P35785 Yellow jacket (V. pensylvanica)

Yellow jacket (V. squamosa)

Yellow jacket (V. vidua)

Yellow jacket (V. vulgaris)

Ves p 5

Ves s 5 P35786

Ves vi 5 P35787

Phospholipase Ves v 1 Hyaluronidase Ves v 2

Ves v 5

L43561 P49369 L43562 P49370 M98858 QO5llO

" References: 1. Used for accessions identified by keyword searching of the reference databases; for accessions listed in Table 8.2 of the reference ).

TABLE IV WHEAT GLUTEN SEQUENCES

Species GP accession SP accession PIR accession

Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin G 1 i a d i n Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin Gliadin

U082S7 KO2068 KO2069 X02.538 X02539 X01130 X02540 KO3075 KO3076 M11074

M10092

M11076

M11075

M11073 X17361 X00627 M36999 M16064 M13713 M11077 M11336 M11335

PO4728 PO4726

PO4727

PO4721

PO4722

PO4723

PO4724

PO4725 P18573 PO2863 P21292 PO8453 PO6659 PO4729

PO4730

SO7361

B22364

C22364

E22364

D22364 A22364 s10015 A03354 JA0153 JS0402 A25632

SO7398

m v1

% TABLE IV (Continued)

Species GP accession SP accession PIR accession

Gliadin Gliadin Gliadin Gliadin Gliadin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin G 1 u t e n i n Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin Glutenin

M16496 M16060

XO4532

X03041 X13306 X12929 x00054 x00055 X03346 X12928 X07747 X.51759 M22209 X61009 X62588 X61026

PO8079 PO2865

PO8488 P10386 P10387 PO2861 PO2862 PO8489 P10388 P10385 P16315

A27319 PS0094 A03356

S52126 A24266 SO4325 SO4832 A03352 A03353 A24107 SO2262 SO1992 SO8683 A30843 S15720 S20853 S18733 S29176 S29177 S29178 S29179 SO6645

M22208 JN0689 JC2099

X13928 X84887 X84959

X84960 X84961 M25536 M25537 502961

P10968 PO2876 P10969 PO1084 P10846 PO1083

PO1085 P16852 P16159 P17314 P16851 P16850

Glutenin Glutenin Agglutinin SO9623 Agglutinin SO9624 Agglutinin A28401 a-Amylaseltrypsin inhib. A01323 a-Amylaseltrypsin inhib. SO501 7 a-Amylase/trypsin inhib. a-Amylase/trypsin inhib. A01322 a-Amylase/trypsin inhib. A01324 a-Amylase/trypsin inhib. D25310 a-Arnylasehypsin inhib. X16733 SO8466 a-Amylase/trypsin inhib. X17574 S10029 a-Amylaseltrypsin inhib. x55454 S13376 a-Amylaseltrypsin inhib. X17575 S10027 a-Amylase/trypsin inhib. X59791 S18241 a-Amylaseltrypsin inhib. S16920 a-Amylase/trypsin inhib. S38955 a-Arnylase/trypsin inhib. S10849 a- Amylaseltrypsin inhib. S10850

'' References: Same as in Table 11.

88 STEVEN M. GENDEL

TABLE V SUMMARY OF ALLERGEN DATABASES

Food allergens Unique sequences 138 GenPept accessions 89 SwissProt accessions 53 PIR accessions 90 Species 15 Proteins 44

Unique sequences 218 GenPept accessions 118 SwissProt accessions 105 PIR accessions 162 Species 65 Proteins 142

Nonfood allergens

reference database. Figure 1 shows an example in which two PIR accessions must be combined to completely match a single SwissProt accession. In these cases, the two (or more) accessions necessary are both listed in the database, separated by a +. In addition, in some cases an entry in one database was an exact match to part of an accession in another database.

1 50 P15322 PAGPFRIPKC RKEFQQAQHL RACQQWLHKQ AMQSGSGPSP QGPQQRPPLL SO1791 PAGPFRIPKC RKEFQQAQHL RACQQWLHKQ AMQSGSGPS. . . . . . . . . . . SO1792 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P QGPQQRPPLL

51 100

P15322 QQCCNELHQE EPLCVCPTLK GASKAVKQQV RQQLEQQGQQ GPHVISRIYQ SO1791 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SO1792 QQCCNELHQE EPLCVCPTLK GASKAVKQQV RQQLEQQGQQ GPHVISRIYQ

101 127 P15322 TATHLPKVCN IPQVSVCPFK KTMPGPS SO1791 . . . . . . . . . . . . . . . . . . . . . . . . . . . SO1792 TATHLPKVCN IPQVSVCPFK KTMPGPS

FIG. 1. An example of the combination of two accessions from one reference database (PIR SO1791 and S01792) to match a single accession from another reference database (SP P15322). These code for the mustard allergen Sin a l , an amylase inhibitor.


For example, SwissProt accession PO2872 for peanut agglutinin contains 236 amino acids that exactly match the middle of the 273 amino acids in PIR accession S24044 and GenPept accession S42352. In these cases, the entry for the short accession is followed by a + in the database.

Multiple nonidentical accessions were found for many proteins during construction of these databases. Most of these sequence differences probably reflect the presence of multiple homologous genes within a single genome and multiple alleles within a population. Because the significance of multiple alleles for food allergy is not known, it is important to include all sequences in the databases used for assessing potential allergenicity. For example, at least six accessions are necessary to include all the variant sequences of chicken ovalbumin that were found in the three reference databases (SwissProt P01014, P01012, P01013; PIR A90455; and GenPept V00383, V00382).

The annotation of the accessions used in the allergen databases revealed three other common problems. First, the sequence present in a single accession may be a consensus sequence derived from multiple sources. The conflicts between the original sequences and the consensus are indicated in the annotation, but the alternative sequences are not available for searching. This is the case for the ovalbumin sequence in GenBank accession V00383 (Fig. 2).

Second, as discussed above, sequencing of multiple clones or cDNAs may reveal the presence of sequence variants present in a single genome or population. These variants may be reported in the annotation, but not in the actual sequence data. For example, the annotation for SwissProt accession P02872, for peanut agglutinin, indicates sequence differences for a minor variant but these differences are not available for sequence searching (Fig. 3). The annotation for SwissProt accession P04405, for soybean glycinin, indicates that both conflicts and sequence variants are present.

Third, although the annotation for some entries cross-reference accessions in the other databases, these cross-references do not always point to identical sequences. For example, SwissProt accession P02666, for &casein, cites GenBank/GenPept accessions M15132 and M55158, neither of which contain exactly matching sequences. However, GenBank/GenPept accession X16645 does match exactly.

Therefore, in constructing the allergen databases, all apparently identical or homologous sequences were directly compared to ensure that all sequence variants present in the three reference databases were identified and duplicate sequences were eliminated. In addition, the annotation for each accession was scanned and the presence of any problem or conflict was indicated in the allergen database.

90 STEVEN M. GENDEL

LOCUS GGALB2 1873 bp RNA VRT 17-MAY-I995 DEFINITION Chicken messenger RNA for ovalbumin. ACCESSION V00383

FEATURES LocationlQualihers

conflict replace(35,"g") /citation=[2] /citation=[3]

/citation=[2] /citation=[3]



/note="A is G in [2]" /citation=[2]

conflict replace( I308,"a") /note="C is A in [2]" /citation=[2]

/citation=[2]

conflict replace(44,"g")

conflict replace(S0,"t")



conflict replace( 1459. ,1474,"gg")

FIG. 2. An example of part of the GenBank annotation for an accession noting conflicts between sequences from different sources (citations). Only the consensus sequence is available in GenPept.

IV. DISCUSSION

Databases of allergen-related amino acid sequences have been constructed by using accessions derived from three large reference databases. Each allergen database was designed to allow identification of a set of unique sequences that includes all accessible alleles and variants of allergenic proteins. These databases will be updated periodically as new allergens sequences become available and as new proteins are identified as allergens and the updated databases will be made available on-line (see Methods).

Sequence matching between the accessions in these databases and proteins used in the production of transgenic foods can be used as part of the safety assessment process for these new food varieties (Metcalfe et al.,


ID AC

DE 0s oc oc

FT FT FT FT FT

SQ

L E C G - M Y STANDARD; PRT; 236 AA PO2872 ;

GALACTOSE-BINDING LECTIN (AGGLUTININ) (PNA) . ARACHIS HYPOGAEA (PEANUT). EUKARYOTR; PLANTA; EMBRYOPHYTA; ANGIOSPERMAE; DICOTYLEDONEAE; FABALES; FABACEAE .

METAL 137 137 MAGNESIUM (BY SIMILARITY) VARIANT 92 92 E - > V (IN MINOR FORM). VARIANT 149 149 K - > A (IN MINOR FORM) . VARIANT 162 162 K - > I (IN MINOR FORM) . VARIRNT 212 213 LG - > RA (IN MINOR FORM) SEQUENCE 236 AA; 25189 MW; 1B4552A8 CRC32;

FIG. 3. An example of part of the SwissProt annotation for an accession noting the presence of variant sequences. The varient sequences are not available for sequence comparisons.

1996). However, it is clear that these databases do not map all of the relevant sequence space. In addition, because little is known about the etiology of food allergies, there are no generally accepted criteria for defining significant matches (Fuchs and Astwood, 1996; Metcalfe et al., 1996). Therefore, sequence matching should be combined with other considerations such as the source of the protein, stability to digestion, and stability to processing in an overall safety assessment, possibly using a scheme similar to that described in Metcalfe et al. (1996).

The problems identified during the construction of these databases highlight the importance of a thorough understanding of the structure and content of any molecular data sources used in public health-related safety assessments. For example, the degree of sequence heterogeneity between the reference databases was unexpected.

These databases are currently being used to test methods of sequence matching to determine the optimal procedure for using sequence information in allergenicity assessment. Further, as more epitopes are identified within allergenic proteins, these databases will be useful for determining whether common structural or sequence features exist in food allergen proteins, and (if they do) for deriving consensus sequences or motifs.

ACKNOWLEDGMENTS

I thank Dr. James Maryanski and Dr. Nega Beru for helpful discussion and comments, and Ted Chambers and Charles Baynard for assistance with the computing. This work was partially supported by Cooperative Agreement no. FD000431 between the FDA and the National Center for Food Safety and Technology.

92 STEVEN M. GENDEL

REFERENCES

Astwood, J., and Fuchs, R. 1996. Allergenicity of food derived from transgenic plants. Monogr.

Bairoch, A., and Apweiler, R. 1996. The SWISS-PROT protein sequence data bank and its

Benson, D., Boguski, M., Lipman, D., and Ostell, J. 1996. GenBank. Nucleic Acids Res. 24,1-5. Bush, R., and Hefle, S. 1996. Food allergens. Crif. Rev. Food Sci. Nutr. 36(S), S119-Sl63. Chandra, R., Gill, B., and Kumari, S. 1995. Food allergy and atopic disease. Clin. Rev. Allergy

FDA. 1992. Statement of policy: Foods derived from new plant varieties. Fed. Reg. 57,22,984-

Feng, D., and Doolittle, R. 1987. Progressive sequence alignment as a prerequisite to correct

Fuchs, R., and Astwood, J. 1996. Allergenicity assessment of foods derived from genetically

Fuchs, R., Re, D., Rogers, S., Hammond, B., and Padgette, S. 1995. Safety evaluation of

George, D., Barker, W., Mewes, H., Pfeiffer, F., and Tsugita, A. 1996. The PIR-international

Hefle, S. 1996. The chemistry of food allergens. Food Technol. 50(3), 86-92. Hingley, A. 1993. Food allergies: When eating is risky. FDA Consumer 27(10), 27-31. King, T., Hoffman, D., Lowenstein, H., March, D., Platts-Mills, T., and Thomas, W. 1994.

Allergen nomenclature. Int. Arch. Allergy Immunol. 10, 224-233. Matsuda, T., and Nakarnura, R. 1993. Molecular structure and immunologic properties of

food allergens. Trends Food Sci. Technol. 4, 289-293. Metcalfe, D. 1992. The nature and mechanisms of food allergies and related diseases. Food

Technol. 46(5), 136-139. Metcalfe, D., Astwood, J., Townsend, R., Sampson, H., Taylor, S., and Fuchs, R. 1996.

Assessment of the allergenic potential of foods derived from genetically engineered crop plants. Crif. Rev. Food Sci. Nutr. 36(S), S165-Sl86.

Nordlee, J., Taylor, S., Townsend, J., Thomas, L., and Bush, R. 1996. Identification of a Brazil- nut allergen in transgenic soybeans. N. Engl. J. Med. 334, 688-692.

O’Mahony, S., and Ferguson, A. 1991. Gluten-sensitive enteropathy (celiac disease) In “Food Allergy: Adverse Reactions to Foods and Food Additives” (D. Metcalfe, H. Sampson, and R. Simon, eds.), pp. 186-1 97. Blackwell Publications, Boston.

Sampson, H. 1992. Food hypersensitivity: Manifestations, diagnosis, and natural history. Food Technol. 46(5), 141-144.

Sampson, H., and Metcalfe, D. 1991. Immediate reactions to foods. In “Food Allergy: Adverse Reactions to Foods and Food Additives” (D. Metcalfe, H. Sarnpson, and R. Simon, eds.), pp. 99-112. Blackwell Publications, Boston.

Smith, T., and Waterman, M. 1981. Comparison of biosequences. Adv. Appl. Math. 2,482-489. Taylor, S. 1992. Chemistry and detection of food allergens. Food Technol. 46(5) , 146-152. Yunginger, J. 1991. Food antigens. I n “Food Allergy: Adverse Reactions to Foods and Food

Additives” (D. Metcalfe, H. Sampson, and R. Simon, eds.), pp. 36-51. Blackwell Publica- tions, Boston.

Allergy 32, 105-120.

supplement TREMBL. Nucleic Acids Res. 24,21-25.

Immunol. 13,293-314.

23,005.

phylogenetic trees. J . Mol. Evol. 25, 351-360.

modified foods. Food Technol. 50(2), 83-88.

glyphosate-tolerant soybeans. Proc. O E C D Workshop Food Safety Eval., 47-55.

protein sequence database. Nucleic Acids Res. 24, 17-20.


DESIGN OF EMULSIFICATION PEPTIDES

DAVID SHEEHAN AND KATHLEEN CAREY

Department of Biochemistry University College

Cork, Ireland

SIOBHAN O’SULLIVAN

TEAGASC Dairy Products Research Centre Moorepark, Fermoy, Co.

Cork. Ireland

I. Introduction 11. Secondary Structure of Peptides

111. Modeling of Peptide Structures A. Graphical Representations of Amphiphilicity B. Computer-Aided Design C. Design of Emulsification Peptides

1V. Synthesis of Designed Peptides V. Testing of Peptide Emulsification Properties

VI. Future Directions References

I. INTRODUCTION

The potential of newly designed peptides for emulsification is described in this chapter. Criteria important in their design, synthesis, and testing are stressed. We emphasize at the outset that there are limitations to this approach at present due to the difficulties in accurately predicting the behavior of a peptide in solution. For instance, it is known that peptide a- helices are more prone to fraying at their termini than are those of proteins (Lyu etaf., 1991). As will be clear from Section 11, structural factors responsible for peptide folding in aqueous solution are still the subject of active

93

Copyright 0 1998 by Academic Press. All rights of rcproduction iii any fiirrn reserved

94 D. SHEEHAN, S. M. O’SULLIVAN, AND K. B. CAREY

research. However, we believe that the general approach outlined here provides an interesting model for the study of emulsification in the laboratory.

Food preparations often contain complex mixtures of water-soluble and oil-soluble components. A good example of this is milk, which contains fat droplets suspended in an aqueous medium. In the absence of emulsifiers, these components may partition out into separate phases. Emulsion instability may also result in creaming or fat separation (Halling, 1981; Kachholz and Schlingmann, 1987; Parker, 1988; Das and Kinsella, 1990). These effects are important in the manufacture and storage of food products. Caseins and albumins are important contributors to the functional characteristics of foods, especially emulsification (Macritchie, 1978). Proteins contain hydrophilic and hydrophobic side chains, and thus can sometimes display “amphiphilicity,” that is, the ability to expose hydrophilic and hydrophobic faces in a single molecule. Indeed, this property is fundamental to protein structure since the presence of hydrophilic side chains on the protein surface and of hydrophobic residues in the protein interior is essential in protein solubility in acqueous solution and also important in protein folding (Creighton, 1992).

Studies on the emulsification properties of proteins in food systems in the past have, of necessity, often been somewhat empirical. This is due to the high degree of structural heterogeneity in the food and protein preparations used. Even in cases where quantitative amounts of a pure protein are available, they may have quite complex structures since they may consist of several domains or, like /3-casein, may be extensively disordered. Attempts have been made to correlate emulsification propensity in proteins with their physiocochemical characteristics such as hydrophobicity (Hague and Kito, 1983; Nakai, 1983), solubility (Wang and Kinsella, 1976; McWatters and Holmes, 1979a,b), water-oil activity index (Elizalde et al., 1988), and stability to heat (Foley and O’Connell, 1990). However, these studies have not been conclusive and it has been suggested that different factors may be important in different proteins (Klemaszewski et al., 1992).

An alternative approach is the preparation of small peptides from the native protein and the study of their emulsification properties. Such an approach has, for instance, been used with peptides of aS1-casein (Shimizu et al., 1983, 1984). It was found that aS1-casein (1-23) gave particularly good emulsification especially in conjunction with other casein peptides (Shimizu et af., 1986; Kaminogawa et al., 1987). Saito et al. (1993) studied emulsification by a trypsin-derived hydrolysate of bovine serum albumin (BSA) that gave better emulsification than BSA alone. These workers found that a 24-kDa peptide (corresponding to residues 377-582) was

DESIGN OF EMULSIFICATION PEPTIDES 95

preferentially adsorbed onto the oil globule surface. However, this peptide alone gave inferior emulsification compared to the whole hydrolysate, suggesting a role for other small peptides in conjunction with this 24-kDa peptide. These studies suggest that peptides are capable of emulsification although binding to oil globules alone does not completely explain this phenomenon. Synergies between peptides appear to be important in such systems for emulsification. Modeling the complex interactions between peptides in such mixtures is likely to be very difficult, however.

Due to developments in peptide synthesis technology it has now become feasible to synthesize peptides at will (Taylor and Kaiser, 1987; DeGrado, 1988). Moreover, thanks to improvements in computer modeling, it has also become possible to design novel peptides, the structure of which could be optimized for any particular property. For example, peptides have been designed to possess modified hormonal (Taylor and Kaiser, 1987), catalytic (Hahn et al., 1990; Atassi and Manshouri, 1993), and antimicrobial (Blond- elle and Houghten, 1992) activities. Thus it is now possible to design peptides with altered amino acid sequences that might be optimized for emulsification properties. Such an approach has been suggested by Bloomberg and his co-workers (Enser et ul., 1990; Bloomberg, 1991). The study of related families of such peptides offers a novel experimental model system for the study of emulsification (Saito et al., 1995). A further relevance of this research, which is beyond the scope of the present chapter, is in the preparation of pharmaceutical formulations for delivery of drugs (Shimizu and Nakane, 1995).

II. SECONDARY STRUCTURE OF PEPTIDES

The major criterion used to distinguish between the terms “protein” and “peptide” is the presence of extensive order in the structure of the former (Creighton, 1984). Important exceptions to this, however, are the p-caseins, which are the most important protein emulsifiers. Due to the presence of Pro and lack of Cys residues these proteins adopt disordered structures in solution, and their adsorption and surface-activity properties have been extensively investigated (Dalgleish and Leaver, 1991; Classon et al., 1995; Leerwakers et ul., 1996).

Peptides often adopt a range of conformations when free in aqueous solution (i.e.. they are often extensively disordered). However, peptides displaying surface-activity properties such as melittin from bee venom (Knoppel et nl., 1979; Inagaki et al., 1989; Kaiser and Kezdy, 1987) and hemolytic toxins (Moellby, 1983) often take up ordered a-helical structures at membrane-water interfaces. A major structural factor in these phenom-


ena is the position of hydrophobic and hydrophilic residues in the peptide sequence. Because a-helices display a high degree of periodicity with 3.6 residues per turn resulting in five complete turns every 18 residues, it is possible for a polypeptide to present highly hydrophilic and hydrophobic “faces” to the surrounding milieu (Taylor and Kaiser, 1987). In proteins, there is a tendency to prefer acidic residues at the N-terminus and basic residues at the C-terminus of a-helices (Chou and Fasman, 1974; Blagdon and Goodman, 1975). The former residues are thought to be more important than the latter as contributors to stability (Dasgupta and Bell, 1993). It has been suggested that this requirement for polar residues at helix termini creates a tendency for apolar residues within the helix, which may explain the amphiphilicity frequently found in such secondary structures (DeGrado et al., 1981; Eisenberg et al., 1982). This phenomenon is known as “helix capping” and it has also been demonstrated in peptides (Forood et al., 1993; see below). Amphiphilic faces will naturally tend to aggregate together (hydrophilic with hydrophilic and vice versa). This can result in the assembly to complex structures either in solution or on contact with lipid bilayers (Terwilliger and Eisenberg, 1982a,b; Vogel and Jahnig, 1986; Smith and Clark, 1992).

Melittin is known to adopt a tetrameric and helical structure at high pH, ionic strength, and peptide concentration (Talbot et al., 1979; Bello et al., 1982; Tenvilliger and Eisenberg, 1982b; Goto and Hagihara, 1992). Hydro- phobic faces of such multimeric assemblies may interact with phospholipid bilayers thus, for example, facilitating pore formation in cell membranes. This is the chemical basis of both hemolysis (Perez-Paya et al., 1994) and antimicrobial peptide activity (Blondelle and Houghten, 1992).

Although it is known that most peptides display little secondary structure when free in aqueous, solution, it is possible to promote a-helix formation experimentally by exposing the peptide to weakly polar alcohols. Methanol and 2,2-dichloroethanol promote helix formation but 2,2,2-trifluoroethanol (TFE) is the most widely used solvent in this regard. The precise basis of this effect is unclear but is thought to be related to its weak basicity, which strengthens hydrogen bonds. It should be noted that while TFE often produces a secondary structure in peptides similar to those found in corresponding parts of the intact protein structure, this is not always the case (Sonnichsen et al., 1992). TFE titration is thought to be a valuable indicator of helical propensity in synthetic peptides, however (Jasanoff and Fersht, 1994).

Comparisons of amino acid sequences of proteins of known secondary structure have allowed the identification of common features that may represent helix stability determinants (Presta and Rose, 1988) and the proposal of helix propensity scales for amino acids (Horovitz et al., 1992;


Blaber et al., 1993). An alternative approach (similar to the one described in the present work) is to prepare peptides with altered stuctural features and then to compare their stabilities experimentally.

The theoretical treatment of helix formation in peptides was first proposed by Zimm and Bragg (1959). These workers treated polypeptides as polymers of generic monomers and predicted that short peptides will not form helices in aqueous solution. While this is true for many situations, there are examples of peptides that do seem to form extensive secondary structure. It has been shown that the 13-residue C-peptide of RNase A shows helicity (Bierzynski et aZ., 1982; Brown and Klee, 1971), which can be increased by suitable substitutions (Shoemaker et al., 1982). From these studies, it has been concluded that side-chain interactions (which are not included in the Zimm-Bragg equation) play a role in a-helix stabilization. Much work has consequently been carried out on side-chain-specific effects such as helix capping and salt bridge formation (Bodkin and Goodfel- low, 1995).

A number of empirical factors have been identified as being of importance of helix stability in peptides. The role of salt bridgeshon pairs was one of the first of these to be investigated (Marquee and Baldwin, 1987). A “host-guest’’ approach studied the helical propensity of a Glu/Lys homopolymer (Lyu et al., 1990). This indicated that Ala, Leu, Met, and Gln (in that order) strongly favored helix stabilization. The precise order of this differed from both previous guest polymer studies (Scheraga, 1978) and the helical propensities of residues in globular proteins (Chou and Fasman, 1974) although these residues are all strong helix formers in both of these systems. The agreement with later work on a-helix propensity in T4 lysozyme is much better, however (Blaber et al., 1993,1994). With the exception of Pro and Gly (which strongly destabilize) and Ala (which strongly stabilizes), these latter workers found a close correlation between a-helix propensity and residue buried surface area for 17 other amino acids. That is, they concluded that hydrophobic stabilization of a-helices may be a major side-chain-associated factor in stabilization of a-helices in small proteins. Forood et al. (1993), meanwhile, investigated the effect of helix capping by introducing residues at the termini of a helical peptide. As with globular proteins, Asp, Asn, Glu, Gln, and Ala stabilized at the N-terminus while Arg, Lys, and Ala stabilized at the C-terminus (Doig et al., 1993).

Ala is known to be the strongest a-helix forming residue while Gly is the strongest helix breaker. A study using synthetic peptides (Chakrabatty et al., 1991) concluded that Ala is 100 times more likely to form a-helix than Gly. Ala-Gly substitutions in the middle of the peptide were much more disruptive of helix than those near the termini, indicating a strong position-dependent effect (Fairman et al., 1991). A study on Barnase (Ser-


ran0 et al., 1992) generally confirmed these conclusions with the further observation that stability differences in this system could not be due to fraying at the helix termini. Interestingly, this study also found a strong correlation between solvent-accessible hydrophobic surface area and stability. This finding is similar to that later reported for T4 lysozyme and discussed above (Blaber et al., 1993, 1994). However, it should be noted that Monte Carlo simulations using a rigid (Creamer and Rose, 1992) or flexible (Creamer and Rose, 1994) polypeptide backbone have suggested that it is loss of side-chain conformational entropy that correlates best with a-helix propensity scales rather than burial of accessible surface area.

Studies have also been performed using peptide models on P-sheet formation in aqueous solution (Zhang et al., 1993; Forood et al., 1995). Peptides with alternating hydrophilic and hydrophobic residues display an especially strong tendency to form P-sheet. Using a host-guest approach in a zinc- finger peptide it was possible to produce a propensity scale for P-sheet formation that agreed well with scales derived from statistical analysis of P-sheet in proteins of known structure (Kim and Berg, 1993). Another P- sheet propensity scale was proposed by Minor and Kim (1994) using the 1gG- binding domain of Protein G. While there appears to be general agreement between these scales, it is noteworthy that the absolute AAG values found in the latter system were an order of magnitude higher than those obtained for the zinc-finger host. By varying the environment (Baumrak et al., 1994) or small chemical substituents (Taylor et al., 1993) of peptides it is sometimes possible to cause them to interconvert between a-helix and P-sheet. It has been pointed out that P-sheet-producing residues such as Ile, Val, and Thr are frequently found in transmembrane a-helices and this has been attributed to the hydrophobicity of the bilayer (Li and Deber, 1992; Li et al., 1995; Deber and Li, 1995).

It has been noted that Leu is a-helix stabilizing while Ile and Val are weakly destabilizing, perhaps due to branching adjacent to the a-carbon. Lyu et al. (1991) synthesized a series of peptides containing “unnatural” (i.e., nonbiologically occurring) aliphatic side chains in derivatives of Ala, Val, Leu, and Ile. These side chains contained two to four carbons. It was concluded from this study that restriction in side-chain conformational freedom is a major factor in a-helix stabilization. @-Branching has been shown to destabilize a-helices in aqueous solution (Padmanabhan et al., 1990; Wojcik et al., 1990; Lyu et al., 1991). This is an environment-dependent affect, however, as it does not occur under hydrophobic conditions (Li and Deber, 1992). The role of hydrophobicity has been tested by varying the length of aliphatic side chains of test residues in a model peptide and comparing these to @-branched nonpolar residues. The aliphatic side chains


all promoted helicity, presumably due to a decrease in conformational entropy (Padmanabhan and Baldwin, 1991).

An interesting extension of this general approach (i.e., the use of residues occurring infrequently or not at all in biology) is the use of highly constrained side chains. These are often powerful promoters of a-helix formation (Balaram, 1992) and can be routinely introduced into peptides by genetic (Noren et al., 1989) or chemical methods (Schnolzer and Kent, 1992). An example is the a,a-dialkylated glycine, a-aminobutyric acid, which occurs naturally in fungal membrane proteins (Karle et al., 1991). Due to the introduction of van der Waals clashes, structurally favored conformational space is limited to helical regions (Karle and Balaram, 1990), which forces the peptide into an a-helix.

In visualizing secondary structures of peptides, therefore, it is important to recognize that, unlike proteins, they are highly flexible molecules capable of adopting a range of conformations in solution. The present chapter describes an approach to rational design of peptides with emulsification properties that takes advantage of the fact that amphiphilicity may be actively designed into such peptides. The overall approach is similar to that of Enser et al. (1990), who pioneered this type of design, and is summarized in Fig. 1. Briefly, naturally amphiphilic peptides and proteins are used as the starting point for the introduction of specific changes in sequence to increase the amphiphilicity of the final structure. The best candidate peptides may be synthesized by solid-phase methods (see Section IV) and the actual structure of the peptide may be investigated in solution. Studies on other properties such as toxicity and solubility may be performed. Last, the behavior of the peptide in a model emulsification system may be determined. This approach allows interpretation of differences in emulsification behavior due to alteration in the peptide primary structure.

Ill. MODELING OF PEPTIDE STRUCTURES

A. GRAPHICAL REPRESENTATIONS OF AMPHIPHILICITY

“Edmundson wheels” provide a simple and convenient means of identifying and visualizing amphiphilicity in a protein or peptide. This involves viewing along the axis of the helix with a residue protruding from a circle every 100”. Examples of this type of representation are shown in Fig. 2. For amphipathic peptides such as apolipoprotein and melittin, hydrophilic residues are seen to cluster to one side of the wheel and hydrophobic residues to the other, thus giving the wheel polar and nonpolar “faces.”


Peptide/protein with emulsification properties

t

Model structure (Energy minimize)

t

Alter sequence to increase amphiphilicity

Synthesize candidate peptides I

Other properties Structure t

Emulsification

FIG. 1. Design of amphiphilic peptides.

Various workers have proposed classifications of amino acids as hydrophilic or hydrophobic. Hopp and Woods (1981) classified Leu, Ile, Val, Tyr, Phe, Trp, Pro, and Met as hydrophobic and Asp, Arg, Lys, and Glu as hydrophilic residues. In addition, Gln, Asn, Thr, Ser, and Gly are re-


8 1 a b

9

18 l1 1 5,. FIG. 2. Graphical representation of amphiphilicity-Edmundson wheels. (a) The first 18 residues of melittin, (b) apolipoprotein IV (181-198). (c) the first 18 residues of aS1-casein (1-23). Strongly hydrophilic (W), weakly hydrophilic (M), neutral (0), weakly hydrophobic (A), and strongly hydrophobic (A) residues are shown as classified by Hopp and Woods (1981).

garded as moderately hydrophilic while Cys, His, and Ala are moderately hydrophobic. The extent of hydrophiiicity/hydrophobicity in a peptide/ protein is therefore directly dependent on amino acid sequence. For helices, Eisenberg et al. (1982) quantitated a hydrophobic moment pH as


where Nis the number of residues in the helix and Hi is the signed hydrophobicity associated with each side chain. By combining a hydrophobicity scale (e.g., that of Wolfenden et al. (1981) with the helix wheel concept of Ed- mundson, the hydrophobic moment can be visually expressed (Fig. 3). Using this type of approach, it has been possible to plot pH against net hydrophobicity H and to demonstrate that transmembrane helices, helices from proteins, and helices that seek interfaces between aqueous and nonpolar phases cluster in different regions of such a plot, thus suggesting a correlation between hydrophobic moment and protein function (Eisenberg et al., 1982).

The best-studied amphipathic structures are those of melittin (Terwilliger and Eisenberg, 1982b), its analogs (DeGrado et al., 1981; Blondelle and Houghten, 1992; Wade et al., 1992; King et al., 1994), apolipoproteins (Se- grest el al., 1994; Butchko et al., 1995), their analogs (Anantharamaiah, 1986), and antimicrobial peptides (Zhong et al., 1995; Merrifield etal., 1995). In the present chapter, these molecules are taken as suitable starting points for the design of emulsification peptides. In addition, aS1-casein (1-23) has also been reported as possessing emulsification properties (Shimizu et al., 1984, 1986). This molecule was also used in the design process described here.

1 a 8 12

15 ? 5

4

l1 q-...

18

7

b

4

11

18

7

6

10 17

1 8 12

16

9

2

13

14

3 17 10

FIG. 3. Hydrophobic moment of (a) lipoprotein IV (181-198) and (b) model peptide 5. The contribution of each residue to hydrophobicity is represented as vectors of solid lines and to hydrophilicity as dashed lines (Eisenberg ef a/., 1982). Hydrophobicity values determined by Wolfenden et al. (1981) were used. The magnitude of the vectors in b are greater, reflecting a larger hydrophobic moment. There is a clear “partitioning” between hydrophobic and hydrophilic faces in these helices.


B. COMPUTER-AIDED DESIGN

Although graphical representation of peptides is useful for simple visualization of amphiphilicity, the wide availability of molecular graphics programs and of computer workstations has allowed more realistic representations of structure (Jameson, 1989; Lesk, 1991; Olsen and Goodsell, 1992; Sun and Cohen, 1993). For example, while the amphiphilicity of melittin is adequately visualized by the representation shown in Fig. 2, this peptide is known to possess a “kink” in its sequence that links two a-helices. This has been observed both in crystals by X-ray crystallography (Tenvilliger and Eisenberg, 1982a) and, when bound to micelles, by two-dimensional NMR (Inagaki et al., 1989). To represent complexities such as this, more sophisticated graphics are required.

Modeling systems generally produce a structure using three basic components (Jameson, 1989). First, the structure is represented as an “energy field” incorporating covalent geometry, electrostatic interactions, van der Waals forces, and hydrogen bonding. This is established with the aid of parameter tables that create mathematical representations of atoms and their possible interactions, such as chemical bonds and nonbond effects. The second component is molecular mechanics, which calculates a local energy minimum for a given structure. Such calculations are not completely accurate with present computing facilities due to the number and complexi- city of calculations required. However, by making appropriate simplifications to bond angles and lengths used, a structure that closely approximates the true one can be calculated. The third component is molecular dynamics (Hermans, 1993), which allows the peptide to be moved in space, thus eliminating steric conflicts and allowing calculation of a global energy minimum (an “energy-minimized” structure).

It is important to realize that structures predicted by this technology are limited in their accuracy by a number of considerations. The protein folding problem (i.e., how one-dimensional information such as an amino acid sequence can dictate a three-dimensional structure such as a folded protein) is still not solved (Lattman and Rose, 1993; Munoz and Serrano, 1994). Moreover, algorithms used to calculate energy minima are based, ulti- mately, on the comparatively small number (approximately 1000) of proteins for which crystal structures are available (Sutcliffe et d., 1987; Ponder and Richards, 1987). These structures include few examples of membrane- bound proteins and may well represent a comparatively small subset of proteins that happen to crystallize fairly readily. The algorithms used also contain simplifications and approximations that may introduce inaccuracies into the final structure. When modeling peptides, in particular, it should also be borne in mind that peptide a-helices are generally longer than


protein a-helices and tend to be frayed at their termini (Lyu et al., 1991). Also, because of their greater flexibility in solution, peptides are capable of taking up a variety of conformations. All of these caveats suggest that care should be taken in interpretation of structures designed by computer graphics. The technique is useful as a guide to design, however, and to the rational introduction of novel features into peptides. As we discuss later (see Sections IV and V), it is necessary to study the actual structure of peptides in aqueous solution by techniques such as circular dichroism and two-dimensional NMR to verify the success or otherwise of the design (Bradley et al., 1990; Fezoui et al., 1995).

A number of structural features (often, ones found in globular protein structures) have been designed de now into synthetic peptides using the approach outlined above (Baglia e f al., 1992; Fezoui et al., 1994, 1995; Kuroda, 1995). Since amphipathic helices occur as structural building blocks in globular proteins or are often found to have surface-active effects, a number of workers have designed such structures either de now or by homology with peptides such as melittin (DeGrado et al., 1981; Moser, 1992; Blondelle and Houghten, 1992; Zhou et al., 1993; Perez-Paya et al., 1994; Epand et al., 1995). Enser et al. (1990) applied this rationale to the design of emulsification peptides. Briefly, it is expected that amphipathic a-helices are likely to display surface-activity properties. By maximizing the amphiphilicity of a peptide, it was hoped to generate a potent emulsifier. This approach was extended by Carey et al. (1994) who designed two novel peptides and demonstrated that they were effective emulsifiers. In the following section, the strategy used in the design of these and other peptides is described.

C. DESIGN OF EMULSIFICATION PEPTIDES

Naturally occurring and candidate synthetic peptides were modeled using INSIGHT I1 (Biosym Technologies) on a personal Iris 4D workstation (Model 891) from Silicon Graphics. The program was used in conjunction with the molecular graphicddynamics package DISCOVER and required IRIX version 3.2 or higher incorporating the UNIX background. This facilitated three-dimensional modeling of peptides and the visualization of hy- dropathies. The structures were energy-minimized using a maximum criterion of 0.001 kCal/A and the steepest descents algorithm, which required no cross terms and no morse terms. The number of iterations and the time necessary for modeling varied for each peptide.

The first candidate structure is that of melittin shown in Fig. 4. This energy-minimized structure clearly shows the “kink” that is seen in the region of residues 10-14 in structures derived from both X-ray crystallogra-


FIG. 4. Energy-minimized structure of melittin and model peptides derived therefrom. The structures were generated by DISCOVER and are shown as “side-on” (i.e., perpendicular to the helix axis) and “end-on” (along the helix axis) views. (a) Melittin, (b) model peptide 1, (c) model peptide 2, (d) model peptide 3, and (e) model peptide 4. Amino acid sequences are detailed in Table 1.


FIG. 4 Continued

phy and two-dimensional NMR studies (Terwilliger and Eisenberg, 1982a; Inagaki et al., 1989). This feature of melittin’s structure (which is thought to be essential for the peptide’s toxicity; Dempsey et al., 1991; Dempsey, 1992) was mentioned in Section IIIB. A number of considerations governed


FIG. 4 Continued

the alterations made to this sequence. These will now be described for melittin but similar considerations governed the engineering of the other candidate sequences (see below). First, as melittin is highly cytotoxic, it was desired to abolish this toxicity. Second, changes were made, where possible, to introduce strongly helix-forming residues into the sequence and, conversely, to remove helix “breakers.” Third, strongly hydrophilid hydrophobic residues (e.g., Leu/Lys) were introduced in place of more neutral or weakly hydrophilic/hydrophobic residues (e.g., Met/Gly).

Beginning with the sequence for melittin shown in Table I, model peptide 1 was energy-minimized. This peptide contains alterations at six positions; 3Gly + Glu, ’Lys -+ Gly, I4Pro + Thr, 22Arg -+ Glu, 24Arg -+ Gln, and 2hGln -+ Thr. These had the effect of removing helix-breaking residues such as Gly and Pro from a number of positions and, as shown in Fig. 4, the modeled energy-minimized structure is somewhat more amphipathic than the parent. This structure still had a discernible kink, however, so further alterations were made to remove it. Model peptide 2 was generated by altering ‘Gly + Ser, 211e -+ Thr, ‘Gly ---f Gin, 4Ala ---f Val, 5Val -+ Thr, ‘Leu -+ Gln, 8Val -+ Leu, ‘Leu -+ Thr, “Thr + Gln, I2Gly + Leu, “Leu + Thr. “Pro + Ser, I5Ala + Gln, ”Ile + Gln, %er + Thr, “Trp -+ Ile, and 2011e -+ Val. These alterations generated a strongly amphipathic structure. It was felt that there still remained some “overlap” between the hydrophilic and hydrophobic faces in this structure, however.

TABLE I AMINO ACID SEQUENCES OF CANDIDATE AND TEST PEPTIDES

Sequence

Peptide 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Melittin Glv Ile Glv Ala Val Leu Lvs Val Leu Thr Thr Glv Leu Pro Ala Leu Ile Model peptide 1 G& Ile G k Ala Val Leu & Val Leu Thr Thr G& Leu Thr Ala Leu Ile Model peptide2 Thr % Val Thr Lys Thr Thr Gln Leu Thr Ser m Leu Gln Model peptide 3 Ser Thr Glu Val Thr Gln Gln Leu Thr Thr Gln Leu Thr Ser Gin Leu Gln Model peptide 4 mr Val Ser Gln Leu Gln Glu Tyr Trp Thr Thr Leu & Gln Ile & Apolipoprotein IV Pro Phe Ala Asn Glu Leu Lvs Glu Lys Phe Asn Gln Asn Met Glu Glv Leu

-(i8i-i98) Modelpeptide 5 Thr Phe Leu Gln Asp Leu Lys Glu Lys &I Gin Gln Leu Thr Glu & Leu aS1-casein (1-23) Arg Pro Lys His Pro Ile Lys His Gln Gly Leu Pro % Val Leu Asn Model peptide 6 Arg & v,l His Thr Val Ile Thr Trp Ala Leu Glu Leu Ile Gln Glu Leu


Accordingly, further alterations were introduced into model peptide 2 to generate model peptide 3: 7Lys + Gln, 22Arg + Gln, 24Arg + Gln, and "Gin + Arg. Again, it was felt that there was some overlap between the two faces and this prompted a further iteration of design. Model peptide 4 retains only three residues at the same positions (10, 11, and 13) as melittin. The alterations made were: 'Gly -+ Thr, 211e + Val, 3Gly + Ser, 4Ala --+ Gln, 'Val + Leu, 6Leu + Gln, 7Lys + Glu, V a l + Tyr, 9Leu + Trp, "Gly + Leu, I4Pro + Ser, *'Ala + Gln, ''Leu + Ile, 1711e -+ Lys, "Ser -+ Thr, I9Trp + Leu, 2011e + Leu, 21Lys + Gln, 22Arg + Gln, 23Lys -+ Ile, 24Arg + Lys, 2'Gln + Thr, and "Gin + Ser. This was the most amphipathic model structure and it was decided to proceed with synthesis of this peptide and to assess it as a potential emulsifier (see Sections IV and V).

This approach allows for modeling a number of peptides and optimizing aspects of their structures for a-helix formation and amphiphilicity. It has been pointed out that one of the great advantages of computer-aided molecular design is the possibility of relatively quickly assessing a variety of structures before committing valuable resources to synthesis (Sun and Co- hen, 1993).

The structure obtained for an amphipathic sequence from apolipoprotein IV (181-198) is shown in Fig. 5. The amino acid sequence of this peptide was altered to enhance amphiphilicity. Apart from 4Asn --+ Gln, 5Glu + Asp, and "Asn -+ Gln, which are conservative mutations, six alterations were made to generate model peptide 5: 'Pro + Thr, 3Ala -+ Leu, 'OPhe + Val, "Asn + Leu, I4Met + Thr, and l'Gly + Ala. These changes had the effect of polarizing hydrophobic/hydrophilic residues on opposite faces of the structure. In addition, the introduction of Leu and Val (strongly hydrophobic) would be expected to make the peptide even more hydrophobic on its hydrophobic face. This is illustrated in Fig. 5. Since Leu is such a strong helix former, it would be expected to help the structure form a helix more readily than the parent peptide.

The third peptide selected for modeling was aS1-casein (1-23), which had previously been reported as capable of emulsification (Shimizu et al., 1983,1984,1986). The energy-minimized structure obtained for this peptide is shown in Fig. 5. It is clear from the "end-on" view of this structure that this peptide is not as amphipathic as that of melittin (Fig. 4) or apolipoprotein IV (181-198; Fig. 5). While there are generally hydrophobic and hydrophilic faces, there are clearly residues present on the "wrong" side of the helix. For example, 'Arg and 3Lys appear to be on the more hydrophobic side of the structure. Of the 23 residues, 19 were altered in several iterations to generate the highly polarized model peptide 6 . The following substitutions were made: 2Pro + Ile, 3Lys -+ Val, 'Pro -+ Thr, 'Ile + Val,


FIG. 5. Energy-minimized structures of apolipoprotein IV (181-198), aS1-casein (1-23), and model peptides derived therefrom. (a) Apolipoprotein IV (181-1Y8), (b) model peptide 5 , (c) aS1-casein (1-23), and (d) model peptide 6. Amino acid sequences are given in Table I.


FIG. 5 Continued

'Lys + Ile, 'His + Thr, 'Gln + Trp, '"Gly + Ala, "Pro + Glu, I3Gln + Leu, I4Glu + Ile, 15Val + Gln, l6Leu + Glu, "Asn + Leu, '"lu + Leu, "Asn + Glu, '"Leu -+ Gln, and 23Phe + Thr. Model peptides 4, 5 , and 6 now appeared to have the desired properties of amphiphilicity and were selected for synthesis and further study.


IV. SYNTHESIS OF DESIGNED PEPTIDES

Solid-phase peptide synthesis was originally developed by Merrified (1963) and allows the rapid production of peptides of high purity and defined sequence (Stewart and Young, 1984; Barany et al., 1987; Fields and Noble, 1990). Because the peptide chain is attached to a resin, the technique lends itself to achieving high-efficiency coupling. Unreacted reagents and by-products of the reaction are conveniently flushed away from the resin. This facilitates high repetitive yields and the synthesis of peptides of maximum lengths 20-30 residues, although longer syntheses have been reported (Ramage et al., 1989). By linking individually prepared peptides together, it is possible to prepare synthetic enzymes, such as RNase, that display enzymatic activity (Gutte and Merrifield, 1969). The basic approach is to establish a system in which reactivity of amino acid substituents is tightly controlled. The first amino acid is attached to a functional group (Stewart and Young, 1984; Barany et al., 1987) on the surface of an insoluble resin (the solid phase, e.g., polystyrene or polyamide). This is achieved by using a chemically activated form of the C-terminal residue. Frequently, an activated ester of the residue (such as a pentafluorophenyl ester) or a symmet- ric or mixed anhydride are used. Undesired side reaction between pairs of the C-terminal derivative that would result in the unwanted formation of a dipeptide is prevented by the presence of a “blocking” group of its -NH2 substituent. “Directionality” of the synthesis is therefore ensured by chemical activation of one group (-COOH) and masking of the other (-NH2). Although Merrifield originally used t-butyloxycarbonyl as an -NH2 blocker, the most commonly used blocking group at present is 9- fluorenylmethoxycarbonyl (F-moc). This may be removed by exposure to trifluoroacetic acid (TFA). The F-moc group and the TFA are washed from the resin and the second residue (i.e., that in the penultimate C-terminal position of the desired sequence) is then added. The activated -COOH group of this residue can now react with the deprotected -NH2 group of the C-terminal residue in a reaction catalyzed by dicyclohexylcarbodiimide or similar reagent. This results in an anchored dipeptide that is now amenable to deprotection and, by repetitive cycles of coupling and deprotection, may be extended to the desired length and sequence. At the end of the synthesis, the peptide is cleaved from the resin by treatment with hydrofluoric acid and recovered.

Note that residues with reactive side chains (i.e., Glu, Lys, Ser) need to have these groups protected during synthesis to prevent unwanted reaction (e.g., with &-NH2 of Lys, thiol group of Cys, hydroxy groups of Ser/Thr, etc.; Meienhofer, 1985; Patek, 1993). These may be removed by treatment with hydrofluoric acid and other reagents after synthesis is complete


(Meienhofer, 1985). It is clear that, since the coupling and deprotection steps require quite different chemical conditions, the chemistry of this synthesis lends itself readily to automation. The model peptides designed in Section IIIC were synthesized in an Applied Biosystems Model 431A synthesizer with F-moc (model peptide 5) and Fast-moc (peptides 4 and 6; Fields and Noble, 1990; Applied Biosystems, 1990) amino blocking groups, respectively.

Two possible problems often arise with solid-phase synthesis of peptides. The first is the formation of deletion peptides due to a synthetic cycle not reacting to completion. This arises from incomplete coupling of a sterically hindered residue where the next residue added to the chain is small and couples efficiently. This can occasionally introduce variability into the coupling efficiency of each cycle of synthesis. The final peptide product may therefore be contaminated with significant amounts of one or more peptides lacking a single residue. This deletion peptide would therefore need to be removed after synthesis and recovery. Incomplete coupling is sometimes overcome by recoupling the difficult residue in a different solvent or by using a different coupling reaction. This phenomenon is sometimes difficult to predict as difficulties in coupling are occasionally sequence-specific. A second problem that may introduce heterogeneity into the peptide preparation is incomplete deblocking. It has been estimated that the efficiency of deblocking is as low as 93% and this is one of the factors that limits the length of peptides with which solid-phase synthesis is possible (Meienhofer, 1985; Patek, 1993).

In order to detect these problems, monitoring of the synthesis is necessary. For example, ninhydrin tests may be carried out to quantitate unreacted -NH2 groups. The F-moc substituent may be conveniently measured spectrophotometrically at 300 nm and gives a useful measure of extent of deblocking and for comparing different cycles in the same synthesis. After synthesis is complete, it is necessary to establish the purity of the peptide by high-resolution techniques such as high-performance liquid chromatog- raphy (HPLC), mass spectrometry (MS), or capillary electrophoresis. MS is especially useful for detecting incomplete deblocking and deletion peptides because of its ability to resolve very small mass differences. It is therefore often possible to specify which coupling cycle did not go to completion or which blocking group was not completely removed from such an analysis and thus to facilitate repeal of the synthesis with greater success. For purposes of illustration, HPLC and MS analysis of model peptide 6 are shown in Fig. 6. This demonstrates a single peptide present on HPLC analysis with a mass of 2324 (estimated mass: 2323). The synthesized peptides were also sequenced by N-terminal microsequencing (Applied Biosys- tems Model 477A) as an aid to assessing the success of synthesis. This is

*280

+Ions

2,000 -

1,000 -

0

0.4

0.2

0

2,324

1 I

I I I

a

i L 0 15 30

Retention (Min.)

3,000 {

1,000 2,000 3,000 4,000

FIG. 6. Analysis of model peptide 6. After peptide synthesis it is essential to characterize the purity of the product. (a) Reversed-phase HPLCon C-8 resin. Buffer A: 0.1% trifluoroacetic acid; buffer B: 70% acetonitrile, 0.065% trifluoroacetic acid. A gradient of 5100% B over 30 min was used and peptide was detected at 280 nm. (b) Time of flight MS analysis. The estimated mass based on amino acid sequence is 2323.


not as useful a criterion as HPLC and MS as the peptide may often be lost from the sequencer when it has been reduced to tetra- or tripeptide length. However, since N-terminal sequencing functions in the opposite “direction” to Merrifield peptide synthesis, it can give useful information on the later cycles of synthesis (which couple the earlier residues sequenced). Small amounts of incomplete deblocking or coupling are not as easy to detect by this technique alone as, for example, by MS, so it is essential that as many high-resolution analytical techniques as possible be applied to the peptides to ensure the investigator of successful synthesis before proceeding to test their secondary structure or performance as emulsifiers (Bradley et al., 1990).

V. TESTING OF PEPTIDE EMULSIFICATION PROPERTIES

Foods are complex biochemical systems that contain lipid and water as well as other nutrients such as proteins, vitamins, and carbohydrates. Water and lipid do not intermix because of their polar and apolar natures, respectively (Tanford, 1980). Emulsifiers are frequently present or are added to food preparations to forestall phase separation (Friberg, 1976). Contact between lipid and water is thermodynamically disfavored but the presence of emulsifiers can allow them to function long enough to give useful shelf- lives (Das and Kinsella, 1990). Formation of an emulsion requires a large increase of the interfacial surface of emulsified droplets. This makes the AG of formation of an emulsion positive (Das and Kinsella, 1990). However, the time required for emulsion breakdown (and phase separation) may be much longer than the shelf-life of the product. Emulsion stability is influenced by the net balance between attractive (van der Waals and London forces) and repulsive (electrostatic, hydration) forces (Parker, 1988; Dickin- son and Stainsby, 1988).

A wide variety of chemical structures possess emulsification properties. For example, the phospholipid molecule, lecithin, and proteins such as BSA. Molecules such as gum guar play a role in emulsification by acting as thickeners resulting in emulsion stabilization. A common feature observed, in cases where detailed structural data are available for emulsifiers, is the presence of distinct nonpolar hydrophobic and polar hydrophilic regions in the structure (i.e., the molecules are amphipathic; Bourrel and Schlechter, 1988; Kachholz and Schlingmann, 1987). However, the molecule may possess a range of charge states being positively or negatively charged or zwitterionic. When mixed in a model system, oil and water separate into two distinct phases with an oil-water interface between them. Emulsifiers lower the surface tension at this interface because of their surface-activity


properties (Fisher and Parker, 1988). These properties are directly conse- quential on the amphipathic structure of the emulsifier (Lang and Wagner, 1987). In forming an emulsion, then, the emulsifier may migrate to the oil-water interface and facilitate intermixing of the phases by stabilizing small droplets of oil (in an oil in water emulsion) or water (in the case of a water in oil emulsion).

A number of experimental methods may be used to study emulsions (Das and Kinsella, 1990). One of the most important measurements is droplet size. This requires a means of microscopically measuring droplets and determining the range of droplets in the emulsion. Interfacial surface tension may be measured in a tensionemeter (Tornberg, 1978).

In the present work, detailed measurements of this type were not carried out. A simple assay was used to determine emulsion stability (Carey et al., 1994). Briefly, emulsifier was added to 10 ml of 1 : 1 rapeseed oil : 20 mM sodium acetate/acetic acid buffer, pH 4.38, and homogenized. The time taken for the phases to separate was noted. Blank emulsions generally separated completely in approximately 2 hr at 6°C while emulsions prepared with known emulsifiers such as lecithin and BSA or an emulsion stabilizer such as gum guar took several days to separate (Fig. 7). Emulsion stabilities with melittin and the three model peptides designed in this work are shown in Fig. 8. These indicate that model peptide 4 is particularly effective by this criterion. This peptide stabilizes approximately an order of magnitude better than the emulsifier/stabilizers shown in Fig. 7 (i.e., 0.2 mg/ml model peptide 4 gives similar emulsion stability to 2 mg/ml BSA). Model peptides 5 and 6, though less effective than model peptide 4, were found to be as effective as conventional emulsifier/stabilisers.

Note that model peptide 4 is also significantly more effective (by approximately 10-fold) than its parent compound, melittin. Eleven of melittin’s 26 residues are located near the N-terminus (giving a relatively large hydrophobic surface area here, which is thought to be important for attacking biological membranes). In model peptide 4, this number was reduced to 8, covering less than a third of the available surface. Moreover, because changes were deliberately introduced to create more clear-cut polarity between the hydrophilic and hydrophobic faces, it was possible to achieve an increase in surface activity while introducing greater net water solubility into the peptide.

Circular dichroism (CD) measurements were carried out to assess the actual secondary structure of the model peptides. Model peptide 6 was found to be only sparingly soluble in aqueous buffers and, accordingly, did not give a very good spectrum. Because of the amphipathic nature of emulsifiers, it is sometimes difficult to strike a balance between functional activity and water solubility. However, it was possible to determine the secondary structure of model peptides 4 and 5 by this method. The spectra


Time (Hrs.)

Time (Hrs.)

Time (Hrs . )

"- 0 1 2 3 4 5

mg/ml

6o

[ c

30 1 I

6o

50

- c -

1 2 3 4 5

mglml

FIG. 7. Emulsification assay. The time required for oil-water separation was determined for a range of concentrations of the following known emulsifiers: (a) BSA, (b) gum guar, and ( c ) lecithin.

118 D. SHEEHAN, S. M. O'SULLIVAN, AND K. B. CAREY

Time (Hrs.)

a 100 -

80 -

60 -

40 -

20 -

Time (Hrs.)

mglml

FIG. 8. Emulsification by amphipathic peptides. The time required for phase separation in a system identical to that of Fig. 7 was measured for a range of concentrations of (a) melittin, (b) model peptide 4, (c) model peptide 5, and (d) model peptide 6.

obtained are shown in Fig. 9 and the data are tabulated in Table 11. Model peptides 4 and 5 form little ordered structure in aqueous buffer. However, the addition of TFE induces the formation of 90-100% order in the structure. Interestingly, only approximately 33-45% of this structure is a-helical while some 60% is P-sheet. This was unexpected and suggests that the peptides are capable of existing in a variety of conformations and, presumably, are capable of interconverting between them. Despite their different emulsification stabilities (Fig. 8), model peptide 5 consistently displayed slightly higher order than model peptide 4. When CD measurements of a- helix were made in an oiYwater system, this observation still held true. This suggests that the two peptides do not assume the same structures in solution as predicted from their energy-minimized structures shown in Figs. 4 and 5 (Carey et al., 1994). This is probably due to the shortcomings of the modeling approach used, which were highlighted earlier. Although


Time (Hrs.)

Time (Hrs . )

mglml

2o I d

mglml

FIG. 8 Continued

more detailed investigations of emulsion stability were not performed on these peptides, we have found with other melittin-derived model peptides that droplets in the size range 1-10 p m are obtained with Coulter counter measurements (Ebeling et nl. 1997).

Using naturally amphipathic structures as starting points, it has been possible to design peptides with enhanced emulsification stability by maximizing the amphipathic nature of the structure. These peptide emulsifiers were as good as or better than melittin, one of the most surface-active peptides known.

VI. FUTURE DIRECTIONS

A number of exciting developments suggest that the approach described in this paper could be fruitfully pursued in the future. The availability of

60,000 -

- 50,000

a

190 210 230 250

b 40,000 -

[el

-2 0,000 -I I I I

190 210 230 250

209000

- 20,000

C

190 21 0 230 250


TABLE I1 CIRCULAR DICHROISM OF MODEL PEPTIDES

Pep t ide &Sheet (%) a-Helix (%) Random coil (%)

Model peptide 4 20 mM sodium acetate, pH 4.8 53 13 (14) 33 Buffer: SO% TFE 58 33 (32) 9

20 mM sodium acetate, pH 4.8 51 5 (0) 44 Buffer: 50% TFE 55 45 (49) 0

- (11) Buffer: Oil - Model peptide S

- (27) Buffer: Oil -

faster and more powerful molecular modeling systems together with greater knowledge of the behavior of model peptides in solution suggest the possibility of improved design of peptides. The commercial availability of blocked, conformationally constrained residues suitable for peptide synthesis (Balaram, 1992) underlines the possibility of generating families of peptides with an even greater propensity to form a-helix or @-sheet structures than natural amino acids. A number of novel chemical options are also now available for peptide design. One of the more interesting of these is the concept of joining helices together into bundles by covalently linking their termini. The introduction of disulfide bridges may be achieved by the judicious introduction of Cys residues (Jackson et af., 1991; Rivett et af., 1996). An alternative is the use of “branching” residues such as ornithine and lysine at helix termini. These may be acylated at non-a amino groups, thus providing amide linkages between different helices (Hahn et af., 1990). We are presently extending the work described here with studies of melittin analogs containing Cys residues. It is possible to generate peptide dimers and oligomers in this way and we have found that this affects surface activity (unpublished observations).

Although CD and MS measurements give adequate structural data, it would be of interest to determine the structure with two-dimensional NMR measurements in solution. A necessary prerequisite of such a study would be the assessment of the effects of variables such as pH and salt concentration on emulsification. These variables have been found to affect the struc-

FIG. 9. Circular dichroism spectra (50% TFE) of (a) model peptide 4, (b) model peptide 5, and (c) model peptide 6. Because of the limited solubility of peptide 6 in aqueous buffers, the spectrum was poor and it was not possible to determine secondary structure accurately. Secondary structures derived for the other two model peptides are tabulated in Table 11.


tures of melittin (Bello et al., 1982) and models derived therefrom (Daggett et al., 1991; Moser, 1992). It should also be noted that surface activity alone, while important, is not the only property desired in a food emulsifier and we have not yet carried out studies on the more precise measures of emulsifier performance such as emulsification activity index, emulsification capacity, and emulsion stability (Das and Kinsella, 1990). In view of the impressive performance of model peptide 4 in the present study, this molecule is a tempting target for such studies.

ACKNOWLEDGMENTS

We are grateful to Professor C. Daly of the National Food Biotechnology Centre (NFBC), University College Cork, for his encouragement of this work and also to Mrs. Aine Healy of the NFBC for her expert help in peptide synthesis and analysis. Circular dichroism measurements were carried out by our collaborators, Dr. N. C. Price and Ms. S. M. Kelly, University of Stirling. Silicon Graphics equipment was a generous gift from Schering-Plough Corp. Innishannon, Co., Cork. DS is grateful to the Spanish government for a sabbatical fellowship at the Consejo Superior de Investigaciones Cientificas, Centro de Investigacion Y Desenvolu- pament (CSIC, CID), Barcelona, which facilitated preparation of the manuscript. We are also grateful to Dr. D. Wilcock (CSIC, CID) for critically reading the manuscript.

REFERENCES

Anantharamaiah, G. M. 1996. Synthetic peptide analogs of apolipoproteins. Mefh. Enzymol.

Applied Biosystems. 1990. FastMocTM chemistry: HBTU activation in peptide synthesis on the model 431A. User Bulletin No. 33, Nov. 1-5.

Atassi, M. Z . , and Manshouri, T. 1993. Design of peptide enzymes (pepzymes): Surface- simulation synthetic peptides that mimic the chymotrypsin and trypsin active sites exhibit the activity and specificity of the respective enzyme. Proc. Natl. Acad. Sci. USA 90,2828- 8286.

Baglia, F. A., Bradford, A. J., and Walsh, P. N. 1992. Fine mapping of the high molecular weight Kininogen binding site on blood coagulation factor IX through the use of rationally designed synthetic analogues. J . Biol. Chem. 267,4247-4252.

Balaram, P. 1992. Non-standard amino acids in peptide design and protein engineering. Curr. Opin. Strut. Biol. 2, 845-851.

Barany, G., Kneib-Cordonier, N., and Mullen, D. G. 1987. Solid phase peptide synthesis: A silver anniversary report. Int. J . Pept. Prof. Res. 30, 705-739.

Baumrak, V., Huo, D., Dukor, R. K., Keiderling, T. A., Leievre, D., and Brack. A. 1994. Conformational study of sequential lys and leu based polymers and oligomers using vibrational and electronic circular dichroism spectra. Biopolymers 34, 11 15-1 121.

Bello, J., Bello, H. R., and Granados, E. 1982. Conformation and aggregation of melittin: Dependence on pH and concentration. Biochemistry 21,461-465.

Bierzynski, A,, Kim, P. S., and Baldwin, R. L. 1982. A salt bridge stabilizes the helix formed by isolated C-peptide of RNA-ase. Proc. Nail. Acad. Sci. USA 79, 2470-2474.

128,626-668.


Blaber, M., Zhang. X.-J.. Lidstrom, J. D., Pepoit, S. D., Baase, W. A,, and Matthews, B. M. 1994. Determination of alpha-helix propensity within the context of a folded protein. J. Mol. Biol. 235, 600-624.

Blaber. M., Zhang. X.-J., and Matthews, B. M. 1993. Structural basis of amino acid alpha- helix propensity. Science 260, 1637-1640.

Blagdon, D. E., and Goodman. M. 1975. Mechanisms of protein and peptide helix initiation. Biopolyniers 14, 241 -245.

Blondelle, S. E., and Houghten, R. A. 1992. Design of model arnphipathic peptides having potent antimicrobial activities. Biochemistry 31, 12,688-12,694.

Bloomberg, G. 1991. Designing proteins as emulsifiers. Food Market. Technol. 5, 14-15. Bodkin, M. T.. and Goodfellow, J. M. 1995. Competing interactions contributing to alpha-

helical stability in aqueous solution. Pror. Sci. 4, 603-612. Boguski, M. S., Elshourbagy, N., Taylor, J. M., and Gordon, J. 1. 1985. Rat apolipoprotein

A IV contains thirteen tandem repetitions of a 22-amino acid segment with amphipathic helical potential. Proc. Natl. Acad. Sci. USA 81, 5021-5025.

Bourrel, M., and Schlechter, R. S. 1988. The R-ratio. In “Microemulsions and Related Sys- tems,” pp. 1-27. Marcel Dekker, New York.

Bradley, E. K., Thomason. J. F., Cohen, F. E., Kosen. P. A., and Kuntz, I. D. 1990. Studies of synthetic helical peptides using circular dichroism and nuclear magnetic resonance. J. Mol. Biol. 215, 607-622.

Brown, J. E.. and Klee, W. A. 1971. Helix-coil transition of the isolated amino terminus of ribonuclease. Biochemistry 10, 470-476.

Butchko, G. W., Rozek, A,, Zhoug, Q., and Cushley, R. J. 1995. Sequence-specific ’H NMR assignments and secondary structure of a lipid-associating peptide from human apolipoprotein C-I: An NMR study of an amphiphilic helix motif. P e p . Res. 8, 86-94.

Carey, K. B., Kelly, S. M., Price, N. C.. O’Sullivan, D. J.. and Sheehan, D. 1994. Two amphiphilic, synthetic peptides display strong emulsification properties. Food Chem. 50, 83-85.

Chakrabatty, A,, Schellman, J. A,, and Baldwin, R. L. 1991. Large differences in the helix propensities of alanine and glycine. Nature 351, 586-588.

Chou, P. Y., and Fasman, G. P. 1974. Prediction of protein conformation. Biochemistry 13,222-245.

Classen. P. M., Blomberg, E.. Froeberg, J. C., Nylander, T., and Arrebant. T. 1995. Protein interactions at solid surfaces. Adv. Colloid Interface. Sci. 57, 161-227.

Creamer. T. P., and Rose. G. D. 1992. Side-chain entropy opposes alpha helix formation but rationalises experimentally determined helix-forming propensities. Proc. Natl. Acad. Sci. USA 89, 5937-5941.

Creamer, T. P., and Rose, G. D. 1994. Alpha-helix forming propensities in peptides and proteins. Prot. Struct. Funct. Genet. 19, 85-97.

Creighton, T. E. 1984. “Proteins, Structures and Molecular Properties.” W. H. Freeman, London.

Creighton, E. 1992. Folding and binding. Curr. Opin. Struct. Bid . 2, 1-5. Daggett. V., Kollman, P. A,, and Kuntz, I. D. 1991. Molecular dynamics simulations of small

Daggett, V., and Levitt, M. 1992. Molecular dynamics simulation of helix denaturation.

Dalgleish, D. G., and Leaver, J. 1991. The possible conformations of milk proteins adsorbed

Das. K. P., and Kinsella, J. E. 1990. Stability of food emulsions: Physicochemical role of

peptides: Dependencc on dielectric model and pH. Biopolymers 31, 285-304.

J . Mol. Bid . 223, 1 12 1-1 138.

on oil/water interfaces. J . Colloid Interface Sci. 141, 288-294.

protein and nonprotein emulsifiers. Adv. Food Nutr. Res. 34, 81-201.


Dasgupta, S., and Bell, J. A. 1993. Design of helix ends. Amino acid preferences, hydrogen

Deber, E. M., and Li, S. C. 1995. Peptides in membranes: Helicity and hydrophobicity.

DeGrado, W. F. 1988. Design of peptides and proteins. Adv. Prof. Chem. 39,51-124. DeGrado, W. F., Kezdy, F. T., and Kaizer, E. T. 1981. Design synthesis and characterization

of a cytotoxic peptide with melittin-like characteristics. J . Am. Chem. SOC. 103,679-681. Dempsey, C. E. 1992. Quantitation of the effects of an internal proline residue on individual

hydrogen bond stabilities in an a-helix: pH-dependent amide exchange in melitin and [Ala-l4] melittin. Biochemistry 31, 4705-4712.

Dempsey, C. E., Bazzo, R., Harvey, T. S., Syperek, I., Boheim, G., and Campbell, 1. D. 1991. Contribution of proline-14 to the structure and actions of melittin. FEES Lett.

Dickinson, E., and Stainsby, G . 1988. Emulsion stability. In “Advances in Food Emulsions and Foams” (E. Dickinson and G . Stainsby, eds.), pp. 1-44. Elsevier, Amsterdam.

Doig, A. J., Chakrabartty, A., and Baldwin, R. L. 1993. Helix N-cap propensities in peptides parallel those found in proteins. Proc. Nafl. Acad. Sci. USA 90, 11,332-11,336.

Ebeling, S. C., Kelly, S. M., O’Kennedy, B. T., Price, N. C., and Sheehan, D. (1997). Surface activity properties of cysteine-substituted C-terminal melittin analogues. Biochemie 79,

Eisenberg, D., Weiss, B. M., and Terwilliger, C. 1982. The helical hydrophobic moment: A measure of the amphiphilicity of a helix. Nature 299, 371-374.

Elizalde, B. E., DeKanterewicz, R. J., Pilosof, A. M. R., and Bartholomai, G . B. 1988. Physico- chemical properties of food proteins related to their ability to stabilize oil-in-water emulsions. J . Food Sci. 53, 845-855.

Enser, M., Bloomberg, G . B., Brock, C. J., and Clark, D. C. 1990. De novo design and structure-activity relationships of peptide emulsifiers and foaming agents. Inf. J. Biol. Macromol. 12, 11 8-124.

Epand, R. M., Shai, Y., Segrest, J. P., and Anantharamaiak, G. M. 1995. Mechanisms for the

bonding and electrostatic interactions. In[. J . Pept. Prof. Res. 41, 499-511.

Biopolymers 37, 295-318.

281,240-244.

503-508.

modulation of membrane bilayer properties by amphipathic helical peptides. Biopolymers 37, 319-338.

Fairman, R., Armstrong, K. M., Shoemaker, K. R., York, E. J., Stewart, I. M., and Baldwin, R. L. 1991. Position effect on apparent helical propensities in the C-peptide helix. J. Mol. Bid. 221, 1395-1401.

Fezoui, Y., Weaver, D. L., and Osterhout, J. J. 1994. De novo and structural characterisation of an alpha-helical hairpin peptide: A model system for the study of protein folding intermediates. Proc. Nafl. Acad. Sci. USA 91, 367553679,

Fezoui, Y., Weaver, D. L., and Osterhout, J. J. 1995. Strategies and rationales for the de novo design of a helical hairpin peptide. Prof. Sci. 4, 286-295.

Fields, G. B., and Noble, R. L. 1990. Solid phase peptide synthesis utilizing 9-fluorenylmethoxycarbonyl amino acids. Inf. J . Pept. Prof. Res. 35, 161-214.

Fisher, L. R., and Parker, N. S. 1988. Effect of surfactants on the interactions between emulsion droplets. In “Advances in Food Emulsions and Foams” (E. Dickinson and G. Stainsby eds.), pp. 45-90. Elsevier, Amsterdam.

Foley, J., and O’Connell, C. 1990. Comparative emulsifying properties of sodium caseinate and whey protein isolate in 18% oil in aqueous systems. J . Dairy Res. 57, 377-391.

Forood, B., Feliciano, E. J., and Nambiar, K. P. 1993. Stabilisation of alpha-helical structures in short peptides via end capping. Proc. Natl. Acad. Sci. USA 90, 838-842.

Forood, B., Perez-Paya, E., Houghten, R. A,, and Blondelle, S. E. 1995. Formation of an extremely stable polyalanine P-sheet macromolecule. Biochem. Biophys. Rex Comrnun. 211, 7-13.


Friberg, S. E. 1976. Emulsion stability. In “Food Emulsions” (S. E. Friberg, ed.), pp. 1-37.

Goto. Y., and Hagihara, Y. 1992. Mechanism of the conformational transition of melittin.

Gutte, B., and Merrifield, R. B. 1969. The total synthesis of an enzyme with ribonuclease A

Habermann, E. 1972. Bee and wasp venoms. Science 177,314-322. Hague, Z., and Kito, M. 1983. Lipophilization of alpha S-1 casein: 1. Palmitoyl protein.

Hahn, K. W., Wieslaw, A,, Klis, A,, and Stewart, J. M. 1990. Design and synthesis of a peptide

Halling, P. J. 1981. Protein-stabilised foams and emulsions. CRC Crit. Rev. Food Sci. Nutr.

Hermans, J. 1993. Molecular dynamics simulations of helix and turn propensities in model peptides. Curr. Opin. Struct. Biol. 3, 270-276.

Hopp, T. P., and Woods, K. R. 1981. Prediction of protein antigenic determinants from amino acid sequences. Proc. Natl. Acad. Sci. USA 78,3824-3828.

Horovitz, A,, Matthews, J. M., and Fersht, A. R. 1992. Alpha-helix stability in proteins. 11. Factors that influence stability at an internal position. J. Mol. Biol. 227,560-568.

Inagaki, F., Shimado, I., Kawaguchi, K., Hirano, M. M., Terasawa, I., Ikura, T., and Go, N. 1989. Structure of melittin bound to perdeuterated dodecylphosphocholine micelles as studied by two dimensional NMR and distance geometry calculations. Biochemistry

Jackson, D. Y.. King, D. S., Chmielewski, J., Singh, S., and Schultz, P. G. 1991. General approach to the synthesis of short alpha-helical peptides. J. Am. Chem. SOC. 113,9391- 9392.

Marcel Dekker, New York.

Biochemisrry 31, 732-738.

activity. J . Am. Chem. SOC. 91, 501-502.

J . Agric. Food Chem. 32,1225-1236.

having chymotrypsin-like esterase activity. Science 248, 1544-1547.

13, 155-120.

28,5985-5991.

Jameson, B. A. 1989. Modelling in peptide design. Nature 341, 465-466. Jasanoff, A., and Fersht, A. R. 1994. Quantitative determination of helical propensities from

trifluoroethanol titration curves. Biochemistry 33, 2129-2135. Kachholz. T., and Schlingmann, M. 1987. Possible food and agricultural application of microbial

surfactants. In “Biosurfactants and Biotechnology” (N. Kosaric, W. L. Cairns, and N. C. C. Grey, eds.), pp. 183-210. Marcel Dekker, New York.

Kaiser, E. T., and Kezdy, F. J . 1987. Peptides with affinity for membranes. Annu. Rev. Biophys. Biophys. Chern. 16,561-581.

Kaminogowa. S., Shimizu, M., Ametani, A., Lee, S. W., and Yamauchi, K. 1987. Proteolysis in structural analysis of alpha S1-casein adsorbed onto oil surfaces of emulsions and improvement of the emulsifying properties of protein. J . Am. Chem. SOC. 64,1688-1691.

Karle, I. L., and Balaram, P. 1990. Structural characterizations of alpha helical peptide molecules containing a-aminobutyric acid. Biochemistry 29, 6747-6756.

Karle. I . L., Flipper-Andersen, J. L., Agarwalla, S., and Balaram, P. 1991. Crystal structure of zervamicin, a membrane ion-channel peptide: Implications for gating mechanism. Proc. Natl. Acad. Sci. USA 88, 5307-5311.

Kim, C. A,, and Berg, J. M. 1993. Thermodynamic /%sheet propensities measured using a zinc-finger host peptide. Nature 362, 267-270.

King, T. P., Wade, D., Coscia. M. R., Mitchell, S., Kochoumian, L., and Merrifield, B. 1994. Structure and immunogenic relevance of melittin, its transposed analogues and D-melittin. J . Immunol. 153, 1124-1131.

Klemaszewski, J. L., Das, K. P., and Kinsella, J. E. 1992. Formation and coalescence stability of emulsions stabilized by different milk proteins. J . Food Sci. 57, 366-372.


Knoppel, E.. Eisenberg, D.. and Wickner, W. 1979. Interaction of melittin, a preprotein model, with detergents. Biochemistry 18,4177-4181.

Kuroda, Y. 1995. A strategy for the de novu design of helical proteins with stable folds. Prof. Eng. 8,286-295.

Lang, S., and Wagner, F. 1987. Structure and properties of biosurfactants. In “Biosurfactants and Biotechnology” (N. Kosaric, W. L. Cairns, and N. C. C. Grey, eds.), pp. 21-45. Marcel Dekker, New York.

Lattman, E. E., and Rose, G . D. 1993. Protein folding-What’s the question? Proc. Nnfl. Acnd. Sci. USA 90,439-441.

Leerwakers, F. A. M., Atkinson, P. J., Dickinson, E., and Horne, D. S. 1996. Self-consistent- field modelling of adsorbed p-casein: Effects of pH and ionic strength on surface coverage and density profile. J . Colloid Interface Sci. 178, 681-693.

Lesk, A. M. 1991. ”Protein Architecture.” IRL Press, Oxford. Li, S. C., and Deber, C. K. 1992. Glycine and P-branched residues support and modulate

peptide helicity in membrane environments. FEBS Lett. 311, 217-220. Li, S. C., Kim, P. K., and Deber, M. 1995. Manipulation of peptide conformation by fine-

tuning of the environment and/or primary sequence. Biopulymers 35, 667-675. Lyu, P. C., Liff, M. I., Marky, L. A,, and Kallenbach, N. R. 1990. Side chain contributions to

the stability of alpha-helical structure in peptides. Science 250, 669-673. Lyu, P. C., Sherman, J. C., Chen, A., and Kallenbach, N. R. 1991. Alpha-helix stabilisation

by natural and unnatural amino acids with alkyl side-chains. Proc. Nntl. Acad. Sci. USA 88,5317-5320.

Lyu, P. C., Wemmer, D. E., Zhou, H. Z.. Pinker, R. J., and Kallenbach, N. R. 1993. Capping interactions in isolated a-helices: Position-dependent substitution effects and structure of a serine-capped peptide helix. Biochemistry 32, 421-425.

Macritchie, F. 1978. Proteins at interfaces. Adv. Prof. Chem. 39, 51-124. Marqusee, S., and Baldwin, R. L. 1987. Helix stabilisation by Glu- . . . . . Lys’ salt bridges

in short peptides of de nuvo design. Proc. Nat. Acad. Sci. U S A 84,8898-8902. McWatters, K. H., and Holmes, M. R. 1979a. Salt concentration, pH and flour concentration

effects on nitrogen solubility and emulsifying properties of peanut flour. J . Food Sci. 44,765-785.

McWatters, K. H., and Holmes, M. R. 1979b. Influence of pH and salt concentration on nitrogen solubility and emulsification properties of soy flour. J . Food Sci. 44, 765-780.

Mercier, J.-C., Groscloude, F., and Ribadcau-Dumas, B. 1971. Structure primaire de la caseine aS1-bovine. Sequence complete. Eur. J. Biochem. 23, 41-51.

Meienhofer, J. 1985. Protected amino acids in peptide synthesis. In “Chemistry and Biochemis- try of the Amino Acids” (G. C. Barrett, ed.), pp. 297-337. Chapman and Hall, London.

Merrifield. R. B. 1963. Solid phase synthesis 1: The synthesis of a tetrapeptide. J . Am. Chem. Soc. 85, 2149-2154.

Merrifield, R. B., Juvvadi. P., Andreu, D., Ubach, J., Boman, A,, and Boman, H. G. 1995. Retro and retroenantio analogs of cecropin-melittin hybrids. Proc. Nail. Acad. Sci. U S A

Minor, D. L., and Kim, P. S. 1994. Measurement of the P-sheet-forming propensities of amino acids. Nature 367, 660-663.

Mollby, R. 1983. Isolation and properties of membrane damaging toxins. In “Staphylococci and Staphylococcal Infections” (C. S. F. Easmon and C. Adlam, eds.), Vol. 2, pp. 619-669.

Moser, R. 1992. Design, synthesis and structure of an amphipathic peptide with pH-inducible haemolytic activity. Prof. Eng. 5, 323-331.

Munoz, V., and Serrano, L. 1994. Elucidating the folding problem of helical peptides using empirical parameters. Nutiire. Strctct. Biol. 1, 399-409.

92,3449-3453.


Nakai, S. C. 1983. Structure-function relationships of food proteins with an emphasis on the

Noren, C. J . , Anthony-Cahill, S. J. , Griffith, M. C., and Schultz. P. G. 1989. A general method

Olsen. A. J. , and Goodsell. D. S. 1992. Macromolecular graphics. Curr. Opin. Sfruct. Biol.

Padmanabhan, S.. and Baldwin, R. L. 199 1. Straight-chain-nonpolar amino acids are good

Padmanabhan, S., Marqusee, S., Ridgeway. T., Laue, T. M.. and Baldwin, R. L. 1990. Relative

Parker, N. S. 1988. Properties and functions of stabilizing agents in food emulsions. CRC Crit.

Patek. M. 1993. Multistep deprotection for peptide chemistry. Int. J . Pept. Prot. Res. 42,97-117. Perez-Paya, E., Houghten, R. A., and Blondelle, S. E. 1994. Determination of the secondary

structure of selected melittin analogues with different haemolytic activities. Biochem. J . 299,587-59 I .

Ponder, J . W., and Richards, F. M. 1987. Tertiary templates for proteins: Use of packing criteria in the enumeration of allowed sequences for different structural classes. J . Mol. Biol. 193, 775-791.

importance of protein hydrophobicity. J . Agric. Food Chem. 31,676-682.

for incorporation of unnatural amino acids into proteins. Science 244, 182-188.

2, 193-201.

helix-formers in water. J . Mol. Biol. 219, 135-137.

helix forming tendencies of non-polar amino acids. Nature 344, 268-270.

Rev. Food Sci. Nutr. 25, 285-296.

Presta, L. G., and Rose, G. R. 1988. Helix signals in proteins. Science 240, 1632-1641. Provencher. S. W.. and Glockner. J. 1981. Estimation of globular protein secondary structure

from circular dichroism. Biochemistry 20, 33-37. Ramage, R., Green, J. , and Ogunjobi, 0. M. 1989. Solid phase peptide synthesis of ubiquitin.

Tetrnhedron Lett. 30, 2149-2152. Rivett, D. E.. Kirkpatrick, A.. Hewish, D. R.. Reilly. W., and Werkmeister, J. A. 1996.

Dimerization of truncated melittin analogues results in cytolytic peptides. Biochem. J . 316,525-529.

Rose, G. D., and Creamer, T. P. 1994. Protein folding: Predicting predicting. Prot. Struct. Funct. Genet. 19, 1-13.

Saito, M., Chikuni, K., Monma, M., and Shimizu, M. 1993. Emulsifying and oil-binding properties of bovine serum albumin and its enzymatic hydrolyzate. Biosci. Biotechnol. Biochem. 57, 952-956.

Saito, M., Ogasawara. M.. Chikuni. K.. and Shimizu, M. 1995. Synthesis of a peptide emulsifier with an amphiphilic structure. Biosci. Biotechnol. Biochem. 59, 388-392.

Scheraga, H. A. 1978. Use of random copolymers to determine the helix-coil stability constants of the naturally occurring amino acids. Pure Appl. Chem. 50, 315-324.

Schnolzer, M.. and Kent, S. B. H. 1992. Constructing proteins by dovetailing unprotected synthetic peptides: Backbone engineered HIV protease. Science 256,221-225.

Scholz, J. M., and Baldwin. R. L. 1992. The mechanism of alpha-helix formation by peptides. Annu. Rev. Biophys. Biomol. Struct. 21, 95-1 18.

Segrest, J. P.. Garber, D. W., Brouillette, C. G., Harvey, S. C., and Anantharamaiah, G. M. 1994. The amphipathic alpha helix: A multifunctional structural motif in plasma apolipoproteins. Adv. Prot. Chem. 45, 303-369.

Serrano, L., Neira, J . L., Sancho, J . , and Fersht, A. R. 1992. Effects of alanine versus glycine in alpha-helices on protein stability. Nature 356, 453-455.

Shimizu, M.. Lee, S. W., Kaminogawa, S. , and Yamauchi. K. 1984. Emulsifying properties of an N-terminal peptide obtained from the peptic hydrolyzate of aS1-casein. J . Food Sci. 49, 1117-1 120.

Shimizu, M., Lee. S. W., Kaminogawa, S., and Yamauchi, K. 1986. Functional properties of a peptide of 23 residues purified from the peptic hydrolyzate of as1 casein: Changes in the emulsifying activity during purification of the peptide. J. Food Sci. 51, 1248-1252.


Shimizu, M., and Nakane, Y. 1995. Encapsulation of biologically active proteins in a multiple emulsion. Biosci. Biotechnol. Biochem. 59,492-496.

Shimizu, M., Takahashi, T., Kaminogawa, S., and Yamauchi, K. 1983. Adsorption onto an oil surface and emulsifying properties of bovine aS1-casein in relation to its molecular structure. J . Agric. Food Chem. 31, 1214-1218.

Shoemaker, K. R., Kim, P. S., and Baldwin, R. L. 1982. Nature of the charged-group effect on the stability of the C-peptide helix. Proc. Nut/. Acad. Sci. USA 82, 2349-2353.

Shoemaker, K. R., Kim, P. S., York, E. J., Stewart, J. M., and Baldwin, R. L. 1987. Tests of the helix dipole model for stabilisation of alpha-helices. Nature 326, 563-567.

Siegel, J. B., Steinemetz, W. E., and Long, G. L. 1980. Acomputer-assisted model for estimating protein secondary structure from circular dichroism spectra. Comparison of animal lactate dehydrogenase. Anal. Biochem. 104, 160-167.

Smith, L. J., and Clark, D. C. 1992. Measurement of the secondary structure of adsorbed protein by circular dichroism. 1. Measurements of the helix content of adsorbed melittin. Biochim. Biophys. Acta 1121, 111-118.

Sonnichsen, F. D., Van Eyk, J. E., Hodges, R. S., and Sykes, B. D. 1992. Effect of trifluoroetha- no1 on protein secondary structure: An NMR and circular dichroism study using a synthetic actin peptide. Biochemistry 31, 8790-8798.

Stewart, J. M., and Young, J. D. 1984. The chemistry of solid phase peptide synthesis. In “Solid Phase Peptide Synthesis,” pp. 1-85. Pierce Chemical Co., Rockford, IL.

Sun, E., and Cohen, F. E. 1993. Computer-assisted drug discovery-A review. Gene 137,

Sutcliffe, M. J., Hayes, F. R., and Blundell, T. L. 1987. Knowledge-based modelling of homologous proteins. Part 11: Rules for the conformations of substituted side chains. Prot. Eng.

Talbot, J. C., Dufourq, J., de Bony, J., Faucon, J. F., and Lussan, C. 1979. Conformation

Tanford, C. 1980. “The Hydrophobic Effect.” Wiley, New York. Taylor, J. W., and Kaiser, E. T. 1987. Structure-€unction analysis of proteins through the

design, synthesis and study of peptide models. Meth. Enzymol. 154,473-498. Taylor, J. W., Shih, I. L., Lees, A. M., and Lees, R. S. 1993. Surface-induced conformational

switching in amphiphilic peptide segments of apolipoproteins B and E and model peptides. Int. J . Pept. Prot. Res. 41, 536-547.

Terwilliger, T. C., and Eisenberg, D. 1982a. The structure of melittin. 1 . Structure determination and partial refinement. J. Biol. Chem. 257,6010-6015.

Terwilliger, T. C., and Eisenberg, D. 1982. The structure of melittin. 2. J. Biol. Chem. 257,6016- 6022.

Tornberg, E. 1978. Functional characterisation of protein-stabilized emulsions: Emulsifying behavior of proteins in a valve homogenizer. J. Sci. Food Agric. 29,867-879.

Vogel, H., and Jaehnig, F. 1986. The structure of melittin in membranes. Biophys. J . 50,

Wade, D., Andreu, D., Mitchell, S. A., Silviera, A. M. V., Boman, A., Boman, H. G., and Merrifield, R. B. 1992. Antibacterial peptides designed as analogues or hybrids of cecrop- ins and melittin. Int. J . Pept. Prot. Res. 40, 429-436.

Wang, J. C., and Kinsella, J. E. 1976. Functional properties of novel proteins: Alfalfa leaf protein. J. Food Sci. 41, 268-292.

Wojcik, J., Altman, K. H., and Scheraga, H. A. 1990. Helix-coil stability constants for naturally- occuring amino acids in water. XXIV. Half-cystine parameters from poly(hydroxybuty1- glutamine)-co-S-methylthio-L-cysteine. Biopolymers 30, 121-1 34.

127-132.

1,385-392.

change and self-association of monomeric melittin. FEBS Lett. 102, 191-193.

573-582.


Wolfenden, R., Anderson, L., Cullis, P. M., and Southgate, C. C. 1981. Affinities of amino acid side chains for solvent water. Biochemistry 20, 849-855.

Zhang, S., Holmes, T., Lockshin, C., and Rich, A. 1993. Spontaneous assembly of a self- complementary oligopeptide to form a stable macroscopic membrane. Proc. Nutl. Acud. Sci. USA 90,3334-3338.

Zhong, L., Putnian, R. J., Johnson, W. C., Jr. , and Rao, A. G. 1995. Design and synthesis of amphipathic antimicrobial peptides. Inr. J. P e p . Prof. Res. 45, 337-347.

Zhou, N. E.. Kay, C. M., Sykes, B. D., and Hodges, R. S. 1993. A single-stranded amphipathic alpha-helix in aqueous solution: Design, structural characterisation and its application for determining a-helix propensities of amino acids. Biochemistry 32, 6190-6197.

Zimm, B. H., and Bragg, J. K. 1959. Theory of the phase transition between helix and random coil in polypeptide chains. J. Chern. fhys. 31, 526-535.

ADVANCES IN FOOD AND NUTRITION RESEARCH, VOL. 42

X-RAY DIFFRACTION OF FOOD POLYSACCHARIDES

RENGASWAMI CHANDRASEKARAN

Whistler Center for Carbohydrate Research Purdue University

West Lafayette, Indiana 47907

1. Introduction A. Importance of Food Polysaccharides B. Structure-Function-Property Correlations Basic Principles of Solving Three-Dimensional Structures A. B. C. D.

111. Molecular Shapes and Interactions A. (1+4)-Linked Polysaccharides B. (1+3)-Linked Polysaccharides C. Alternating (1+3)- and (1+4)-Linked Polysaccharides D. The Gellan Family of Polysaccharides E. Branched Polysaccharides F. Arabinan

11. Single Crystals, Oriented Fibers and Powder Specimens X-Ray Diffraction from Ordered Molecules Structural Analysis of Helical Polysaccharides Model Building and Refinement Techniques

IV. Mixed Polysaccharides

VI. Summary V. Morphology to Macroscopic Properties

References

I. INTRODUCTION

A. IMPORTANCE OF FOOD POLYSACCHARIDES

Carbohydrates are molecules composed of carbon, hydrogen, and oxygen atoms. They usually denote simple monosaccharides such as glucose and galactose, disaccharides such as sucrose and lactose, a spectrum of low- molecular-weight oligosaccharides, and a large number of high-molecular- weight polysaccharides that encompass starch and glycogen. In one form

131

1043-4526198 $25 011 Copyright 0 1998 by Academic Press.

All rights of reproduction in any fnrm reserved.

132 RENGASWAMI CHANDRASEKARAN

or another, carbohydrates constitute an important component of human diet: breakfast cereal, bread, snack, and dessert included. Polysaccharides are abundant in plant life. Some examples are cellulose, xylan, glucan, xyloglucan, starch, and pectate. Carrageenan and alginate are copiously produced from marine algae. Likewise, xanthan, curdlan, and gellan are of bacterial origin. Primarily, cellulose is water insoluble, but the rest are water soluble and hence referred to as hydrocolloids. Their aqueous solutions often are either viscous or able to form gels. Such physical properties can be controlled by the addition of co-solutes, by temperature, and also by the amount and type of cations in the case of anionic polymers. If viscosity is exploited for thickening soups, gravies, sauces, and salad dressings, gelation is useful for desserts, confectionery, jams, jellies, and pet foods. Overall, polysaccharides as minor ingredients, i.e., used in small quantities, are essential for texture, stability, and shelf life of a very large number of food products. Polysaccharides of plant, algal, and bacterial origin have so far found excellent practical applications in the food industry (Whistler and BeMiller, 1993, 1997; Ross-Murphy, 1994; Stephen, 1995; Eliasson, 1996). The focus of this chapter is the molecular architecture of these polysaccharides as related to their observed functional properties, which may result from junction zone formation or through specific interactions with the surrounding molecules, as determined by the polymer shapes.

B. STRUCTURE-FUNCTION-PROPERTY CORRELATIONS

Proteins, nucleic acids, and polysaccharides are the three major classes of biopolymers, the building blocks of which are amino acids, nucleotides, and saccharides, respectively. Both proteins and nucleic acids are linear polymers. The polymerization is always from the N-terminal to the C- terminal end in proteins that are composed of 20 different a-L-amino acids, from alanine to lysine, having distinct side chains in terms of size and charge. Depending on the composition and sequence of amino acids, the polypeptide backbone of a protein chain folds a certain way to form a unique molecular shape. In most instances, protein molecules whose functions range from enzyme activity, to polymerization of polynucleotides, to carbohydrate binding, to name a few, take up compact globular forms. In general, the polypeptide chain morphology utilizes a-helical segments, pleated P-sheets and reverse turns, in conjunction with a few loops and some other nonrepetitive conformations in a well-orchestrated combina- torial manner as dictated by its sequence and environment. On the other hand, a spectrum of muscle proteins, collagen, actin, and myosin, adopt long helical forms compatible with their function as structural proteins (Branden and Tooze, 1991).

X-RAY DIFFRACTION OF FOOD POLYSACCHARIDES 133

The skeleton of nucleic acids is the sugar-phosphate backbone. It always proceeds from the 5' to the 3' end, the five-membered sugar ring being a 0-D-ribose in RNA and a similar 2'-deoxyribose in DNA. One of the four nitrogenous bases (adenine, guanine, cytosine, and thymine or uracil) is a side chain attached to the sugar ring through a glycosyl bond from atom C1 '. While DNAs and some RNAs have a preponderance to form antiparallel double helices stabilized by Watson-Crick base pairs, messenger RNAs are known to adopt folded structures with loops and turns, and transfer RNAs occur as fully compact molecules (Saenger, 1984).

Unlike the unidirectional chemical links connecting successive residues in proteins and nucleic acids, carbohydrates utilize multidirectional links. For example, every pyranosyl unit in a hexopyranose-containing polysaccharide may contain up to four (1-n) links, n being 2, 3, 4, or 6. This versatility results in complex branch-on-branch structures similar to that of glycogen that contains only (1-4) and (1-6) links. Further, both L-

and D-saccharides occur commonly; and in each case both a- and p- conformers exist. This quadruples the number of available saccharide units as building blocks and enables generation of a large variety of structures. Consequently, it is believed that carbohydrates including polysaccharides are excellent molecules for storage of biological and structural information, perhaps even superior to proteins and nucleic acids.

The shapes of biopolymers at the atomic level, as well as the interactions among biopolymers and the molecules in the environment, are the major determinants of observed functional behavior. Thus, a knowledge of the three-dimensional structures is extremely helpful for understanding the molecular basis of the useful and functional properties. In the case of food polysaccharides, this information is important for enhancing the desired properties at a relatively low cost by chemical modification, addition of co- solutes, or other options.

It. BASIC PRINCIPLES OF SOLVING TH R EE-DI MENSIONAL STRUCTURES

The spatial arrangement of atoms gives a recognizable shape that is unique to each molecule. There are several experimental techniques that involve the interaction of radiation with atoms in molecules and provide structural details to varying extents. For example, solution light scattering could lead to estimates of the molecular weight and macroscopic dimensions such as the end-to-end distance and radius of gyration (Brownsey et al., 1984; Burchard, 1994; Gunning and Morris, 1990). Ultraviolet and infrared spectroscopy are helpful for diagnosing the presence of ordered structure


and the existence of hydrogen bonds, if any, in the polymer (Arndt and Stevens, 1996; Crescenzi et al., 1987; Morris, 1994). Imaging by electron microscopy can be used to probe peripheral morphologies to a low resolution of about 100 A, which is relevant for the visualization of gel networks (Hermansson and Langton, 1994). Specifically, the scanning tunneling microscope and the atomic force microscope are gaining momentum in imaging the surfaces of macromolecules at a slightly better resolution (Kirby and Morris, 1994). Since the interatomic distances in a molecule are of the order of an angstrom, diffraction experiments employing electrons, neutrons, and x-rays as radiation sources whose wavelengths are in the region of interest are the tools of choice for exploring the relative positions of atoms in molecules. Among the three, x-ray diffraction is the most powerful experimental technique to date for the determination of three- dimensional structures of biopolymers at high resolution. The fundamentals of x-ray diffraction and results on selected food-related polysaccharides and proteins are reviewed by Clark (1994).

X-ray diffraction is at its best when the polymer specimen is in the crystalline solid state. In other words, it is necessary to induce organization of molecules in a periodic fashion in all three principal directions in space so that the signal-to-noise ratio is high and measurable. Disruption of periodicity as in semi- and noncrystalline regions, or lack of organization within or between molecules as in amorphous regions and in the solution state, is detrimental to the success of the diffraction method.

Like all biopolymers, polysaccharides are living molecules. X-ray diffraction in the solid state helps to elucidate the preferred molecular structure and also the packing arrangement. Most linear polysaccharides consisting of simple repetitive sequences prefer to adopt helical conformations. Unless influenced by external forces in the surroundings, barring minor perturba- tions, these helical structures are quite stable, particularly over short distances, in solid, solution, and gel states. Gels are composed of an intricate, randomly linked, network of junction zones having tremendous water hold- ing capacity. Each junction zone is a pillar of strength such that a gel can stand on its weight, but a sol cannot. To form a junction zone takes few, say 2 to 10, helices with short-range lateral organization, the details of which are revealed by the three-dimensional structure. The rest of this chater is devoted to the determination and description of the anatomies of polysaccharide helices that are related in some way to food systems.

A. SINGLE CRYSTALS, ORIENTED FIBERS, AND POWDER SPECIMENS

In order to understand some fundamental differences among the best experimental specimens that can be prepared in the laboratory for x-ray


diffraction study, it is necessary to have a good idea about the primary building block, known as the unit cell, of the specimen. The parallelepiped enclosing one or more molecules is called a unit cell. The unit cell shown in Fig. 1 contains a disaccharide molecule. In general, six parameters are needed to define the unit cell: the lengths of the three principal edges a, b, and c in angstrom and the three interaxial angles a, p, and y in degrees. A large number of building blocks stacked periodically side to side in three dimensions generates a macroscopic single crystal (Fig. 2). The seven crystal systems based on the unit cell dimensions are listed in Table I. Most of the food polysaccharide structures belong to monoclinic, orthorhombic, tetragonal, and hexagonal systems.

Slow evaporation of its saturated solution is a successful approach to growing single crystals for small molecules such as glucose, sucrose, maltose, and some oligopeptides, oligosaccharides, and oligonucleotides. This technique is equally applicable for globular macromolecules, which include proteins and spherical viruses. In general, production of single crystals from concentrated polymer solutions at controlled temperature, pH, humidity, and the like requires almost isotropic growth rate in three dimensions.

/y

- /

b

FIG. 1. Schematic drawing of a unit cell whose origin is point 0. The volume of the unit cell, defined by lengths a, b, and c and angles a (between b and c), /3 (between c and a) , and y (between n and b ) , is given by V = abc (1 - cos’a - cos2/3 - cos2y + 2 cosa cosp cosy)”’. The c-axis coincides with the helix axis of the polysaccharide.


FIG. 2. Periodic stacking of unit cells in three dimensions generates a single crystal.

Unfortunately, crystalline growth of helix-forming polysaccharides is much faster along the longitude than along the transverse direction. This differential development limits the lateral dimensions of crystallites to only a few unit cells and impedes the formation of single crystals. However, under

TABLE I THE SEVEN CRYSTAL SYSTEMS

System Sides Angles

Triclinic a # b f c f f # P f Y

Orthorhombic a f b f c (y = p = y = $lo"

Tetragonal a = b # c a = p = y = 9 0 "

Monoclinic a f b f c (Y = = 90"; y # 90"

Trigonal a = b = c (Y = p = y # 90" Hexagonal a = h f c (Y = p = 90"; y = 120" Cubic a = b = c = p = = 900


appropriate crystallizing conditions, polycrystalline fibers or films can be obtained. In the case of fibers, a few drops of the polymer solution are placed in the gap of a couple of millimeters between two glass rods in a fiber puller, allowed to dry slowly until a semisolid state is reached, and then gradually stretched to twice or more the original length so that preferential parallelism among several microcrystallites along the molecular axis is induced (Chandrasekaran et al., 1994b). In the case of films, large drops of the polymer solution are placed on Teflon blocks and allowed to dry until films are formed. Since orientation is nonexistent, narrow strips of the required size (about 2 X 5 mm) are cut and each is suspended in a humidity chamber under hanging weight (a few grams), which helps to invoke preferred alignment due to stretching to almost twice the original length of the strip over a few hours. The extent of longitudinal and lateral organization among the unit cells in such an oriented and polycrystalline specimen is schematically represented in Fig. 3a. This clearly shows that crystallites are nearly aligned parallel along the longitude, but they are rotated to different extents about this axis. Hence, a stationary fiber is considered to be equivalent to a rotating single crystal. On the other hand, when the lateral dimensions of microcrystallites do not exceed a few unit cells, stacking of microcrystallites along the molecular axis dominates (Fig. 3b) and this gives rise to an oriented and noncrystalline specimen. If the microcrystallites in the specimen are randomly oriented in space, the stretching procedure serves no useful purpose at all (Fig. 3c); this is often referred to as a powder sample. In principle, therefore, fibrous polymers may exhibit “polycrystallinity and orientation” to varying extents. The primary goal in specimen preparation is to achieve the most of both in order to collect good quality x-ray data suitable for structure analysis. However, the degree of ordering that can be induced in biopolymer fibers extends over a wide range encom- passed by the three distinct types depicted in Fig. 3.

B. X-RAY DIFFRACTION FROM ORDERED MOLECULES

When a narrow beam of x-rays produced by a sealed tube, microfocus or rotating anode, generator is incident on a fiber, electrons surrounding the nuclei in atoms scatter x-rays in specific directions that are dependent on the unit cell dimensions (Fig. 1). This phenomenon is referred to as x- ray diffraction. If the unit cells are stacked as in a single crystal (Fig. 2) , the diffraction consists of a series of reflections as shown in Fig. 4. They all conform to the Bragg’s law 2d sin 6’ = nh, where d is an interplanar spacing, 26’ is the angle between the incident and diffracted beams, n is a positive integer, and A is the wavelength of x-rays. Note that 6’ is known as the Bragg angle. Each reflection corresponds to a specific d-spacing that


n

FIG. 4. Bragg reflections of varying intensities are related to the unit cell dimensions.

A section of the diffraction pattern from a single crystal. The positions of discrete

is a function of the six unit cell parameters. The intensity of a reflection is distinct and related to the spatial arrangement of atoms in the specimen. The distribution of intensities in a diffraction pattern is commensurate with both the molecular shape and the packing arrangement within and between unit cells. X-ray crystallography deals with the determination of crystal structure from the experimentally measured intensities. It is a common practice to use nickel-filtered CuKa radiation of x-rays (A = 1.5418 A) in most crystallographic investigations.

All the fundamental principles of crystallography are fully valid in the case of fibers and films relevant for the polysaccharide structures. Due to a lesser degree of organization described earlier, the number of independent Bragg reflections possible for these systems are unfortunately far less than those from single crystals. Some special care (described in the next section), therefore, must be taken to solve fiber structures. In general, fiber diffraction

FIG. 3. Organization of unit cells in less-ordered systems: (a) The crystallites are preferentially aligned along, and differentially oriented about, the fiber axis in a polycrystalline and orientated specimen, (b) crystallites have preferred orientation but little lateral organization in a noncrystalline and oriented specimen, and (c) microcrystallites are randomly oriented in a powder specimen.


FIG. 5. X-ray fiber diffraction patterns from samples displaying (a) random orientations of microcrystallites as in arabinan, (b) noncrystallinity and orientation for the sodium salt of S- 657, and (c) polycrystallinity and orientation for the calcium salt of welan.


patterns are recorded on photographic films in a flat-film camera, typical fiber-to-film distance being 4 cm. During exposure to x-rays, which could range from a few hours to days, a slow and steady stream of helium gas is first bubbled through a chosen saturated salt solution and then flushed through the specimen chamber. This enables not only retention of the fiber at a desired constant relative humidity (r.h.), but also reduction of the otherwise film-fogging due to air scattering.

The diffraction patterns from three different types of polysaccharide fibers shown in Fig. 5 highlight the individual and joint influence of the two parameters, polycrystallinity and orientation, on the x-ray data. When the microcrystallites are randomly oriented as in Fig. 3c, the pattern consists of a series of concentric rings (Fig. 5a) corresponding to the d-spacings characteristic of the unit cell of the specimen. The intensity is the same within a ring, but it differs among rings. An oriented, noncrystalline sample depicted in Fig. 3b, on the other hand, gives rise to a series of layer lines normal to the fiber axis. As seen in this diffraction pattern (Fig. 5b), the intensity varies along each layer line. The d-spacing of the first layer line

FIG. 5. Continued


represents the c-dimension of the unit cell. In the case of an oriented and polycrystalline specimen as in Fig. 3a, the diffraction pattern displays an array of sharp Bragg reflections (Fig. 5c) at discrete positions where rings intersect layer lines.

Some frequently used terms in fiber diffraction are sketched in Fig. 6. The meridian is the vertical line passing through the center of the pattern where the incident x-ray beam strikes the film and it is parallel to the fiber axis. The horizontal lines are called layer lines, the numbers of which start at 1 = 0 on the equator passing through the center and increase upward and decrease downward. The occurrence of meridional reflections only on 1 = 2 5 means that the polymer exists as a fivefold helix of pitch c. The two helical parameters, N the number of repeats per turn and h the axial rise per repeat, in this example, are 5 and c/5, respectively. Alternately, t the turn angle per repeat is 360/N in degrees. By convention, t is positive for right-handed and negative for left-handed helices, and h is always positive. Any stereochemically satisfactory model for consideration must satisfy the observed helical parameters.

In terms of diffraction geometry, it is sufficient to say that the positions of reflections and their distances from the center of the pattern are dependent on the fiber-to-film separation and the unit cell dimensions. This distance information is used not only to determine the unit cell dimensions,

FIG. 6. Some important terms used in fiber diffraction refer to the equator and meridian, which are the horizontal and vertical lines, respectively, passing through the center of the pattern. Each layer line is roughly parallel to the equator and contains Bragg reflections.


but also to assign a set of three integers, h, k and 1, as Miller indices (1 being the layer line number) that uniquely specify the diffraction direction of each reflection. Subsequently, the intensities of reflections are measured using a microdensitometer or by digitizing with an optical film scanner and then converted into structure amplitudes after applying appropriate geometrical corrections for Lorentz and polarization effects (Millane and Arnott, 1985a,b; Millane, 1989). Note that intensity is proportional to the square of amplitude. This experimental x-ray data set is the key to determining and refining the molecular and packing models and to conducting objective comparisons among alternatives.

Stout and Jensen (1989), for instance, have elegantly described the details of x-ray crystallography to solve structures. The basic principles are mentioned here briefly for convenience. The structure factor F, for reflection h, k, 1 is the resultant of M waves scattered by the M atoms in the unit cell. Each wave has an amplitude proportional to the respective atomic scattering factor and a phase E relative to the unit cell origin. The structure factor is hence calculated using the expression

F,. (h k 1 ) = xfi exp 2ni (hxj + ky j + lz,), (1) i

where f i is the atomic scattering factor of the jth atom positioned at frac- tional coordinates x,, yj, and z, in the unit cell and all M atoms in it are included in the summation. The scattering factor depends on Bragg angle 8 and, starting from a value of 2 (atomic number) at 8 = 0, f decreases as 8 increases. Equation (1) can be rewritten in the form

F, (h k I) = IF, (h k f)lexp i6 = A + iB, (2)

which gives the structure amplitude IF, (h k 1)1 = (A2 + B2)”* and phase of the reflection 6 = tan-’ BIA.

Alternately, if amplitudes and phases are known for all reflections, the electron density p at any point x, y, z in the unit cell of volume V can be calculated by using the Fourier summation

p (x, y , z ) = (1/V) F (h k 1 ) exp -2ni (hx + ky + lz ) . (3) h k l

An important extension of Eq. ( 3 ) is called the difference Fourier or difference electron density ( A p ) map, which can be computed and used to locate missing atoms or relocate wrongly placed atoms. In this synthesis,


the coefficients are AF = (IFo/ - IF,[) exp is, where 6 is based on the current model, and IF, land IF,/ are the observed and calculated structure amplitudes, respectively. Thus,

Ap (x, y, z) = (W) 2x2 AF (h k 1) exp -2ni (hx + ky + Iz). (4) h k 1

C. STRUCTURAL ANALYSIS OF HELICAL POLYSACCHARIDES

Even the best fiber diffraction pattern seldom contains over 100 reflections up to about 3 A resolution, in contrast to a couple of thousand reflections from a single crystal having comparable unit cell dimensions. This drastic reduction is due to the following reasons. First, since the fiber is equivalent to a rotating crystal, systematic superposition of equivalent reflections, from the four hk-quadrants (++, +-, -+ and --), as in an orthorhombic system, on any layer line 1 gives rise to a single spot instead of four. Second, fortuitous overlap of two or more independent reflections with different intensities having the same 1 and the same Bragg angle 0 cannot be avoided. Third, a few adjacent reflections might be too close to separate analytically. Under these circumstances, all overlapping reflections have to be considered individually in structure factor calculation and compounded properly for comparison with the observed composite reflection. Unobserved reflections that are too weak to see are assigned threshold values based on the lowest measured intensities. Still, the available x-ray 'data alone are far less than the number of atomic coordinates in a repeat of the helix, and hence inadequate to solve a fiber structure.

This pitfall is somehow overcome by appropriately embedding stereochemical information derived from the crystal structures of related monomers into the polymer as well (Scott and Arnott, 1972). This includes bond lengths, bond angles, and some conformation angles. On this basis, an average or a standard geometry for a sugar ring is an excellent starting point for polysaccharide structure analysis. Thus, when the ring shape in a monosaccharide is known, only three conformation angles and the bond angle at the glycosyl bridge oxygen atom between the sugar rings are required to describe the helix geometry, instead of 33 coordinates for 11 nonhydrogen atoms as variables. The conformation angles are 4 and $ around the glycosyl bonds and x, which defines the orientation of hydroxymethyl group about the C5-C6 bond. An order of magnitude reduction in the number of variables significantly increases the data-to-parameter ratio and fiber structure analysis becomes meaningful.


D. MODEL BUILDING AND REFINEMENT TECHNIQUES

Linked-atom least-squares (LALS) analysis (Smith and Arnott, 1978) and the variable virtual bond (PS79) method (Zugenmaier and Sarko, 1980) are two major computer programs developed on the same basic principles for the generation and refinement of helical structures using x-ray data from fibers. The LALS program has been instrumental in unraveling well over 100 structures including polysaccharides, polypeptides, polynucleotides, and polyesters. This program requires information on precursor atom, scattering type, bond length, bond angle, and conformation angle to describe a tree geometry for every atom in one repeat plus three adjoining atoms in the next repeat. The extra three-atom plane is essential for establishing helix connectivity. In addition to refining the main chain and other relevant conformation angles, the positioning of the repeat is simultaneously adjusted by refining three Eulerian angles and a distance from the helix axis (z-axis) for a conveniently chosen root atom, until adjacent repeats conform to the desired N and h values of the helix.

The function minimized by the LALS program is given by

= x + Y + E + C + L . (6)

The first term ( X ) on the right-hand side accounts for the sum of squares of the differences AF,,, between observed (F,) and calculated (F,) x-ray structure amplitudes of Bragg reflections. The second term ( Y ) accounts for the sum of squares of the differences AT,, between observed (To) and calculated (T,) x-ray structure amplitudes of continuous diffraction. Either or both terms can be used as necessary. The third term ( E ) minimizes the differences A Oi between expectedktandard values ( 0,) and corresponding conformation and bond angles (Oc) in the model. The fourth term (C) includes both intra- and interchain hydrogen bonds, and the differences Acj between acceptable (do) and calculated nonbonded distances (d,) for those contacts that are smaller than the acceptable limiting values; this is designed to keep the model free from steric compression. The weights associated with these four types of observations are w , u, e, and ki, respectively. Finally, the fifth term imposes constraints ( Gh, with Lagrange multipliers Ah) for ring closure and helix connectivity, and it reaches zero when the constraints are obeyed. For an unobserved reflection to be included in the refinement, its F, must be larger than F,.

Alternative molecular models to be examined in order to select the best include (a) both right- and left-handed helices and (b) single as


well as multistranded helices with parallel and antiparallel strands. The packing arrangement in the unit cell is considered next. If two or more helices can account for the measured fiber density, their positions, orientations and relative polarities are individually tested. Hamilton’s significance test is used to assess the relative merits of competing models on the basis of a, E, X Y, or C, or in terms of the crystallographic R-values, R = xlFo-Fc I /gFo and R = { ~ I F o - F c 1 2 / ~ F ~ ~ } 1 ’ 2 (Hamilton, 1965).

If good Bragg data are available, difference Fourier maps in the last stages of structure analysis are helpful in locating ordered water molecules and/or cations responsible for the integrity of helix and in the association of helices. The positions of guest atoms are refined to enhance the fit of the augmented crystal structure with the x-ray data. Also, sugar rings are flexed by refining the endocyclic conformation angles along with the endocyclic bond angles and bond angle at the bridge oxygen atom. The accuracies of the final atomic coordinates are within a few tenths of an angstrom and the R-values are typically around 0.25 in many polysaccharide structures determined by this methodology. It should be pointed out that most structures reported prior to 1980 did not include hydrogen atoms mainly because of inadequate computing power.

I l l . MOLECULAR SHAPES AND INTERACTIONS

The three-dimensional structures of a number of food polysaccharide helices of interest have been determined to date. The repeating units range from a simple monosaccharide to a branched hexasaccharide. This includes only mono, di, tetra, penta, and hexasaccharide repeats. On the basis of linkage, it is convenient to sort the polymers into five distinct groups. The first deals with (1+4)-linked polysaccharides, the second with (1+3), the third with alternating (1+3) and (1+4), the fourth with the gellan family of gel-forming polysaccharides, and the fifth with some branched polysaccharides. Some of these results have been extrapolated to related systems that are more complex in terms of chemical repeat and type of linkage.

The monosaccharide unit shown in Fig. 7 is to help with atom labels. Descriptions of the molecular features in the following sections include specific details on the helical parameters, unit cell dimensions, number of helices per unit cell and the relationship between them, conformation angles in the main chain and side chain, orientation of the hydroxymethyl or carboxylate group in the repeating unit, hydrogen bonds within and between helices, and much more. The subsequent stereo drawings are intended to reveal clearly many of these aspects; hydrogen bonds are indicated by


H6 1

H3

FIG. 7. Atom labeling in a monosaccharide unit.

dashed lines and hydrogen atoms, even when available in the original papers, are omitted to minimize crowding of atoms and maximize clarity. Each major conformation angle cited in the text refers to the dihedral angle between planes ABC and BCD in a four-atom segment A-B-C-D; the eclipsed cis conformation sets this angle to zero and clockwise rotation of the farthest bond while looking along the middle bond is reckoned positive (IUPAC-IUB, 1971). For the (l+n) linkage, n being 2, 3 or 4, 4 and + refer to 05-C1-0nr-Cn’ and C1-On’-Cn’-Cn+l’, respectively, where the primes denote atoms belonging to the reducing end of the chain. How- ever, for the (1+6) linkage, 4 and + are 05-C1-06-C6 and C1-06-C6- C5, respectively. Unless otherwise specified, the hydroxymethyl group orientation, as in glucose, refers to C4-C5-C6-06; the carboxylate group orientation, as in glucuronate, refers to C4-C5-C6-061. For sulfate attachment in carrageenan to atom On, the two dihedral angles are 0, = Cn+l- Cn-On-S and 0, = Cn-On-S-OS1. However, if n = 6, = Cn-l-Cn- On-S. The bond angle C1-On’-&’ at the glycosyl oxygen atom, referred to as the bridge bond angle T, is usually refined during structure analysis. Unless otherwise specified it is 116.5’.

A. (1-+4)-LINKED POLYSACCHARIDES

Polysaccharides with monosaccharide repeating units may be divided into three distinct groups. The first group includes cellulose and mannan, which are structural polysaccharides that display similar extended ribbon- like molecular morphologies. If the structure of cellulose is important with regard to those of microcrystalline cellulose and some derivatives of cellulose amply used in many food products, the structure of mannan is equally important in the context of an entire series of galactomannans that are


utilized extensively as thickening agents in soups and sauces. The second group includes amylose and amylopectin, which are the basic constituents of starch deposited in the form of granules of varying shapes and sizes depending on the botanical source. A series of amylose derivatives adopt numerous polymorphic forms. The third group contains alginate and pectate, which are gel-forming polysaccharides used in the food industry. All polymers carry 4C1 pyranose rings in their main chains with the exception of polyguluronic acid, which consists of 'C4 pyranose rings. Their three- dimensional structures are presented below and correlated with the observed physical properties.

1. Cellulose

The repeating unit of cellulose is +4)-P-~-glucose-( l+. Native cellulose fibers from Valonia, ramie, cotton, and woods are referred to as cellulose I. The material obtained from native cellulose either by a solution regenera- tion process or by mercerization involving a swelling treatment with alkali is referred to as cellulose 11. Two other minor forms, I11 (Sarko et al., 1976) and IV (Gardiner and Sarko, 1985), are derived by heat or alkali treatment from I and 11, respectively, and their structural details remain elusive. Historically, the first x-ray diffraction study (Meyer and Misch, 1937) established that (a) cellulose chain forms a twofold helix of pitch 10.3 A; (b) the monoclinic unit cell, a = 8.35, b = 7.9, c = 10.3 A, and y =

96", accommodates two chains (total of four glucosyl units) passing through (0 0 0) and (a/2 b/2 0), respectively; and (c) an antiparallel packing is preferred over a parallel alignment of both chains. Since then, these molecular features have been widely accepted and confirmed by other independent studies. However, the original packing arrangement has been criticized, specifically in relation to the biosynthesis of cellulose.

a. Cellulose I. Gardner and Blackwell (1974) used the LALS program to define the molecular structure and to propose the correct packing arrangement of Valonia cellulose I. The ribbon-like twofold helix ( t =

180") of pitch 10.38 A (h = 5.19 A) is stabilized by a series of 03H. . .05 (2.75 A) hydrogen bonds formed across each glycosidic oxygen atom (7 = 114.8'). The main chain conformation angles ( 4 = -98', + = -143') are near the global energy minimum for an isolated cellulose chain. The x-ray data consist of 36 of 39 reflections that fit the monoclinic unit cell (a = 8.17, b = 7.86, c = 10.38 A, and y = 97.0"). On the basis of a significantly lower R-value of 0.18, compared to 0.21 for the antiparallel packing, the parallel model was judged to be the best for cellulose I.


The lateral ordering of the parallel chains produces hydrogen-bonded sheets parallel to the ac-plane as shown in Fig. 8a. Each hydroxymethyl group adopts a tg (i.e., C6-06 trans to C5-05 and gauche minus to C5-C4) conformation (x = -81') so that intrachain 02H.-.06 (2.87 A) and interchain 06H...03 (2.79 A) hydrogen bonds are maintained simultaneously. An axial projection of the unit cell is shown in Fig. 8b. The relative displace- ment of the corner from center chains (sheets) along the fiber axis is 0 .27~ and it leads to excellent stacking between adjacent pyranosyl rings. These

a

b

FIG. 8. Parallel packing arrangement of the twofold cellulose I helices. (a) Stereo view of two unit cells roughly normal to the ac-plane (c-axis is vertical). The corner chains (thin bonds), separated by a, in the back are connected by hydrogen bonds to form a sheet. The center chain is drawn in thick bonds. (b) Projection of the unit cell along the c-axis, with a down the page. Only van der Waals attraction is present between the corner and center chains.


structural features are responsible for the self-association of cellulose chains, which in turn explains the strength and rigidity displayed by the crystalline regions of cellulose. Similar results were simultaneously published by Sarko and Muggli (1974). Comparison shows that differences, if any, are only marginal.

b. Cellulose ZI. Fiber diffraction analysis of this crystalline form (Kol- pak and Blackwell, 1976) has produced a twofold helix ( t = 180" and h =

5.18 A) as its molecular structure and its backbone (T = 114.8", 4 = -9@, i,!i = -145") is isomorphous with that of cellulose I. Its unit cell is also monoclinic (a = 8.01, b = 9.04, c = 10.36 A, and y = 117.1'). Two cellulose chains pass through the unit cell the same way as in cellulose I except that they run antiparallel. Both chains maintain 03H.e.05 (2.69 A) hydrogen bonds. The x-ray results (final R-value = 0.17 for 44 observed reflections) show that the hydroxymethyl orientations are gt and tg, respectively, for the corner (x = 174") and center (x = -70") chains. Consequently, 02H. . .06 (2.73 A) hydrogen bonds are possible only in the "down" center chains, which form a sheet in the 020 plane and exhibit interchain 06H.e.03 (2.67 A) hydrogen bonds (Fig. 9a). The "up" corner chains, connected by 06H. . .02 (2.76 A) hydrogen bonds, also form a sheet structure; their 0 2 H groups make a new set of 02H. . . 02 (2.77 A) hydrogen bonds in the 110 (diagonal) plane with a neighboring center chain. In other words, adjacent antiparallel sheets are now connected. Due to this additional linkage, cellulose I1 is energetically more stable than cellulose I. The relative orientations of the corner and center chains, which are significantly different from those of cellulose I, can be visualized from the axial projection of the unit cell (Fig. 9b). Results of the same sort have also been published by Stipanovic and Sarko (1976). It is now well known that the transformation of cellulose I to cellulose 11, from parallel to antiparallel packing, is irreversible. There- fore, I and I1 are the metastable and stable forms of cellulose.

c. Cellulose Derivatives. Despite conformational similarity in the backbone, which remains rigid due to regular interresidue 03H- . .05 hydrogen bonds, there are structural differences between cellulose I and I1 in fibers. For example, due to freedom of rotation about the C5-C6 bonds, the hydroxymethyl groups are not restricted to the same orientation, but are found in two distinct domains; also, cellulose chains are able to aggregate in parallel and antiparallel modes and are stabilized by alternate intra- and interchain hydrogen bonds. In any case, this self-association is so strong that cellulose is insoluble in most solvents.

However, there are ample food applications for cellulose in one form or another (Coffey et al., 1995). The noncrystalline region of cellulose in wood


a

b

FIG. 9. Antiparallel packing arrangement of the twofold cellulose I1 helices. (a) Stereo view of two unit cells roughly normal to the uc-plane (c-axis is vertical). The corner chains (thin bonds). separated by a, in the back are connected by hydrogen bonds to form a sheet. The center chain is drawn in thick bonds. (b) Projection of the unit cell along the c-axis, with a down the pagc. The corner and center chains are joined by 02H.- .02 hydrogen bonds.

pulp is hydrolyzed into fringed crystallites, which in turn are converted to microcrystalline particles less than 1000 A in diameter by spray drying or mechanical shearing. This type of microcrystalline cellulose (MCC) is used in shredded cheese to carry flavor, in high-fiber bakery products for reduced calories, and in the production of extruded snacks. Also, dispersion of MCC in water develops appreciable viscosity. Derivatization of cellulose at atoms


0 2 , 0 3 , or 0 6 lends water solubility. While the reactivity is high and nearly the same at 0 2 and 06 , it is very low at 0 3 , perhaps due to its involvement in intrachain hydrogen bond with 05. Methyl, hydroxypropyl, and carboxymethyl substituents, as ether groups, are introduced to one or more sites by treating alkali cellulose with methylchloride, propylene oxide, and sodium salt of chloroacetic acid, respectively. By controlled reaction, the degree of substitution (DS) is monitored and a variety of derivatives with DS in the range 0.02 to 0.95 are prepared for use as water-soluble food gums (Desmarais and Wint, 1993; Feddersen and Thorp, 1993; Grover, 1993). Hydroxypropylmethyl cellulose (HPMC) is prepared by reacting alkali cellulose with propylene oxide and methyl chloride at the same time and is also referred to as methyl cellulose (MC). As the four panels in Fig. 10 indicate, the substituents in MC, caboxymethyl cellulose (CMC), hydroxypropyl cellulose (HPC), and HPMC stick out laterally from the main chains to increasing extents. For example, the width of a cellulose chain is only about 7.9 A compared to 9.1, 10.3, 10.7, and 10.9 A in the four derivatives, respectively. In the case of HPC and HPMC, the hydroxyl moiety in hydroxypropyl group is also a possible additional substitution site that will further increase the lateral dimension to values greater than 11.0 A. Thus, due to bumps along the chain at the modified sites, the strong intermolecular hydrogen bonds between the main chains are no longer possible and the association of polymer molecules becomes weaker, leaving the interstices accessible to solvent molecules. In all cases, the interactions between the substituents and the environment will dominate and result in solubilizing the polymer. These food gums are used in ice cream, pancake syrup, onion rings, sandwich spreads, extruded products, and frozen desserts, to name a few.

2. Mannan

The repeating unit of mannan is +4)-p-~-mannose-( l+. Microfibrils of mannan are the major constituents of the cell walls of ivory nut endosperm and date seeds; mannan also occurs in a granular form in siphoneous green algae (Frei and Preston, 1968). As in the case of cellulose, two distinct crystalline forms, I and 11, are known for the native and alkali-treated materials, respectively. They both have orthorhombic unit cells with quite different dimensions and hence their packing arrangements are not the same. Antiparallel chains are the common features. The mannan chains are twofold helices, very similar to those of cellulose, although the 0 2 H group in mannose is axial instead of equatorial as in glucose.

Early x-ray studies (Frei and Preston, 1968; Nieduzynski and Marchessault, 1972) followed by more recent electron diffraction

a. Mannan 1.


a

C d

FIG. 10. Side views of two turns of cellulose derivatives. (a) 02-Methyl cellulose, (b) 03- caboxyinethyl cellulose, (c) 06-hydroxypropyl cellulose, and (d) 06-hydroxypropyl-02- methyl cellulose.

(Chanzy et al., 1987) and x-ray (Atkins et al., 1988) analyses show that the mannan I helices ( t = 180" and h = 5.14 A) crystallize in a two-chain orthorhombic unit cell ( a = 8.92, b = 7.21, and c = 10.27 A). The chains passing through (0 0 0) and (a/2 b/2 0) are related by twofold screw axes perpendicular to the chain direction and hence are antiparallel. According to the x-ray results, the backbone conformation angles are 4 = -90" and $ = -149", and 7 is 117.0"; the hydroxymethyl group orientation is gt (gauche, trans), but x alternates between 175 and -175" in successive residues. The molecule is stabilized by two intrachain hydrogen bonds, 03H. - .05 (2.58 A) and 0 6 H - - . 0 3 (3.0 A). The molecular features and intermolecular interactions are shown in Fig. l l a . The corner and center chains are con-


b

FIG. 11. Antiparallel packing arrangement of the twofold helices of mannan I. (a) Stereo view of two unit cells approximately normal to the bc-plane. The two corner chains (thin bonds) in the back, separated by b, are not hydrogen bonded to each other. Instead, the antiparallel center chain (thick bonds) in the middle is linked to the corner chains by 02H. . . 05 hydrogen bonds. (b) Projection of the unit cell along the c-axis; the n-axis is down the page. This highlights the hydrogen bonds between corner and center chains.

nected by 02H. . .05 (2.93 A) hydrogen bonds. The axial projection (Fig. l l b ) shows that the sugar rings are somewhat rotated from the bc-plane.

b. Mannan II . Based on two x-ray analyses (Frei and Preston, 1968; Millane and Hendrixson, 1994), mannan I1 crystallizes in a four-chain orthorhombic unit cell (a = 9.00, b = 16.65, and c = 10.35 A) that is twice as l aye as that of mannan 1. There are four helices ( t = 180" and h = 5.18 A), two up and two down, related by crystallographic dyads normal to the c-axis, and all are conformationally identical ( T = 116.5", 4 = -88",


Ic, = -153"). The molecular structure is essentially the same as that of mannan I except that the hydroxymethyl group is in the tg conformation (x = -23"). The antiparallel strands are connected by 06H.. .06 (2.6 A) hydrogen bonds and they form sheets almost parallel to the bc-plane. In addition, there is a water molecule (W) on a twofold rotation axis per mannose residue that forms an intersheet 02+. .W- . .02 bridge (O...W =

2.9 A) and provides additional stability to the mannan I1 structure.

3. Galactornannans

Since self-association of mannan helices in the solid state is as strong as that in cellulose, it is not a bit surprising that mannan also is insoluble in most solvents. Of course, derivatization of mannan is a way of inducing solubility, but nature has provided a better course by producing an abundant supply of a number of galactomannans in plants. Water solubility increases with the amount of substitution of (l-+6)-linked a-D-galactose residues as random side chains on the mannan backbone. Depending on its source, the galactose/mannose ratio varies from 0.3 in carob gum, 0.45 in tara gum, 0.6 in guaran, to 0.95 in water-insoluble fenugreek. The galactomannans produce extremely viscous aqueous solutions and hence are used extensively as thickening agents in food products. A mixture of carob gum and xanthan, however, forms gels.

X-ray fiber diffraction patterns from any of these galactomannans resemble that of mannan, implying twofold helix symmetry ( t = 180") and a pitch of 10.3 or h = 5.15 A (Chien and Winter, 1985). Guaran, for instance, has an orthorhombic unit cell (a = 9.3, b = 30.8, and c = 10.3 A) and the four helices in it are related by twofold screw axes parallel to the a- and b-axes. According to a recent x-ray study (Chandrasekaran et al., 1998), the guaran helix has a mannan-like backbone ( 4 = -loo", Ic, = -149", and x = -109") and the galactose side chain conformation is 4 = 32" and Ic, = -123". The polymer chain with galactosyl unit on alternate mannose residues, shown in Fig. 12, is stabilized by a series of 03H(Man)...OS(Man), 03H(Man)...05(Gal), and 06H(Gal)...03(Man) hydrogen bonds. The four galactomannan chains in the unit cell are packed similar to those of mannan I1 and are stabilized by hydrogen bonds between the side chains. The galactose residues are further important in hindering the otherwise self-association of the mannan skeletons. Also, the side chains are in orientations that promote favorable interactions with the surrounding solvent molecules. This is the structural reason for the solubility of some of the galactomannans in water that results in highly viscous solutions.

The food applications of galactomannans are very extensive (Maier et al., 1993). Guaran is easily water soluble at room temperature. Because of


FIG. 12. helix axis and the side chains appear on one side.

Stereo view of two turns of a twofold galactomannan helix. The vertical line is the

the thickening caused by enhanced viscosity, this gum is used as the most economical thickening agent in a variety of food products, including dips, salad dressings, and cheese, which require an acidic environment, and sauces, soups, and related canned food items. Carob, also known as locust bean gum, due to its lower G/M ratio, is only slightly water soluble even at an elevated temperature of 8 5 T , but develops good viscosity when cooled down to 25°C. This behavior leads to its ample use in dairy and frozen food desserts, including ice cream, cheese spreads, sour cream- based dips, and yogurts. Tara gum, nearly 70% soluble in water at room temperature, has viscosity intermediate between carob and guaran. Its limited availability has restricted its use mainly to processed seafoods, frozen meats, ice cream, and frozen yogurt, where water binding and emulsion stability are important. Fenugreek, with its G/M content nearing 1.0, is not known for its viscosity like the other galactomannans. Instead, the ground seed is used as a spice in curry powder, chutneys, and corn bread. Fenugreek extract is added to maple syrup for flavor.

4. Amylose

Starches in large quantities are commercially produced from corn, potatoes, rice, sorghum, and tapioca, to name a few. They are water soluble, form viscous dispersions, and are gel formers at low concentrations. In


addition to being the major food for human consumption, native as well as modified starches are commonplace in a variety of food applications (BeMiller, 1993). Each starch granule in plants stores the polysaccharide made up of a-D-glucose. Its two major components are the linear amylose or (1+4)-a-~-glucan and the branched polysaccharide amylopectin, which contains both (1+4) and (1+6) links. The relative amount of the two components varies with species and hybrid starch. For example, high- amylose corn starch has up to 70% amylose, and waxy maize starch has about 98% amylopectin and less than 2% amylose, while potato starch has 21% amylose.

The linear polymer amylose has a molecular weight ranging from 160,000 to 2.6 million Da and the degree of polymerization (DP) varies from 1000 to 16,000 in a single chain. In contrast, amylopectin is a branch-on-branch polymer in which the (1+6) links are the branch points from the main chain, each of which may also be connected to a bunch of short stretches of amylose having only 20 to 30 units. The molecular weight can be as high as 400 million Da, which accounts for a net DP of 2.5 million. Within the starch granules, it is believed that the three-dimensional ordering of amylopectin segments is responsible for the observed crystallinity, whereas the amylose segments account for the amorphous domains (Imberty et al., 1991).

Two principal types of crystalline starch granules are the A-starch of cereals and B-starch of tubers. On the basis of x-ray diffraction patterns, Wu and Sarko (1978a,b) were the first to suggest that starch molecules exist as helices in both cases and that their packing arrangements are different. A third type, C-starch, found in some plants such as arrow root, pea, and tapioca, yielding distinctive diffraction patterns, is proposed to be a mixture of A- and B-forms, but remains to be fully characterized. Further, amylose can be regenerated in the presence of solvents or complexed with molecules such as alcohols, fatty acids, and iodine; the molecular structures and crystalline arrangements in these materials are classified under V- amylose. When amylose complexes with alkali or salts such as KBr, the resulting structures (Sarko and Zugenmaier, 1980) are surprisingly far from those of V-amyloses.

a. A-Amylose. Wu and Sarko (1977) prepared a pure sample of A- amylose by solid-state deacetylation of oriented and crystalline amylose triacetate fibers with 0.2 M KOH in 75% ethanol, followed by exposure to 80% or higher r.h. at temperatures of 85°C or higher. A good quality x-ray diffraction pattern from this specimen was indexed (Wu and Sarko, 1978a) on an orthorhombic unit cell (a = 11.9, b = 17.7, and c = 10.52 A). Consistent with the presence of meridional reflection on the third layer


line, a sixfold, right-handed ( t = 60°), “half-staggered parallel double helix” of pitch 2c or 21.04 A (h = 3.51 A) was proposed as a possible molecular model. In this type of double helix, the two helices are coaxially interwound; sliding one along the helix axis by c leads to exact superposition on the other. One problem with Wu and Sarko’s model is the relatively low bond angle ( T = 105O) at the glycosidic bridge oxygen atom. They proposed that the two double helices passing through (0 0 0) and (a/2 b/2 0) of the cell were packed in an antiparallel fashion such that each unit cell accommodated half-a-turn per double helix and eight water molecules. Another problem is the antiparallel packing of helices, which has been severely criticized because of biosynthetic considerations that favor only the all- parallel amylose chains (French, 1984; Wild and Blanshard, 1986).

a

FIG. 13. Parallel packing arrangement of A-amylose. (a) Stereo side view of less than two turns of two sixfold double helices separated by 10.62 A along a. The two strands in each helix are in thin and thick bonds for clarity, and the helix axis is vertical. Atom 0 6 mediates both intra- and inter-double helix hydrogen bonds. (b) A c-axis projection of the unit cell contents, with n down and b across the page. A water molecule (crossed circle) per trisaccharide bridges three surrounding helices.


To set the record straight, Imberty et al. (1988) collected their own x- ray powder and electron microcrystal diffraction data and used them along with the x-ray fiber data of Wu and Sarko (1978a) to reexamine the previous model. Powder patterns were from pure amylose (DP 15) obtained after mild hydrolysis of potato starch. Electron diffraction data were from micro single crystals grown from the same material. According to the revised results, the reindexing of the original fiber data is compatible with a monoclinic unit cell (a = 21.24, b = 11.72, c = 10.69 A, and y = 123.5') that accommodates 12 glucose residues and 4 water molecules in order to be compatible with the fiber density of 1.51 g/ml. The final structure corresponds to a sixfold, left-handed ( t = -60"), parallel double helix of pitch 21.38 A (h = 3.56 A). The inner and outer diameters of the helix are 3.4 and 10.4 A, respectively. Two double helices, having the same polarity, pass through (0 0 0) and (a/2 0 c/2) per cell. A side view of this pair is shown in Fig. 13a. The crystal structure has been refined against the observed x- ray intensities, which include 34 fiber data and 21 powder peaks extracted from the one-dimensional 20 scan; the corresponding final R-values are 0.27 and 0.21. The electron diffraction data have not been used in the refinement.

FIG. 13. Continued


Since Imberty et al. (1988) described the A-amylose double helix in terms of a maltotriose repeating unit that corresponds to t = -180" and h = 10.69 A, the three glucose residues, confined to identical 4C1 ring geometries, are no longer required to retain conformational identity. Three sets of four angles (7, 4, I), x) given by (117.0, 92, -153,50), (113.6,86, -145,62), and (123.0, 93, -150, 56) in degrees define the resulting helix. There is no intrachain hydrogen bonding, but the gg (gauche, gauche) orientations of all hydroxymethyl groups result in interchain hydrogen bonds between atoms 0 2 and 0 6 within the double helix and among atoms 02 , 03 , and 0 6 between double helices. The one and only water molecule per maltotriose is also involved in hydrogen bonds with three double helices surrounding it as shown in Fig. 13b. The hydrogen bond distances range from 2.6 to 2.9 A.

b. B-Amylose. Material preparation similar to that for A-amylose, but through two stages of 3 days each of 80 and 100% r.h., at room temperature and a third stage of annealing in hot water at 90°C for 1 hr produces B-amylose from amylose triacetate fibers (Wu and Sarko, 1978b). X-ray diffraction patterns correspond to a hexagonal unit cell (a = b = 18.5, c = 10.4 A, and y = 120"). Similar to their results on A-amylose (Wu and Sarko, 1978a), these authors proposed an antiparallel packing of two sixfold, right-handed ( t = 607, half-staggered, parallel double helices ( h = 3.47 A) passing through (2d3 b/3 0) and (a/3 2b/3 0) of the unit cell. Half-a-turn per double helix and 36 water molecules were located in the cell. Again, as in the case of A-amylose, the antiparallel packing was criticized from the point of view of starch biosynthesis (French, 1984; Wild and Blansh- ard, 1986).

Subsequently, Imberty and PCrez (1988) used the x-ray data of Wu and Sarko (1978b) to demonstrate that the correct B-amylose helix essentially resembles the revised A-amylose structure (Imberty et al. 1988) described above, but the packing arrangement differs. The pitch of the half-staggered, parallel, sixfold, left-handed double helix is 20.8 A. The helix is described in terms of a maltose repeating unit in which t = -120" and h = 6.93 A are twice those for a monomer. The inner and outer diameters of the helix are 3.0 and 10.6 A, respectively. The final R-value after refinement of this structure against 34 reflections is 0.15. Two sets of (T,$, I), x), namely (115.8, 84, -144,53) and (116.3, 84, -144, 68) in degrees, define the disaccharide conformation. As Fig. 14a shows, hydrogen bonds are absent within, but are present between chains connecting atoms 0 2 and 0 6 and they stabilize the double helix. Two double helices of the same polarity pass through (2~13 b/3 0) and (a/3 2b/3 0) of the unit cell. Six water molecules per disaccharide repeat, or 36 per cell, have been located. The two helices in


the hexagonal cell are related by a crystallographic twofold screw parallel to c passing through (a/2 b/2 0). This setup leads to a hexagonal network of amylose helices as shown in Fig. 14b. The channel in the middle, almost as wide (10 A) as the helix itself, is filled with ordered water molecules that are hydrogen bonded to one or more of the six surrounding helices. Exchange of these water molecules with another amylose double helix is all it takes to reproduce the A-amylose packing arrangement shown in Fig. 13b.

With regard to molecular morphology, in both A- and B-amylose, there is no room for a water or similar molecule of radius about 2 A to pass through the inner core of the relatively extended helix. Coaxial intertwining of two such rod-like helices obviously leads to a tight complex and hence low solubility, as observed.

c. Amylopectin. The (1+6) branching point of amylopectin has been modeled based on a 6-glucosylmaltotriose unit by energy calculations (Im- berty and PCrez, 1989). The glucose side chain is attached to the second residue in the backbone. The maltotriose conformation is minimally altered from the A- or B-helix when the side chain folds back on the main chain in an orientation in which the branched tetrasaccharide unit falls in place to form a parallel double helical segment, as during the process of twinning of short amylopectin chains. Such a finding is consistent with the cluster model proposed for amylopectin aggregation in the growth of starch granules (French, 1984).

d. V-Amylose. Following an initial x-ray study of Zobel et al. (1967), later investigations under suitable experimental conditions have characterized the structures of a family of sixfold, left-handed ( t = -60"), single helices for V-amylose with very low h values around 1.35 A. They include certain allomorphs for amylose fibers obtained from solution in Me2S0 (French and Zobel, 1967; Winter and Sarko, 1974a; Rappenecker and Zu- genmaier, 1981), iodine (Bluhm and Zugenmaier, 1981), or butanol (Booy et al., 1979) in anhydrous, hydrous, or intermediate states. The structure of anhydrous V-amylose (Winter and Sarko, 1974a) is chosen here to represent the entire family. Its unit cell is orthorhombic (a = 12.97, b = 22.46, and c = 7.91 A). The left-handed sixfold helix ( t = -60" and h = 1.32 A) is very shallow as shown in Fig. 15a. The main chain conformation angles, 4 = 115" and I) = -131", are up to 30" away from those in A- and B-amylose, and 7 is 119.7'. The hydroxymethyl group has gt orientation (x = -168"). The inner and outer diameters of the helix are about 7.1 and 13.7 A, respectively. The helix looks like a hollow cylinder (Fig. 15a) and is stabilized by intrachain 03H.. .02 (2.8 A) and 02H...06 (2.9 A) hydrogen


a

b b

FIG. 14. Parallel packing arrangement of B-amylose. (a) Stereo side view of slightly less than two turns of two sixfold double helices 10.7 A apart along the long diagonal of the ah- plane. The two strands in each helix are in thin and thick bonds for clarity, and the helix axis is vertical. Atom 0 6 mediates both intra- and inter-double helix hydrogen bonds. (b) A c- axis projection of the unit cell contents plus four other helices surrounding a cluster of six water molecules (crossed circles) per disaccharide in the middle.

bonds that link adjacent residues and turns, respectively. Related by crystallographic twofold screw axes parallel to a and b, the two helices in the unit cell are antiparallel and connected by intermolecular 02H. . .02 (2.9 A) hydrogen bonds (not shown). An axial projection of the unit cell contents is shown in Fig. 15b. Variations among the known V-amylose molecular structures are marginal. The helical cavity is wide enough to accommodate guest molecules such as water, ion, alcohol, lipid chain, and others of similar size. Depending on the complexing agent or solvent used, the full range of flexibility available to the hydroxymethyl group is exploited. As a consequence, there are changes in packing arrangements of helices among the various examples studied to date (Zobel et al., 1967; Hinkle and Zobel, 1968; Winter and Sarko, 1974a,b; Zugenmaier and Sarko, 1976; Booy et al., 1979; Bluhm and Zugenmaier, 1981; Rappenecker and Zugenmaier, 1981).


FIG. 14. Continued

Further reduction of pitch from that of V-amylose is stereochemically disallowed, but a sudden jump to zero is quite appropriate in the case of a short amylose chain with a DP of 6. This doughnut-like circular structure corresponds to a-cyclodextrin, which retains the central cavity, 5 A wide, similar to that in V-amylose and forms inclusion complexes with various guest molecules that range from acetic acid, benzoic acid, iodine, hexanol, to tyrosine (Clarke et al., 1988). Similar cyclic structures having DP 7 and 8 have increasing cavity sizes (6.2 to 8.0 A) and are known as P- and y- cyclodextrin, respectively. Although cyclodextrins are not approved for food use in the United States, they are of immense importance and interest to the pharmaceutical industries. For example, the strong binding property of P-cyclodextrin to cholesterol is used to remove the latter from dairy products.

Investigation of the complex of amylose with lipids is highly relevant toward understanding the structure and interactions of starch with food constituents (Biliaderis, 1991). In the presence of fatty acids, single-helical


b

FIG. 15. Packing arrangement of shallow, sixfold, V-amylose helices. (a) Stereo view of two unit cells approximately normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two helices (thin bonds) back at the corners. Intrachain hydrogen bonds (03H...02 and 02H...06) stabilize each helix. (b) A c-axis projection of the unit cell shows that three water molecules (crossed circles) per monosaccharide, inside and between helices, are involved in the organization of helices.

V-amylose is apparently the active form, according to several independent investigations. For example, thermal behavior of the complex is consistent only with V-amylose allomorphs (Biliaderis et al., 1985; Biliaderis and Galloway, 1989). X-ray powder diffractograms and differential scanning calorimetry coupled with molecular modeling indicate that a lipid inclusion


complex consists of a left-handed sixfold Vh-helix whose interior cavity is large enough to accommodate the aliphatic portion of the lipid in a trans conformation; the polar group cannot enter the cavity at all due to steric and electrostatic repulsion (Godet et al., 1993a,b, 1995). Thus, although the double helix represents the native form of starch, it is now well known that the single-chain V-helix prevails when amylose interacts with fatty acids to form an inclusion complex.

e. KOH-Amylose Complex. In sharp contrast to the shallow V-helix, amylose can be induced to adopt an extended single helix under certain circumstances. For instance, a sixfold, left-handed ( t = -60") helix with h = 3.74 A is formed when amylose complexes with KOH (Sarko and Biloski, 1980). The unit cell is orthorhombic with dimensions a = 8.84, b =

12.31, and c = 22.41 A. The repeating motif is compatible with a maltotriose repeat in the final model, similar to that in A-amylose (Imberty et al., 1988). Each trimer is associated with a KOH and three water molecules. The main chain conformation angles C$ and I,!I for the three saccharides are less than 1" off from 93" and -150", respectively. The corresponding values of T are 115.6, 116.3, and 113.7". Because of structural similarity, it is not surprising that the amylose chain in KOH complex (Fig. 16a) is almost superimposable on a chain in the A- or B-amylose double helix. Due to its slightly larger pitch, the inner and outer diameters of the single helix (3.0 and 10.2 A) are a bit smaller than those in the latter. The orientations of the three hydroxymethyl groups are two gg (x = 57 and 68") and one tg (x = -68"). There are two helices organized in an antiparallel mode in the unit cell (Fig. 16b) and they are related by crystallographic twofold screw axes parallel to a and b. Adjacent helices are linked by a potassium ion involving two oxygen atoms as ligands from each helix. The close packing of helices is further stabilized by a water molecule per monosaccharide.

f Amylose Derivatives. Several amylose derivatives in which all three potential hydroxyl groups are substituted are of special interest in relation to the structure of chemically modified starch and further extends its practical applications to food products. They include amylose triacetate (Sarko and Marchessault, 1967), trimethylamylose (Zugenmaier et al., 1977), triethylamylose (Bluhm and Zugenmaier, 1979a), and its complexes with solvents (Bluhm and Zugenmaier, 1979b). All of them generally conform to fourfold, left-handed helices ( t = -9OO) whose h values vary from 3.75 to 4.05 A. Since intrachain hydrogen bonds are not possible in these helices, their packing arrangements are controlled by van der Waals interactions only. The structure of triethylamylose (Bluhm and Zugenmaier, 1979a) having h = 3.87 A is described here as an example. Its main chain conformation


a

h b

h


angles are 4 = 64" and 4 = -162" and T is 122.1'. While x is -94", all three ethyl groups adopt roughly extended conformations, as shown in Fig. 17a. The orthorhombic unit cell (a = 15.36, b = 12.18, and c = 15.48 A) accommodates two antiparallel helices that are related by crystallographic twofold screw axes parallel to a and b. An axial projection of the unit cell contents (Fig. 17b) shows clearly the interdigitation of substituents from neighboring helices. Consistent with the large t (-90'7 and regular substitution, the inner and outer diameters of the helix are 2.8 and 13.4 A, respectively. The closure of the helical cavity shows that it has no room to capture guest molecules as the V-form does.

The polymorphism of amylose helices in oriented fibers is a remarkable observation. The spectrum of structures covers single and double helices, all of which are left-handed; h values range from 1.35 to 3.8 A; the shallow helices are excellent candidates for inclusion complexes, but the extended helices are not. The variety of intra- and interchain hydrogen bonds is clear testimony to the structural flexibility for transition between helical states and for transition from helix to coil under the influence of new interactions that amylose chains might encounter within their surroundings. These events in starch-based foods involve amylose, amylopectin, cations, water molecules, lipids, proteins, and the rest, and they undergo continuous changes in a dynamic fashion due to variations in temperature, pressure, shear, and so on. The structural details preceding gelation, gelatinization, and retrogradation of starch are indeed complex to visualize, but worthy of understanding through further investigation.

5. Pectin

Pectin is a heterogeneous complex structural plant polysaccharide that is part of the human diet. Depending on its source (e.g., apple pomace or citrus peel) and isolation procedure, pectin is polydisperse and polymolecu- lar. The backbone of pectin is made up of (l+4)-linked a-D-galacturonic acid repeating units and some of the carboxyl groups are methyl esterified. Often, (1+2)-linked a-L-rhamnose residues interrupt the regular main chain. The high viscosity and cation-dependent gelling properties of pectin are exploited by the food and pharmaceutical industries. Moreover, the

FIG. 16. Packing arrangement of extended, sixfold, KOH-amylose helices. (a) Stereo view of two unit cells approximately normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two helices (thin bonds) back at the corners. Water molecules (open circles) and hydroxyls from amylose helices are ligands to potassium ions (crossed circles). (b) A c-axis projection of the unit cell shows that amylose helices are packed tightly aided by ions and water molecules.

168

a


b

FIG. 17. Antiparallel packing arrangement of extended, fourfold, 2,3,6-tri-O-ethylamylose helices. (a) Stereo view of two unit cells approximately normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two helices (thin bonds) at the back corners. There is no intra- or interchain hydrogen bond, and only van der Waals forces stabilize the helices. (b) A c-axis projection of the unit cell shows that the ethyl groups extend into the medium in radial directions.


rheological properties of pectin can be controlled by the degree of esterifi- cation (Rolin, 1993). X-ray studies on pectic acid (polygalacturonic acid), sodium pectate, calcium pectate (salts of polygalacturonic acid), and pectinic acid (100% methyl ester of pectic acid) have led to their structural details to varying extents.

a. Sodium Pectate. The x-ray diffraction pattern of sodium pectate (Walkinshaw and Arnott, 1981a) shows that the polymer forms a threefold helix of pitch 13.36 A as evident from meridional reflections on the third and sixth layer lines. The unit cell is monoclinic ( a = 8.39, b = 14.27, c = 13.36 A, and y = 90"). The 1,4-diaxial links, coupled with 4C1 chair conformation for the galacturonate repeat, generate a right-handed helix ( t = 120") with h = 4.45 A that fits the x-ray data (26 reflections) much better ( R = 0.23) than a competing left-handed ( t = -120") model. Its conformational parameters 7, 4, and CC, are 116.5, 80, and 90", respectively. The three residues per turn have, however, slightly different carboxylate group orientations: 115,97, and 99" for x. The helix is stabilized by intrachain 02H.*.061 hydrogen bonds (2.8 A) across each bridge oxygen atom and its inner and outer diameters are 0.4 and 7.0 A, respectively. There are two water molecules and one sodium ion (near the carboxylate group) per monomer. The two rod-like helices in the unit cell are packed antiparallel and related by crystallographic twofold screw symmetry parallel to the b- axis; they are laterally separated by 8.4 A and their carboxylate groups are bridged by sodium ions and water molecules. These structural features are shown in Fig. 18.

b. Pectic Acid. According to the x-ray study on the acid form (Walkin- shaw and Arnott, 1981a), pectic acid crystallizes in a monoclinic unit cell (a = 9.9, b = 12.3, c = 13.3 A, and y = 90"). Similar to sodium pectate, it forms a threefold helical structure ( t = 120" and h = 4.43 A) stabilized by 02H...061 (2.62 A) hydrogen bonds between adjoining residues. The conformational parameters of a monomer unit are 7 = 116.5", 4 = 73", CC, = 97", and ,y = 93". The two helices in the unit cell are antiparallel and related by twofold screw symmetry along the b-axis. Due to the absence of ions in this case, however, the interhelical association is directly through 03H.. .062 (2.8 A) and 061H...061 (2.8 A) hydrogen bonds (Fig. 19a). The resulting packing arrangement (Fig. 19b) is somewhat different from that of sodium pectate (Fig. 18b) due to changes in lateral dimensions of their unit cells.

c. Calcium Pectate and Pectinic Acid. Neither calcium pectate nor pectinic acid has produced decent diffraction patterns for detailed structure


a

b ,... Q

FIG. 18. Antiparallel packing arrangement of threefold sodium pectate helices. (a) Stereo view of two unit cells roughly normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two in the front (thin bonds). Intrachain hydrogen bonds stabilize each helix. Sodium ions (crossed circles) and water molecules (open circles) connect adjacent helices. (b) A c-axis projection of the unit cell contents shows that ions and water molecules are located between helices.

analysis. Preliminary x-ray analysis of calcium pectate ( Walkinshaw and Arnott, 1981b) favors the same molecular structure and packing arrangement as pectic acid except that the 061H...061 hydrogen bond is replaced by an 061-..Ca2+...061 interaction. Pectinic acid is also a threefold helix, but it has a hexagonal packing arrangement with only one helix per trigonal unit cell (a = b = 8.37 and c = 13.0 A). In this idealized 100% methylester


a

b

FIG. 19. Antiparallel packing arrangement of threefold pectic acid helices. (a) Stereo view of two unit cells roughly normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two in the front (thin bonds). Intrachain hydrogen bonds stabilize each helix. Association of helices is through direct hydrogen bonds involving the carhoxyl groups. (b) A c-axis projection of the unit cell contents highlights the interactions among three helices.

form of pectinic acid, the unit cell has a triangular column of methyl (hydrophobic) groups, enclosed by three helices, and a similar channel filled with water molecules hydrogen bonded to the surrounding polymer chains (Walkinshaw and Arnott, 1981b). The network formed by alternating hydrophobic and hydrophilic channels embedded among the helices provides an interesting motif in the formation of junction zones, which are implicated in the gelation process. If reducing the degree of methylation would increase


the amount of water within the network, addition of co-solute would not only compete for this water but also disrupt the formation of the hydrophobic cage. Both strategies, in conjunction with pH and temperature as variables, are commercially exploited to alter the gelation properties to suit current needs (Rolin, 1993).

Jams and jellies containing desired levels of sugar are always associated with pectins, fruits and plants being their respective origins. Pectins are also used in fruit sauces, fruit yogurts, milk drinks, and heat-resistant bakery glazing. In all cases, the pH is very acidic (3.0 to 5.5).

6. Alginic Acid

Alginic acid constitutes one of the structural polysaccharides of marine brown algae and some bacteria. Although a homopolymer of D-mannuronic acid (M) is the starting fermentation product, mannuronan C5 epimerase converts 6-D-mannuronic acid to a-L-guluronic acid (G) in certain blocks (Clare, 1993). Therefore, the (1+4)-linked polymer contains blocks of p- D-mannuronic acid alternating with blocks of a-L-guluronic acid of varying size. The composition and block lengths are species dependent. For example, the sample from Fucus or Asophyllum species is 97% polymannuronic acid and that from Lamineria hyperborea is rich in polyguluronic acid (Atkins et al., 1973a). Alginic acid is useful to the food industries because of its cation-dependent gelling properties. Since M and G adopt 4C1 and 'C4 chair conformations, respectively, and since the linkages are diequatorial for M and diaxial for G, morphologies of the two polymers are quite unrelated.

a. Poly(p-D-Mannuronic Acid). The x-ray diffraction pattern (Atkins et al., 1973a) recorded from a bundle of fibers prepared from Fucus vesicu- losius is reminiscent of those from mannan I. The orthorhombic unit cell dimensions are a = 8.6, b = 7.6, and c = 10.4 A. The molecule is an extended ribbon-like twofold helix (t = 180" and h = 5.2 A) similar to that of mannan stabilized by 03H...05 hydrogen bonds (2.7 A) across each glycosidic bridge oxygen atom. The conformational parameters T, 4, $, and ,y are 117.1, -94, -145, and 94", respectively. The two helices located at (a/4 0 0) and (3a/4 bi2 0) in the unit cell are antiparallel and related by crystallographic twofold screw axes parallel to a and b as shown in Fig. 20a. One of the carboxylate oxygen atoms is hydrogen bonded (2.7 A) to atom 0 3 of a neighboring chain of the same polarity, separated by the b- axis. The antiparallel chains are connected by 02H-a.05 hydrogen bonds (3.0 A). The c-axis projection (Fig. 20b) shows that the sugar rings are oriented roughly parallel to the bc-plane. In all respects, this molecular


a

b

FIG. 20. Antiparallel packing arrangement of twofold poly(mannuronic acid) helices. (a) Stereo view of two unit cells roughly normal to the bc-plane. The helix at the center (thick bonds) is antiparallel to the two (thin bonds) in the back. Intrachain hydrogen bonds stabilize each helix. Association of helices through direct hydrogen bonds involves carboxylate groups for parallel, and the axial 2-hydroxyl groups for antiparallel chains. (b) A c-axis projection of the unit cell contents highlights hydrogen bonds between helices.

structure and packing arrangement are very similar to those previously described for mannan I (Atkins et al., 1988).

a. Poly(a-L-Guluronic Acid). The x-ray diffraction pattern (Atkins et al., 1973b) shows a meridional reflection on the second layer line, which is diagnostic of a twofold helix ( t = 180" and h = 4.35 A). The unit cell is orthorhombic (a = 10.7, b = 8.6, and c = 8.7 A). The 8.7 A pitch is 1.7 A shorter than that of cellulose or mannan. The guluronate residue is in the preferred 'Cq chair conformation. Because of the lP-diaxial C-0 bonds, the helix is a buckled ribbon (Fig. 21a) that does not resemble the poly(mannuronic acid) structure. The conformational parameters T, 4, $, and ,y are 117.6, -108, -134, and 99", respectively. The helix is stabilized by


a

b

FIG. 21. Antiparallel packing arrangement of twofold poly(gu1uronic acid) helices. (a) Stereo view of two unit cells roughly normal to the hc-plane. The helix at the center (thick bonds) is antiparallel to the two (thin bonds) in the back. lntrachain hydrogen bonds stabilize each helix. Association of antiparallel helices involves carboxylate groups and water molecules (crossed circles). (b) A c-axis projection of the unit cell contents shows that helices are connected via water molecules. Due to a buckled shape of the helix, there is complete overlap between adjacent sugar rings.

02H...061 (2.7 A) hydrogen bonds across every glycosyl bridge oxygen atom. The two chains in the unit cell, located at (a/4 0 0) and (3d4 b/2 0) and related by crystallographic twofold screw symmetry parallel to a and b, are antiparallel. Adjacent helices are connected by water molecules, one per guluronate residue, via 0 2 . . - W...O3 and 0 5 . a . W...O3 interactions. Figure 21b suggests that adjacent sugar rings in the chain overlap completely when projected down the helix axis.

The carboxylate oxygen atoms are farthest from the helix axis at 3.7 in both poly(mannuronic acid) and poly(gu1uronic acid) so that they can


easily interact with ions and water molecules in the environment. It is believed that pairs of carboxylate groups may be connected by calcium ions. The painvise association of helices mediated by an array of calcium ions in the middle bears certain resemblance to a Styrofoam box packed with eggs. This similarity is the basis for the "egg-box'' model (Grant et al., 1973), which depicts the junction zone in guluronate-rich alginate gels. According to computer modeling, a pair of chains can associate in parallel as well as antiparallel mode with the help of calcium ions (Mackie el al., 1983). The thickening effect and calcium-dependent gelation properties of alginate are extensively utilized in pimento-stuffed olives, salad dressings, fruit drinks, and coatings on food to prevent bacterial contamination (Clare, 1993).

B. (1-+3)-LINKED POLYSACCHARIDES

p-Glucan and scleroglucan are the two polysaccharides in this group having food applications. The former, known as curdlan, has excellent gelling properties and is used in numerous Japanese foods. It adopts at least three different helical structures. The latter is (1+6)-branched and its backbone retains one of the curdlan structures.

1. P-Glucan

Curdlan, the extracellular microbial polysaccharide having (1+3)-linked P-D-glucose as its repeating unit, forms resilient gels from aqueous solutions at 95°C. X-ray diffraction patterns from oriented specimens (Marchessault et nl., 1980) identified three distinct forms, I, 11, and 111, representing the native specimen, the hydrated form after annealing at 140"C, and the dehydrated form after annealing, respectively. This investigation assigned similar triple helical structures to all three, but Curdlan I is now known to exist as a single helix (Okuyama et al., 1991).

a. CurdLon I . According to a recent x-ray reexamination (Okuyama et af., 1991), the unit cell of form I is monoclinic (a = 28.8, b = 18.6, c =

22.8 A, and y = 90"). Consistent with a meridional reflection on the sixth layer line in the diffraction pattern, the molecular structure is a sixfold, right-handed, single helix (t = 60" and h = 3.8 A). Its conformational parameters T, 4, $, and x are 116.5, -71, 126, and -54", respectively. As illustrated in Fig. 22a, the helix is rather extended (pitch = 22.8 A) and is stabilized by 04H...05 (3.14 A) hydrogen bonds connecting adjacent residues. With reference to atomic centers, the inner diameter is too small (2.0 A) to hold any guest molecules, but the outer diameter of the helix is


a

FIG. 22. Parallel packing arrangement of sixfold, curdlan I helices. (a) Stereo view of two unit cells approximately normal to the bc-plane. The helix is stabilized by intrachain 04H...05 hydrogen bonds. There are only van der Waals interactions between helices. (b) A c-axis projection of the unit cell shows large gaps between helices which are alleged to be filled with 250 water molecules.

large (14.3 A). Two helices pass through (0 0 0) and (a/2 b/2 0) of the unit cell in a parallel mode. There are no interhelical hydrogen bonds and the large interstitial space in the unit cell, clearly seen in Fig. 22b, is reported to be filled with about 250 water molecules. The final R-value is 0.14 for 40 reflections.

b. Curdlan ZZZ. The dehydrated form 111 is structurally well organized and its x-ray analysis (Deslandes et al., 1980) reveals that the polymer chains crystallize in a hexagonal cell (a = b = 14.41, c = 5.87 A, and y =

120"). The meridional reflection occurs on the second layer line and there are six glucose residues per cell. A sixfold, parallel, triple helix of pitch 3c = 17.61 A (h = 2.94 A), in which the three chains are related by c translation along the helix axis, is consistent with these observations. The pitch (17.61 A) is 5.2 A shorter than that of curdlan I. A right-handed helix ( t = 60") corresponding to a final R-value of 0.23 for 21 reflections is superior to any left-handed alternative. The conformational parameters T, 4, 4, and x of the triple helix shown in Fig. 23a are 110.6, -92, 126, and -82", respectively. Both its inner and outer diameters are 1 A larger than those in curdlan I, so that triads of interchain 02H- . .02 (2.72 A) hydrogen


b

d I

FIG. 22. Continued

bonds are formed at successive levels separated by h in the central wider core. As a result, the triple helix appears like a solid cylinder. The intrachain 04H- . .05 distance of 3.18 A between adjacent residues represents a weak hydrogen bond. The packing of molecules in the unit cell (Fig. 23b) is facilitated by a series of strong hydrogen bonds (2.70 to 2.75 A) between neighboring triple helices involving atoms 0 4 and 06.

b

>

P


a

FIG. 23. Structure of the sixfold anhydrous curdlan I11 helix. (a) Stereo view of a full turn of the parallel triple helix. The three strands are distinguished by thin bonds, open bonds, and filled bonds, respectively. In addition to intrachain hydrogen bonds. the triplex shows a triad of 0 2 H . . . 0 2 interchain hydrogen bonds around the helix axis (vertical line) a t intervals of 2.94 A. (b) A c-axis projection of the unit cell contents illustrates that 0 6 H . . . 0 4 hydrogen bonds between triple helices stabilize the crystalline lattice.


c. Curdlan 11. The hydrated form I1 is not quite like the ordered structure of curdlan 111. According to its x-ray analysis (Chuah et al., 1983), the hexagonal unit cell (a = b = 15.56, c = 18.78 A, and y = 120") of I1 is somewhat larger than, and the intensity distribution in the diffraction pattern is slightly different from, that of 111. In order to account for the 18.78 A layer line, the final model for curdlan I1 compromises with a triple helical backbone as in 111, and distinct orientations covering the entire range of staggered domains for the 18 hydroxymethyl groups in one turn. This arrangement is feasible by revising the conformational parameters 7, 4, and +!I to 104.6, -87, and 127", respectively. Six water molecules per monosaccharide are involved in connecting triple helices in the unit cell through hydrogen bonds. The final R-value is 0.17 for 79 reflections.

The water-mediated hexameric aggregation of curdlan triple helices is strong enough to result in micelle formation. The interaction of these micelles with each other is believed to be responsible for gelation (Fulton and Atkins, 1980). The dependence of gelation on high temperatures is an attractive feature of curdlan for food applications (Harada et al., 1993). For instance, gels formed by heat treatment at 120°C have remarkable fat- adsorbing properties. Curdlan helps to improve the texture of soy bean curd, sweet bean paste jelly, boiled fish paste, Japanese noodles, sausage, jellies, and jams. It is also used as a stabilizing agent in ice cream and a thickener in sauces and soup.

2. Scleroglucan

Scleroglucan has a branched tetrasaccharide repeating unit composed of a trisaccharide fragment like that of curdlan and an (1+6)-linked /3-D- glucoside as side chain. X-ray diffraction of oriented fibers of scleroglucan coupled with molecular modeling (Bluhm et al., 1982) suggests considerable morphological similarity with curdlan 111. The unit cell is hexagonal (a =

b = 17.3, c = 6.0 A, and y = 120") but the poor quality of the x-ray data has not permitted an independent structure determination. Based on a conformational analysis of gentobiose, the preferred orientation of the side chain is given by x = 60", 4 = 60", and +!I = -130". Incorporation of a side chain in this conformation on every third glucoside residue in a curdlan I11 triple helix generates a plausible model for scleroglucan as shown in Fig. 24. The peripheral side chains increase the outer diameter to 24.2 A from 15.3 A for the main chains alone, and thus hinder main chain-main chain association of scleroglucan helices. However, aggregation is promoted by side chains and a large number of water molecules are trapped between helices. This structural feature is important for the observed gelation properties.


FIG. 24. A stereo view of a full turn of the sixfold scleroglucan triple helix. The three strands are distinguished by thin bonds, open bonds, and filled bonds, respectively. As in curdlan 111, the triad of 02H...02 interchain hydrogen bonds stabilize the triple helix. Peripheral side chains shield the main chains appreciably.

Scleroglucan is used as a stabilizer in frozen and aerated desserts, creams, and sauces. As in the case of curdlan, Japan is the foremost country to use scleroglucan in cakes, steamed foods, rice crackers, and bakery products (Brigand, 1993).

C. ALTERNATING (1+3) AND (1+4)-LINKED POLYSACCHARIDES

This group includes carrageenans and agarose, which are useful as gelling agents. These polysaccharides are exclusively built with galactopyr- anosyl units with disaccharide motifs, which utilize both L- and D-sugars; and a- and /3-anomers. While carrageenans are sulfated to varying extents, agarose and related polymers are not sulfated at all. Both polymers tend to be double helical. In contrast with the monotonous monosaccharide repeat and the same type of linkage in the previously described polysaccharides, due to their disaccharide repeats, carrageenan and agarose helices display greater conformational flexibility and hence more interesting morphology.

1. Currugeenuns

Carrageenans belong to a family of gel-forming sulfated polysaccharides found in the marine red algae Rhodophyceae. Two principal members, known as L- and K-carrageenan, chemically differ in degree of sulfation (Therkelsen, 1993). The disaccharide repeat (A-B) for each case is


A B L-carrageenan: -3)-P-~-Ga14sO,--( 1+4)-3,6-anhydro-a-o-Ga12S03--( 1 4 K-carrageenan: -3)-P-o-Ga14S03--( 1+4)-3,6-anhydro-cr-o-Gal-( 1-

Two sets of parameters are needed to describe the disaccharide conformation: TI, $1, $1 refer to A(1-+4)B and 7 2 , $2, & to B(1-+3)A linkage; xA and xB are the orientations of the hydroxymethyl groups. Note that in the case of K-carrageenan, the anhydrogalactose residue B is not sulfated. Both polymers form thermally reversible gels that exhibit substantial differences in their physical properties. For example, &-carrageenan forms very clear and elastic gels that neither synerese nor undergo hysteresis effects. In contrast, K-carrageenan gels are hazy and brittle and exhibit syneresis as well as hysteresis effects. Consistent with their chemical differences, their molecular structures are not the same, resulting in distinct physical properties in the presence of a variety of ions.

a. L-Carrageenan. X-ray diffraction patterns from polycrystalline and well-oriented fibers of Ca2+ &-carrageenan have meridional reflections on the third and sixth layer lines (Arnott et af., 1974a). The unit cell is trigonal (a = h = 13.73, c = 13.28 A, and y = 120"). The x-ray data are best fitted by a threefold, right-handed ( t = 120"), parallel, half-staggered, double helix. The pitch is 2c = 26.56 A so that h = 8.85 A. The helix, shown in Fig. 25a, incorporates the preferred 4CI and 'C4 chair geometries in residues A and B, respectively, and contains both sulfate groups on its outside. The molecule is extremely sinuous due to its disaccharide motif. The conformational parameters ( T ~ , $,, $,) and ( T ~ , &, &) are (116.5, -87, 94) and (116.5,75,79), respectively, and xA = 176, all in degrees. The sulfate group orientations (01, 02) are ( l lSo , -159") in A and (-140", -165") in B. The galactose residues are connected by six interchain 06H.a.02 (2.7 A) hydrogen bonds per turn. The trigonal unit cell accommodates only one helix. In terms of coordinates published for nonhydrogen atoms alone (Arnott et al., 1974a), its inner and outer diameters are 3.4 and 14.2 A, respectively. A packing arrangement with up- and down-pointing molecules distributed randomly at each cell corner, similar to that shown in Fig. 25b, explains the presence of both Bragg reflections and layer line streaks in the diffraction pattern. The two sulfate groups on the anhydrogalactose residues of adjacent helices, linked through direct SO4-..-Ca2+.-.SO4- interactions, are believed to be responsible for polymer aggregation during gel formation.

b. K-Carrageenan. Fibers prepared from the potassium salt of K-

carrageenan are less oriented and less crystalline than those of L-

carrageenan. The diffraction pattern exhibits only continuous intensities


a

h

b

9lk &’

FIG. 25. (a) Stereo view of slightly over a turn of the threefold double helix of 6-carrageenan. The two chains are distinguished by thin and thick bonds for clarity. The vertical line is the helix axis. Six interchain hydrogen bonds per turn among the galactose residues stabilize the double helix. The sulfate groups lined up near the periphery are crucial for intermolecular interactions. (b) An axial projection of the unit cell contents. The helix at each corner can be either “up-” or “down-pointing’’ in terms of the x-ray data. All are, however, up in this diagram so that a calcium ion (crossed circle) is connected to the sulfate groups in three surrounding helices.


on layer lines, indicating that there is no lateral organization of polymer molecules in the fiber (Millane et af., 1988). The first layer line spacing (25 A) is roughly twice that observed for L-carrageenan (i.e., c = 13.28 A), and meridional intensity is seen on the sixth and ninth layer lines. On the basis of modest similarity in overall intensity distribution with that of L,

and as a reasonable fit with the observed x-ray intensities, K-carrageenan is also double helical. As shown in Fig. 26, the two chains in the threefold, right-handed ( t = 120"), parallel, double helix of pitch 25 A (h = 8.33 A) are offset from the half-staggered position by 1.1 A along, and 28" about, the common helix axis. Due to this offset, the 25 A layer line is not extin- guished in the diffraction pattern, and there are only half the number of interchain 06H. . . 02 (2.5 A) hydrogen bonds connecting the galactose residues. The main chain conformational parameters (T,, 41, and (Q, 42, +2) are(116.5, -98,108) and(116.5,61,81),respectively,andXA = -178, all in degrees. The two angles 8, and 0, for the sulfate group orientation in residue A are 118" and -155", respectively. Due to a shorter pitch, the outer diameter of the helix is about 1 A larger than that in &-carrageenan. Details of the interactions among helices and the structural role of potassium ions are beyond the scope of this analysis.

FIG. 26. Stereo view of about a turn of the threefold double helix of K-carrageenan. The two chains are distinguished by thin and thick bonds for clarity. The vertical line is the helix axis. Only three interchain hydrogen bonds per turn among the galactose residues stabilize the helix. The peripheral sulfate groups are crucial for intermolecular interactions.


Consistent with their chemical differences, the molecular structures of L- and K-carrageenans are not identical. A shorter pitch and an offset positioning of the two chains in the K-heliX is compatible with the absence of sulfate groups on its anhydrogalactose residues. The variations in molecular structures mirror the types of junction zones formed by these polymers and relate to differences in the observed gelation properties (Therkelsen, 1993). Carrageenans, known also for stabilizing and viscosity building beyond gelation, are routinely used in dairy products, including ice cream, chocolate milk, and custards to name a few. Water desserts incorporating fruit constituents, glazing of hot tart berry fillings, processed cooked ham, and canned meat and fish are other examples of foods that contain carrageenans exclusively or in combination with other food gums.

2. Agarose

Agarose, a member of the agar family of galactan polysaccharides found in red seaweeds, is an excellent gel-forming agent extracted from Rhodo- phyceae. The only difference in covalent structure between agarose and carrageenan is in inversion of the anhydrogalactose residues from D to L.

Also agarose carries no sulfate groups. Its disaccharide repeat (A-B) is --+3)-P-~-Gal-( 1+4)-3,6-anhydro-a-~-Gal-( 1+. As in the case of K-

carrageenan, agarose films and fibers used in the x-ray study are oriented and noncrystalline. Therefore, its diffraction patterns consist of only continuous intensities on layer lines. The first layer line spacing c is 9.5 A and meridional intensity is on the third layer line. Arnott et al. (1974b) proposed a threefold, left-handed ( t = - 120"), half-staggered, parallel, double helix of pitch 2c = 19.0 A ( h = 6.33 A) as the best model and it contains both monomers in the same 4C1 chair conformation. This helix, as shown in Fig. 27, is much shorter than the carrageenan helices (Figs. 25a and 26). Its inner and outer diameters are 4.2 and 13.6 A, respectively, in terms of nonhydrogen atoms alone. The conformational parameters (T~, 41, and (q, &, &) are (116.5, -124, -113) and (116.5, -52, 157), respectively, and xA = -69, all in degrees. There are no hydrogen bonds either within or between chains and the double helix is stabilized only by van der Waals forces. The inner cavity is just wide enough to hold water molecules, which can mediate interchain hydrogen bonds with the oxygen atoms 0 2 A and 05B positioned in the interior.

Although sharper diffraction patterns from agarose films dried at about 100°C are in favor of extended single helices, h ranging from 8.9 to 9.7 A (Foord and Atkins, 1989), such models are incompatible with chiroptical data (Shafer and Stevens, 1995) that are consistent only with the double helical model derived from the original x-ray analysis (Arnott et al., 1974b)


FIG. 27. Stereo view of about a turn of the threefold double helix of agarose. The two chains are distinguished by thin and thick bonds for clarity. The vertical line is the helix axis. Only van der Waals forces stabilize the double helix.

or from a later energy minimization procedure (Jimenez-Barber0 et aL, 1989).

Despite the absence of internal hydrogen bonds in the double helix, agar gels have a high melting point (75-85°C). This superiority over carrageenan gels is particularly useful in bakery products, for instance, cake icing, cream cheeses, candies, cookies, and pie fillings. Although nondigestible, agar jelly noodles are widely consumed in Japan (Selby and Whistler, 1993; Stanley, 1995).

D. THE GELLAN FAMILY OF POLYSACCHARIDES

Up to eight anionic polysaccharides, including gellan, welan, S-657, and rhamsan, are secreted by unrelated bacteria and, barring substituents, their main chains have the same tetrasaccharide repeat (A-B-C-D) as the parent polymer gellan (Chandrasekaran and Radha, 1995). Chemically, barring the difference between CH20H and COOH groups, 75% of the gellan backbone (i.e., A-B-C) is cellulose-like. As shown below, the native polymer secreted by Pseudomonas elodea has both glyceryl and acetyl substituents and is referred to as native gellan. Commercial deesterification of native gellan by hot alkali treatment removes the substituents and the resulting material is called gellan. The branched polysaccharide welan (with


side chain E, Rha:Man being 2:l) is an excellent viscosifier. Since it has the potential for food applications, it is included in this review.

Crystal structures are now well established for native gellan, gellan, and welan. Most other members including S-657 and rhamsan are branched like welan. The side chains in them are flexible disaccharides and their diffraction patterns are of poor quality. Hence structural details are elusive, but through computer modeling, they are believed to exist as gellan-like helices (Lee and Chandrasekaran, 1991).

f-L-glycerate

2 Native gellan:-b3)-P-D-Glc-( 1 + 4)-P-D-GlcA-( 1 + 4)-P-D-Glc-( 1-+4)-a-~-Rha-( 1 +

i 0-acetate (0.5 occupancy)

Gellan: + 3)-P-D-Glc-( I+ 4)-P-u-GlcA-(l+ 4)-P-~-Glc-( 1 +4)-a-~-Rha-( I+

A B C D

Welan: -+ 3)-P-~-Glc-( 1+4)-P-o-GlcA-( I--* 4)-P-D-Glc-( 1 +4)-a-~-Rha-( 1+

3 0-acetate (0.85 occupancy) i

a-L-Rha or a-L-Man

The high-acyl native gellan and the acyl-free gellan are both gel-forming and texturing polysaccharides. While the former forms only soft and spongy gels, the latter forms hard and brittle gels. By combining the two in proper ratios, and by the use of appropriate cations, gels of desired quality can be prepared. X-ray analyses reveal only subtle differences in molecular morphology between the two polymers. Altered packing arrangements, however, are able to explain their respective gelling behaviors.

a. Potassium Native Gellan. An x-ray study (Chandrasekaran et al., 1992) revealed that native gellan helices are packed in a trigonal unit cell (a = b = 16.47, c = 28.42 A, and y = 120"). The molecule forms a parallel, threefold, left-handed ( t = -120") double helix of pitch 2c = 56.84 A (h = 18.95 A) as shown in Fig. 28a. Among the four monomers, a-L-rhamnose adopts 'C4, and the other three have 4C1 chair geometry. The inner and outer diameters of the helix are 1.2 and 20.7 A, respectively, implying that there is no hole along the helix axis. Four sets of conformational parameters (in degrees) at the linkages A-B, B-C, C-D, and D-A define the main chain geometry. They are specified by ( T ~ , 4,, GI) = (116.1, -99, -150),


( T ~ , &, &) = (116.6, -134, -148), (q, A, I/J3) = (115.9, -141,98), and ( T ~ , 41~, I/J4) = (116.4, -119,64), respectively; the hydroxymethyl or carboxylate orientations in the four saccharides are xA = -82", ,ye = 24", xc = 62", and xD = 91". Parameters defining the orientation of the glyceryl group at 0 2 A are: 01(C3-C2-02-C9) = 42", 02(C2-02-C9-C10) = -166", 03(02-C9- C10-Cll) = 78", and 04(C9-C10-C11-Oll) = 96". Those for the acetate group at 0 6 A are: 05(C5-C6-06-C7) = 159" and 06(C6-06-C7-C8) = 73". The nearly extended polysaccharide chain exhibits one hydrogen bond across every glycosyl bridge oxygen: cellulose-like 03HB.. .05A (2.73 A), 02HB.. .06C (2.92 A), 03HD- . .02C (2.99 A), and 04HA- . .05D (3.10 A). The glycerate group also participates in two intrachain hydrogen bonds: 010HA...061B (2.59 A) and Ol lHA. . . 03D (3.08 A). Crucial to stability of the double helix are the three interchain hydrogen bonds 06HC.- .062B (3.08 A), 06HC- . -010A (2.55 A), and 02HB.. .010A (3.11 A), of which the latter two involve the glycerate moiety. This setup leads to partial shielding of the carboxylate group so that the potassium ion located nearby has only half occupancy; also the ion has just three ligands, instead of about six as normally observed for potassium. They are atoms 062B, 05B, and 0 3 C in disaccharide B-C. Unable to reach out both chains in the helix, each ion dangles to one chain only. Two double helices pass through (a/3 2h/3 0) and (2a/3 h/3 0) of the unit cell in an antiparallel fashion (Fig. 28b), the lateral separation between them being 9.5 A. There are two ordered water molecules per tetrasaccharide repeat that form water bridges and help to stabilize the packing arrangement, which includes only a few direct hydrogen bonds. The final R-value of this structure is 0.17 for 42 reflections. Since their positions are not conducive to linking the carboxylate groups of neighboring helices, the ions are unable to strengthen the junction zones in the case of native gellan.

The structural roles of the acetyl and glyceryl groups can be readily extracted from these results. For example, the axial view of the unit cell contents (Fig. 28b) indicates that the acetyl groups on the periphery pro- trude into neighboring unit cells, but do not interfere with the association of double helices. On the other hand, each glycerate group screens an adjacent carboxylate group from efficient ion binding, thereby making ion- mediated interhelical association less probable or fragile at best. Hence, it seems that native gellan's soft and spongy gelling behavior stems from its glycerate and not acetate groups. This is further corroborated by experimental observations (Baird et af., 1992).

b. Potassium Gellan. Diffraction patterns from polycrystalline and well-oriented fibers of potassium gellan (Chandrasekaran et al., 1988b) contain nearly 40 sharp Bragg reflections up to 3 A resolution that can be


a

FIG. 28. (a) Stereo view of about a turn of the threefold double helix of native gellan. The two chains are drawn in thin and thick bonds for distinction. Both intra- and interchain hydrogen bonds stabilize the helix. The vertical line is the helix axis. Potassium ions (crossed circles) have only half occupancy at each site. (b) A c-axis projection of the trigonal unit cell contents showing an antiparallel packing arrangement of two double helices drawn in thin and thick bonds.

indexed on a trigonal unit cell (a = b = 15.75, c = 28.15 A, and y = 120"). The x-ray analysis has (a) confirmed that the previously determined molecular structure and packing arrangement of lithium gellan (Chandra- sekaran et al., 1988a) are valid for potassium gellan also, and (b) yielded structural details on several ordered water molecules in the unit cell and on the tight binding of potassium ion to gellan helix. This information was


beyond the reach of the lithium gellan investigation owing to the very low scattering power of lithium ion. With its slightly shorter axial rise per tetrasaccharide repeat ( h = 18.78 A), the potassium gellan helix ( t = -120") shown in Fig. 29a is isomorphous to that of native gellan described before. The parallel, half-staggered, double helix is stabilized by both intra- and interchain hydrogen bonds. The flexibility of the chain is tempered by three bonds per tetrasaccharide, 03HB..-OSA (2.92 A), 02HA.-.061B (2.51 A), and 02HB..-06C (2.45 A), but there are none across C-D or D-A. Apart from the interchain 06HC...062B (3.04 A), both chains in the double helix are strongly linked through six ligands to a potassium ion with full occupancy near each carboxylate group. The roughly octahedral coordination is generated by atoms 02A, 061B, and 062B, in one chain, 02C and 06C in the other, and an ordered water molecule (W). This ion cage seems to be a symbol of robustness of the gellan molecule in the presence of monovalent ions such as lithium, sodium, potassium, and rubidium. The inner and outer diameters of this rod-like helix are 1.1 and 15.8 A, respectively, and the latter corresponds to atom 06A. The backbone geometry and the side group orientations are given by ( T I , 41, 41) = (116.5, -101, -136), ( T ~ , 4 2 , &) = (116.5, -154, -144), (q, &, &) = (115.9, -150, 86), and ( T ~ , 44, t,h4) = (117.7, -124, 88); XA = -7Y, XB = lo", xc = 58", and xD = 77". The R-value for the final model is 0.18 for 51 reflections, of which 38 are observed.

Two double helices pass through (2a/3 b/3 0) and (a/3 2b/3 0) per cell in an antiparallel fashion (Fig. 29b). They are laterally separated by 9.1 A, which is 0.4 A less than in native gellan. Four structured water molecules per tetrasaccharide repeat have been located in the unit cell. Each ion cage


a

FIG. 29. (a) Stereo view of about a turn of the threefold double helix of potassium gellan. The two chains are drawn in thin and thick bonds for distinction. Both intra- and interchain hydrogen bonds stabilize the helix. The vertical line is the helix axis. Octahedrally coordinated potassium ions (crossed circles) and triply hydrogen-bonded water molecules (open circles) located above each ion are part and parcel of gellan. (b) A c-axis projection of the trigonal unit cell contents showing an antiparallel packing arrangement of two double helices, drawn in thin and thick bonds.

is juxtaposed to another in a neighboring helix so that distance between the two cations is only 4.3 A. This allows interhelical association through strong carboxylate-potassium-water-potassium-carboxylate interactions in every tetrameric repeat, which in turn strengthens the junction zones

responsible for gelation. Extrapolation of these experimental results by computer modeling (Chandrasekaran and Thailambal, 1990) shows that calcium gellan double helices would associate even stronger due to direct carboxylate-calcium-carboxylate interactions. This appears to be the molecular basis for the strong and brittle gelation behavior of calcium gellan even at low ionic strength.

Although gellan gum was discovered in 1977, approval for food applications in the United States and several other countries was not granted until the middle of this decade. It is not just another gelling agent, it is also a texturizing, stabilizing, film-forming, viscosifying, and flavor-releasing polysaccharide. Its versatility is obvious from its use in puddings, dessert gels, frostings, beverages, dairy products, fruit spreads, bakery fillings, glazes, confections, sauces, batters, and breading (Kang and Pettitt, 1993).

X-ray diffraction patterns from well-oriented and polycrystalline fibers of calcium welan (Fig. 5c) closely resemble those of gellan and contain excellent intensity data, about 80 reflections up to 2.5 A resolution (Chandrasekaran et al., 1994a). The unit cell is trigonal (a = h = 20.83, c = 28.69 A, and y = 120") and much larger than that of gellan. Detailed structure analysis reveals that welan al:o exists as a gellan-like double helix whose pitch, 2c = 57.38 A, is 1 A higher (h = 19.1 A) than that of gellan. The backbone is stiffened by a total of six intrachain hydrogen bonds per pentasaccharide repeat, including 03HB...05A (2.46 A), 03HD...02C (3.06 A), and 04HA...05D (2.93 A) within the main chain. The other three are through the rhamnose/mannose side chain



a

FIG. 30. (a) Stereo view of one turn of the threefold welan double helix. The two chains are drawn in thin and thick bonds for distinction. The vertical line represents the helix axis. Both intra- and interchain hydrogen bonds and side chains hydrogen bonded to carboxylate groups stabilize the double helix. Calcium ions (crossed circles) are present near the carboxylate groups, but outside the helix in order to make inter-double helical connections. (b) A c-axis projection of the trigonal unit cell contents shows the packing arrangement of three welan double helices. The helix drawn in solid bonds is antiparallel to the remaining helices (thin and open bonds). Note that calcium ions are positioned between helices and each water molecule (large open circle) shown here is connected to all three surrounding helices. Several other ordered water molecules (not shown) have been located in the interstitial space.


(E), namely, 03HE.e.062B (3.04 A), 04HE.a.062B (2.58 A), and 02HC-a-05E (2.84 A). Gellan-like 06HC...062B (3.08 A) and side chain-promoted 03HE-..05C (3.16 A) and 02HB...07A (2.49 A) are interchain hydrogen bonds. Consequently, the welan double helix (Fip. 30a) is very sturdy. Its inner and outer diameters are 1.2 and 13.0 A, respectively. Surprisingly, in spite of its side chains, the welan helix is marginally slimmer than native gellan and gellan helices. The five sets of conformational parameters in degrees are given by (q, $1, t+bl) =

(116.6, -90, -157), (72, 42, &) = (116.6, -147, -157), (73, &, &) =

(115.9, -155, 96), (T~, 44, &) = (117.6, -104, 90), and ( T ~ , 45, t+b5) = (116.5, -35, 168) for the side chain; the hydroxymethyl or carboxylate orientations are xA = -70", ,ye = 37, xc = 85", XD = 84", and XE =

37". The orientation of the acetyl substituent at 0 2 A is given by 81(C3-C2-02-C7) = -66" and &(C2-02-C7-07) = -21". The R- value for the final model is 0.22 for 102 reflections, of which 76 are observed.

b

FIG. 30. Continued


Three welan helices pass through the unit cell as shown in Fig. 30b. Two of them, I and 11, respectively at (2d3 h/3 0) and (a/3 2h/3 0), are antiparallel as in gellan and the third, 111, at (0 0 0), is new and parallel to the first. They are equidistant (12.0 A) from each other, 2.9 A farther than in gellan. The unit cell accommodates 9 pentasaccha- rides, 6 calcium ions, and 75 ordered water molecules. The experimental results (Chandrasekaran et al., 1994a) show that welan helices interact via side chains and surrounding guest molecules. Specifically, I and I1 are linked only by side chains-eside chain hydrogen bonds involving atoms 0 2 and 0 4 . I and 111, aligned parallel, are connected by carboxylate-calcium-water-carboxylate interactions in addition to hydrogen bonds between residues A and E. I1 and 111, aligned antiparallel, are held together by very strong carboxylate-calcium-carboxylate interactions in addition to side chaineaemain chain hydrogen bonds. Largely facilitated by ordered water molecules, these interactions are responsible for the intense viscosifying behavior of welan in solutions up to 130°C. Thus, the crystal structure of welan is an excellent comprehensive visualization of the constructive role of side chains to the stability of a branched polysaccharide helix whose favorable associative interactions at the molecular level relate to its macroscopic properties.

E. BRANCHED POLYSACCHARIDES

Side chains in polysaccharides can have stabilizing as well as destabilizing influences on structural organization. The structure of welan just described attests to the stabilizing effect of side chains on secondary structure and intermolecular ordering. The gel-forming, doubly branched, capsular polysaccharide (CPS) from Rhizohium trifolii having a hexasaccharide repeat also has an ordered helical structure: the side chains stabilize the molecular structure but not its packing arrangement. The monosaccharide side chains in galactomannans are flexible (Section III,A,3). Xanthan has a pentasaccharide repeat that includes a trisaccharide side chain that is far more flexible than the monosaccharide side chain in welan or galactomannan, and hence its structure remains elusive. Long side chains are susceptible to conformational fluctuations, resulting in less ordered structures. Such systems are invariably difficult to solve.

1. Rhizobium trifolii CPS

The doubly branched galactose-rich gel-forming capsular polysaccharide from R. trifolii consists of a hexamer repeat composed of main chain


-A-B-C- and two side chains, a monosaccharide D and a disaccharide F-E; both D and E are attached to A as shown below:

D a -~ -Ga l

i 2

+ 4)-a-D-Glc-( 1+3)-a-D-Man-( 1 + 3)-P-D-Gal-( 1- A6 B C t I

i P-D-Gal-( I-4)-P-D-Gal

F E

The neutral polysaccharide is water soluble and thermoreversible gels are formed in the temperature range 42 to 49°C in the absence of ionic co- solutes. For this reason it has the potential for food applications. Consistent with the x-ray data consisting of predominantly continuous intensities on layer lines and one or two Bragg reflections, the polymer forms a right- handed twofold helix ( t = 180") of pitch 20.2 A (h = 10.1 A). The molecular structure has been determined and refined against the continuous intensities to a final R-value of 0.28 (Lee and Chandrasekaran, 1992). As illustrated in Fig. 31a, the helix is very sinuous and it progresses with a right-handed twist. The six sets of conformational parameters in degrees at A-B, B-C, C-A, D-A, E-A, and F-E are given by (T,, +,, = (116.6, 67, lOl), (Q, 42, &) = (116.8,65, loo), (73, &, $3) = (116.9, -84, -135), (74, 44, $4) = (116.7, 101, -94), (Q, &, I,!I~) = (116.6, -90, -loo), and (76, 46, $6) =

(116.6, -132, 113), respectively; the hydroxymethyl orientations are xA =

85", X B = -172", xc = 70", X D = -65", xE = 79" and xF = 83". There are eight hydrogen bonds per repeat, of which four are bifurcated. The interior main chain is stabilized by 03HA.- -05C/05D (2.72/2.79 A) and 06HB...02C (3.00 A). The peripheral mono- and disaccharide side chains are connected to successive repeats by 02HD.e.03F (3.03 A) and 04HF. . .02D/03D (3.03/3.08 A). The disaccharide side chain is internally stiffened by 06HE.v-05F (2.80 A) and 02HF. . .03E (2.52 A). The extensively hydrogen-bonded polymer chain has the appear$nce of a double helix whose inner and outer diameters are 2.2 and 19.8 A, respectively.

A putative packing arrangement in a monoclinic unit cell (a = 16.8, b = 9.7, c = 20.2 A, and y = 90') is shown in Fig. 31b. It reveals that there are no direct interactions among main chains of adjacent helices, but side chains may generate hydrogen bonds. Such an ordered picture is, however,

-

196

a


FIG. 31. (a) Stereo view of two turns of the right-handed, twofold helix of R. frifolii CPS. The mono- and disaccharide side chains (thick bonds) generate intrachain hydrogen bonds so that the molecule has the appearance of a double helix composed of two, inner and outer, strands. The vertical line represents the helix axis. (b) A c-axis projection of a monoclinic unit cell shows the putative packing arrangement of helices promoted by hydrogen bonds involving side chains.

not supported by the diffraction pattern. Perhaps the side chains in this CPS promote helix stability, but not interhelical association.

2. Xanthan

The microbial polysaccharide xanthan from Xanthomonas campestris has a pentasaccharide repeat composed of a cellobiose (disaccharide) in the main chain and an anionic trisaccharide side chain as shown below:


b

FIG. 31. Continued

+ 4)-P-D-Glc-( 1 + 4)-P-D-Glc-( I -+

i’ 1

P-~-Man4,6-Pyr-( 1 - 4)-P-D-GlcA-( 1 -+ 2)-a-~-Man6-acetate

Xanthan is extensively used in many food applications because of its unique rheological properties (Kang and Pettitt, 1993; Morris, 1995). For example, its very high viscosity in aqueous solution is independent of pH in the range 2 to 12, and insensitive to salt effects and temperature from 18 to 80°C. Xanthan is added as an emulsifier in salad dressings and as a dispersion agent in desserts, gravies, sauces, and beverages. It is used in syrups, toppings, and relishes for enhancing consistency and flow properties.


Batters for baked goods are often mixed with xanthan in order to achieve required pseudoplastic properties.

X-ray diffraction patterns from xanthan fibers show much diffuse continuous intensity on layer lines and very few Bragg reflections (Moorhouse et al., 1977; Okuyama et al., 1980). Although the intensity data are not conducive to conducting detailed structure analysis, meridional spots on the 5n layer lines and their spacings indicate that the polymer has five-fold helix symmetry and a pitch of 47.0 A; the cellobiose repeat in the main chain has h = 9.4 A, which is 0.8 A shorter than that of cellulose. From a knowledge of the helical parameters alone, several speculative models have been proposed so far. They include single and double helices; parallel and antiparallel chains; left- and right-handed chiralities; and up to four chains in the unit cell. However, none of the models provide an acceptable x-ray fit. Consequently, the molecular details of xanthan remain unknown. A right-handed ( t = 72") antiparallel duplex, generated by molecular modeling, is shown in Fig. 32 as an example of the various possibilities. Since the side chain is longer than the main chain in each repeat, it has greater flexibility to wrap around the main chain, reach out neighboring side chains, or extend out radially as much as possible. The range of interactions among adjacent side chains is perhaps so overwhelming that it may never be feasible to trap xanthan molecules in an ordered state for detailed diffraction study.

F. ARABINAN

Arabinan, a plant polysaccharide commonly found in large quantities in sugar beet, various seeds and roots, has a main chain composed of (1-5)- linked a-L-arabinofuranose residues. It occurs as a side chain attached to a long pectate chain through a rhamnose residue. Arabinan itself is a branched polymer containing arabinofuranose monomers as side chains at random. The enzymatically debranched linear polymer (DP 50-80), called araban, has very low water solubility, forms a spreadable gel similar to maltodextrins, and exhibits rheological properties like those of high-fat products and thus can be used in food applications as fat mimetics (Cooper et al., 1992). It has also been reported that an L-arabinan obtained from apple juice is essentially linear (Churms et al., 1983). Conformational analysis has led to six probable models, two-, three-, four-, and six-fold helices with distinct values of pitch, for arabinan (Cros et al., 1994).

Oriented fibers of araban diffract like a microcrystalline powder sample, as shown in Fig. 5a. The spacings of the concentric rings in this pattern are consistent with a monoclinic unit cell of dimensions a = 5.55, b = 6.14, c = 8.74 A, and y = 90", which could accommodate only two monomers (Chandrasekaran et al., 1994b). This implies that a twofold helix of pitch


FIG. 32. Stereo view of about a turn of an antiparallel, right-handed, double helix as a probable molecular model of xanthan. The two strands are in thin and thick bonds for clarity. Both intra- and interchain hydrogen bonds are present. The side chains are turned toward the nonreducing end and aligned almost parallel to the helix axis (vertical line). The helix is further stabilized by hydrogen bonds between side chains.

a, b, or c is a good possibility if the araban chain adopts a regular structure. Conventional structure determination of a helical polymer is not possible with powder diffraction data. Hence, molecular modeling based on energy calculations has been employed to explore both C2’-endo and C3’-endo pentose conformations as viable options; the three conformation angles 4(C2-Cl-O5’-C5’), @(Cl-O5’-CS’-C4’), and x(OS-CS-C4-C3) are the other variables. This has given rise to seven low-energy molecular models, all with the same pitch ( c = 8.74 A). There are considerable differences among them in terms of molecular geometry and crystal packing arrangement. Three of them correspond to C2’-endo and the other four to C3’-endo sugar rings. Two selected models, having the lowest energy within each


kind, are shown in Fig. 33. The C2'-endo helix (Figs. 33a, 33b) characterized by (7, 4, $, x) = (116.5,116,132, -66") is stabilized by intrachain hydrogen bond 02H.. .04 (2.82 A) across the glycosidic oxygen atom. It has inner and outer diameters of 1.4 and 6.0 A, respectively. Crystalline packing is facilitated by interhelix hydrogen bonds 03H...04/02 (2.70/2.82 A). On the other hand, the C3'-endo helix (Figs. 33c, 33d) characterized by (7, 4, I), x) = (116.5,54, -119, -59") does not have intrachain hydrogen bonds. Its inner and outer diameters are 0.7 and 6.5 A, respectively. Its crystalline packing arrangement is facilitated by 03H...02 (3.0 A) hydrogen bonds. Similar variations are present in the other twofold helical models (Radha and Chandrasekaran, 1997).

Due to the small energy difference between any pair, transition from one allomorph to another could be easily induced by breaking old and making new hydrogen bonds within and between helices. This polymorphism and tendency to form microcrystals in the solid state, as evident from the x-ray diffraction pattern, might confer upon arabinan or araban the observed fat mimetic properties.

IV. MIXED POLYSACCHARIDES

The quest for new polysaccharides superior to those now in use for food applications in this country is greatly hindered by factors such as the time (years) and expense (millions of dollars) it takes to get the FDA's approval for human consumption. A sensible approach for both cost effectiveness and quick availability is to tailor mixed polysaccharide systems having desired functional properties. This exploits the synergism between inexpensive materials and the necessity of using an expensive ingredient already in vogue.

Several laboratories have reported interesting results from a series of physicochemical (Dea et al., 1977: Cheetham and Mashimba, 1991; Doub- lier and Llamas, 1991; Shatwell et al., 1991a,b,c; Lundin and Hermansson, 1995) and x-ray (Cairns et al., 1986; Brownsey et al., 1988; Millane and Wang, 1990) studies during the past 20 years on the synergistic interactions of at least four major complexes: K-carrageenan:galactomannan, K-carrageenan:glucomannan, agarose:glucomannan, and xanthan:galacto-

FIG. 33. Two probable molecular structures and packing arrangements of twofold araban helices. Stereo view of two unit cells roughly normal to the bc-plane. The helices in the front are in thick bonds and those in the back are in thin bonds: (a) CZ'-endo and (c) C3'-endo. A c-axis projection of the unit cell: (b) C2'-endo and (d) C3'-endo.


a

b

C

d


mannan systems. In each case, an anionic and a neutral polysaccharide are involved. No specific details on the molecular interactions have yet emerged. For the xanthan:galactomannan complex, it has long been believed that the backbone of xanthan helix associates with the naked region of galactomannan chain as inferred from schematic drawings (Cairns et al., 1986; Cheetham and Mashimba, 1991). A recent modeling study guided by limited x-ray data has revealed preliminary molecular details of this complex for the first time (Chandrasekaran and Radha, 1997). The packing arrangement of cellulose I (Fig. 8) is one of the options that evokes backbone-backbone interactions between xanthan and galactomannan if both take up twofold helices like cellulose and mannan, respectively. An alternative is a fivefold double helix of pitch 47.4 A in which a xanthan chain intertwines a galactomannan chain. Whether the strands are parallel or antiparallel, and whether the helix is right- or left-handed, are unresolved. One of the putative helices (Chandrasekaran and Radha, 1997) suggests that the hybrid may closely resemble the xanthan model depicted in Fig. 32. Since neither xanthan nor locust bean gum is known for gel formation, this type of intimate association is attractive with regard to onset of gelation observed for the complex.

V. MORPHOLOGY TO MACROSCOPIC PROPERTIES

The extended ribbon structures of cellulose and mannan enable them to strongly associate among themselves through hydrogen bonds. Such an effect prevents them from intimate interactions with solvents of any kind. This insolubility is essential for the integrity of cell walls in plants. Either structure is a valid model for the related glucomannan (konjac mannan), which is a P(1+4)-linked random copolymer of glucose and mannose. In fact, the mannan I1 packing arrangement is viable for glucomannan also (Millane and Hendrixson, 1994). Introduction of side chains as in galactomannans and xanthan generates bumps along their main chains at regular or random intervals, creating big holes between neighboring polymer molecules to such an extent that polymer-solvent interactions exceed polymer- polymer interactions. As a result, an aqueous solution of either polymer becomes very viscous and is suitable for food applications. The same phenomenon is applicable to several cellulose derivatives, the strategy for the insoluble to soluble polymer conversion being the breakdown of the otherwise strong self-association.

It is quite remarkable that several polysaccharides, namely pectin, alginate, curdlan, carrageenan, and gellan, are known for their gelation proper-


ties and most of them are used in food applications. In principle, any gelling polysaccharide would do the job as well as any other; preferences are based on subtlety in behavior, availability, and above all economics.

Compatible with its function as a storage polysaccharide, amylose and its derivatives adopt a variety of structures including single and double helices. The ability to form a short and wide helix or a ring structure such as cyclodextrin is intended for capturing a spectrum of small molecules. The double helix is vital for the crystalline arrangement of amylopectin that gives robustness to starch granules.

Mixed-linkage (1+3), (1+4)-P-~-glucans, found in oat and barley cereals, are called soluble fibers or dietary fibers. While a detailed secondary structure of P-glucan is almost impossible to determine, it is not unthinkable that the (1+4) and (1+3) blocks might resemble the corresponding cellulose and curdlan structures previously described. A short stretch of an extended sixfold single helix of curdlan I (Okuyama et af., 1991) is a probable motif for the latter block. The triple helix of curdlan I11 (Deslandes et af., 1980) would require a cooperative association of three strands in the dietary fiber in proper register. The hexameric association of curdlan triple helices helps to form micelles (Fulton and Atkins, 1980). The existence of such a state in dietary fibers is realistic and consistent with the finding of micelle-like aggregation from light scattering experiments (VArum et af., 1992).

A comparison of the ordered structures of welan and R. trifofii CPS with those of galactomannans and predicted for xanthan gives mixed messages on the correlation between side chains and molecular morphology. The myth is that side chains interfere with helix formation, lateral organization, or both. This is consistent with the difficulty in inducing order in fibers of xanthan in which the main chains prefer extended ribbon structures and hydrogen-bonded sheets, but side chains have to be somehow accommodated. On the other hand, galactomannan, welan, and R. trifofii CPS helices require their side chains to promote helix stability and/or packing arrangement. These observations, taken together, point to the dual (constructive and destructive) roles of side chains to polymer geometry as well as to broader practical applications. Since flexibility prevails for long side chains, it is likely that backbone conformation is regular, but those of side chains are not. This leads to a partially ordered system. Similarly, several alternatives exhibiting different levels of ordering exist. Therefore, it is nearly impossible to correctly predict the structural behavior of branched polysaccharides in the solid state or in solution with currently available experimental and theoretical methods.


VI. SUMMARY

The morphologies of food polysaccharides described in this chapter illus- trate the power of x-ray fiber diffraction in conjunction with computer modeling and sophisticated refinement techniques. On the other hand, the lack of information on structures such as xanthan reflects the inadequacy of the experimental techniques used to date. But the demands from academic and industrial sectors to investigate the molecular interactions in multicomponent systems, including protein-protein, protein-polysaccharide, polysaccharide-polysaccharide, and other complexes, are high and growing, because they have important food applications. These complexes are structurally more difficult than those solved in the past 40 years and it is improbable that any chosen system will be amenable for crystallographic investigation, crystals or fibers. Modern research facilities that include two- dimensional area detectors, millisecond exposures with synchrotron x-ray radiation, interactive computer graphics, sophisticated molecular dynamics calculations, unbelievably fast and inexpensive computers, and our own intellectual abilities are indispensable tools for the future of structural science in general and food polysaccharides in particular.

ACKNOWLEDGMENT

This work was supported by the Industrial Consortium of the Whistler Center for Carbohy- drate Research.

REFERENCES

Arndt, E. R., and Stevens, E. S. 1996. Vacuum ultraviolet circular dichroism of gellan-family polymer films from water and dimethyl sulfoxide. Carbohydr. Res. 280, 15-26.

Arnott, S., Scott, W. E., Rees, D. A., and McNab, C. G. A. 1974a. &-Carrageenan: Molecular structure and packing of polysaccharide double helices in oriented fibers of divalent cation salts. J. Mol. Biol. 90, 253-267.

Arnott, S., Fulmer, A., Scott, W. E., Dea, I. C. M., Moorhouse, R., and Rees, D. A. 1974b. The agarose double helix and its function in agarose gel structure./. Mol. Biol. 90,269-284.

Atkins, E. D. T., Nieduszynski, I. A., Mackie, W., Parker, K. D., and Smolko, E. E. 1973a. Structural components of alginic acid. I. The crystalline structure of poly-P-D-mannuronic acid. Results of x-ray diffraction and polarized infrared studies. Biopolymers 12, 1865- 1878.

Atkins, E. D. T., Nieduszynski, I. A., Mackie, W.. Parker, K. D., and Smolko, E. E. 1973b. Structural components of alginic acid. 11. The crystalline structure of poly-a-L-guluronic acid. Results of x-ray diffraction and polarized infrared studies. Biopolymers 12, 1879- 1887.

Atkins, E. D. T., Farnell, S., Mackie, W., and Sheldrick, B. 1988. Crystalline structure and packing of mannan I. Biopolymers 27,1097-1105.


Baird, J. K., Talashek, T. A., and Chang, H. 1992. Gellan gum: Effect of composition on gel properties. In “Gums and Stabilisers for the Food Industry” (G. 0. Phillips, P.A. Williams, and D. J. Wedlock, eds.), pp. 479-487. IRL Press, Oxford.

BeMiller, J. N. 1993. Starch-based gums. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 579-600. Academic Press, New York.

Biliaderis, C. G. 1991. The structure and interactions of starch with food constituents. Can. J. Physiol. Pharmucol. 69, 60-78.

Biliaderis, C. G., and Galloway, G. 1989. Crystallization behavior of amylose-V complexes: Structure-property relationships. Curbohydr. Res. 189, 31 -48.

Biliaderis, C. G., Page, C. M., Slade, L., and Sirett, R. R. 1985. Thermal behavior of arnylose- lipid complexes. Curbohydr. Polym. 5,367-389.

Bluhm, T. L., and Zugenrnaier, P. 1979a. The crystal and molecular structure of tri-0-ethylamylose (TEA 3). Curbohydr. Res. 68,15-21.

Bluhm, T. L., and Zugenmaier, P. 1979b. Conformation and packing analysis of polysaccharides and derivatives: 6. Detailed refinement of tri-0-ethylamylose-solvent complexes TEA1- C1 and TEAl-DCMl. Polymer 20,23-30.

Bluhm, T. L., and Zugenmaier, P. (1981). Detailed structure of the Vh-amylose-iodine complex: a linear polyiodine chain. Curbohydr. Res. 89, 1-10,

Bluhm, T. L., Deslandes, Y., Marchessault, R. H., PCrez, S., and Rinaudo, M. 1982. Solid- state and solution conformation of scleroglucan. Curbohydr. Res. 100, 117-130.

Booy, F. P., Chanzy, H., and Sarko, A. 1979. Electron diffraction study of single crystals of amylose complexed with n-butanol. Biopolymers 18,2261-2266.

Branden, C., and Tooze, J. (eds.) 1991. “Introduction to Protein Structure.” Garland Publish- ing, New York.

Brigand, G. 1993. Scleroglucan. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 461-474. Academic Press, New York.

Brownsey, G. J., Silvers, G. R., I’Anson, K., and Morris, V. J. 1984. Some observations (or problems) on the characterization of gellan gum solutions. Int. J . Biol. Mucromol.

Brownsey, G. J., Cairns, P., Miles, M. J., and Morris, V. J. 1988. Evidence for intermolecular binding between xanthan and the glucomannan konjac gum. Curbohydr. Res. 176,329-334.

Burchard, W. 1994. Light scattering techniques. In “Physical Techniques for the Study of Food Biopolymers” (S. B. Ross-Murphy, ed.), pp. 151-213. Blackie A & P, London.

Cairns, P., Miles, M. J., and Morris, V. J. 1986. Intermolecular binding of xanthan and carob gum. Nature 322, 89-90.

Chandrasekaran, R., and Radha, A. 1995. Molecular architectures and functional properties of gellan gum and related polysaccharides. Trends Food Sci. Technol. 6, 143-148.

Chandrasekaran, R., and Radha, A. 1997. Molecular modeling of xanthan:galactornannan interactions. Curbohydr. Polym. 32, 201-208.

Chandrasekaran, R., and Thailambal, V. G. 1990. The influence of calcium ions, acetate and L-glycerate groups on the gellan double-helix. Curbohydr. Polym. 12, 431-442.

Chandrasekaran, R., Millane, R. P., Arnott, S., and Atkins, E. D. T. 1988a. The crystal structure of gellan. Curbohydr. Res. 175, 1-15.

Chandrasekaran, R., Puigjaner, L. C., Joyce, K., and Arnott, S. 1988b. Cation interactions in gellan: An x-ray study of the potassium salt. Curbohydr. Res. 181, 23-40.

Chandrasekaran. R., Radha, A., and Thailambal, V. G. 1992. Roles of potassium ions, acetyl and L-glyceryl groups in native gellan double helix: An X-ray study. Curbohydr. Res. 224, 1-17.

6, 211-214.


Chandrasekaran, R., Radha, A,, and Lee, E. J. 1994a. Structural roles of calcium ions and side chains in welan: An x-ray study. Carbohydr. Res. 252, 183-207.

Chandrasekaran, R., Radha, A., Lee E. J . , and Zhang, M. 1994b. Molecular architecture of araban, galactoglucan and welan. Carbohydr. Polym. 25, 235-243.

Chandrasekaran, R., Radha, A,, and Okuyama, K. 1998. Morphology of galactomannans: An x-ray structure analysis of guaran. Carbohydr. Rex, in press.

Chanzy, H.. Perez, S., Miller, D. P., Paradossi, G., and Winter, W. T. 1987. An electron diffraction study of the mannan I crystal and molecular structure. Macromoleciiles 20,

Cheetham, N. W. H., and Mashimba, E. N. M. 1991. Conformational aspects of xanthan- galactomannan gelation. Further evidence from optical rotation studies. Carbohydr. Po- lym. 14, 17-27.

Chien, Y. Y., and Winter, W. T. 1985. Accurate lattice parameters for tara gum. Macromole- cules 18, 1357-1359.

Chuah, C. T., Sarko, A., Deslandes, Y., and Marchessault, R. H. 1983. Triple-helical crystalline structure of curdlan and paramylon hydrates. Macromolecules 16, 1375-1382.

Churms, S. C., Merrifield, E. H., Stephen, A. M., Walwyn, D. R., Polson, A., van der Menve, K. J., Spies, H. S. C., and Costa, N. 1983. An L-arabinan from apple-juice concentrates. Carbohydr. Rex 113,339-344.

Clare, K. 1993. Algin. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 105-143. Academic Press, New York.

Clark, A. H. 1994. X-ray scattering and diffraction. In “Physical Techniques for the Study of Food Biopolymers” (S. B. Ross-Murphy, ed.), pp. 65-149. Blackie A & P, London.

Clarke, R. J., Coates, J. H., and Lincoln, S. F. 1988. Inclusion complexes of the cyclomaltooligo- saccharides (cyclodextrins). Adv. Carbohydr. Chem. Biochem. 46,205-249.

Coffey, D. G., Bell, D. A,, and Henderson, A. 1995. Cellulose and cellulose derivatives. In “Food Polysaccharides and Their Applications” (A. M. Stephen, ed.), pp. 123-153. Marcel Dekker. New York.

Cooper, J. M., McCleary, B. V., Morris, E. R., Richardson, R. K., Marrs, W. M., and Hart, R. J. 1992. Preparation, physical properties and potential food applications of enzymically- debranched araban from sugar beet. In “Gums and Stabilizers for the Food Industry 6” (G. 0. Phillips, P. A. Williams, and D. J. Wedlock, eds.), pp. 451-460. IRL Press, Oxford.

Crescenzi, V., Dentini. M., Coviello, T., Paoletti, S., Cesiro. A,, and Delben. F. 1987. On the solution and gelling behaviour of some bacterial polysaccharides. Gazerra Chim. Zral. 117, 611-616.

Cros, S., Imberty, A,. Bouchemal. N., Penhoat, C. H. D., and PCrez, S. 1994. Modeling of arabinofuranose and arabinan. 11. NMR and conformational analysis of arabinobiose and arabinan. Biopolymers 34, 1433-1437.

Dea, I. C. M., Morris, E. R.. Rees, D. A., Welsh, E. J., Barnes, H. A., and Price, J. 1977. Associations of like and unlike polysaccharides: Mechanisms and specificity in galactomannans, interacting bacterial polysaccharides, and related systems. Carbohydr. Res. 57,249-272.

Deslandes, Y., Marchessault, R. H., and Sarko, A. 1980. Triple-helical structure of (1-+3)+- o-glucan. Macromoleciiles 13, 1466-1471.

Desmarais, A. J., and Wint, R. F. (1993). Hydroxyalkyl and ethyl ethers of cellulose. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 505-535. Academic Press, New York.

Doublier, J-L., and Llamas, G. 1991. Flow and viscoelastic properties of mixed xanthan gum + galactomannan systems. In “Food Polymers, Gels and Colloids” (E. Dickinson, ed.), pp. 349-356. Royal Society of Chemistry, London.

2407-24 13.


Eliasson, A.-C. (ed.) 1996. “Carbohydrates in Food.” Marcel Dekker, New York. Feddersen, R. L., and Thorp, S. N. 1993. In “Industrial Gums: Polysaccharides and Their

Derivatives“ (R . L. Whistler and J. N. BeMiller, eds.). pp. 537-578. Academic Press, New York.

Foord, S . A.. and Atkins. E. D. T. 1989. New x-ray diffraction results from agarose: Extended single helix structures and implications for gelation mechanism. Eiopolymers 28, 1345- 1365.

Frei, E., and Preston, R. D. 1968. Non-cellulosic structural polysaccharides in algal cell walls. 111. Mannan in siphoneous green algae. Proc. R. Soc. London Ser. B 169, 127-145.

French, A. D., and Zobel, H. F. 1967. X-ray diffraction of oriented amylose fibers. I. Amylose dimethyl sulfoxide complex. Eiopolymers 5, 457-464.

French, D. 1984. Organization of starch granules. In “Starch: Chemistry and Technology” (R. L. Whistler, J. N. BeMiller, and E. F. Paschall, eds.), pp. 183-247. Academic Press, New York.

Fulton, W. S . , and Atkins. E. D. T. 1980. The gelling mechanism and relationship to molecular structure of microbial polysaccharide curdlan. Am. Chem. Soc. Symp. Ser. 141, 385-410.

Gardiner, E. S . . and Sarko, A. 1985. Packing analysis of carbohydrates and polysaccharides. 16. The crystal structure of celluloses IVI and IVII. Can. J . Chem. 63, 173-180.

Gardner, K. H., and Blackwell, J. 1974. The structure of native cellulose. Biopolymers 13,1975- 2001.

Godet, M. C., Bulkon, A., Tran, V., and Colonna, P. 1993a. Structural features of fatty acid- amylose complexes. Curbohydr. Polym. 21, 91-95.

Godet, M. C., Tran. V.. Delage, M. M., and Buleon, A. 1993b. Molecular modelling of the specific interactions involved in the amylose complexation by fatty acids. Inr. J. Eiol. Macromol. 15, 11-16.

Godet. M. C., Bizot, H.. and Buleon, A. 1995. Crystallization of amylose-fatty acid complexes prepared with different amylose chain lengths. Cnrbohydr. Polym. 27, 47-52.

Grant. G. T., Morris, E. R., Rees, D. A,, Smith, P. J. C., and Thorn, D. 1973. Biological interactions between polysaccharides and divalent cations: The egg-box model. FEES Lett. 32, 195-198.

Grover, J. A. 1993. Methyl cellulose and its derivatives. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 475-504. Academic Press, New York.

Gunning, A. P.. and Morris, V. 1. 1990. Light scattering studies of tetramethylammonium gellan. Int. J . Eiol. Macromol. 12,338-341.

Hamilton, W. C. 1965. Significance tests on crystallographic R factor. Acru Crystnllogr. 18, 502-5 13.

Harada, T., Terasaki, M., and Harada, A. 1993. Curdlan. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller. eds.), pp. 427-445. Academic Press, New York.

Hermansson, A.-M., and Langton, M. 1994. Electron Microscopy. In “Physical Techniques for the Study of Food Biopolymers” ( S . B. Ross-Murphy, ed.), pp. 277-341. Blackie A & P, London.

Hinkle, M. E., and Zobel, H. 1968. X-ray diffraction of oriented amylose fibers. 111. The structure of amylose-n-butanol complexes. Eiopolymers 6, 11 19-1128.

Imberty, A,, and Perez, S . 1988. A revisit to the three-dimensional structure of B-type starch. Eiopolywers 27, 1205-1221.

Imberty, A,, and Perez. S . 1989. Conformational analysis and molecular modelling of the branching point of amylopectin. Inr. J . Eiol. Macromol. 11, 177-185.


Imberty, A., Chanzy, H., PCrez, S., BulCon, A,, and Tran, V. 1988. The double-helical nature

Imberty, A., BulCon, A., Tran, V., and Perez, S. 1991. Recent advances in knowledge of starch

IUPAC-IUB Commission on Biochemical Nomenclature. 1971. Arch. Biochem. Biophys.

Jimenez-Barbero, J., Bouffar-Roupe, C., Rochas, C., and Ptrez, S. 1989. Modelling studies of solvent effects on the conformational stability of agarobiose and neoagarobiose and their relationship to agarose. Int. J . Biol. Macromol. 11, 265-272.

Kang, K. S., and Pettitt, D. J. 1993. Xanthan, gellan, welan and rhamsan. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 341-397. Academic Press, New York.

of the crystalline part of A-starch. J . Mol. Biol. 201, 365-378.

structure. StarcWStarke 43, 375-384.

145,405-421.

Kirby, A. R., and Morris, V. J. 1994. New microscopes feel their way. Biologist 41, 9-13. Kolpak F. J., and Blackwell, J. 1976. Determination of the structure of cellulose. 11. Macromole-

cules 9, 273-278. Lee, E. J., and Chandrasekaran, R. 1991. X-ray and computer modeling studies on gellan-

related polymers: Molecular structures of welan, S-657, and rhamsan. Curbohydr. Res.

Lee, E. J., and Chandrasekaran, R. 1992. The “pseudo double-helical’’ structure of the gel- forming capsular polysaccharide from Rhizobium trifolii. Curbohydr. Res. 231,171-183.

Lundin, L., and Hermansson, A-M. 1995. Supermolecular aspects of xanthan-locust bean gum gels based on rheology and electron microscopy. Curbohydr. Polym. 26,129-140.

Mackie, W., PCrez, S., Rizzo, R., Taravel, F., and Vignon, M. 1983. Aspectsof the conformation of polyguluronate in the solid state and in solution. Int. J . Biol. Macromol. 5, 329-341.

Maier, H., Anderson, M., Karl, C., Magnuson, K., and Whistler, R. L. 1993. Guar, locust bean, tara, and fenugreek gums. In “Industrial Gums. Polysaccharides and Their Deriva- tives” (R. L. Whistler and J . N. BeMiller, eds.), pp. 181-226. Academic Press, New York.

Marchessault, R. H., Deslandes, Y., Ogawa, K., and Sundararajan, P. R. (1980). X-ray diffraction data for @-(1+3)-~-glucan. Can. J . Chem. 55, 300-303.

Meyer, K. H., and Misch, L. 1937. Positions des atomes dans le nouveau modde spatial de la cellulose. Helv. Chim. Acta 20, 232-244.

Millane, R. P. 1989. Relating reflection boundaries in x-ray fiber diffraction patterns to specimen morphology and their use for intensity measurement. J. Mucromol. Sci. -Phys. B

Millane, R. P., and Arnott, S. 1985a. Background removal in x-ray fiber diffraction patterns. J . Appl. Crystallogr. 18, 419-423.

Millane, R. P., and Arnott, S. 1985b. Digital processing of x-ray diffraction patterns from oriented fibers. J . Macromol. Sci. -Phys. B 24, 193-227.

Millane, R. P., and Hendrixson, T. L. 1994. Crystal structures of mannans and glucomannans. Carbohydr. Polym. 25,245-251.

Millane, R. P., and Wang, B. 1990. A cellulose-like conformation accessible to the xanthan backbone and implications for xanthan synergism. Curbohydr. Polym. 13, 57-68.

Millane. R. P., Chandrasekaran, R., Arnott, S., and Dea, I. C. M. 1988. The molecular structure of kappa-carrageenan and comparison with iota-carrageenan. Carbohydr. Res. 182,l-17.

Moorhouse, R., Walkinshaw, M. D., and Arnott, S. 1977. Xanthan gum-Molecular conformation and interactions. Am. Chem. Soc. Symp. Ser. 45, 90-102.

Morris, E. R. 1994. Chiroptical methods. In “Physical Techniques for the Study of Food Biopolymers” ( S . B. Ross-Murphy, ed.), pp. 15-64. Blackie A & P, London.

Morris, V. J. 1995. Bacterial polysaccharides. In “Food Polysaccharides and Their Applica- tions” (A. M. Stephen, ed.), pp. 341-375. Marcel Dekker, New York.

214, 11-24.

28, 149-166.


Nieduzynski, I., and Marchessault, R. 1972. The crystalline structure of poly-P,~(1+4‘)mannose: Mannan I. Can. J. Chem. 50,2130-2138.

Okuyama, K., Arnott, S., Moorhouse, R., Walkinshaw, M. D., Atkins, E. D. T., and Wolf- Ullish, Ch. 1980. Fiber diffraction studies of bacterial polysaccharides. Am. Chem. SOC. Symp. Ser. 145,411-427.

Okuyama, K., Otsubo, A., Fukuzawa, Y., Ozawa, M., Harada, T., and Kasai, N. 1991. Single- helical structure of native curdlan and its aggregation state. J. Carbohydr. Chem. 10,

Radha, A., and Chandrasekaran, R. 1997. X-ray and conformational analysis of arabinan. Curbohydr. Res. 298,105-115.

Rappenecker, G., and Zugenmaier, P. 1981. Detailed refinement of the crystal structure of Vh-amylose. Curbohydr. Res. 89, 11 -19.

Rolin, C. 1993. Pectin. In “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 257-293. Academic Press, New York.

Ross-Murphy, S. B. (ed.) 1994. “Physical Techniques for the Study of Food Biopolymers.” Blackie A & P, London.

Saenger, W. 1984. In “Principles of Nucleic Acids Structure” (C. R. Cantor, ed.). Springer- Verlag, Berlin.

Sarko, A., and Biloski, A. 1980. Crystal structure of the KOH-amylose complex. Curbohydr. Res. 79, 11-21.

Sarko, A., and Marchessault, R. H. 1967. The crystalline structure of amylose triacetate. I. A stereochemical approach. J. Am. Chem. SOC. 89,6454-6462.

Sarko, A,, and Muggli, R. 1974. Packing analysis of carbohydrates and polysaccharides. Ill. Valonia cellulose and cellulose 11. Macromolecules 7,486-494.

Sarko, A,, and Zugenrnaier, P. 1980. Crystal structures of amylose and its derivatives. Am. Chem. Soc. Symp. Ser. 141,459-482.

Sarko, A,, Southwick, J., and Hayashi, J. 1976. Packing analysis of carbohydrates and polysaccharides. 7. Crystal structure of cellulose 111, and its relationship to other cellulose polymorphs. Macromolecules 9, 857-863.

Scott, W. E., and Arnott, S. 1972. Accurate x-ray diffraction analysis of fibrous polysaccharides containing pyranose rings. Part I. The linked-atom approach. J. Chem. SOC. Perkin Trans. 11, 324-335.

Selby, H. H., and Whistler, R. L. 1993. Agar. I n “Industrial Gums: Polysaccharides and Their Derivatives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 87-103. Academic Press, New York.

Shafer, S. E., and Stevens, E. S. 1995. A re-examination of the double-helix model for agarose gels using optical rotation. Biopolymers 36, 103-108.

Shatwell, K. P., Sutherland, I. W.. Ross-Murphy, S. B., and Dea, I. C. M. 1991a. Influence of the acetyl substituent on the interaction of xanthan with plant polysaccharides-I. Xanthan-locust bean systems. Carbohydr. Polym. 14,29-51.

Shatwell, K. P., Sutherland, 1. W., Ross-Murphy, S. B.. and Dea, 1. C. M. 1991b. Influence of the acetyl substituent on the interaction of xanthan with plant polysaccharides-11. Xanthan-guar gum systems. Carbohydr. Polym. 14,115-130.

Shatwell, K. P., Sutherland, I. W., Ross-Murphy, S. B., and Dea, I. C. M. 1991~. Influence of the acetyl substituent on the interaction of xanthan with plant polysaccharides-111. Xanthan-konjac mannan systems. Carbohydr. Polym. 1% 131-147.

Smith, P. J. C., and Arnott, S. 1978. LALS: A linked-atom least-squares reciprocal-space refinement system incorporating stereochemical restraints to supplement sparse diffraction data. Actu Crystalfogr. A 34, 3-11.

645-656.


Stanley, N. F. 199.5. Agars. In “Food Polysaccharides and Their Applications” (A. M. Stephen, ed.), pp. 187-204. Marcel Dekker, New York.

Stephen, A. M. (ed.) 1995. “Food Polysaccharides and Their Applications.” Marcel Dekker, New York.

Stipanovic, A. J., and Sarko, A. 1976. Packing analysis of carbohydrates and polysaccharides. 6. Molecular and crystal structure of regenerated cellulose 11. Macromolecules 9,851-857.

Stout, G. H., and Jensen, L. H. 1989. “X-Ray Structure Determination: A Practical Guide.” Wiley Interscience, New York.

Therkelsen, G. H. 1993. Carrageenan. In “Industrial Gums: Polysaccharides and Their Deriva- tives” (R. L. Whistler and J. N. BeMiller, eds.), pp. 145-180. Academic Press, New York.

Vhrum, K. M., Smidsr~id, 0.. and Brant. D. A. 1992. Light scattering reveals micelle-like aggregation in the (1+4), (1-+3)-P-~-glucans from oat aleurone. Food Hydrocolloids 5,497-511.

Walkinshaw, M. D., and Arnott, S. 1981a. Conformations and interactions of pectins. I. X- ray diffraction analysis of sodium pectate in neutral and acidified forms. J. Mol. Biol. 153, 1055-1073.

Walkinshaw, M. D., and Arnott, S. 1981 b. Conformations and interactions of pectins. 11. Models for junction zones in pectinic acid and calcium pectate gels. J. Mol. Biol. 153, 1075-1085.

Whistler, R. L., and BeMiller, J. N. (eds.) 1993. “Industrial Gums: Polysaccharides and Their Derivatives.” Academic Press, New York.

Whistler, R. L., and BeMiller, J. N. (eds.) 1997. “Carbohydrate Chemistry for Food Scientists.” Eagan, Minnesota.

Wild, D. L., and Blanshard, J. M. V. 1986. The relationship of the crystal structure of amylose polymorphs to the structure of the starch granule. Carbohydr. Polym. 6, 121-143.

Winter, W. T., and Sarko, A. 1974a. Crystal and molecular structure of V-anhydrous amylose. Biopolymers 13, 1447-1460.

Winter, W. T., and Sarko, A. 1974b. Crystal and molecular structure of the amylose-DMSO complex. Biopolymers 13, 1461-1482.

Wu, H. C. H., and Sarko, A. 1977. The crystal structure of A-starch: Is it double helical? Carbohydr. Res. 54, C3-C6.

Wu, H. C. H., and Sarko, A. 1978a. The double-helical molecular structure of crystalline A- amylose. Carbohydr. Res. 61, 27-40.

Wu, H. C. H., and Sarko, A. 1978b. The double-helical molecular structure of crystalline B- amylose. Carbohydr. Res. 61, 7-25.

Zobel, H. F., French, A. D., and Hinkle, M. E. 1967. X-Ray diffraction study of oriented amylose fibers. 11. Structure of V amyloses. Biopolymers 5, 837-845.

Zugenmaier, P., and Sarko, A. 1976. Packing analysis of carbohydrates and polysaccharides. IV. A new method for detailed crystal structure refinement of polysaccharides and its application to V-amylose. Biopo1ymer.s 15, 212 1-2136.

Zugenmaier, P., and Sarko, A. 1980. The variable virtual bond: Modeling technique for solving polymer crystal structures. Am. Chem. Soc. Symp. Ser. 141, 225-237.

Zugenmaier, P., Kuppel, A., and Husemann, E. 1977. Detailed refinement of trimethylamylose. Am. Chem. Soc. Symp. Ser. 48, 115-132.

ADVANCES IN FOOD AND NUTRITION RESEARCH, VOL. 42

CELLULAR SIGNAL TRANSDUCTION OF SWEETENER-INDUCED TASTE

MICHAEL NAIM AND BENJAMIN J. STRIEM

Institute of Biochemistry, Food Science and Nutrition Faculty of Agricultural, Food and Environmental Quality Sciences

The Hebrew University of Jerusalem Rehovot 76-100, Israel

MICHAEL TAL

Sigma Israel Chemicals Ltd. Park Rabin

Rehovat 76-100, Israel

I. Introduction 11. Recognition Stage at the Taste-Receptor Cell

A. Chemical Aspects of the Sweet Molecule B. Possible Receptors at the Plasma Membrane

111. Components of the Downstream Transduction Pathway A. Involvement of the Adenylyl Cyclase Cascade in Sugar-Taste

Transduction B. Role of the Phosphoinositide Transduction Pathway in Taste Induced by

Saccharin and SC45647 Sweeteners IV. Involvement of Gustducin/Transducin in Sweet-Taste Transduction V. Amiloride-Sensitive Sweet-Taste Transduction

VI. The Hypothesis of Receptor-Independent Activation of Sweet Taste by Amphipathic Nonsugar Sweeteners

References VII. Summary and Research Needs

I. INTRODUCTION

In humans, the response to sweet taste is believed to be innate. Studies examining facial expressions of neonates following taste stimulation indi-

21 1

1043-4526/98 $2.5 00 Copyright 0 1998 by Academic Press.

All rights of reproduction in any forin reserved.

212 M. NAIM, B. J. STRIEM, AND M. TAL

cated appealing responses to sweet taste and aversion to bitter taste (Steiner, 1973). Newborns prefer sugar solutions to water and their preference increases with sugar concentration (Desor et al., 1973). Furthermore, their preference for different sugars corresponds to adult perception (Maller and Desor, 1974). The high threshold for the appealing taste of sugar (above 10 mM) and actual preference level (0.3-0.5 M ) (Cagan and Maller, 1974) lead mammals to select high-caloric foods and consequently have had a significant effect on survival during phylogenetic development. Moreover, the preference for sugars may change with physiological state. Human subjects prefer a 5% sucrose solution to a 30% sucrose solution, but can be induced to prefer the more concentrated solution once blood glucose level is decreased to 50 mg% following insulin injection (Mayer-Gross and Walker, 1946). Similar results were found in rats (Jacobs, 1958). A strong preference for sweet high-fat foods and fat is associated with opioids (Drew- nowski, 1991; Drewnowski et al., 1992; Marks-Kaufman and Kanarek, 1981; Marks-Kaufman and Lipeles, 1982). For example, administration of naloxone, an opioid antagonist, reduced the hedonic preference for sugar in high-fat food (Drewnowski et al., 1995). In fact, oral stimulation by sweet- taste substances may, via a cephalic phase, release insulin, which in turn may participate in stimulating eating (Brand et al., 1982). Individual human subjects whose eating was most responsive to cues associated with food showed the largest insulin release in response to the sight and smell of steaks being grilled (Rodin, 1978), and insulin release was higher after food presentation in obese than in nonobese subjects (Sjostrom et al., 1980). The selection of sweet tasting and other foods is therefore complex: it may be governed by innate and learned responses, and it is often controlled by the physiological state of the organism.

High sugar intake, which has increased significantly in the present century, has been linked to metabolic disorders such as hypertension, diabetes, and obesity as well as to dental caries, suggesting that sugar intake should be limited (e.g., Grand, 1974). As a result, chemical studies were initiated to explore alternative sweeteners and research, especially in the past 10 years, has led to the availability of very potent ones (e.g., Tinti and Nofre, 1996). In fact, in the West low-calorie soft-drink consumption has increased significantly in recent years, creating a multibillion dollar economic target. However, the sweet taste of sugars, especially that of sucrose, is regarded as pure, whereas many nonsugar sweeteners possess inferior sweet quality. Almost all of the latter have undesirable sensory properties such as slow taste onset and lingering (sweet persistence) aftertaste. (Birch et al., 1980; DuBois et al., 1981; Larson-Powers and Pangborn, 1978; Naim et al., 1986; Schiffman et al., 1979). To date, the molecular basis for these phenomena is unknown. One may hypothesize that modifications in sweet intensity-time

SWEETENER-INDUCED TASTE TRANSDUCTION 213

relationships involve events at the taste-cell level, either at the receptor or along the signal transduction chain. Therefore, studies initiated in recent years on taste transduction (see below) may lead to the identification of factors responsible for temporal sweet taste.

Many mammals prefer the sweet taste of sugars. With other sweeteners, however, variability among species is evident (Fisher et al., 1965). Aspar- tame, monellin, thaumatin, and neohesperidin dihydrochalcone (NHD) are potent taste stimuli for humans and old-world monkeys (Table I), whereas they elicit little or no taste responses in new-world monkeys, guinea pigs, or rats (Brouwer et al., 1973; Glaser et al., 1978, 1992, 1996; Hellekant et al., 1976; Naim et al., 1982). This implies a phylogenetic relationship between the responses to sweeteners of old-world monkeys and humans and suggests multiple mechanisms for sweet-taste transduction in different mammalian species.

Taste-papillae-containing taste buds located on the dorsal surface of mammalian tongues and additional taste buds located in other parts of the oral cavity (e.g., the palate) are the organelles responsible for taste chemoreception (Kinnamon, 1987; Miller and Spangler, 1982). Taste buds contain about 50-150 cells, some of which are sensory cells (Fig. 1). A taste pore is located at the apical end of each taste bud and it has long been suggested that the initial stage in taste chemoreception is the interaction between a taste stimulus and microvilli located at the apical end of the taste-receptor cells. Axons of specific sensory neurons enter the buds through special openings at the bottom, where they form a synapse with the taste-receptor cells. Following the chemical interaction between a taste stimulus and the taste-receptor cell, the cell responds with membrane depolarization, which leads to the release of a putative neurotransmitter at the synapse. The taste nerves carry the signals to neurons of the solitary tract nucleus. Subsequently, these signals are carried via nerve projections to

TABLE I RESPONSES OF SOME MAMMALIAN SPECIES TO CERTAIN SWEETENERS

Monellin Aspartame NHD Saccharin SC45647 Sucrose

Humans ++ ++ ++ ++ ++ ++

+- +- ?? ++ ++ ++ ++

Old-world monkeys + + ++ ++ ++ ++ ++ New-world monkeys - - _ _

_ _ - _ - _ Rats

Nore. f, Positive; -, negative; + -, partial behavioral and/or electrophysiological response. Sources: Brouwer et al. (1973); Glaser et al. (1978, 1992); Hellekant and Danilova (1996); Hellekant ef al. (1976); Naim et al. (1982).


FIG. 1. Diagrammatic representation of mouse taste bud. B, basal cells; D, dark cells; I , intermediate sensory cells; L, light sensory cells. Synaptic contacts (s) are present between dark, intermediate, and light cells and nerve fibers (N). Reprinted from Kinnamon (1987). In T. E. Finger, ed., “Neurobiology of Taste and Smell,” pp. 227-297, by kind permission of John Wiley & Sons, Inc.

specific nuclei and cortical areas of the central nervous system where they are interpreted as a specific taste sensation. The objective of this review is to discuss the complex cellular events related to mechanisms of sweet-taste transduction, focusing on the biochemical processes leading to the release of chemical signals known as second messengers. Additional taste qualities and taste cellular electrophysiology have been reviewed (Brand and Feigin, 1996; Kinnamon and Margolskee, 1996; Lindemann, 1996a,b).

II. RECOGNITION STAGE AT THE TASTE-RECEPTOR CELL

An understanding of the mechanisms via which sweeteners are recognized by taste-receptor cells requires a close examination of both the functional groups in the sweetener molecule that form the pharmacophore (chemical aspects) needed to elicit the sweet taste and the receptor molecule(s) in the sweet-taste-receptor cell (biochemical and physiological aspects) that may interact with a given pharmacophore. Surprisingly, very little interaction has occurred, to date, between scientists in these two research areas.


A. CHEMICAL ASPECTS OF THE SWEET MOLECULE

Many chemical studies have been conducted on the chemical structure- sweet taste relationships of various sweeteners (see DuBois, 1997, for a recent review). The main objective of these studies was to explore a pharmacophore(s) that stimulates some common sweet-taste receptor(s). Sweet taste substances include a large collection of diverse synthetic and naturally occurring organic compounds such as sugars, sulfamates, oximes, amino acids, peptides, proteins, guanidines, and terpenoids (Fig. 2) as well as inorganic salts, especially when used at low concentrations. Interestingly, the sweet potency of these compounds is also extremely diverse.

Whether a single or multiple receptor mechanism is responsible for eliciting sweet-taste sensation by these structurally diverse compounds is a pecu- liar and unsolved question. Most of the structure-function relationship studies have proposed a single sweet-taste receptor for all sweeteners. The ShallenbergedAcree AH-B model (Shallenberger and Acree, 1967) suggested that a sweet molecule must contain a hydrogen-donor group (AH) and a hydrogen-acceptor group (B) and that these two groups interact via hydrogen bonding with the corresponding groups on the sweet-taste receptor. The fact that some D-amino acids are sweet and L-amino acids are bitter led Kier (1972) to assume that a third binding site, characterized by a dispersion binding group (X), is needed, hence the AH-B-X triangle model of sweetness with 2.5 A between A H and B binding groups, 5.5 A between B and X, and 3.5 A between AH and X groups. Belitz and colleagues (Rohse and Belitz, 1991) proposed that e-n (electrophile- nucleophile) groups should replace the AH-B designation since some sweet-tasting compounds do not exhibit H-bond-donating or H-bond- accepting groups. Based on the AH-B-X model, van der Heijden and colleagues (1985a,b) proposed several different receptor sites for sweetness due to some variation in the distances between AH, B, and X groups of different sweeteners. Tinti and Nofre, often regarded as the “sweet people” due to the extremely potent sweeteners they have synthesized, proposed the multipoint attachment (MPA) model of sweetness (Nofre and Tinti, 1996; Tinti and Nofre, 1996) (Fig. 3). This theory postulates the existence of eight optional cooperative recognition sites in the sweet receptor that are able to interact with sweeteners via ionic and hydrogen bonds, as well as via hydrophobic interactions. These occur as a set of three receptor recognition sites (B, AH, and XH) in which ionic and hydrogen bonds are involved and a second set of four sites (Gl, G2, G3, and G4), involved in the steric fit with the sweet molecules. The eighth D receptor recognition site is a hydrogen-donor group. A sweetener need not bear all eight binding sites to elicit the sweet taste. High sweet-taste potency may result, however,


SUCROSE

HCCC

Y o H-N

w-3 ASPARTAME

SACCHARIN CYCLAMATE

&i&I

NEOHESPERIDIN DIHYDROCHALCONE

H O W

H @ OH

GLYCYRRHlZlN

FIG. 2. The diverse chemical

GUANlDlNEACmC ACID (SC4647)

structures of some sweeteners.


W

FIG. 3. Spatial arrangement of the eight sites of interaction according to the multipoint attachment theory. Reprinted from Nofre and Tinti (1996) Food Chern. 56,263-274, by kind permission of The Lancet Ltd.

from the cumulative presence of these binding sites. Recently, Birch and colleagues (1993, 1996) introduced the importance of water and molar volumes in sweet-taste perception. Molecular volumes have been shown to be important in the efficacy of drug actions (McGowan and Mellors, 1986) and in sulfamate sweeteners (Spillane and McGlinchey, 1981). Using experimental biophysical parameters and computer modeling, Birch and colleagues (1993) have proposed that the interaction of a sweet stimulus with the sweet receptor requires specific volumes of the stimulus in water and biophase. Apparent specific volume defines taste quality and specific volumes define hydrostatic packing of the sweet molecules among the water molecules, whereas the related intrinsic viscosities define the hydrodynamic behavior of the sweet molecules.

Among the aformentioned models, the Nofre-Tinti model appears to have made the most significant contribution, leading to the discovery of highly potent sweeteners such as superaspartame (10,000 X sucrose) (Nofre and Tinti, 1987), guanidineacetic acid sweeteners (200,000 X sucrose), N - acylglutamylanilide sweeteners (9000 X sucrose), and N-alkyldipeptide sweeteners (10,000 X sucrose) (Nofre and Tinti, 1994).

B. POSSIBLE RECEPTORS A T THE PLASMA MEMBRANE

An initial study by Dastoli and Price (1966) was the first to claim the isolation of a sweet-sensitive protein from bovine tongue epithelium. Addi- tional studies (e.g., Hiji and Sato, 1973; Shimazaki et al., 1986) proposed the existence of such proteins in rats and monkeys. The binding of sugars and nonsugar sweeteners to plasma membranes prepared from bovine and human taste papillae was also demonstrated (Cagan, 1971; Cagan and

218 M. NAIM, B. J . STRIEM, A N D M. TAL

Morris, 1979). Moreover, the application of proteolytic enzymes such as pronase E to receptor membranes blocked the sweet response (Hiji, 1975). Hence, in analogy to hormones and neurotransmitters, it has been suggested that at least some sweeteners interact with specific receptors located on the plasma membranes, and that intake of stimulus into the cell is not necessary for the recognition stage. However, the low affinity (KD - lo-’-

M) of sugars to taste receptors (Cagan, 1971) made binding experiments very difficult to conduct. Although the affinity of some nonsugar sweeteners is much higher (KD - lo-’ M ) (Cagan and Morris, 1979), complications arose in the preparation of sufficient amounts of membrane for these experiments and in the “noise” often produced by the nonspecific binding of stimuli to sites other than taste receptors. Studies in recent years (see discussion below) have suggested the involvement of heterotrimeric guanine nucleotide-binding proteins (GTP-binding proteins, G-proteins) in a putative signal transduction chain in sweet-responsive cells corresponding to: receptor -+ G-protein + effector enzyme + intracellular signal molecule += ionic channels + cell depolarization .+ neurotransmitter release. As a result, molecular biology approaches have been used in an attempt to clone the putative seven transmembrane taste receptors. Some polymerase chain reaction (PCR) experiments on rat lingual epithelia and bovine taste tissue (Abe et al., 1993a,b; Matsuoka et al., 1993) identified such receptors that either exhibited significant homology with putative olfactory G-protein-coupled receptors (Buck and Axel, 1991) or were also expressed in nonsensory epithelium. An additional seven transmembrane receptor related to neuropeptides was found to be expressed in gustatory tissue (Tal et al., 1995). A collection of circumvallate (CV) papillae-specific PCR-derived cDNAs, encoding a putative G-protein-coupled receptor, were identified by distinguishing PCR products amplified from the cDNA of nonsensory epithelial sheets lacking taste buds. A novel G-protein- coupled receptor gene, expressed in the CV taste buds but not in the nonsensory sheets, which may or may not encode a taste receptor, was cloned. Ligands to the above receptors have not yet been identified and, to date, no sweet-taste receptor has been identified or cloned. Thus, the putative sweet-taste receptors may differ in sequence from known G- protein-coupled receptors, or may not be coupled to G-proteins at all; if so, a new approach needs to be taken.

As already indicated, chemical studies in general have concluded that a single sweet-taste receptor mediates sweet sensation for all sweeteners. However, cross-adaptation studies in humans (Lawless and Stevens, 1983; Schiffman and Cahn, 1981), electrophysiological recordings in monkeys and rodents (Faurion et al., 1980; Hellekant, 1975; Jakinovich, 1982), and biochemical sweet-taste transduction studies (see below) leave little doubt


that either multiple receptors or multiple transduction mechanisms exist for sweet taste.

Ill. COMPONENTS OF THE DOWNSTREAM TRANSDUCTION PATHWAY

The hypothesis that sweet-taste receptors occur on the surface of the plasma membrane of taste cells is generally accepted. However, due to the aforementioned difficulties in identifying and isolating such taste receptors, available information is limited. Neurotransmitters and hormones, as well as some sensory signals of vision, olfaction, and taste, bind to specific membrane receptors, then initiate biochemical events through membrane components leading to the formation of intracellular signals (Berridge, 1984; Gilman, 1984; Pace et al., 1985; Striem et al., 1989; Stryer, 1986). In many cases, G-protein is a coupling component between receptors and effector enzymes (Simon et al., 1991). G-proteins transmit either stimulatory (e.g., via G,) or inhibitory (e.g., via Gi) signals from receptors to adenylyl cyclase, cGMP phosphodiesterase, phospholipases, and several types of ion channels (Schultz et al., 1990; Yatani ef al., 1987). A significant feature of such mechanisms is that the G-proteins temporally uncouple the detection of external signals at the receptor from the activation of the effector. This allows the system to amplify the signal manyfold. In vision, a single photoexcited rhodopsin molecule can activate hundreds of phosphodiesterase (PDE) molecules, due to facilitated binding of GTP to a large number of transducin molecules (Stryer, 1986). The cellular amplification of transduction signals following binding at the receptor makes the measurement of the transduction response more sensitive, and perhaps more relevant than binding experiments. Furthermore, components along the signal transduction chain may, in fact, be the site at which taste sensation is primarily modulated. Taste adaptation may be a type of receptor desensitization, whereas sweet intensity and persistence may be dependent on the modulation of transduction systems.

A. INVOLVEMENT OF THE ADENYLYL CYCLASE CASCADE IN SUGAR-TASTE TRANSDUCTION

The idea that cyclic adenosine monophosphate (CAMP) is involved in chemosensory transduction as a second messenger was advanced following results showing its involvement in the transduction of hormones (Robison et al., 1971). Available knowledge at that time indicated that adenylyl cyclase catalyzes the conversion of ATP to CAMP, that a specific PDE

220 M. NAIM, B. J . STRIEM, AND M. TAL

breaks cAMP down to 5’-AMP, and that both membrane enzymes alter the intracellular level of CAMP. High adenylyl cyclase (Kurihara and Koyama, 1972) and PDE (Kurihara, 1972) activity was found in the gustatory epithelium. The presence of these enzymes in the microvilli of cells in rabbit taste buds was also shown (Asanuma and Nomura, 1982). Furthermore, cAMP content could increase in intact bovine taste papillae in response to sucrose stimulation (Cagan, 1974). However, due to the very limited information available at the time on the significant role of G-proteins in the adenylyl transduction pathway, these studies provided no direct evidence for the involvement of cAMP as a second messenger in sweet-taste transduction. The tools needed to test the involvement of the adenylyl cyclase cascade directly became available when the signal-chain components were identified, leading to controlled biochemical experiments (e.g., adding ATP, guanine nucleotides, and other effectors) using plasma membrane preparations (Striem et al., 1989). In addition, the patch-clamp technique became available to record electrophysiological responses in single taste cells following the intracellular administration of cAMP (e.g., Avenet et al., 1988).

A few basic criteria needed to be met prior to drawing a conclusion regarding the involvement of the adenylyl cyclase cascade (e.g., Suther- land, 1972) in sweet-taste transduction. The involvement of G-protein in the mediation of cellular cAMP formation had to be shown. The response also had to be tissue and stimulus specific. In fact, due to the species specificity of the sweet responses in mammals, biochemical studies were needed to follow behavioral and electrophysiological responses where taste responses by a given animal to specific sweeteners had been proven. Time-course experiments needed to be conducted to show that the increase in cAMP or other messengers occurs in real time. Indeed, various sugars stimulated the activity of adenylyl cyclase in crude taste- membrane preparations from the apical anterior of rat tongues (Figs. 4C and 4E) and the formation of cAMP in response to sugar stimulation was dependent on the presence of GTP (Fig. 4E), suggesting that the formation of cAMP is mediated by G-proteins (Striem et al., 1989). Because the reaction mixture contained the PDE inhibitor 3-isobutyl-l- methylxanthine (IBMX), the possible involvement of the G, type of G- protein, which is present in taste cells (McLaughlin et al., 1993), was proposed. Similar results were found with membranes prepared from pig CV papillae (Naim et al., 1991), and in neither rats nor pigs did sucrose stimulate adenylyl cyclase activity in membranes prepared from nonsensory lingual epithelium. In a similar taste-membrane preparation from rats, sucrose stimulation of adenylyl cyclase activity was inhibited by Cu2+ and Zn2+ but not by Cd2+, Co2+, or Ni2+ (Striem et aL, 1989), in agreement with the effect of these ions on the neural taste responses


to sugar stimulation in rats and mice (Kasahara et al., 1987; Yamamoto and Kawamura, 1971). Furthermore, methyl-4,6-dichloro-4,6-dideoxy-a- D-galactopyranoside (MAD), an inhibitor (probably competitive) of the sucrose stimulation of the chorda tympani nerve in gerbils and rats (Blochaviak and Jakinovich, 1985; Jakinovich, 1983), and a suppressor of the perception of several sweeteners’ sweet intensity in humans (Schiffman et af., 1987), also inhibited sucrose stimulation of adenylyl cyclase activity in taste membranes (Fig. 4C), in complete agreement with the chorda tympani electrophysiological responses (Striem et al., 1990b). As the concentration of MAD increased, stimulation of adenylyl cyclase by sucrose decreased, with the effect of sucrose being completely abolished in the presence of 60 m M MAD. The inhibitor itself had no effect on adenylyl cyclase activity.

The above results were in line with the required criteria that sugar stimulation of adenylyl cyclase activity in lingual membranes be GTP dependent, tissue specific, and inhibited by sweet-taste inhibitors. However, the taste membranes prepared from the tip of the tongue of rats, although containing >90% of the tongue’s fungiform taste papillae (Miller, 1976; Miller and Spangler, 1982), or membranes prepared from the whole CV papillae of pigs, (Naim et af., 1991) containing about 6000 taste buds (Tuck- erman, 1888), were contaminated with membranes derived from epithelial nonsensory, connective, and perhaps even muscle membranes.

An additional study in rats was aimed at obtaining much cleaner sensory tissue and intact taste cells, in which changes in the intracellular level of cAMP in response to sweetener stimulation could be evaluated (Striem et al., 1991). A CV papilla containing a large number of taste buds, located within an epithelial layer covering connective and muscle tissue, was selected. Following subepithelial collagenase treatment, the taste-bud sheet containing condensed taste buds was separated from the connective and muscle tissues. A similar size sheet, separated from the surrounding epithelium of the same papilla, but lacking taste buds, was used as the control nonsensory epithelial tissue. Exposure of CV taste-bud sheets for 6 min to sucrose resulted in a two to threefold increase in cAMP accumulation (Fig. 4F). There was no such response in the nonsensory epithelial tissue derived from the same taste papilla of the same animals. The accumulation of cAMP in these taste buds in response to sucrose was dose dependent and inhibition (65%) of sucrose-dependent cAMP formation was observed in the intact tissue after application of 50 mM MAD. The use of IBMX in these biochemical experiments enables the measurement of cAMP formation on a scale of minutes, evidently beyond the expected below-1-sec real time for taste responses (Bernhardt et al., 1996; Spielman et al., 1996). Thus, the temporal dynamics of cAMP signals, as shown in the neural circuit (Hempel


- c 180

c

2

160 0

W 140

Y 120

5 100

5 2 80

-I

100 - c m 5 80 a

: 60

3 20

- 5 40

a

U 0

FIG. 4.

cAMP -ATP

NO MA0

002M MAD

OOBM MAD

00 025 0 5 0.725 1 0

SUCROSE (M)

RAT

E T

SUC GTP SUC

GTP

PIG

SUC GTP SUC

GTP

Non-sugar B e

SUC SACC SC45647

F

4 BAS SUC

__

SUC SUC

MA0

Biochemical data supporting the model (A) for adenylyl cyclase-induced cAMP formation in sugar-taste transduction. Sucrose (SUC) stimulates adenylyl cyclase activity in taste-membrane preparations of rats and pigs (C,E). The sucrose-stimulated adenylyl cyclase activity is GTP-dependent (E) and is inhibited by the sweet-taste inhibitor MAD (C). In taste-bud sheets, cellular cAMP is increased above basal (BAS) levels (D,F); it is inhibited by MAD (F). cAMP through the activation of PKA induces phosphorylation of the K' channel, leading to decreased conductance. This transduction pathway appears not to be


etal., 1996), has to be determined in taste cells. Indeed, preliminary observations indicated a sucrose-stimulated cAMP formation in CV taste-bud sheets within the 500 msec time course (Naim et al., unpublished). Interest- ingly, stimulation by SC45647, a nonsugar sweetener (Nofre and Tinti, 1987), in humans and rats (Hellekant and Walters, 1993), like saccharin, did not stimulate intracellular formation of cAMP in rat CV taste-bud sheets (Naim el al., 1996) (Fig. 4D).

Important support for the suggestion that cAMP may be a second messenger in sweet-taste transduction came from cellular electrophysiological studies. Lindemann and co-workers (Avenet et al., 1988), using whole-cell recordings from isolated taste cells of frogs (inside-out membrane patches) indicated that cAMP causes depolarization via the action of CAMP- dependent protein kinase A (PKA), which inactivates potassium channels. At the same time, Tonosaki and Funakoshi (1988) found that cGMP, and to a lesser extent CAMP, injected into mouse taste cells decreases potassium conductance, leading to depolarization, and sucrose stimulation depolarized the same cells via potassium conductance. More recently, Kinnamon and co-workers (Cummings ef al., 1993) showed in hamsters that fungiform taste buds, which responded to stimulation by sucrose and some nonsugar sweeteners by inducing action potentials, were those which responded in the same way to stimulation by membrane-permeant analogs of cAMP and cGMP. Similar results were also observed with hamster-isolated taste cells (Cummings et al., 1996). Based on these results, a model for sugar-taste transduction has been hypothesized (Fig. 4A). The transduction pathway involves sugar stimulation of specific receptors located on the taste cell’s apical membrane. The adenylyl cyclase (AC) cascade is then stimulated via the mediation of G,-type G-proteins, leading to the production of CAMP, which, via PKA, induces phosphorylation of potassium channels located in the basolateral membranes. The reduced potassium conductance causes membrane depolarization, eventually leading to neural transmission.

affected by saccharin (SACC) or SC45647 sweeteners (B, D). Abbreviations: R, putative receptors; G. G-protein with the cY,py-subunits; AC. adenylyl cyclase; PKA, protein kinase. (C) Adapted from Striem et a/. (1990b) Chem. Senses 15, 529-536, by kind permission of Oxford University Press: (D) reproduced from Naim et n/. (1996). In W. Pickenhagen, C.-T. Ho, and A. M. Spanier, eds. “The Contribution of Low- and Nonvolatile Materials to the Flavor of Foods.” pp. 65-75, by kind permission of Allured Publishing Corporation; (E) adapted from Striem ef ul. (1989) Biochern. J. 260, 121-126, by kind permission of The Biochemical Society and Portland Press; (E) adapted from Naim el al. (1991) Comp. Biochern. Physiof. B 100,455-458, by kind permission of Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington OX5 lGB, UK; (F) adapted from Striem eta/. (1991) Cell. Physiol. Biochern. 1, 46-54, by kind permission of S. Karger AG. Basel.


B. ROLE OF THE PHOSPHOINOSITIDE TRANSDUCTION PATHWAY IN TASTE INDUCED BY SACCHARIN AND

SC45647 SWEETENERS

The above-described results suggested a role for cAMP in sugar-taste transduction in rats and pigs, and electrophysiological experiments (Cum- mings et al., 1996) suggested that some artificial sweeteners elicit action potentials in hamster taste cells and that these responses are mimicked by membrane-permanent analogs of cAMP and cGMP. However, as already indicated, neither SC45647 nor saccharin sweeteners stimulated cAMP formation in intact taste-bud sheets prepared from the CV papilla of rats (Fig. 4D). Such observations have led to the conclusion that an additional signal transduction pathway may be operative for sweet taste. Since psychophysi- cal and chemical studies have suggested that sweet and bitter tastes are interrelated (Bartoshuk, 1975; Birch and Mylvaganam, 1976), the possible stimulation of the phosphoinositide pathway was tested next (Bernhardt et al., 1996).

1. Phospholipase C Activity

The phosphoinositide pathway is known to involve receptor stimulation of phospholipase C (PLC) via G,- or Gi-type G-proteins to stimulate the formation of inositol trisphosphate ( IP3) and diacylglycerol (DAG) as second messengers (Berridge, 1984; Simon et al., 1991). Its involvement in the transduction of bitter taste in rats and mice has also been proposed (Hwang et al., 1990; Spielman et al., 1994,1996). IP3 may, in turn, release Ca2+ from intracellular Ca2+ stores via IP3-sensitive receptors (Hwang et al., 1990), as found following stimulation by the bitter tastant, denatonium (Akabas et al., 1988). In contrast to the case of CAMP, an efficient inhibitor of intracellular IP3 degradation is not available. In view of the fact that the release of second messengers in response to olfactory and taste stimulations observed in membrane experiments is very rapid (Breer et al., 1990; Spielman et al., 1996), a fast-reaction timer instrument was used, thus enabling the monitoring of IP3 level in intact taste cells in close to real time, less than 1 sec (Bernhardt et al., 1996).

Incubation of intact CV taste-bud sheets with SC45647 and saccharin for 0.5 sec significantly stimulated the formation of intracellular IP3 (Fig. 5C) whereas no significant IP3 response was observed in the nonsensory sheets. Moreover, the response was, as expected, time dependent, with maximal IP3 formation within 0.5 sec under the experimental conditions (Bernhardt et al., 1996). The effects of both SC45647 and saccharin on IP3 formation were concentration dependent, validating the assumption of a physiological


phenomenon. In contrast to SC45647 and saccharin, sucrose, which is known to stimulate CAMP, produced only very small amounts of IP3 (Fig. 5C). The phenomenon whereby sugars use CAMP as an intracellular messenger whereas nonsugar sweeteners such as saccharin and SC45647 use IP3, the latter having been proposed to mediate bitter-taste signaling, has raised a significant question related to taste specificity. The following Ca-imaging experiments conducted by Lindemann and co-workers (Bernhardt et al., 1996) were helpful in assessing this question.

2. Imaging of Intracellular Ca2+ Following Stimulation by Sweeteners

Since exposure of CV taste buds to the nonsugar sweeteners SC45647 and saccharin stimulated intracellular IP3 formation, one would suspect that IP3, in turn, will release Ca2+ from intracellular stores by acting on IP3-sensitive receptors (Irvine et al., 1986). The permeant fluorescent indicator fura 2 acetoxymethyl ester (Tsien, 1983) was used to monitor changes in cellular Ca2+ content in CV taste buds isolated from the CV taste-bud sheets of rats by a brief low-calcium treatment (Bernhardt et al., 1996). Under a fluorescence microscope, single taste buds were subjected to digital fluorescence ratio imaging following superfusion with sweeteners. The imaging technique provided significant information because the Ca2+ level could be monitored in living taste cells in response to stimulation by a few taste stimuli applied in sequence, each stimulus followed by a Tyrode wash.

Stimulation of isolated taste buds by SC45647 increased the cellular content of Ca2+ (Fig. 5D). This release of Ca2f must be from intracellular sources because the same response was observed when Ca2+ ions were removed from the extracellular medium by ethylene glycol bis(P- aminoethyl ether) N,N'-tetraacetic acid (EGTA; Fig. 5D, hatched bars). The same taste cells that responded to SC45647 also responded to stimulation by saccharin with Ca2+ release, again independent of the presence of Ca2' in the extracellular medium. Thus these nonsugar sweeteners, which stimulated the formation of IP3 in CV taste-bud cells, also stimulated the release of Ca2t from intracellular stores of these cells. Unexpectedly, sucrose, which did not (or only slightly) stimulate IP3 formation in the biochemical experiments, also stimulated an increase in Ca2+ content in the same taste cells that responded to saccharin and SC45647 (Fig. 5D, open bars). However, when Ca2' ions were removed from the extracellular medium by EGTA, the increase in cellular Ca2' content in response to sucrose was no longer seen (Fig. 5D, hatched bars). Hence, the same cells, which could now be classified as sweet-responsive cells, responded to all sweeteners by elevating cellular Ca2+ levels. However, sucrose increased Ca2+ from an external source, whereas the nonsugar sweeteners increased Ca2+ from

CAMP P * T P

A

800

- c ._ d 600

g rn I

400

- ! 5c 200 a

0

P P 6 6

Non-sugar

n@

C

T

SUC SACC SC45647

SUC FORSK SACC SC-45647

FIG. 5. Data supporting the model for sugar-induced CAMP-dependent Ca2' entry from extracellular medium (A) and for nonsugar-sweetener-induced IP3-dependent Ca2+ elevation from intracellular stores (B). Stimulation of isolated CV taste buds by saccharin (SACC) and SC45647 sweeteners, but not by sucrose, increased cellular IP3 content (C) . Stimulation by


intracellular stores. When the membrane-permeant diterpene forskolin, a direct activator of adenylyl cyclase (Bouhelal et al., 1985; Seamon et al., 1981), was applied to the sweet-responsive cells (i.e., those cells that responded to SC45647, saccharin, and sucrose), the cellular Ca2+ level was elevated, and as with sucrose, this elevation depended on the presence of Ca2' in the extracellular medium (Fig. 5D). This result supports the notion that the sucrose-induced entry of Ca2+ is mediated by CAMP. Interestingly, taste cells that responded with elevated Ca2+ following stimulation by the sweeteners sucrose, saccharin, and SC45647 did not respond to stimulation by the bitter tastant denatonium. The opposite was also true: cells that responded with elevated Ca2t to stimulation by denatonium never responded to these sweeteners. Thus, although saccharin, SC45647, and denatonium stimulate the same transduction pathway in rats (i.e., IP3), Ca2+- imaging of individual cells suggests that these events occur in two different subpopulations of taste cells. These results could explain sweet- and bitter- taste specificity and are in line with recent results in chimpanzees indicating that sweet-responding fibers of the chorda tympani nerve did not respond to bitter compounds (Hellekant and Ninomiya, 1994). The slight bitterness that accompanies the sweetness of saccharin applied at high concentrations might therefore be due to the generation of IP3 in bitter-responsive cells, in addition to the IP3 generated in sweet-responsive cells.

Thus in a sweet-responsive cell, in addition to the adenylyl cyclase cascade, which is apparently stimulated by sugars and which induces CAMP- dependent Ca'' entry from the extracellular medium (Fig. 5A), a second pathway for sweet taste may be occurring (Fig. 5B). The proposed second pathway involves nonsugar sweeteners that stimulate the phosphoinositide transduction pathway to form IP3 via G,-like G proteins. IP3 then acts on intracellular Ca2+ stores to release Ca2'.

IV. INVOLVEMENT OF GUSTDUClN/TRANSDUClN IN S W EET-TASTE TRANS DUCT ION

G-proteins such as G,, Gi, G,, GI4, and transducin have been identified in taste buds, and a taste-specific G-protein, gustducin, was cloned by

sucrose (SUC), forskolin (FORSK), saccharin, or SC45647 increased the cellular content of Ca" (D, open bars). However, only saccharin and SC4.5647 increased cellular Ca" in the absence of Ca" in the extracellular medium (D, hatched bars). Abbreviations: R, putative receptors, G , G-protein with the a,&-subunits; AC, adenylyl cyclase; PKA, protein kinase A; PLC. phospholipase C; PIP2, phosphatidylinositol bisphosphate. (C, D) Adapted from Bernhardt et al. (1996) J . Physiol. 490, 325-336, by kind permission of Cambridge Univer- sity Press.


Margolskee and co-workers (McLaughlin ef al., 1992, 1993). Gustducin has recently been found to be expressed in intestinal cells (Hofer et al., 1996). It is closely related to the transducins, and both activate a PDE purified from taste cells (Kolesnikov and Margolskee, 1995; Ruiz-Avila et al., 1995) in a manner similar to the activation of rod transducin in vision. A PDE activation is expected to reduce cellular cyclic nucleotides, as proposed for bitter stimulation (Kurihara, 1972; Price, 1973), rather than increase in cAMP content via the sugar-stimulated adenylyl cyclase activity described earlier. Alternatively, PDE activation may suppress the sugar-stimulated cellular cAMP formation, thereby playing a role in the extinction of the sweet-taste response. There are some data supporting the role of gustducin andlor transducin in both bitter- and sweet-taste transduction. The bitter tastant denatonium, in the presence of bovine CV papillae membranes activated a-transducin but not Gi-protein, suggesting that the putative reduction in cAMP is related to the PDE pathway rather than to the inhibitory pathway of adenylyl cyclase (Ruiz-Avila et al., 1995). In frogs, the nonsugar sweeteners saccharin and NC-01 appear to stimulate the transducin-PDE pathway, suggesting that such a sweet-transduction pathway occurs in a subpopulation of cells different from that containing the adenylyl cyclase pathway; alternatively in frogs these sweeteners stimulate bitter-taste transduction (Kolesnikov and Margolskee, 1995). Most important, Margolskee and co-workers (Wong et al., 1996) have recently found that a-gustducin knockout mice (mutant mice in which the gene encoding the a-subunit of gustducin has been eliminated) exhibit reduced (though not completely eliminated) behavioral and electrophysiological responses to both bitter- and sweet- (see the effect on sweet taste; Fig. 6), but not salty- or sour- taste stimulations. These results suggest that gustducin is involved in both bitter- and sweet-taste transduction. This conclusion appears to contradict the aforementioned data suggesting that the adenylyl cyclase and PLC pathways mediate sweet-taste transduction. However, as was hypothesized by Kinnamon (1996), the a-subunit of gustducin may activate the proposed pathways of bitter-taste transduction (e.g., PDE-reduced CAMP), while the &subunits may activate PLC following nonsugar stimulation. Such a hypothesis is in line with data indicating dual signaling via G-proteins, whereby the a-subunit may involve one pathway, for example, adenylyl cyclase (Zhu et al., 1994), while the &complex stimulates PLC (Gierschik and Camps, 1993; Katz et al., 1992). One problem with the a-gustducin knockout mice is the possibility that By-subunits may be in excess. In the a-gustducin knockout mice, excess &subunits may interfere indirectly with a-subunits of other G-proteins (Kinnamon and Margolskee, 1996; Neer, 1995), leading to impaired bitter- and sweet-taste responses. Gust- ducin, as already mentioned, may also play a role in sweet-taste extinction,


SUCROSE (mM) SC45647 (mM)

FIG. 6. Relative chorda tympani nerve responses to lingual sweet-taste stimulation (A,B) and 48-hr two-bottle (Sweetener vs distilled water) preference tests (C,D) of control (open circles) and a-gustducin knockout (filled circles) mice. Reproduced from Wong et QZ. (1996) Nature 381, 796-800, by kind permission of Macmillan Magazines Ltd.

for example, by reducing the CAMP level after it has been elevated due to sugar stimulation. Such questions will probably be resolved by detailed tirne-course measurements of second messengers released by the proposed pathways.

V. AMILORIDE-SENSITIVE SWEET-TASTE TRANSDUCTION

DeSimone and co-workers (1981) proposed a role for active transepithe- lial ion transport in taste transduction. Such a mechanism was proposed


for sugar sweet-taste transduction in dogs (Mierson et af., 1988; Simon et af., 1986). In response to sugar stimulation, the mucosa of dog tongue has been found to generate a macroscopic inward current through the apical membrane of taste receptor cells. Amiloride (a heterocyclic carboxy guani- dinium compound that is a reversible blocker of epithelial channel-mediated Na' transport) caused partial blockage of the sugar-evoked current and also suppressed the neural response to sugar-taste stimulation (Mierson et al., 1988). Furthermore, the presence of NaCl in the dog-tongue mucosa enhanced the taste response to sucrose (Kumazawa and Kurihara, 1990), although its presence is not mandatory for the sweet response. The application of amiloride reduces sweet-taste intensity in humans as well (Schiffman et al., 1983). Amiloride inhibited the sensory nerve responses to sugars in dogs (Mierson et af., 1988) but not in rats, gerbils, or hamsters (Herness, 1987; Jakinovich, 1985). This sugar-taste transduction pathway is proposed to be occurring in dogs, rabbits (Simon et af., 1986), and perhaps humans, but not in rats, and vice versa: the cAMP pathway proposed for sugar taste transduction in rats (Striem et al., 1989) may not be involved in the canine response to sucrose (Simon et af., 1989). It should be noted, however, that amiloride may interact with G-proteins and affect the inhibitory pathway of adenylyl cyclase ( Anand-Srivastava, 1989) and, therefore, may modify cellular cAMP content.

VI. THE HYPOTHESIS OF RECEPTOR-INDEPENDENT ACTIVATION OF SWEET TASTE BY AMPHIPATHIC

NONSUGAR SWEETENERS

The understanding of signal-transduction mechanisms of taste is further complicated by the hypothesis that some bitter and sweet stimuli may initiate taste via receptor-independent mechanisms (Koyama and Kuri- hara, 1972; Naim et al., 1994; Spielman et al., 1992). It is proposed that some amphipathic (i.e., having both hydrophobic and hydrophilic domains) tastants may activate G-proteins directly, act directly on effector enzymes along the transduction pathway, or even act directly on ionic channels. The idea of a direct activation of G-proteins may be of particular relevance. The heterotrimeric G-proteins act as switches that regulate information-processing circuits connecting cell surface receptors to a variety of enzyme effectors (Simon et al., 1991). These proteins contain three subunits, G, and G,,, which link together. Following a receptoral signal, G-proteins bind GTP via the G, subunit which dissoci- ates from the &subunits. The intrinsic GTPase activity of the G,- subunit terminates the activation and converts the G-protein to an

SWEETENER-INDUCED TASTE TRANSDUCTION 23 1

inactive G,p,-GDP-bound form. As discussed earlier, it is now evident that both the G,- and Gp,-subunits are capable of activating effector enzymes such as adenylyl cyclase and PLC (Clapham and Neer, 1993; Katz et al., 1992). A variety of amphipathic neuropeptides, venom peptides (e.g., substances P, bradykinin, mastoparan), and nonpeptide substances can activate G proteins directly, leading to cellular responses such as histamine and insulin release (Avidor et al., 1993; Higashijima et al., 1990; Mousli et al., 1990). The amphipathic properties of such compounds allow them to penetrate deep into the plasma membrane and activate G-proteins directly, thereby mimicking receptoral stimulation of cellular responses (Mousli et al., 1990). Bitter stimuli can depolarize N-18 mouse neuroblastoma cells, which are unrelated to taste (Kumazawa et al., 1985), and some bitter drugs (e.g., propranolol) and neuropeptides (bradykinin is bitter; Spielman et al., 1992), are direct activators of G- proteins (Hagelueken et al., 1994). In contrast to sugars, nonsugar sweeteners are, like bitter stimuli, chemically diverse, that is, they are aromatic compounds, sulfamates, dipeptides, and guanidines, and they are also amphipathic (Fig. 2). Indeed, some sweet amphipathic tastants such as saccharin, NHD (e.g., Fig. 7), cyclamate, and the bitter tastant quinine have been found to activate the GTPase of either purified transducin or a mixture of purified Gi/Go-proteins (and Gi/Go-proteins reconstituted into phospholipid vesicles) in vitro in a concentration- dependent manner (Naim et al., 1994). The concentrations of the taste substances that activated G,/Go-proteins and transducin correlated closely with those needed to elicit taste. If these tastants could bypass the taste receptors and permeate the plasma membrane, then taste specificity might be achieved at intracellular targets (e.g., G-proteins) and/or be dependent on the permeation of the amphipathic tastant through the plasma membrane, the composition of which may vary among mammals (Maddy, 1966). Changes in lipid composition affected the interaction of some bitter tastants with liposomes (Kumazawa et al., 1988).

The above in vitro experiments are consistent with the slow taste onset and lingering aftertaste that are common among nonsugar sweeteners (Birch et al., 1980; Larson-Powers and Pangborn, 1978) and that may result from a process of stimulus penetration through the plasma membrane. Furthermore, some of these taste compounds elicit taste and taste nerve responses (Bradley, 1973; Fishberg et al., 1933; Hellekant et al., 1987) following intravenous or intralingual administration, independent of stimulation of putative receptors at the apical surface of the tongue. For example, Fishberg and colleagues (1933) used intravenous injection of sodium saccharin to measure the blood circulation time in humans from the time of saccharin injection into the peripheral vein until subjects indicate that they


Gi /Go

( 0 ) SACC 400

= 300

c f2 8 200

c

Y- O

100

c 2 I- 0

Transducin

( 0 )

-1 150

E 0 50 100 150 200

NHD

6

160

140

120

100

250

200

150

100 u 100 0 5 10 15 20

CONCENTRATION (mM)

FIG. 7. In vitro stimulation of GTPase of Gi/G,-proteins (filled circles) and transducin (open circles) by saccharin (SACC) and neohesperidin dihydrochalcone (NHD). Reproduced from Naim ef nl. (1994) Biochem J. 297, 451-454, by kind permission of The Biochemical Society and Portland Press.

taste the compound. Saccharin modifies adenylyl cyclase activity in membranes derived from the muscle and liver, which are unlikely to contain taste receptors (Striem et al., 1990a), and it has been proposed to affect the adenylyl cyclase catalytic unit in fat cells (Dib et al., 1996). Moreover, the nonsugar sweetener acesulfame K acts directly on pancreatic islets, potentiating insulin release (Liang et al., 1987). These observations have generally been interpreted as stimulation of putative receptors in various tissues. However, the lack of tissue specificity for such tastants may also be indicative of receptor-independent mechanisms. Thus, if some of these amphipathic sweeteners are proven to be taste-cell permeants, they may affect taste-transduction pathways independently of receptors located on the plasma membrane.


VII. SUMMARY AND RESEARCH NEEDS

In addition to the innate strong preference of most mammals for sweet stimuli, oral sweet-taste signals of a particular food are coupled with postin- gestive signals of that food to participate in the control of the release of some hormones, such as endorphins and insulin, and in the regulation of diet selection. As such they have become nutritionally important. From a nutritional and food research point of view, an understanding of the molecular mechanisms of sweet taste is even more essential as sweetness has always been associated by humans with an appealing sensation. Because a restriction in the intake of sugar (sucrose), bearing the most appealing sweet taste, has been recommended, the need to discover new, alternative sweeteners with a desirable sensation, close to that of sucrose, is thus imperative. The ability to synthesize very high-potency sweeteners appears to be available in terms of advanced chemical considerations, but very little is known about molecular factors affecting slow taste onset and lingering aftertaste, which are common in nonsugar sweeteners and which reduce sweet-taste quality. Recent progress in the understanding of sweet-taste transduction at the taste-cell level has opened the way to monitoring the signal-transduction output following stimulation by new sweeteners. A collaboration between chemists and taste physiologists has therefore become relevant to identify sites affecting sweet-taste quality.

Despite research efforts in recent years, the initial step in sweet-taste sensation, that is, identification or cloning of the receptor molecule(s) in the taste cell responsible for sweet specificity, has not yet been performed. Studies employing molecular biology techniques are expected to be the main tool to overcome this important obstacle. Do certain amphipathic nonsugar sweeteners bypass the taste receptors in vivo to act on intracellular targets along the downstream transduction pathway? This is a hypothesis that has been raised for bitter taste as well, but requires much more investigation, though in vitro experiments have suggested that some sweeteners are direct activators of G-proteins. The permeation of amphipathic tastants through and/or their interacting with components of the plasma membrane of taste-receptor cells still needs to be shown.

Accumulated data suggest that multiple transduction pathways at the taste-cell level may be activated by sweet-taste stimulation. The understanding of such mechanisms is particularly complicated by the diversity of species specificity in the taste response to nonsugar sweeteners and the different mechanisms that may be operative for sugar and nonsugar sweeteners in different mammals. For example, in rats and apparently in hamsters, measurements of cellular chemical signals (i.e., second messengers) and cellular electrophysiological studies suggest the involvement of cyclic nucle-


otides such as cAMP (via the adenylyl cyclase cascade) and perhaps cGMP, as second messengers of sugar-taste transduction. The phosphoinositide pathway (e.g., formation of IP3) is activated by saccharin and SC45647 and perhaps by other nonsugar sweeteners. However, the amiloride-sensitive sweet-transduction pathway may be operative in dogs and rabbits, and perhaps in humans, but not in rats. Ca-imaging experiments suggest that the cAMP and IP3 pathways occur in the same sweet-responsive cell; both pathways elevate cellular Ca level. Could cross-talk between these pathways, as shown in other systems, be needed for optimal sweet-taste response? Additional studies are required to simultaneously quantify the time course of the release of these two second messengers, to determine whether they are sequentially co-released.

A major challenge in the near future is to explore the pathway by which gustducin and/or transducin, two G-proteins known to activate PDE, are involved in sweet-taste transduction. The recent results showing that sweet and bitter sensation are impaired in a-gustducin knockout mice suggest such an involvement and appear to contradict the proposed cAMP and IP3 pathway hypothesis. One may suggest that gustducin activates an additional, different pathway(s). A more likely possibility is that the gustducin- PDE components participate in the aforedescribed mechanisms. Perhaps gustducin is important in steps occurring later in the transduction sequence rather than at the initial stage, for example, in metabolic and other functions (neurotransmitter release) of the stimulated taste cell. The a-gustducin- PDE system may be important in reducing cAMP level after it has been elevated via sugar-induced adenylyl cyclase activity. Furthermore, a- gustducin may be responsible for maintaining a low level of cAMP under resting conditions, thus allowing the cAMP level to be elevated following sugar-induced adenylyl cyclase activity. If so, the resting cAMP level in taste cells of the a-gustducin knockout mice may be too high, resulting in an impaired sugar-stimulated cAMP signal. As suggested (Lindemann, 1996a), this phenomenon may also be related to the impaired sweet response to stimulation by SC45647 (stimulating the PLC-IP3 pathway) in the a-gustducin knockout mice. If cross-inhibition occurs between the cAMP and IP3 pathways, as in other systems (Liu and Simon, 1996), then the IP3 level may increase in response to SC45647 stimulation only if the cAMP level is kept low.

In summary, based on studies utilizing rodent taste tissue, possible transduction pathways may be proposed (Fig. 8). In a sweet-responsive cell, there are at least two pathways for sweet-taste transduction. The first involves sugar stimulation of putative receptors to stimulate the adenylyl cyclase cascade, which forms cAMP via the mediation of a G,-type of G- protein (Fig. 8, left). CAMP, in turn, depolarizes the taste cell by reducing


FIG. 8. Proposed pathways of sweet-taste transduction in a sweet-responsive cell in the rat circurnvallate taste papilla. Abbreviations: R, putative receptors; G, G-protein with a,& subunits; AC, adenylyl cyclase; PKA, protein kinase A; PDE, phosphodiesterase; PLC, phospholipase C; PIP*, phosphatidylinositol bisphosphate.

potassium conductance. The resulting depolarization may lead to the entry of Ca2+ from the extracellular medium via voltage-dependent Ca2+ channels. The a-subunit of gustducin, acting on PDE, may keep the CAMP resting level low before and after sugar stimulation. The second proposed pathway involves nonsugar sweeteners, which stimulate the phosphoinositide transduction pathway to form IP3 via the mediation of the a-subunits of G,-like G-proteins or by the &-subunits of gustducin. IP3 then acts on intracellular Ca2+ stores to release Ca2+. It is still possible that both transduction pathways are initiated by a single taste receptor. Recent studies list a number of cloned, G-protein-coupled receptors (for thyrotropin, luteinizing hormone, calcitonin, parathyroid hormone, vasopressin, catecholamines) that are able to activate both adenylyl cyclase and PLC (Gierschik and Camps, 1993; Zhu et al., 1994). In at least some of these cases, the a-subunit of the heterotrimeric G-proteins activates adenylyl cyclase while, at a higher ago- nist concentration, the &-complex activates PLC. Amphipathic nonsugar sweeteners, in addition to their action on putative receptors, may permeate

M. NAIM, B. J. STRIEM, AND M. TAL

the membrane and act directly on G-proteins or other targets along the transduction pathways.

The increase in the cellular content of Ca2’, either from the extracellular medium due to sucrose or from intracellular stores due to the nonsugar sweeteners SC45647 and saccharin, probably leads to the release of a neurotransmitter in the synapse with sensory nerve fibers that carry the signal to the brain.

ACKNOWLEDGMENTS

We thank Bernd Lindemann for critical reading of the manuscript, Oded Naim for excellent computerized drawings, and Robert F. Margolskee for kindly providing data related to Fig. 6. The editing by Mrs. Camille Vainstein is highly appreciated. Supported by the German-Israeli Foundation for Scientific Research and Development (GIF 1-100-181) and the Deutsche Forschungsgemeinschaft (SFB 246-Cl).

REFERENCES

Abe, K., Kusakabe, Y., Tanemura, K., Emori, Y. , and Arai, S. 1993a. Multiple genes for G protein-coupled receptors and their expression in lingual epithelia. FEBS Lett. 316,

Abe, K., Kusakabe, Y . , Tanemura, K., Emori, Y., and Arai, S. 1993b. Primary structure and cell-type specific expression of a gustatory G-protein-coupled receptor related to olfactory receptors. J. Biol. Chern. 268, 12,033-12,039.

Akabas, M. H., Dodd, J., and Al-Awqati, Q. 1988. A bitter substance induces a rise in intracellular calcium in a subpopulation of rat taste cells. Science 242, 1047-1050.

Anand-Srivastava, M. B. 1989. Amiloride interacts with guanine nucleotide regulatory proteins and attenuates the hormonal inhibition of adenylate cyclase. J . Biol. Chern. 264, 9491- 9496.

Asanuma, N., and Nomura, H. 1982. Histochemical localization of adenylate cyclase and phosphodiesterase activites in the foliate papillae of the rabbit. 11. Electron microscopic observations. Chern. Senses 7, 1-9.

Avenet, P., Hofmann, F., and Lindemann, B. 1988. Transduction in taste receptor cells requires CAMP-dependent protein kinase. Nature 331, 351-354.

Avidor, M., Rajmilevich, G., Beaven, M. A,, and Sagi-Eisenberg, R. 1993. Activation of exocytosis by the heterotrimeric G protein GJ. Science 262, 1569-1571.

Bartoshuk, L. M. 1975. Taste mixtures: Is mixture suppression related to compression? Physiol. Behav. 14,643-649.

Bernhardt, S. J., Naim, M., Zehavi, U., and Lindemann, B. 1996. Changes in IP3 and cytosolic Ca2+ in response to sugars and non-sugar sweeteners in transduction of sweet taste in the rat. J. Physiol. 490,325-336.

Berridge, M. J. 1984. Inositol trisphosphate and diacylglycerol as second messengers. Biochern.

Birch, G. G., Karim, R., Lopez, A., and Morini, G . 1993. Sweetness, structure and specific volume. In “Sweet Taste Chemoreception” (M. Mathlouthi, J. A. Kanters, and G. G. Birch, eds.), pp. 129-140. Elsevier Applied Science, London.

253-256.

J. 220,345-360.


Birch, G. G., Latymer, Z. , and Hollaway, M. 1980. Intensity/time relationships in sweetness: Evidence for a queue hypothesis in taste chemoreception. Chem. Senses 5, 63-78.

Birch, G. G., and Mylvaganam, A. R. 1976. Evidence for the proximity of sweet and bitter receptor sites. Nature 260, 632-634.

Birch, G. G., Parke, S., Siertsema, R., and Westwell, J. M. 1996. Specific volumes and sweet taste. Food Chem. 56, 223-230.

Blochaviak, P., and Jakinovich, W., Jr. 1985. Suppression of the rat’s chorda tympani responses by taste inhibitors. I n “Seventh Annual Meeting of the Association for Chemoreception Sciences, Sarasota, FL,” Abstract 102.

Bouhelal, R., Guillon, G., Homburger, V., and Bockaert, J. 1985. Forskolin-induced change of the size of adenylate cyclase. J. Biol. Chem. 260, 10,901-10,904.

Bradley, R. M. 1973. Electrophysiological investigations of intravascular taste using perfused rat tongue. Am. J. Physiol. 224,300-304.

Brand, J. G., Cagan, R. H., and Naim, M. 1982. Chemical senses in the release of gastric and pancreatic secretion. Annu. Rev. Nutr. 2, 249-276.

Brand, J. G., and Feigin, A. M. 1996. Biochemistry of sweet taste transduction. Food Chem.

Breer, H., Boekhoff, I., and Tareilus, E. 1990. Rapid kinetics of second messenger formation in olfactory transduction. Namre 345,65-68.

Brouwer, J. N., Hellekant, G., Kasahara, Y., van der Wel, H., and Zotterman, Y. 1973. Electrophysiological study of the gustatory effects of the sweet proteins monellin and thaumatin in monkey, guinea pig and rat. Acru Physiol. Scand. 89, 550-557.

Buck, L., and Axel, R. 1991. A novel multigene family may encode odorant receptors: A molecular basis for odor recognition. Cell 65, 175-187.

Cagan, R. H. 1971. Biochemical studies of taste sensation. I. Binding of ‘‘C-labeled sugars to bovine taste papillae. Biochim. Biophys. Acta 252, 199-206.

Cagan, R. H. 1974. Biochemistry of sweet sensation. In “Sugars in Nutrition” (H. L. Sipple and K. E. McNutt, eds.), pp. 19-36. Academic Press, New York.

Cagan, R. H., and Maller, 0. 1974. Taste of sugars: Brief exposure single-stimulus behavioral method. J . Comp. Physiol. Psychol. 87, 47-55.

Cagan, R. H., and Morris, R. W. 1979. Biochemical studies of taste sensation: Binding to taste tissue of 3H-labeled monellin, a sweet-tasting protein. Proc. Nutl. Acud. Sci. USA

Clapham, D. E., and Neer, E. J. 1993. New roles for G-protein By-dimers in transmembrane signalling. Nature 365, 403-406.

Cummings, T. A., Daniels, C., and Kinnamon, S. C. 1996. Sweet taste transduction in hamster: Sweeteners and cyclic nucleotides depolarize taste cells by reducing a K’ current. J. Neurophysiol. 75, 1256-1263.

Cummings, T. A,, Powell, J., and Kinnamon, S. C. 1993. Sweet taste transduction in hamster taste cells: Evidence for the role of cyclic nucleotides. J. Neurophysiol. 70, 2326-2336.

Dastoli, F. R., and Price, S. 1966. Sweet-sensitive protein from bovine taste buds: Isolation and assay. Science 154, 905-907.

DeSimone, J. A,, Heck, G. L., and DeSimone, S. K. 1981. Active ion transport in dog tongue: A possible role in taste. Science 214, 1039-1041.

Desor, J. A,, Maller, O., and Turner, R. E. 1973. Taste in acceptance of sugars by human infants. J . Comp. Physiol. Psychol. 84,496-501.

Dib, K., Oget, I., Wrisez, F., El Jamali, A,, Aguie-Aguie, G., Correz, C., and Lambert, B. 1996. Effects of sodium saccharin diet on fat-cell lipolysis: Evidence for increased function of the adenylyl cyclase catalyst. Inf. J. Obes. 20, 15-20.

56,199-207.

76,1692-1696.


Drewnowski, A. 1991. Fat and sugar: Sensory and hedonic aspects of sweet, high-fat foods. In “Chemical Senses: Appetite and Nutrition” (M. 1. Friedman, M. G. Tordoff, and M. R. Kare, eds.), Vol. 4, pp. 69-83. Marcel Dekker, New York.

Drewnowski, A., Krahn, D. D., Demitrack, M. A,, Nairn, K., and Gosnell, B. A. 1992. Taste responses and preferences for sweet high-fat foods. Evidence for opioid involvement. Physiol. Behav. 51, 371-379.

Drewnowski, A., Krahn, D. D., Demitrack, M. A., Nairn, K., and Gosnell, B. A. 1995. Naloxone, an opiate blocker, reduces the consumption of sweet high-fat foods in obese and lean female binge eaters. Am. J . Clin. Nurr. 61, 1206-1212.

DuBois, G. E. 1997. New insights on the coding of the sweet taste message in chemical structure. In “Olfaction and Taste. A Century for the Senses” (G. Salvadori, ed.), pp. 32-95. Allured Publishing Corp., Carol Stream, IL.

DuBois, G. E., Crosby, G. A.,Lee, J. F., Stephenson, R. A,, and Wang, P. C. 1981. Dihydrochal- cone sweeteners. Synthesis and sensory evaluation of a homoserine-dihydrochalcone conjugate with low aftertaste, sucrose like organoleptic properties. J . Agric. Food Chem.

Faurion, A., Saito, S., and MacLeod, P. 1980. Sweet taste involves several distinct receptor mechanisms. Chem. Senses 5, 107-120.

Fishberg, A. M., Hitzig, W. M., and King, F. H. 1933. Measurement of the circulation time with saccharin. Proc. Soc. Exp. Biol. Med. 30, 651-652.

Fisher, G. L., Pfaffmann, C., and Brown, E. 1965. Dulcin and saccharin taste in squirrel monkeys, rats and men. Science 150,506-507.

Gierschik, P., and Camps, M. 1993. Stimulation of phospholipase C by G protein /3y subunits. In “GTPases in Biology: Handbook of Experimental Pharmacology” (B. Dickey and L. Birnbaumer, eds.), Vol. 108/11, pp. 251-264. Springer, Heidelberg.

29, 1269-1276.

Gilman, A. G. 1984. G proteins and dual control of adenylate cyclase. CeN 36, 577-579. Glaser, D., Hellekant, G., Brouwer, J. N., and van der Wel, H. 1978. The taste responses in

primates to the proteins thaumatin and monellin and their phylogenetic implications. Folia Primatol. 29, 56-63.

Glaser, D., Tinti, J.-M., and Nofre. C. 1996. Gustatory responses of non-human primates to dipeptide derivatives or analogues, sweet in man. Food Chem. 56, 313-321.

Glaser, D., van der We], H., Brouwer, J. N., DuBois, G. E., and Hellekant, G. 1992. Gustatory responses in primates to the sweetener aspartame and their phylogenetic implications. Chem. Senses 17,325-335.

Grand, F. 1974. Sugars in cardiovascular disease. In “Sugar in Nutrition” (H. L. Sipple and K. W. McNutt, eds.), pp. 401-437. Academic Press. New York.

Hagelueken, A., Gruenbaum, L., Nuernberg, B., Harhammer, R., Schunack, W., and Seifert, R. 1994. Lipophilic /3-adrenoceptor antagonists and local anesthetics are effective direct activators of G-proteins. Biochem. Pharmacol. 47, 1789-1795.

Hellekant, G. 1975. Different types of sweet receptors in mammals. In “Olfaction and Taste V” (D. A. Denton and J. T. Coghlan, eds.), pp. 15-21. Academic Press, New York.

Hellekant, G., and Danilova, V. 1996. Species differences toward sweeteners. Food Chem.

Hellekant, G., Glaser, D., Brouwer, J. N., and van der Wel, H. 1976. Gustatory effects of miraculin, monellin and thaumatin in the Sagitinus midas tamarin monkey studied with electrophysiological and behavioral techniques. Acta Physiol. Scand. 97, 241-250.

Hellekant, G., and Ninomiya, Y. 1994. Bitter taste in single chorda tympani taste fibers from chimpanzee. Physiol. Behav. 56, 1185-1188.

Hellekant, G., Roberts, T., af Segerstad, C. H., and van der Wel, H. 1987. Intralingual stimulation with sweet proteins in rhesus monkey and rat. Ann. New York Acad. Sci. 510, 356-358.

56,323-328.


Hellekant, G., and Walters, D. E. 1993. An example of phylogenetic differences in sweet taste: Sweetness of five high-potency sweeteners in rats. In “Sweet-Taste Chemoreception” (M. Mathlouthi, J. A. Kanters and G . G . Birch, eds.), pp. 373-386. Elsevier, London.

Hempel, C. M.. Vincent, P., Adams, S. R., Tsien, R. Y., and Selverston, A. I. 1996. Spatio- temporal dynamics of cyclic AMP signals in an intact neural circuit. Nature 384, 166-169.

Herness, M. S. 1987. Effect of amiloride on bulk flow and iontophoretic taste stimuli in the hamster. J . Comp. Physiology A 160, 281-288.

Higashijima, T.. Burnier. J., and Ross, E. M. 1990. Regulation of G, and Go by mastoparan, related amphiphilic peptides, and hydrophobic amines. J. Biol. Chem. 265,14,176-14,186.

Hiji. Y. 1975. Selective elimination of taste responses to sugars by proteolytic enzymes. Nature 256,427-429.

Hiji, Y., and Sato, M. 1973. Isolation of the sugar-binding protein from rat taste buds. Nature 244,91-93.

Hofer, D.. Puschel. B., and Drenckhahn, D. 1996. Taste-receptor-like cells in the gut identified by expression of a-gustducin. Proc. Natl. Acad. Sci. USA 93, 6631-6634.

Hwang, P. M., Verma, A., Bredt, D. S., and Synder, S. H. 1990. Localization of phosphatidylinositol signaling components in rat taste cells: Role in bitter taste transduction. Proc. Natl. Acrid. Sci. USA 87, 7395-7399.

Irvine, R. F., Letcher, A. J., Lander, D. J., and Berridge, M. J. 1986. Specificity of inositol phosphate-stimulated Ca” moblization from Swiss-mouse 3T3 cells. Biochem. J. 240,

Jacobs, H. L. 1958. Studies on sugar preference. I. The preference for glucose and its modifica-

Jakinovich. W.. Jr. 1982. Stimulation of the gerbil’s gustatory receptors by saccharin. J. Neu-

Jakinovich, W., Jr. 1983. Methyl 4,6-dichloro-4,6-dideoxy-a-D-galactopyranoside: An inhibitor

Jakinovich, W., Jr. 1985. Stimulation of the gerbil’s gustatory receptors by methyl glycopyrano-

Kasahara, T.. Iwasaki. K., and Sato, M. 1987. Ingestive responses to some heavy metal salts

Katz, A., Wu, D., and Simon. M. I. 1992. Subunits Py of heterotrimeric G protein activate

Kier, L. M. 1972. A molecular theory of sweet taste. J. Pharm. Sci. 61, 1394-1397. Kinnamon, J. C. 1987. Organization and innervation of taste buds. In “Neurobiology of Taste

Kinnamon, S. C. 1996. A bitter-sweet beginning. Nature 381, 737-738. Kinnamon, S. C., and Margolskee, R. F. 1996. Mechanisms of taste transduction. Curr. Opin.

Neurobiol. 6, 506-513. Kolesnikov, S. S., and Margolskee, R. F. 1995. A cyclic-nucleotide-suppressible conductance

activated by transducin in taste cells. Nalure 376, 85-88. Koyama, N.. and Kurihara, K. 1972. Mechanism of bitter taste reception: Interaction of

bitter compounds with monolayers of lipids from bovine circumvallate papillae. Biochim. Biophys. Acta 288, 22-26.

Kumazawa, T.. Kashiwayanagi, M., and Kurihara, K. 1985. Neuroblastoma cell as a model for a taste cell: Mechanism of depolarization in response to various bitter substances. Brain Res. 333, 27-33.

Kumazawa, T., and Kurihara, K. 1990. Large enhancement of canine taste responses to sugars by salts. J. Gen. Physiol. 95, 1007-1018.

301-304.

tions by injections of insulin. J. Comp. Physiol. Psychol. 51, 304-310.

rosci. 2, 49-56.

of sweet taste responses in gerbils. Science 219,408-410.

sides. Chem. Senses 10, 501-604.

in mice and inhibition of taste nerve responses by metals. Chem. Senses 12, 295-305.

p2 isoform of phospholipase C. Nature 360, 686-689.

and Smell” (T. E. Finger, ed.), pp. 227-297. Wiley, New York.

240 M. NAIM, B. J. STRIEM, A N D M. TAL

Kumazawa, T., Nomura, T., and Kurihara, K. 1988. Liposomes as model for taste cells: Receptor sites for bitter substances including N- C=S substances and mechanism of membrane potential changes. Biochemistry 27, 1239-1244.

Kurihara, K. 1972. Inhibition of cyclic 3‘, 5‘-nucleotide phosphodiesterase in bovine taste papillae by bitter taste stimuli. FEBS Lett. 27, 279-281.

Kurihara, K., and Koyama, N. 1972. High activity of adenyl cyclase in olfactory and gustatory organs. Biochenz. Biophys. Res. Commun. 48,30-34.

Larson-Powers, N., and Pangborn, R. M. 1978. Paired comparison and time-intensity measurements of the sensory properties of beverages and gelatins containing sucrose or synthetic sweeteners. J . Food Sci. 43, 41-46.

Lawless, H. T., and Stevens, D. A. 1983. Cross adaptation of sucrose and intensive sweeteners. Chem. Senses 7,309-315.

Liang, Y., Maier, Y., Steinbach, G., Lali, L., and Pfeiffer, E. F. 1987. The effect of artificial sweetener on insulin secretion. 11. Stimulation of insulin release from isolated rat islets by acesulfame K (in vitro experiments). Horm. Metabol. Res. 19, 285-289.

Lindemann, B. 1996a. Chemoreception: Tasting the sweet and the bitter. Curr. Biol. 6,1234- 1237.

Lindemann, B. 1996b. Taste reception. Physiol. Rev. 76,719-766. Liu, M., and Simon, M. I. 1996. Regulation by CAMP-dependent protein kinase of a G-

protein-mediated phospholipase C. Nature 382, 83-87. Maddy, A. H. 1966. The chemical organization of the plasma membrane of animal cells. Int.

Rev. Cytol. 20, 1-65. Maller, O., and Desor, J. A. 1974. Effect of taste on ingestion by human infants. In “Oral

Sensation and Perception: Development in the Fetus and Infant” ( J . Bosma, ed.), pp. 279-291. U.S. Govt. Printing Office, Washington, D.C.

Marks-Kaufman, R., and Kanarek, R. B. 1981. Modifications in nutrient selection induced by naloxone in rats. Psychopharmacology 74,321-324.

Marks-Kaufman, R., and Lipeles, B. J. 1982. Pattern of nutrient selection in rats orally self- administering morphine. Nutr. Behav. 1, 33-46.

Matsuoka, I., Mori, T., Aoki, J., Sato, T., and Kurihara, K. 1993. Identification of novel members of G-protein coupled receptor superfamily expressed in bovine taste tissue. Biochem. Biophys. Res. Commun. 194, 504-51 1.

Mayer-Gross, W., and Walker, J. W. 1946. Taste and selection of food in hypoglycemia. Br. J. Exp. Pathol. 27, 297-298.

McGowan, J. C., and Mellors, A. (eds.) 1986. “Molecular Volumes in Chemistry and Biology”. Ellis Horwood, New York.

McLaughlin, S. K., McKinnon, P. J., and Margolskee, R. F. 1992. Gustducin is a taste-cell- specific G protein closely related to the transducins. Nature 357,563-569.

McLaughlin, S. K., McKinnon, P. J., Robichon, A., Spickofsky, N., and Margolskee, R. F. 1993. Gustiducin and transducin: A tale of two G proteins. In “The Molecular Basis of Smell and Taste Transduction” (D. Chadwick, J. Marsh and J. Goode, eds.), pp. 186-200. Wiley, Chickester, UK.

Mierson, S., DeSimone, S. K., Heck, G. L., and DeSimone, J. A. 1988. Sugar-activated ion transport in canine lingual epithelium: Implications for sugar taste transduction. J. Gen. Physiol. 92, 87-1 11.

Miller, I. J., Jr. 1976. Taste bud distribution and regional responsiveness on the anterior tongue of the rat. Physiol. Behav. 16,439-444.

Miller, I . J., Jr., and Spangler, K. M. 1982. Taste bud distribution and innervation on the pallate of the rat. Chem. Senses 7, 99-115.


Mousli, M., Bueb, J.-L., Bronner, C., Rouot, B., and Landry, Y. 1990. G protein activation: A receptor-independent mode of action for cationic amphiphilic neuropeptides and venom peptides. Trends Pharmacol. Sci. 11,358-362.

Naim, M., Bernhardt, S. J., Zehavi, U., Levinson, M., and Lindemann, B. 1996. Molecular aspects of sweet taste transduction. In “The Contribution of Low- and Nonvolatile Materi- als to the Flavor of Foods” (W. Pickenhagen, C.-T. Ho, and A. M. Spanier, eds.), pp. 65-75. Allured Publishing C o p , Carol Stream, IL.

Naim, M., Dukan, E., Zehavi, U., and Yaron, L. 1986. The water-sweet aftertaste of neohesperidin dihydrochalcone and thaumatin as a method for determining their sweet persistence. Chem. Senses 11,361 -370.

Naim, M., Rogatka, H., Yamamoto, T., and Zehavi, U. 1982. Taste responses to neohesperidin dihydrochalcone in rats and baboon monkeys. Physiol. Behav. 28,979-986.

Naim, M., Ronen, T., Striem, B. J., Levinson, M., and Zehavi, U. 1991. Adenylate cyclase responses to sucrose stimulation in membranes of pig circumvallate taste papillae. Comp. Biochem. Physiol. B 100,455-458.

Naim, M., Seifert, R., Niirnberg, B., Griinbaum, L., and Schultz, G. 1994. Some taste substances are direct activators of G-proteins. Biochem. J. 297,451-454.

Neer, E. 1995. Heterotrimeric G proteins: Organizers and transmembrane signals. Cell 80, 249-257.

Nofre, C., and Tinti, J.-M. 1987. Sweetening Agents. “United States Patent 4,645,678. Nofre, C., and Tinti, J.-M. 1994. “Sweetening Agent Derived from L-Aspartic or L-Glutamic

Acid.” United States Patent 5,310,908. Nofre, C., and Tinti, J.-M. 1996. Sweetness reception in man: The multipoint attachment

theory. Food Chem. 56,263-274. Pace, U., Hanski, E., Salomon, Y., and Lancet, D. 1985. Odorant-sensitive adenylate cyclase

may mediate olfactory reception. Nature 316, 255-258. Price, S. 1973. Phosphodiesterase in tongue epithelium: Activation by bitter taste stimuli.

Nature 241, 54-55. Robison, G. A., Butcher, R. W., and Sutherland, E. W. (eds.) 1971. “Cyclic AMP.” Academic

Press, New York. Rodin, J. 1978. Has the distinction between internal versus eternal control of feeding outlived

its usefulness? In “Recent Advances in Obesity Research” (G. Bray, ed.), Vol. 2, pp. 75-85. Newman, London.

Rohse, H., and Belitz, H.-D. 1991. Shape of sweet receptors studied by computer modelling. In “Sweeteners: Discovery, Molecular Design and Chemoreception” (D. E. Walters, F. T. Orthoefer, and G. E. DuBois, eds.), pp. 171-192. American Chemical Society, Washington, D.C.

Ruiz-Avila, L., McLaughlin, S. K., Wildman, D., McKinnon, P. J., Robichon, A., Spickofsky, N., and Margolskee, R. F. 1995. Coupling of bitter receptor to phosphodiesterase through transducin in taste receptor cells. Nature 376, 80-85.

Schiffman, S. S., and Cahn, H. 1981. Multiple receptor sites mediate sweetness: Evidence from cross adaptation. Pharmacol. Biochem. Behav. 15, 377-388.

Schiffman, S. S., Lockhead, E., and Maes, F. W. 1983. Amiloride reduces the taste intensity of Na’ and Lib salts and sweeteners. Proc. Natl. Acad. Sci. USA 80, 6136-6140.

Schiffman, S. S., Reilly, D. A,, and Clark, T. B. 1979. Qualitative differences among sweeteners. Physiol. Behav. 23, 1-9.

Schiffman, S. S., Sugarman, D., Jakinovich, W., Jr., Paikin, A., and Crofton, V. 1987. Inhibition of sweet taste in humans by methyl 4,6-dichloro-4,6-dideoxy-a-D-galactopyranoside. Chem. Senses 12, 71-76.

242 M. NAIM, B. J. STRIEM, A N D M. TAL

Schultz, G., Rosenthal, W., Hescheler, J., and Trautwein, W. 1990. Role of G proteins in calcium channel modulation. Annu. Rev. Physiol. 52, 275-292.

Seamon, K. B., Padgett, W.. and Daly, J. W. 1981. Forskolin: Unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc. Null. Acud. Sci. USA 78,3363- 3367.

Shallenberger, R. S., and Acree, T. E. 1967. Molecular theory of sweet taste. Nature 216,

Shimazaki, K., Sato, M., and Nakao, M. 1986. Photoaffinity labeling of thaumatin-binding protein in monkey circumvallate papillae. Biochim. Biophys. Acta 884, 291-298.

Simon, M. I., Strathmann, M. P., and Gautam, N. 1991. Diversity of G proteins in signal transduction. Science 252, 802-808.

Simon, S. A., Labarca, P., and Robb, R. 1989. Activation by saccharides of a cation-selective pathway on canine lingual epithelium. Am. J. Physiol. 256, R394-R402.

Simon, S. A., Robb, R., and Garvin, J. L. 1986. Epithelial responses of rabbit tongues and their involvement in taste transduction. Am. J. Physiol. 251, R598-R608.

Sjostrom, L., Garrellick, G., Krotkievski, M., and Luyckx, A. 1980. Peripheral insulin in response to the sight and smell of food. Metabolism 29, 901-909.

Spielman, A. I., Huque, T., Nagai, H., Whitney, G., and Brand, J. G. 1994. Generation of inositol phosphates in bitter taste transduction. Physiol. Behuv. 56, 1149-1 155.

Spielman, A. I., Huque, T., Whitney, G., and Brand, J. G. 1992. The diversity of bitter taste signal transduction mechanisms. In “Sensory Transduction” (D. P. Corey and S. D. Roper, eds.), pp. 307-324. The Rockefeller University Press, New York.

Spielman, A. I., Nagai, H., Sunavala, G., Dasso, H., Breer, H., Boekhoff, I., Huque. T., Whitney, G., and Brand, J. G. 1996. Rapid kinetics of second messenger formation in bitter taste. Am. J. Physiol. 270, C926-C931.

Spillane, W. J.. and McGlinchey, G. 1981. Structure-activity studies on sulfamate sweeteners. 11: Semiquantitative structure-taste relationship for sulfamate (RNHS03-) sweeteners-The role of R. J . Phurm. Sci. 70,933-935.

Steiner, J. 1973. The human gustofacial response: Observation of normal and anencepahalic newborn infants. In “Fourth Symposium on Oral Sensation and Perception” (J . F. Bosma, ed.), pp. 254-278. U S . Department of Health, Education and Welfare, National Institutes of Health, Bethesda, MD.

Striem, B. J., Naim, M., and Lindemann, B. 1991. Generation of cyclic AMP in taste buds of the rat circumvallate papilla in response to sucrose. Cell. Physiol. Biochem. 1, 46-54.

Striem, B. J., Naim, M., Zehavi, U., and Ronen, T. 1990a. Saccharin induces changes in adenylate cyclase activity in liver and muscle membranes in rats. Life Sci. 46, 803-810.

Striem, B. J., Pace, U., Zehavi, U., Naim, M., and Lancet, D. 1989. Sweet tastants stimulate adenylate cyclase coupled to GTP-binding protein in rat tongue membranes. Biochrm.

Striem, B. J., Yamamoto, T., Naim, M., Lancet, D., Jakinovich, W., Jr., and Zehavi, U. 1990b. The sweet taste inhibitor methyl 4,6-dichloro-4,6-dideoxy-a-D-galactopyranoside inhibits sucrose stimulation of the chorda tympani nerve and of the adenylate cyclase in anterior lingual membranes of rats. Chem. Senses 15, 529-536.

480-482.

J. 260, 121-126.

Stryer, L. 1986. Cyclic GMP cascade of vision. Annu. Rev. Neurosci. 9, 87-119. Sutherland, E. W. 1972. Studies on the mechanism of hormone action. Science 177,401-408. Tal, M., Ammar, D. A., Karpuj, M.. Krizhanovsky, V., Naim, M., and Thompson, D. A. 1995.

A novel putative neuropeptide receptor expressed in neural tissue. including sensory epithelia. Biochem. Biophys. Res. Commun. 209, 752-759.

Tinti, J.-M., and Nofre, C. 1996. New high-potency sweeteners. In “The Contribution of Low- and Nonvolatile Materials to the Flavor of Foods” (W. Pickenhagen. C.-T. Ho, and

A. M. Spanier, eds.), pp. 77-93. Allured Publishing Corp., Carol Stream, IL.


Tonosaki, K., and Funakoshi, M. 1988. Cyclic nucleotides may mediate taste transduction.

Tsien, R. W. 1983. Calcium channels in excitable cell membranes. Annu. Rev. Physiol. 45,

Tuckerman, F. 1888. Note on the papilla foliata and other taste areas of the pig. Anat. Ant .

van der Heijden. A,, van der Wel, H., and Peer, G. H. 198Sa. Structure-activity relationships in sweeteners. I. Nitroanilines, sulphamates, oximes, isocoumarins and dipeptides. Chem. Senses 10, 57-72.

van der Heijden, A,, van der Wel, H., and Peer, G. H. 198Sb. Structure-activity relationships in sweeteners. 11. Saccharins, acesulfames, chlorosugars, tryptophane and ureas. Chem. Senses 10, 73-88.

Wong, G. T., Cannon, K. S., and Margolskee, R. F. 1996. Transduction of bitter and sweet taste by gustducin. Nature 381, 796-800.

Yamamoto, T.. and Kawamura, Y. 1971. Inhibitory effect of cupric and zinc ions on sweet taste response in the rat. J . Osaka Univ. Dent. Sch. 11,99-104.

Yatani, A,, Codina, J., Imoto, Y., Reeves, J. P., Birnbaumer, L., and Brown, A. M. 1987. A G protein directly regulates mammalian cardiac calcium channels. Science 238,1288-1292.

Zhu, Z., Gilbert, S., Birnbaumer, M., and Birnbaumer, L. 1994. Dual signaling potential is common among G,-coupled receptors and dependent on receptor density. Mol. Pharma-

Nature 331, 354-356.

341-358.

3, 69-73.

COI. 46,460-469.


ANTIOXIDANT ACTIVITY OF THE LABIATAE

SUSAN L. CUPPETT

Department of Food Science and Technology University of Nebraska-Lincoln

Lincoln, Nebraska 68583

CLIFFORD A. HALL 111

Department of Food and Nutrition North Dakota State University

Fargo, North Dakota 58105

I. 11.

111. IV. V.

VI. VII.

VIII. IX.

Introduction Evolution of Labiatae as Antioxidant Sources Plant Tissue Studies Labiatae Essential Oils as Antioxidants Rosemary Extracts Isolation and Identification of Rosemary Compounds Compound Activities Rosemary Synergism(s)and Heat Stabilities Health Implications References

I. INTRODUCTION

There are a wide range of valuable products that originate from the plant kingdom, including dyes, flavors, medicines, and cosmetic and food preservatives (Veljkovic and Stankovic, 1993). Although there is no defined distinction between spices and herbs, herbs generally include plants used for their leaves, stems, flowers, and roots while spices include herbs as well as aromatic seeds. During the past 40 years there has been interest in defining the benefits of plant secondary metabolites as food preservatives, that is, antioxidants. However, most of the research conducted on identifi-

245

1043-4526198 $25.00 Copyright Q 1998 by Academic Press.

All rights of reproduction in any form reserved.

246 SUSAN L. CUPPETT AND CLIFFORD A. HALL 111

cation, isolation, and testing of active compounds from plants has been conducted within the past 20 years.

There are 31 plant families that have been investigated for active compounds (Veljkovic and Stankovic, 1993); of these, a primary family that has attained a great deal of notarity for active compounds is the Labiatae (Lamiaceae). The Labiatae family consists of approximately 3500 species that are native chiefly to the Mediterranean area, although some have origins in Australia, Southwest Asia, and South America (Veljkovic and Stankovic, 1993). Of these, approximately 500 species have traditionally been used throughout the world for their wide range of medicinal properties (Grieve, 1971; Morton, 1981). Some of the Labiatae species that have been studied are shown in Table I and include balm, basil, hyssop, marjoram, mint, oregano, rosemary, sage, savory, and thyme. Variants can exist within a species, for example, there are 9 variants within oregano (Origanum vulgare L.), 17 within rosemary (Rosmarinus officinalis L.), 7 within sage (Salvia officinalis L.), and 2 within thyme (Thymus vulgaris L.) (Veljkovic

TABLE I LABIATAE SPECIES USED FOR THE

ACTIVE COMPOUNDS

Common name Scientific name

Balm Basil Hyssop Marjoram

Sweet Spanish

Spearmint Peppermint

Oregano Greek Turkish Spanish

Rosemary Sage Savory

Mint

Thyme

Melissa officinalis L. Ocimum basilicum L. Hyssopus officinalis L.

Origanum majorana L. Thymus mastichina L.

Mentha spicata Mentha piperita L.

Origanum vulgare L. Origanum onites L. Coridothymus capitatus L. Rosmarinus officinalis L. Salvia officinalis L. Satureja hortensis L. or S.

montana L. Thymus vulgaris L.

and T. zygis L.

Nore. Adapted from Veljkovic and Stankovic (1993).

LABIATAE ANTIOXIDANT ACTIVITY 247

and Stankovic, 1993). Researchers need to be aware that between variants there can be a wide range of concentrations for specific compounds. For instance, Duke (1992) reported that the concentration of rosmarinic acid from rosemary (R. oficinalis) could range from 3500 to 38,507 ppm.

The driving force for the development of plant-derived compounds for use in the food industry is the increasing concern over the safety of the synthetic antioxidants butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), and tertiary butylated hydroquinone (TBHQ) (Gray et al., 1988; Banias et al., 1992; Wong et al., 1995); this concern impacts consumer acceptance. Consumers have been made very aware of and are sensitive to the use of food additives and respond favorably to the concept of food being “natural.”

II. EVOLUTION OF LABIATAE AS ANTIOXIDANT SOURCES

Prior to the 1950s, there were several limited reports on the use of plants (herbdspices) as antioxidants. The first in-depth study of spices as antioxidants was conducted by Chipault’s group. Chipault et al. (1952) investigated the antioxidant activity of 32 spices as well as extracts of these spices. Of the spices tested, rosemary and sage had very good antioxidant activity. Petroleum ether extracts of most of the spices showed activity, but they were not as effective as the whole spice. Chipault et al. (1956), in a study involving oil-in-water emulsions, found clove to be an outstanding antioxidant. Other spices exhibiting antioxidant activity included allspice, cardamon, cassia, cinnamon, ginger, mace, nutmeg, oregano, black pepper, white pepper, rosemary, sage, thyme, and turmeric.

Since Chipault’s work, researchers throughout the world have investigated a wide range of plants for antioxidant activity with the Labiatae family generally being the source of greatest activity. Studies have evaluated: (1) ground plant tissue; (2) essential oil extracts; (3) crude extracts; and (4) isolated, purified, and identified compounds. The following sections will present information relative to the Labiatae family as sources of antioxidant activity with emphasis on rosemary, sage, oregano, and thyme.

Ill. PLANT TISSUE STUDIES

Palitizsch et al. (1969, as cited by Griffiths and McDonald, 1985) showed that whole rosemary, sage, and nutmeg had significant antioxidant activity in commercial lard when added at levels of 0.1 to 0.2%. Further improve-


ments were obtained by adding ascorbic acid (0.1%) or tocopherols (0.05%), which acted synergistically with the whole spices.

Bishov et al. (1977) reported that the addition of 2.5% ground oregano, thyme, marjoram, and spearmint to a freeze-dried model system of corn oil and carboxymethyl cellulose held at 65°C showed protection factors (ratio of induction period of control to induction period of treatment samples) of 4, 2, 3, and 2.5, respectively. Gerhardt and Blat (1984) reported that thyme had a protection factor of 4.6 in pork fat, which was equal to those of rosemary and sage, while marjoram had a factor of only 1.7 in the pork fat system.

Pizzocaro et al. (1985, as cited by Gray et al., 1988) reported that the addition of ground thyme or oregano to minced sardine muscle stored at 0°C did not show any antioxidant activity while basil had slight activity. Rosemary and sage were tested but no data on their effectiveness were given. Korczak et al. (1988) studied the effect of spices (rosemary, sage, and marjoram) in pork-blend meatballs that were deep fat fried and then held frozen (-18°C). They found that rosemary and sage were effective in reducing oxidation; however, marjoram exhibited strong prooxidant activity. Previous work (Korczak et al., 1987, as cited by Korczak et al., 1988) had shown marjoram to have some antioxidant activity at a temperature of 60°C. These authors felt that the antioxidant activity of marjoram was affected by temperature.

Dried leaves of rosemary added to cooked minced pork meatballs re- tarded the development of warmed over flavor (WOF) during cold storage (Huisman et al., 1994). Rosemary added at a level of 0.05% of the total product weight was found to be acceptable by a sensory panel. Huisman et al. (1994) also studied the effect of the combination of rosemary addition and packaging atmospheres on the development of WOF. Cooked meatballs, with or without rosemary addition, were packaged in different atmospheres, including normal air, 5% air/95% Nz, 3% oz/97% Nz, 1% O2/99% Nf, and 100% N2, and stored at 5°C. Samples were monitored for oxidation by 2-thiobarbituric acid-reactive substance (TBARS) values and sensory evaluation. The combination of rosemary and reduced oxygen in the package resulted in significantly lower TBARS values and significantly higher sensory scores.

Hall et al. (1962) found that the addition of sage to pork sausages treated with sodium chloride was able to inhibit the oxidative effect of the salt. Sage also helped maintain high flavor (sensory) scores of the pork sausage.

Tsimidou ef al. (1995), in a series of studies, evaluated the antioxidant activity of ground oregano in an unsaturated lipid system (mackerel oil) compared to ground rosemary, BHA, or TBHQ. All samples were stored at 40°C in the dark. In the first study, oregano at 1% (w/w) was found to


be equivalent to 200 pprn BHA in controlling oxidation of the mackerel oil. In the second study, oregano and rosemary at 0.5% (w/w) inhibited oxidation of the mackerel oil for about 15 days. When oregano was tested at 1% (w/w), it was found to have activity comparable to 200 ppm TBHQ.

IV. LABIATAE ESSENTIAL OILS AS ANTIOXIDANTS

In the plant, essential oils serve as insect attractants for pollination or they can act as a deterrent to microbial and/or insect attack (Tsimidou and Boskou, 1994). The role of essential oils in food preservation is limited due to their strong odor. The chemical composition of essential oils is very complex and they contain on an average 50-100 components; however, some may contain a primary constituent at very high levels (i.e., 80-90%). A wide range of compounds have been isolated and identified from essential oils, including monoterpenes, sesquiterpenes, and their oxygenated derivatives (Tsimidou and Boskou, 1994).

Farag et af. (1989) tested the essential oils from spices, including some from the Labiatae family, for antioxidant activity in a linoleic acid emulsion system. In addition they identified and tested the basic component of each oil for activity in the same system. Results showed the antioxidant effectiveness of the essential oils from the spices to be: Clove > thyme > rosemary > cumin > sage > caraway. The activities of the primary components of essential oils (Fig. 1) were concentration dependent, and it was found that thymol (thyme) and eugenol (clove) at 1200 ppm were between 0.6 and 0.7 times as effective as 200 pprn BHT while carvone (caraway), thujone (sage), cumaldehyde (cumin), and borneol (rosemary) had little antioxidant activity. In addition, Farag et al. (1989) found that thyme and clove essential oil extracts had antioxidant activity in cottonseed oil. These researchers tested the impact of 50 to 1200 ppm of thyme and clove oil on the sensory (odor) characteristics of cottonseed oil and found no impact.

Thymol and its isomer carvacrol (Fig. 2) are found together but the final flavor impact is affected by their relative concentrations. Thymol is characteristic of thyme while carvacrol is predominant in oregano. The antioxidant activity of the essential oils of oregano has been investigated. Lagouri et af. (1993) tested the essential oils from four different oregano plants common to Greece. Results indicated the antioxidant activity of the oils was related to the carvacrol and thymol levels found in the oils. At 1000 pprn the essential oils were equivalent to 200 ppm BHT in lard.

Many spices and herbs that display antioxidant activity also display antimicrobial activity, both activities have been shown to be related to their phenolic structure (Tsimidou and Boskou, 1994). A relationship between

250 SUSAN L. CUPPEIT AND CLIFFORD A. HALL I11

CH2 I

OH I

QocH3 CH2 I

CHO 0 \

A B C

\

CH CH / \

CH3 H3C CH3 / \

H3C D E F

FIG. 1. Structure of monoterpenes from caraway (A; carvone); clove (B; eugenol); cumin ( C cumaldehyde); rosemary (D; borneol); sage (E; thujone); and thyme ( F thymol). From Tsimidou and Boskou (1994).

A B

FIG. 2. Structures of thymol (A) and carvacrol (B).

LABIATAE ANTIOXIDANT ACTIVITY 25 1

the antioxidant activity and chemical composition was proposed by Farag et al. (1989). Structural features required for activity of monoterpenes (Fig. 1) have been shown to include the presence of a phenolic ring containing an electron repelling group. Data indicate that eugenol and thymol have greater antioxidant activity due to the presence of a hydroxyl group on the aromatic ring while carvone, thujone, and borneol lack this structure, explaining their lack of activity (Farag et al., 1989).

While it is accepted that the antioxidant activity of monoterpenes is related to the presence of a hydroxyl group on the aromatic ring there is also evidence that the presence of an ethylidene group, as found in linalool and linayl acetate (Fig. 3), has the ability to interrupt the oxidative free radical chain (Farag et al., 1989). It is believed that the ethylidene side group has the ability to react with lipid free radicals forming a stable allylic tertiary free radical. It has been shown that linalool has a slight prooxidant activity in heated soybean oil while linayl acetate does not display this type of activity (Farag er al., 1989).

The structure-activity of the monoterpene compounds is not dissimilar to those shown to be important for activity in phenolic acids. Cuppett et al. (1997) reported that antioxidant activity of phenolic acids was related to the presence of two hydroxy groups orrho to each other on the aromatic ring. In addition, cinnamic-based dihydroxyphenolic acids, such as caffeic, ferulic, sinapic, and p-coumaric acids, have greater antioxidant activity than dihydroxy derivatives of benzoic acid (Fig. 4) such as p-hydroxybenzoic, vanillic, syringic, and 3,4-dihydroxybenzoic acids (Marinova and Yanis- hieva, 1994).

V. ROSEMARY EXTRACTS

Although a wide variety of plant materials and their extracts have been studied for antioxidant activity, rosemary and sage have received the great-

A B

FIG. 3. Structures of linalool (A) and linalyl acetate (B).

252 SUSAN L. CUPPETT AND CLIFFORD A. HALL I11

FIG. 4. tives of benzoic acid (lower row).

Structures of cinnamic acid based phenolic acids (upper row) and dihydroxy deriva-

est level of attention. The first use of an extract of rosemary leaves as an antioxidant was reported by Rac and Ostric-Matijasevic in 1955 (as cited by Chen et al., 1992). Berner and Jacobson (as cited by Chen et al., 1992) obtained a patent in 1973 for the production of an antioxidant extract from rosemary using oil as the solvent. Chang et al. (1977) reported a process for the extraction of rosemary and sage, followed by vacuum steam distillation in an edible oil or fat to obtain a colorless, odorless natural antioxidant. Today rosemary is being commercially exploited and sage is being developed as a source of antioxidants.

At present, rosemary extracts are available as the original colored and flavored extract that can be produced in either a lipid-soluble or a water- soluble form (Duxbury, 1989). In addition a decolorized, deodorized product is available (Tsimidou and Boskou, 1994). The process for production of an alcoholic extract of rosemary was described by Loliger (1989). The process involves an initial steam distillation of the rosemary leaves to remove the essential oils. The stripped leaves are then extracted with food grade alcohol. Bracco et al. (1982) described the production of an extraction process using peanut oil as the solvent. This process involved extracting the ground rosemary leaves with oil followed by micronization, heat treatment, and molecular distillation. The condensate from the molecular distillation contains the antioxidants. This process produced an odorless and colorless system.


More recently another technique, supercritical carbon dioxide extraction, has been used to produce extracts of rosemary and sage. This process has been reported to extract more than 60% of the phenolic diterpenes from rosemary and sage. Carnosic acid is the major (77%) component within this extract (Gerard et al., 1995). Supercritical carbon dioxide extraction has the advantage of not using excessive heat (Djarmati et al., 1991), which affects the production of artifacts in the rosemary and sage extracts (Hall and Cuppett, 1997).

The development of commercialized rosemary extracts was based on a great deal of research beginning with work by Chang et al. (1977) who extracted rosemary and sage with a series of solvents: hexane, benzene, ethyl ether, chloroform, ethylene dichloride, dioxane, and methanol. Resultant extracts were then tested, at a 0.02% level, in a prime steam lard system held at 60°C in the dark. These researchers found that the greatest antioxidant activity was located in the methanol extract of rosemary. The methanol extract was further purified and the resultant fraction was tested for activity in potato chips fried in sunflower oil and held at 60°C in the dark for 60 days. Chang et al. (1977) found the purified fraction from the methanol extract of rosemary had outstanding antioxidant activity; however, they did not identify the active compound(s) at that time.

Research has been conducted to evaluate the effectiveness of rosemary extracts on the stability of a variety of food systems. MacNeil et al. (1973) found that a rosemary extract at a 0.01% level was as effective as a polyphos- phate (0.5%) treatment; but a 0.05% level of the extract was needed to be equivalent to a mixture (0.075%) of BHT and citric acid in a cooked mechanically deboned poultry meat system stored at 3°C for 11 days.

Barbut et al. (1985) studied the effectiveness of a rosemary oleoresin (RO) in turkey breakfast sausages composed of 25% mechanically deboned meat. These reseachers found that RO was as effective as the combination of BHA/BHT/citric acid in suppressing oxidative rancidity. In addition, it was found that rosemary-treated samples did not produce the same profile of volatile compounds.

Lai et al. (1991) studied the effectiveness of an oleoresin of rosemary (OR) alone or in combination with sodium tripolyphosphate (STPP) vs STPP plus TBHQ in controlling lipid oxidation in restructured chicken nuggets during refrigerated and frozen storage. Lipid oxidation was monitored by TBARS, sensory evaluation, and chromatographic analysis. TBARS and sensory data showed that STPP/OR was comparable to STPP/ TBHQ in preventing WOF. STPP or OR alone was less effective than the STPP/OR combination in both studies. STPP/OR and STPP/TBHQ prevented oxidation of polyunsaturated fatty acids during frozen storage.


Stoick et af. (1991) studied the oxidative stability of restructured beef steaks processed with OR, TBHQ, and STPP; they found that the addition of the OR gave no benefit over STPP. The OWSTPP combination was equivalent to the TBHQ/STPP treatment in preventing oxidation.

Liu et af. (1992) studied the effectiveness of both a water-soluble rosemary extract and an oil-soluble OR in combination with STPP vs TBHQ and STPP in controlling lipid oxidation in restructured pork steaks. They found OR, either water-soluble or oil-soluble in combination with STPP, did not increase lipid stability over STPP alone.

Boyd et af. (1993) studied the effectiveness of TBHQ, ascorbic acid, and rosemary extract in inhibiting oxidation in cooked frozen fish flakes. These researchers found that the combination of TBHQ with ascorbic acid, was the most effective followed by a mixture of TBHQ, ascorbic acid, and rosemary extract; rosemary extract alone was third in activity.

VI. ISOLATION AND IDENTIFICATION OF ROSEMARYCOMPOUNDS

Concurrent with the evaluation of rosemary extracts as antioxidants to inhibit lipid oxidation in food systems, research was also focused on isolating, identifying, and testing the active compounds contained in the extracts. Wu el af. (1982) reported the fractionation and identification of urosolic acid and carnosol from a methanol extract of rosemary. They found that urosolic acid was not an effective antioxidant in prime steam lard, but carnosol was more effective than BHT. Note that carnosol had been previously isolated and identified by Breiskorn et af. (1964).

Other active compounds have been identified from rosemary. Nakatani and Inatani (1981) identified rosmanol and carnosol (Fig. 5 ) and showed that both were more effective than a-tocopherol, BHT, and BHA using an active oxygen method (AOM). They also reported that rosmanol (Fig. 5) had greater antioxidant activity than carnosol. Inatani et af. (1982) reported a “new” antioxidant phenolic diterpene from rosemary that in reality was a rediscovery of rosmanol and carnosol. Nakatani and Inatani (1983) reported the isolation of rosmadial (Fig. 5) from rosemary. Inatani et al. (1983) reported the antioxidant effectiveness of rosmanol, rosmadial, carnosol, and their derivatives, including diacetylcarnosol, four methyl derivatives of carnosol, triacetylrosmanol, and two monomethyl derivatives of rosmanol. Inatani et af. (1983) also studied two flavones, 5-hydroxy-7,4’-dimethyl flavone and 5,4’-dihydroxy-7-methyl flavone. Three tests were used for antioxidant activity: ferric thiocyanate (FeCN), TBARS, and AOM. When tested by the AOM, carnosol and rosmanol(O.O2% each) had greater activity


OH

I

OH

I11

OH

OH

I1

m /

IV

VI V

0 II

HO

H

VII

FIG. 5. Antioxidant compounds identified from rosemary and sage: carnosol ( I ) , carnosic acid ( II ) , rosmanol (III), rosmaridiphenol (IV), rosmadial (V), rosmariquinone (VI), and rosmarinic acid (VII).

256 SUSAN L. CUPPEIT AND CLIFFORD A. HALL I11

than a-tocopherol, BHA, and BHT, and at 0.01% rosmanol was more active than carnosol (0.01%) while carnosol was comparable to BHA. In the FeCN test, carnosol and rosmanol were equivalent but were slightly less active than BHT, and rosmadial was moderately active. In the TBARS test rosma- no1 and carnosol were equivalent to BHT and slightly better than tocopherol, and rosmadial was moderately active. The derivatives, triacetyl- and dimethylrosmanol, showed weak acti-, ity in FeCN and TBARS tests, but monomethylated derivatives were as strong as rosmanol at a 0.02%.

Rosmarinic acid (RA) (Fig. 5) was reported by Gerhardt and Schroter (1983) to be the second most frequently occurring caffeic acid ester, following chlorogenic acid, and to have antioxidant activity equivalent to that of caffeic acid. Gerhardt and Schroter (1983) detected RA in balm, rosemary, sage, thyme, oregano, marjoram, savory, peppermint, and for the first time in basil, and the levels detected ranged from 0.07 to 0.84%. However, no correlation between antioxidant activity of the herb and its RA content was found. Houlihan et al. (1984) reported the isolation and identification of rosmaridiphenol (Fig. 5), a unique compound having a seven carbon ring in its structure. They also found rosmaridphenol to be more active than BHA in lard and only equivalent to BHT in this test system. Shahidi and Naczk (1995) reported that rosmaridiphenol(O.O2%) had activity equivalent to BHT (0.02%) in steamed lard. Nakatani and Inatani (1984) isolated and identified isorosmanol and epirosmanol from rosemary. When tested, using AOM and FeCN methodologies, both compounds showed high activity in both lard and linoleic acid; in lard they were four times more active than BHA and BHT. Houlihan et al. (1985) isolated and identified rosmariquinone (Fig. 5), which was found to be superior to BHA and equivalent to BHT in controlling the oxidation of a lard system. Rosmariquinone had also been isolated from Salvia rnilitorrhiza Bunge by Hayashi et al. (1970). These researchers named the compound militrone.

VII. COMPOUND ACTIVITIES

Brieskorn and Domling (1969) first reported on the antioxidant activity of carnosic acid and carnosol compared to BHT in inhibiting O2 uptake of methyllinolenate using a Barcroft-Warburg apparatus. Data showed that carnosic acid and carnosol were as effective as BHT and that their effectiveness was concentration dependent.

Chen et al. (1992) investigated methanol, acetone, and hexane extracts of rosemary as well as a bleached methanol rosemary extract for their ability to inhibit lipid oxidation (Rancimat) and lipoxygenase. The extracts were analyzed and found to contain carnosol, carnosic acid, and urosolic


acid. The hexane extract (4.2% yield) contained 100.3 mg/g carnosic acid, 16.4 mg/g carnosol, and 13.7 mg/g urosolic acid; the methanol extract (26% yield) contained only a trace of carnosic acid, 24.1 mg/g carnosol, and 76.6 mg/g urosolic acid; and the acetone extract (13.8% yield) contained 58.4 mg/g carnosic acid, 36.5 mg/g carnosol, and 88.5 mg/g urosolic acid. Bleached methanol extract (5% yield) contained trace amounts of carnosic acid, 42.7 mg/g carnosol, and 180 mg/g urosolic acid. When tested in lard using a Rancimat system it was found that 0.02% carnosol and carnosic acid were superior to BHA and BHT; urosolic acid had minimal activity. The hexane and acetone extracts had better activity than BHA and BHT, but the two methanol extracts were equivalent to BHA and BHT. When tested for inhibition of lipoxygenase, urosolic acid was an effective inhibitor followed by carnosol and carnosic acid. Within the extracts tested, the bleached methanol was best at inhibiting lipoxygenase followed by methanol, acetone, and hexane extracts.

Richheimer et al. (1996) studied the activity of carnosol, carnosic acid, 12-methoxycarnosic acid, 7-methoxyrosmanol, rosmanol, and 7-epimethoxyrosmanol vs BHT, BHA, and TBHQ in soybean oil. These researchers also quantitated the level(s) of these compounds in rosemary leaves. Car- nosic acid was found to be more potent than BHT and BHA but less potent than TBHQ. Rosemary leaves contained between 2 and 3% carnosic acid and small amounts of 12-methoxycarnosic acid and carnosol.

Relative antioxidant activity levels were reported as follows (Richheimer et al., 1996): With carnosic acid assigned a level of 1, carnosol was at 0.44, 7-epimethoxyrosmanol was at 0.42, and 12-methoxycarnosic acid was at 0.1. These values were in agreement with those of Chen et al. (1992) who reported that when added to lard at the 200 ppm level, carnosic acid was about 1.2 times more active than carnosol in the Rancimat test. Relative to synthetic controls, carnosic acid had approximately 7 times the activity of BHT and BHA and a little less than half the activity of TBHQ (Chen et al., 1992). The low level of activity of 12-methoxycarnosic acid found by Richheimer et al. (1996) was thought to be due to its lack of the two ortho phenolic groups adjacent to the isopropyl group. Brieskorn and Domling (1969) had noted that the activity of carnosol and carnosic acid was due to the cooperation of their two ortho phenolic groups with their isopro-

Cuvelier et al. (1994) isolated six major antioxidant compounds from a oleoresin of sage (S. oficinalis): rosmanol, epirosmanol, carnosol, rosmadial, carnosic acid, and methylcarnosate. The researchers questioned the source of methylcarnosate since it was not found in all extracts. To determine the source of the methylcarnosate, the researchers tested the stability of carnosic acid and showed that when carnosic acid in a methanol solution

PYl group.


was monitored over time (temperature not given) there was a gradual loss of carnosic acid and increases in carnosol, methylcarnosate, and rosmadial, and in that order of concentration.

Frankel et al. (1996) studied carnosol, carnosic acid, and rosmarinic acid for their ability to inhibit hydroperoxide decomposition in tocopherol- stripped corn (bulk) oil and in a corresponding corn oil-in-water emulsion. In the bulk oil system carnosic acid and rosmarinic acid were more active than carnosol; however, in the emulsion (pH 4.8-5.0) system, carnosol and carnosic acid were more active than rosmarinic acid. When the effect of pH (4.0, 5.0, and 7.0) of the emulsion system on activity of the three compounds was investigated, it was found that carnosol and carnosic acid were more active at pH 4-5 than at pH 7. The difference in activity at different pHs was attributed to the concept that at the lower pHs the compounds may be more stable and/or may have better reducing capacity while at the higher pH they may be partitioning in such a way as to be lost from the phase interface. The effect of pH on the activity of rosmarinic acid was that it had minimal activity at pH 5 and no activity at pH 4 or 7.

The antioxidant activity of rosemary and sage extracts has been shown to be due to the presence of carnosic acid and carnosol, with carnosic acid being the more potent of the two compounds (Chen et al., 1992; Schwarz and Ternes, 1992; Cuvelier et al., 1994; Richheimer et al., 1996). However, carnosic acid has also been reported to be unstable and artifact compounds can be produced during extract production or during isolation and analysis of the active compounds from rosemary and sage (Hall and Cuppett, 1997). Conversion of carnosic acid to carnosol has been shown to occur via the formation of a semiquinone-quinone intermediate (Fig. 6). However, the conversion of carnosic acid to rosmanol is not a direct conversion, as proposed by Wenkert et al. (1965) (Fig. 7 ) , but requires carnosol as an intermediate (Fig. 8), which then converts to rosmanol (Gonzalez et al., 1992).

VIII. ROSEMARY SYNERGISM(S) AND HEAT STABILITIES

The potential for rosemary extracts to have synergism with other known antioxidants, such as tocopherol, has been investigated, and conflicting results have been reported. Wada and Fang (1992) evaluated the antioxidant effectiveness of 0-tocopherol (AT) (0.05%) and rosemary extract (RE) (0.02%), alone or in combination vs BHA, in a sardine oil model system and in frozen-crushed fish meat. The AT-RE mixture had the strongest activity [as measured by TBARS and peroxide value (PV)]; it delayed the onset of rancidity 5 days longer than the individual compounds and this

LABIATAE ANTIOXIDANT ACTIVITY 2.59

OH

0

/

rosmanol

FIG. 6. by Wenkert ef RI. (1965).

Oxidation mechanism involved in conversion of carnosic acid to rosmanol as proposed

level of activity was comparable to that of BHA. The AT content remained 5 days longer when used in combination with RE than when used alone.

In contrast to the findings of Wada and Fang (1992), Wong et al. (199.5) evaluated the antioxidant effectiveness of a-tocopherol, a rosemary extract, and a sage extract, alone and in combination, in a cooked beef homogenate.


OH

rosmanol

FIG. 7. Proposed mechanism for conversion of carnosol to rosrnanol (Gonzalez et al., 1992).

Their results showed that the combination of the herb extracts with tocopherol were comparable to the activity exhibited by the tocopherol alone, indicating that no synergism was occurring.

In another study Fang and Wada (1993) evaluated the antioxidant activity of AT and RE, alone or in combination, in Fe2+/ascorbic acid and hemopro- tein catalyzed oxidation in a sardine oil or bonito dark muscle model system. Results showed that the mixture of AT and RE extended the induction

LABIATAE

R O J a

ANTIOXIDANT ACTIVITY 261

or R = H; R' = CH, OH

OH

rosmanol

FIG. 8. Proposed mechanism for conversion of carnosic acid to rosmanol via 6,7-dehydrocar- nosic acid (Gonzalez et ul., 1992).

period of the sardine oil oxidation catalyzed by Fe*'/ascorbic acid for 10 and 16 days longer than AT or RE alone, respectively. In addition the AT lifetime in the mixture treatment in the oil system was 10 days longer than when used alone. In the dark muscle system, again the mixture of AT and RE showed greater activity than AT or RE alone.


These researchers (Fang and Wada, 1993) theorized that the mechansim for the synergism between RE and AT was that AT acts as a radical scavenger, combining with free radicals and stabilizing them. RE acts to regenerate the AT by H donation. When the RE was depleted, that is, totally oxidized, then the AT began to oxidize and was lost from the system.

Synergism between rosmariquinone (RQ) and mixed tocopherols in both auotoxidation and photo-sensitized oxidation of stripped soybean oil was investigated by Hall (1996). Data showed that the mixture of RQ (200 ppm) with tocopherols (825 ppm) extended the induction period to reach a PV of 20 meq/kg in a stripped soybean oil undergoing autoxidation to over 200 hr as opposed to RQ alone (75 hr) or tocopherols alone (50 hr), indicating a strong synergism. It was found that during autoxidation RQ acted to spare the tocopherol content of the oxidized oil. In photo-sensitized oxidation, the combination of RQ and tocopherols did not show synergism in controlling oxidation. The differences in activity of the mixture of RQ and tocopherols was attributed to the fact that in autoxidation RQ had the ability to spare the loss of tocopherol, probably by donation of hydrogens to regenerate the tocopherol. However, no such regenerating effect was found in the photo-sensitized system, indicating that tocopherol was being oxidized by singlet oxygen to the tocopheryl quinone, which could not be regenerated (Hall, 1996). The mechanisms/structure-activities of the active compounds that have been identified in rosemary and sage have been reviewed by Hall and Cuppett (1997) and evidence is that carnososic acid, carnosol, rosmanol, rosmarinic acid, and rosmariquinone act primarily as hydrogen donors. Aruoma et al. (1992) reported that carnosic acid and carnosol showed the ability to chelate iron and were effective radical scavengers of peroxy radicals.

The heat stability of rosemary extracts has been studied. Wada and Fang (1992) reported that when AT and RE were tested, alone or in combination, at a higher temperature (60 vs 30°C) in the same system there was an increase in the rate of oxidation and RE showed no activity, having the same induction period as the control at the higher temperature. These resarchers indicated that this lack of activity could be due to instability of RE at the higher temperature or to an increased interaction between the free radicals and the antioxidant compounds of RE. In addition there was a weakened synergism between AT and RE at the higher temperature.

Gray et al. (1988) reported on the antioxidant activity of a rosemary extract that had been heated. When a RO was heated under vacuum to either 204" or 260"C, the samples at 204°C maintained activity for 18 hr of heating and the samples at 260°C maintained activity at the end of 1 hr of heating. However, when the RO was heated in open vials at 204"C, there


was a gradual loss of activity; after 4 hr of heating only 40% of the original level of activity remained.

Schwarz et al. (1992) evaluated the heat stability of diterpenes found in rosemary extracts under three sets of conditions: heating in lard at (1) 100°C over time (up to 90 hr); or (2) at temperatures between 100 and 170°C for 10 hr; and (3) exposed to steam at 200°C and a pressure of <1 mbar for 2 hr. Compounds that were monitored included rosmanol, epirosmanol, 7-methylepirosmanol, and carnosol. In the constant 100°C system there was a continual degradation of the four compounds and an appearance of four unidentified substances that were described as de novo synthesis products of the degradation process. During the 90-hr test period at 1OO"C, rosmanol was reduced from 11.4% to 1.0% of the diphenolic compounds; epirosmanol was reduced from 4% of the compounds to not being present at 90 hr; 7-methylepirosmanol was also absent at 90 hr after starting at 8% in the initial sample; and carnosol went from 78% of the initial content to 5% after 90 hr of heating at 100°C.

In samples heated for 10 hr at temperatures between 100 and 170°C, the four identified compounds, rosmanol, epirosmanol, 7-methylepirosmanol, and carnosol, were monitored as a group and it was found that as the level of heat increased there was a concomitant decrease in the concentration of these compounds. The greatest level of degradation occurred at 170°C where there was a decrease from accounting for 67% of the active components at Time 0 to accounting for only 2% at 10 hr. Similarly, in the steam- heated system, the four compounds were studied as a group and it was found that after 4 hr of heating there was a reduction from 66 to 18% of the diphenolic compound content in the test samples. Even with this reduction of diphenolic compound content, the test sample still exhibited antioxidant activity, but at a much reduced level.

Richheimer er al. (1996) noted that most commercial rosemary extracts are deodorized and bleached; both of these processes can involve heating, which converts carnosic acid to carnosol. In addition, bleaching involves the use of activated charcoal, which can convert carnosic acid to carnosol. Richheimer et al. (1996) found that when carnosic acid or carnosol was heated in methanol, 7-methylrosmanol and 7-methylepirosmanol were formed with 7-methylrosmanol being the primary product.

IX. HEALTH IMPLICATIONS

Beckstrom-Sternberg and Duke (1994) summarized the number of active compounds in 27 spicedherbs and the diseases they have been reported to impact. Rather than reproducing this listing, the goal here is to show the


wide range of activities of four of the plants receiving attention for use in food systems: rosemary, sage, thyme, and oregano. The following indicates the potential health benefit of spices and herbs and the number of active compounds (in parentheses) that had been identified in each at the time of publication. Cancer prevention: oregano (38), rosemary (26), sage (13), and thyme (5). Antiseptic: rosemary (14), oregano ( l l ) , sage (5 ) , and thyme (3 ) . Antiinflammatory: oregano (21), rosemary (15), and sage (6). Fungicide: oregano (lo), rosemary (8), and sage (4). Bactericide: oregano (19), rosemary (19), sage (6), and thyme (5). Spasmolytic: oregano (16), rosemary (14). sage (9, and thyme (5). Antiviral: oregano ( l l ) , rosemary (lo), and thyme (3 ) . Choleretic: oregano (1 1). Sedative: rosemary (6). Anesthetic: rosemary (6) . Antiaggregant: thyme (2). Antitumor: rosemary (9) and sage (7). Hypocholesterolemic: sage (4). Hepatoprotective: rosemary (6). Hypo- tensive: oregano (7). Analgesic: rosemary (7), oregano (6), and thyme (2). Antidepressant: thyme (2). Antiasthmatic: sage (4). Antihistaminic: oregano (6). Diuretic: oregano (6) and rosemary (6). Acetylcholinesterase and Anti- lipoperoxidant: rosemary (5 and 3, respectively). Antihepatotoxic: sage (5) and rosemary (3). Anti-Crohn S: thyme (2). Antiacne, Antiatherosclerotic, Anticold, Antidandruff, Anti-insomniac, and Antiosteoporotic: thyme (2-3). Antioxidant: oregano (14), rosemary (12), sage (7), and thyme (4).

Rosemary has been suggested as a potential treatment for Alzheimer’s disease because of its anticholinesterase activity; Alzheimer’s may result from the degradation of acetylcholine, a necessary neurotransmitter. Rose- mary has been shown to contain five compounds capable of inhibiting acetylcholinesterase: carvacrol, fenchone, limoene, thymol, and 1$-cineole (Beckstrom-Sternberg and Duke, 1994).

The value of rosemary, sage, thyme, and oregano in the medical arena is just beginning to be defined by researchers. Information on the biological activity of rosemary and some of its active compounds has begun to appear in the literature over the past 5 years.

Huang and Ferraro (1992) reported strong inhibitory activity of rosemary extract against 12-0-tetradecanoylphorbol-13-acetate (TPA)-induced inflammation, ornithine decarboxylase (ODC) activity, and tumor promotion. The major component of this extract, carnosol, demonstrated activity similar to the rosemary extract.

Ho et al. (1994) reported that topical application of a methanol extract of rosemary, prepared using the procedure of Wu et al. (1982), 5 min prior to treatment with [3H]BP (benzopyrene) to the backs of CD-1 mice inhibited the formation of [3H]BP-metabolite-DNA adduct in the mouse epidermis. There was a dose response to the level of rosemary extract used. Rosemary was shown to inhibit skin tumor initiation by BP in CD-1 mice. Rosemary inhibited skin tumor initiation by 7,12-dimethylbenz[a]anthra-


cene (DMBA) by up to 92%. In addition, it was also shown that rosemary extract was not a tumor initiator on mouse skin. Topical application to the backs of mice of rosemary extract in combination with 5 nmol TPA inhibited the TPA induction of ODC; this occurred in a dose-dependent manner. Inhibition ranged from 66 to 100%. Rosemary also inhibited TPA-induced skin inflammation and hyperplasia.

Huang et af. (1994) reported that a methanol extract of rosemary was studied for its effects on mouse skin tumor initiation and promotion by benzo[a]pyrene (BAP) and DMBA topically applied to mouse skin. The rosemary extract was found to inhibit the covalent binding of BAP and inhibited tumor promotion by BAP, TPA, and DMBA. There was also an inhibition of ODC, TPA-induced inflammmation, arachidonic acid-induced inflammation, TPA-induced hyperplasia, and TPA-induced tumor promotion. Topical application of carnosol or ursolic acid inhibited TPA-induced ear inflammation, ODC activity, and tumor promotion. They also showed activity toward inhibiting DMBA-intiated tumorigenesis.

Orally administered rosemary extract (1 %) was found to reduce the incidence of DMBA-induced mammary cancer (Singletary and Nelshoppen, 1991); these authors also reported that rosemary extract inhibited in vivo binding of DMBA to mammary epitheila cell DNA and the formation of DNA adducts. Ho el al. (1994) found that dietary (2%) rosemary extract inhibited BP-induced forestomach and lung tumors in A/J mice as well as azoxymethane-induced colon tumors in Sencar treated interperitoneally with DMBA.

Munnunni et al. (1992) reported that rosemary extract that had carnosol and carnosic acid as its primary ingredients was shown to inhibit tert-butyl hydroperoxide and H202 induced mutagenesis of Ames tester strain TA102.

Offord et af. (1995) reported that in animal model systems, a rosemary extract was found to inhibit the initiation and tumor promotion phases of carcinogensis. This study focused on the mechanisms of rosemary components for blocking caricogenesis induced by the pro-carcinogen BAP in human bronchial cells (BEAS-2B) by monitoring the inhibition of cyto- chrome P450 (CUP) 1Al activity and CYP 1Al mRNA expression and DNA adduct formation. Results showed that when rosemary and its components, carnosol and carnosic acid, at 6 pg/ml or equivalent levels of pure compounds,respectively, were incubated with human bronchial cells in the presence of 1.5 p m BAP there was an 80% reduction in DNA adduct formation, a 50% reduction in CYP 1Al mRNA expression and a 70-90% inhibition of CYP 1Al activity. It was conjectured that the observed reduction in DNA adduct formation by the rosemary/rosemary components was due to the inhibition of BAP conversion to its ultimate active metabolites. Carnosol was found to affect the expression of the phase I1 enzyme


glutathione-S-transferase (GST), which is known to detoxify the proximate carcinogen BAP metabolite. When BEAS-2B cells were treated with 1 mg/ml carnosol for 24 hr there was a three- to fourfold induction of GST T mRNA; carnosol also induced the expression of another important phase I1 enzyme, NAD(P)H:quinone reductase. Offord et al. (1995) concluded that rosemary and carnosol show promise in chemoprevention due to their ability to decrease activation and increase detoxification of BAP, an important human carcinogen.

Carnosol and urosolic acid inhibited TPA-induced ear edema, ODC activity, and tumor promotion in mouse skin. Topical application of carnosol together with TPA (5 nmol) to backs of mice previously initiated with DMBA reduced the formation of skin tumors. Urosolic acid has also been shown to reduce TPA-induced skin tumors on mice skin (Ho et al., 1994).

Chan et al. (1995) found that carnosol inhibited lipopolysaccharide and interferon-y induced nitrite production by mouse peritoneal calls by greater than 50%. Paris et al. (1993) studied carnosol, carnosic acid, and rosmanol, and their semisynthetic derivatives (7-0-methylrosmanol, 7-0-ethylrosma- nol, and 11,12-o-o-dimethylcarnosol) for their potential to inhibit HIV protease. Carnosic acid showed the strongest activity against the HIV- 1 protease; carnosic acid also showed potential for inhibiting the HIV- 1 replication.

Cuppett et al. (1996) showed that RQ was found to inhibit the mutagenicity of N-methyl-N-nitrosourea and N-nitrobis(2-oxopropy1)amine in V79 cells. Complete inhibition was obtained with a dose level of 250 nM, indicating that RQ has power as a chemopreventative agent. RQ also inhibited ODC activity in mouse skin treated with TPA.

Haragguchi et al. (1995) showed that carnosol, carnosic acid, rosmanol, and epirosmanol inhibited superoxide anion production in an xanthine/ xanthine oxidase system. Carnosic acid was the most effective inhibitor of superoxide anion generation by xanthine oxidase, reducing production by 70% at 9 pM concentration.

These compounds were also tested for their ability to reduce microsomal and mitochondrial lipid peroxidation induced by NADH or NADPH oxidation (Haragguchi et al., 1995). Rosmanol and epirosmanol were the most effective in reducing microsomal oxidation, producing complete inhibition of oxidation at the 3 pM level. Carnosol and carnosic acid reduced oxidation by 88 and 53%, respectively. Carnosic acid and carnosol achieved full inhibition at 9 pM while rosmanol and epirosmanol required a concentration of 30 pM to achieve complete inhibition of mitochondrial oxidation. Carnosic acid was effective in protecting human erythrocytes against oxidative hemolysis (Haragguchi et al., 1995).


Rutherford et al. (1992) found that ethanolic extracts of sage inhibited t-butylcyclophosphoro[35S]thionate binding to the chloride channel of rat cerebrocortical membranes. Active compounds were shown to be carnosol and carnosic acid.

Although much data have been and are being collected on the potential of compounds from the Labiatae plant family, additional data are needed to fully evaluate the range of antioxidant activities, including possible synergisms with each other and other established natural antioxidants, for instance, tocopherols and ascorbic acid.

REFERENCES

Aruoma, 0. I., Halliwell, B., Aeschbach, R., and Loliger, J. 1992. Antioxidant and prooxidant properties of active rosemary constituents: Carnosol and camosic acid. Xenbiotica

Banias, C., Orepoulou, V., and Thompoulos, C. D. 1992. The effect of primary antioxidants and synergists on the activity of plant extracts in lard. J. Am. Oil Chem. SOC. 69,520-524.

Barbut, S., Josephson, D. B., and Maurer, A. J. 1985. Antioxidant properties of rosemary oleoresin in turkey sausage. J. Food Sci. 50, 1356-1363.

Beckstrom-Sternberg, S. M., and Duke, J. A. 1994. Potential for synergistic action of phytochemicals in spices. In “Spices, Herbs and Edible Fungi,” (G. Charlambous, ed.), pp. 201-223. Elsevier, New York.

Bishov, S. J., Masuoka, Y., and Kapsalis, J. G. 1977. Antioxidant effect of spices, herbs and protein hydrolyzates in freeze-dried model systems: Synergistic action with synthetic phenolic antioxidants. J. Food Process. Preserv. 1, 153-166.

Boyd, L. C., Green, D. P., Giesbrecht, F. B., and King, M. F. 1993. Inhibition of oxidative rancidity in frozen cooked fish flakes by tert-butylhydroquinone and rosemary extract. J. Sci. Food Agric. 61, 87-93.

Bracco, U., Loliger, J., and Viret, J. L. 1982. Production and use of natural antioxidants. J. Am. Oil Chem. SOC. 58,686-690.

Brieskorn, C. H., and Domling, H-J. 1969. Carnosolaure, der wichtige antioxydativ wirksame Inhaltsstoff des Rosmarin-und Salbeiblattes. Z. Lebensmittel. Untersuchung. Forsch.

Brieskorn, C. H., Fuchs, A,, Bredenberg, J. B., McChesney, J. D., and Wenkert, E. 1964. The structure of carnosol. J . Organ. Chem. 29,2293-2298.

Chan, M. M., Ho, C-T., and Huang, H. I. 1995. Effects of three dietary phytochemicals from tea, rosemary and tumeric on inflammation-induced nitrite production. Cancer Lett.

Chang, S. S., Ostric-Matijasevic, B., Heish, 0. L., and Huang, C-L. 1977. Natural antoxidants from rosemary and sage. J. Food Sci. 42, 1102-1106.

Chen, Q., Shi, H., and Ho, C-T. 1992. Effects of rosemary extracts and major constituents on lipid oxidation and soybean lipoxygenase activity. J. Am. Oil Chem. SOC. 69,999-1002.

Chipault, J. R., Mizuno, G. R., Hawkins, J. M., and Lundberg, W. 0.1952. The antioxidant properties of natural spices. Food Res. 17, 46-55.

Chipault, J . R., Mizuno, G. R., and Lundberg, 0. W. 1956. The antioxidant properties of spices in foods. Food Technol. 10,209-211.

22,257-268.

141,lO-16.

96, 23-29.


Cuppett, S., Hall, C. 111, Conway, H., Birt, D., and Lawson, T. 1996. Inhibition of the mutagenicity of alkylating agents by rosmariquinone, a component of Rosmarinus officinalis. Cancer Prevent. Int. 2,25-31.

Cuppett, S . L., Schnepf, M., and Hall, C. 111. 1996. Natural antioxidants-Are they a reality? I n “Natural Antioxidants” (F. Shahidi, ed.), pp. 1-11. American Oil Chemists’ Society, Champaign, IL.

Cuvelier, M-E., Berset, C., and Richard, H. 1994. Antioxidant constituents in sage (Salvia officinalis). J. Agric. Food Chem. 42,665-669.

Djarmati, Z., Jankov, R. M., Schwirtlich, E., Djulinac, B., and Djordjevic, A. 1991. High antioxidant activity of extracts obtained from sage by supercritical C02 extraction. J. Am. Oil Chem. SOC. 68,731-734

Duke, J. A. 1992. In “Handbook of Phytochemical Constituents of GRAS Herbs and Other Economic Plants,” p. 532. CRC Press, Boca Raton, FL.

Duxbury, D. D. 1989. Decolorized, deflavorized rosemary extract’s antioxidant properties retard oxidative rancidity. Food Process. May, 62-63.

Economou, K. D., Oreopoulou, V., and Thomopoulos, C. D. 1991. Antioxidant activity of some plant extracts of the family Labiatae. J. Am. Oil Chem. SOC. 68, 109-113.

Fang, X., and Wada, S. 1993. Enhancing the antioxidant effect of a-tocopherol with rosemary in inhibiting catalyzed oxidation caused by Fez+ and hemeoprotein. Food Res. Int. 26,

Farag, R. S., Badel, A. Z. M. A., and Baroty, G. S. A. 1989. Influence of thyme and clove essential oils on cottonseed oil oxidation. J. Am. Oil Chem. SOC. 66,800-804.

Frankel, E. N., Huang, S-W., Aeschbach, R., and Prior, E. 1996. Antioxidant activity of a rosemary extract and its constituents, carnosic acid, carnosol, and rosmarinic acid, in bulk oil and oil-in-water emulsion. J. Agric. Food Chem. 44, 131-135.

Gerard, D., Quirin, K-W., and Schwarz, E. 1995. COZextracts from rosemary and sage. Food Market. Technol. Oct., 46-52.

Gerhardt, U., and Blatt, P. 1984. Dynamishce messmethode zur ermittlung der fettstabilitat. Einfluss von gewurzen und zusatzstoffen. Fleischwirtschraft 64,484-486.

Gerhardt, U., and Schroter, A. 1983. Rosmarinic acid-An antioxidant occurring naturally in herbs. Fleischwirtschraft 63, 1628-1630.

Gonzalez, A., Andres, L., Aguiar, Z., and Luis, J. 1992 Diterpenes from Salvia mellifera and their biogenetic significance. Phytochemistry 31,1297-1305.

Gray, J. I., Crackle, R. L., Cook, R. J., Gastel, A. L., Evans, R. J., and Buckley, D. J. 1988. “Observations on the Antioxidant Properties and Thermal Stability of an Oleoresin Rosemary.” Presented at the Third International Conference on Ingredients and Addi- tives, London, Nov. 15-17.

Grieve, M. 1971. In “A Modem Herbal,” Vol. 11, pp. 700-707. Dover Publications, New York. Griffiths, B., and McDonald, B. 1985. The natural answer. Food May, 44-47. Hall, C. A. 111.1996. “A Mechanistic Study to Determine and Define the Antioxidant Activity

of Rosmariquinone.” Doctorate thesis, University of Nebraska, Lincoln. Hall, C. A. III., and Cuppett, S. L. 1997. Structure-activities of natural antioxidants. In

“Antioxidant Methodology: In Vivo and In Vifro concepts” (0. I. Auroma and S. L. Cuppett, eds.), pp. 141-172. American Oil Chemists’ Society, Champaign, IL.

Hall, J. L., Harrison, D. L., and Mackintosh, D. L. 1962. Counter effect of sodium chloride and sage on development of peroxide in frozen stored sausage. Food Technol. 16(3), 102-104.

Haragguchi, H., Saito, T., Okamura, N., and Yugi, A. 1995. Inhibition of lipid peroxidation and superoxide generation by diterpenoids from Rosmarinus oficinalk. Planta Med.

405-411.

61-333-336.

LABIATAE ANTIOXIDANT ACTIVITY

Hayashi, T., Kakisawa, H., Hsu, H.-Y., and Chen, Y. 1970. The structure of militrone, a new diterpenoid quinone. Chem. Commun. 1, 299-305

Ho, C-H., Ferraro, T., Chen, Q., Rosen, R. T., and Huang, M-T. 1994. Phytochemicals in teas and rosemary and their cancer- preventing properties. In “Food Phytochemicals for Cancer Prevention 11. Teas Spices and Herbs” (C-T. Ho, T. Osawa, M-T. Huang, and R. T. Rosen, eds), pp. 7-16. American Chemical Society, Washington, D.C.

Houlihan, C. M., Ho, C-T., and Chang, S. S. 1984. Elucidation of the chemical structure of a novel antioxidant, rosmaridiphenol, isolated from rosemary. J. Am. Oil Chem. SOC.

Houlihan, C. M., Ho, C-T., and Chang, S. S. 1985. The structure of rosmariquinone-A new antioxidant isolated from Rosmarinus officinalis L. J. Am. Oil Chem. SOC. 62,96-98.

Huang, M-T., and Ferraro, T. 1992. Phenolic compounds in food and cancer prevention. In “Phenolic Compounds in Food and Their Effects on Health. 11. Antioxidants & Cancer Prevention,” (M-T. Huang, C-T. Ho, and C. Y. Lee, eds), pp. 24-25. American Chemical Society, New York.

Huang, J. D., Ho, C-T., Wang, Z. Y., Ferraro, T., Lou, Y. R., Stauber, K., Ma, W., Georgiadis, C., Laskin, J. D., and Conney, A. H. 1994. Inhibition of skin tumorigenesis by rosemary and its constituents carnosol and ursolic acid. Cancer Res. 54,701-708.

Huisman, M., Lindberg Madsen, H., Skibsted, L. H., and Bertelsn, G. 1994. The combined effect of rosemary (Rosmarinus officinalis L.) and modified atmosphere packaging as protection against warmed over flavour in cooked minced meat. Z. Lebensmittel. Untersu- chung. Forsch. 198,57-59.

Inatani, R., Nakatani, N., and Fuwa, H. 1983. Antioxidative effect of the constituents of rosemary (Rosmarinus officinalis L.) and their derivatives. Agric. Biol. Chem. 47,521-528.

Inatani, R., Nakatani, N., Fuwa, H., and Seto, H. 1982. Structure of a new antioxidant phenolic diterpene isolated from rosemary (Rosmarinus officinalis L.). Agric. Biol. Chem. 46,1661- 1666.

Korczak, J., Flaczyk, E., and Pazola, Z. 1988. Effects of spices on stability of minced meat products kept in cold storage. Fleischwirtschrafr 68, 64-66.

Lagouri, V., Blekas, G., Tsimidou, M., Kokkini, S., and Boskou, D. 1993. Composition and antioxidant activity of essential oils from oregano plants grown wild in Greece. Z. Lebens- mittel. Untersuchung. Forsch. W7, 20-23.

Lai, S-H., Gray, J. I., Smith, D. M., Booren, A. M., Crackel, R. L., and Buckley, D. J. 1991. Effects of oleoresin rosemary, tertiary butylhydroquinone, and sodium tripolyphosphate on the development of oxidative rancidity in restructured chicken nuggets. J. Food Sci.

Loliger, J. 1989. Natural antioxidants. In “Rancidity in Foods” (J. C. Allen and R. J. Hamilton, eds), pp. 113-123. Elsevier, New York.

Liu, H. F., Booren, A. M., Gray, J. I., and Crackel, R. L. 1992. Antioxidant efficacy of oleoresin rosemary and sodium tripolyphosphate in restructured pork steaks. J. Food Sci. 57,803-806.

MacNeil, J. H., Dimick, P. S., and Mast M. G. 1973. Use of chemical compounds and a rosemary spice extract in quality maintenance of deboned poultry meat. J. Food Sci. 38,1020-1021.

Marinova, E. M., and Yanishieva, N. V. 1994. Effect of lipid unsaturation on the antioxidant activity of some phenolic acids. J. Am. Oil Chem. SOC. 71,427-434.

Morton, J. 1981. In “Atlas of Medicinal Plants of Middle America,” pp. 780-784. Charles C. Thompson, Springfield, IL.

Munnunni, M., Wolleb, U., Mueller, O., Pfeifer, A., and Aeschbacher, H. U. 1992. Natural antioxidants as inhibitors of oxygen species induced mutagenicity. Mutur. Res. 269,

61,1036-1039

56,616-620.

193-200.


Nakatani, N. 1992. Natural antioxidants from spices. In “Phenolic Compounds in Food and Their Effects on Health. 11. Antioxidants & Cancer Prevention” (M-T. Huang, C-T. Ho, and C. Y. Lee, eds.), pp. 72-85. American Chemical Society, New York.

Nakatani, N., and Inatani, R. 1981. Structure of rosmanol, a new antioxidant from rosemary (Rosmarinus officinalis L.) Agric. Bid . Chem. 45,2385-2386.

Nakatani, N., and Inatani, R. 1983. A new diterpene lactone, rosmadial, from rosemary (Rosmarinus officinalis L.). Agric. Biol. Chem. 47, 353-358.

Nakatani, N., and Inatani, R. 1984. Two antioxidative diterpenes from rosemary (Rosmarinus officinalis L.) and a revised structure for rosmanol. Agric. Bid. Chem. 48, 2081-2085.

Offord, E. A., Mace, K., Ruffieux, C., Malone, A.,and Pfeifer, A. M. 1995. Rosemary components inhibit benzo[a]pyrene-induced genotoxicity in human bronchial cells. Carcinogene- sis 16, 2057-2062.

Paris, A., Strunkelj, B., Renko, M., Turk, V., Pukl. M., Umek, A,, and Korant, B. D. 1993. Inhibitory effect of carnosic acid on HIV-1 protease in cell-free assays. J. Nut. Prod.

Richheimer, S. L., Bernart, M. W., King, G. A,, Kent, C., and Bailey, D. T. 1996. Antioxidant activity of lipid-soluble phenolic diterpenes from rosemary. J. Am. Oil Chem. Soc. 73,

Rutherford, D. M., Nielsen, M. P., Hansen, S. K., Witt, M. R., Bergendorf, 0.. and Sterner, 0.1992. Isolation and identification from Salvia officinalis of two diterpenes which inhibit t-butylcy~lophosphoro[35~]thionate binding to chloride channel of rat cerebrocortical membranes in vitro. Neurosci. Lett. 135, 224-226.

Schwarz, K., and Ternes, W. 1992. Antioxidative constituents of Rosmarinus officinalis and Salvia officinalis. I1 Isolation of carnosic acid and formation of other diterpenes. Z. Lebensmittel. Untersuchung. Forsch. 195, 99-103.

Schaarz, K., Ternes. W., and Schmauderer, E. 1992. Antioxidative constituents of Rosmarinus officinalis and Salvia officinalis. 111. Stability of phenolic diterpenes of rosemary extracts under thermal stress as required for technological processes. Z . Lebensmittel. Untersu- chung. Forsch. 195, 104-107.

Shahidi, F., and Naczk, M. 1995. Antioxidant properties of food phenolics. In “Food Phenolics: Sources, Chemistry Effects, Applications,” pp. 268-269. Technomic Pub. Co., Basel, Swit- zerland.

Singletary, K. W., and Nelshoppen, J. M. 1991. Inhibition of 7,12-dimethylbenz[a]anthracen- e(DMBA)-induced mammary tumorigenesis and of in vivo formation of mammary DMBA-DNA adducts by rosemary extract. Cancer Lett. 60, 169-175.

Stoick, S. M., Gray, J. I., Booren, A. M., and Buckley, D. J. 1991. Oxidative stability of restructured beef steaks processed with oleoresin rosemary, tertiary butylhydroquinone and sodium tripolyphosphate. J. Food Sci. 56, 597-600.

Tsimidou, M., and Boskou, D. 1994. Antioxidant activity of essential oils from the Lamincae family. In “Spices, Herbs and Edible Fungi” (G. Charalmbous, ed), pp. 273-284. Elsevier, New York.

Tsimidou, M., Paprver. E., and Boskou, D. 1995. Evaluation of oregano antioxidant activity in mackerel oil. Food Res. Int. 28,431-433.

Veljkovic, V. B., and M. Z., Stankovic. 1993. Herbs and their uses. In “Encyclopaedia of Food Science, Food Technology and Nutrition” Volume 4 (R. Macrae, R. K. Robinson, and M. J. Sadler, eds.), Vol. 4, pp. 2334-2348. Academic Press, New York.

Wada, S., and Fang, X. 1992. The synergistic antioxidant effect of rosemary extract and a- tocopherol in sardine oil model and frozen-crushed fish meat. J. Food Process. Preserv.

56, 1426-1430.

507-514

16,263-274.


Weng, X . C., and Gordon, M. H. 1992. Antioxidant activity of quinones extracted from tanshen (Salvia rniltiorrhiza bunge) J. Agric. Food Chem. 40, 1331-1336.

Wenkert, E., Fuchs, A., and McChesney, J. D. 1965. Chemical artifacts from the family Lahiatae. J . Organ. Chem. 30, 2931-2934.

Wong, J. W., Hashimoto, K., and Shibamoto, T. 1995. Antioxidant activities of rosemary and sage extracts and vitamin E in a model meat system. J. Agric. Food Chem. 43,2707-2712.

Wu, J. W., Lee, M-N., Ho, C-T., and Chang, S. S. 1982. Elucidation of the chemical structures of natural antioxidants isolated from rosemary. J. Am. Oil Chem. SOC. 59, 339-345.

INDEX

A

Adenylyl cyclase cascade, involvement in sugar-taste transduction, 219-223

Agarose, molecular shapes and interactions,

Alginic acid, molecular shapes and interactions, 172-175

Allergenicity, proteins, see Amino acid sequence alignments, allergenicity potential assessment

intolerance

184-1 85

Allergies, see Food allergies and

Amiloride, sweet-taste transduction,

Amino acids 229-230

hydrophilicity/hydrophobicity, 100-102 sequence alignments, allergenicity

potential, 64-65 algorithms, 46 assessment, 45-91 BLAST, with transgene proteins,

databases, 47-48, 59, 63-91 56-57,59

food allergen sequences, 69-74 nonfood allergen sequences, 75-84 number of accessions, 67 summary, 88 wheat gluten sequences, 85-87

extending local alignments, 57-59

with transgene proteins, 53-56 methods, 47-50,65-67 Needleman and Wunsch algorithm,

positive control sequence construction, 48

transgene test sequences, 48-50

FASTA, 50,52-53,59-60

50-51

Amphiphilicity, graphical representations,

Amylopectin, molecular shapes and interactions, 161

Amylose, molecular shapes and interactions, 156-168

99-102

amylopectin, 161 A-amylose, 157-160 B-amylose, 160-163 V-amylose, 161-165 derivatives, 165, 167-168 KOH-amylose complex, 165-166

Analphylactoid reactions, 10-11 Anaphylactic shock, 5 Antioxidant activity, see Labiatae Apolipoprotein IV, energy-minimized

structures, 109-111 Arabinan, molecular shapes and

interactions, 198-201 Asthma, sultite-induced, 12-13 Avoidance diet, 6

B

Balsam of Peru, contact sensitivity, 24-26 BLAST, with transgene proteins, 56-57, 59

C

Calcium pectate, molecular shapes and interactions, 169-172

CAMP, as second messenger in sweet-taste transduction, 223

Caraway, monterpenes, 250 Carnosic acid, conversion to rosmanol, 259,

Carnosol, conversion to rosmanol, 260 261

273

274 INDEX

Carrageenans, molecular shapes and

Carvacrol, 249-250 aS1-Casein, 94

interactions, 180-1 84

energy-minimized structures, 109-111

Ca*+, intracellular, imaging following stimulation by sweeteners, 225, 227

Celiac disease, 9 Cell-mediated allergies, 8-9 Cellulose, molecular shapes and

interactions, 148-153 Challenge tests, 32, 34-35 Chemoprevention, rosemary and carnosol,

Circular dichroism measurements, 266

emulsification peptides, 116, 118, 120-121

Contact dermatitis, spices, 29 Contact sensitivities, 9-10, 24-26 Cumin, monterpenes, 250 Curdlan, molecular shapes and interactions,

175- 179

D

Difference Fourier maps, 143, 146 Diterpenes, heat stability, 263 Double-blind, placebo-controlled food

challenge, 34

E Edmundson wheels, 99, 101-102 Emulsification assay, 116-1 17 Emulsification peptides, 93-122

amino acid sequences, 107-109 circular dichroism measurements. 116,

design, 104-111 future directions, 119, 121-122 model peptide 4, 116, 118 model peptide 5, 116, 119 model peptide 6, 113-115, 116, 119 synthesis, 112-115 testing of properties, 115-121

Emulsifiers, 115-1 16 Emulsion instability, 94 Emulsion stability, 115-116

118,120-121

Essential oils, role in food preservation, 249

F

FASTA, 50, 52-53, 59-60 with transgene proteins, 53-56

Favism, 11 Flavor enhancers, intolerances, 30 Flavoring substances, 1-37, see also Food

allergies and intolerance acting as haptens, 28 allergic reactions, published examples,

appropriate diagnostic tests, sensitivity 24-28

to, 31-36 assessment strategy, 33 challenge tests, 32, 34-35 patch tests, 36 skin tests, 35

classes, I5 formation, 15 labeling, 20-21, 24 manufacturing, 16-18 mucosal contact sensitization reactions,

nature and composition, 14-15 occupational sensitivities, 31 physical classification, 13-14 physical or technical functional effects,

protein content, 17-20 sulfur-containing, 17-18 usage levels, 20, 22-23 volatile components, 14

allergic reactions

28-29

21, 24

Food allergies and intolerance

cell-mediated, 8-9 delayed-type, 3-4 IgE-mediated, 4-8, 26-27 incidence, 63 likelihood. flavoring substances,

oral allergy syndrome, 5, 7 28-29

avoidance diet, 6 classification, 3 common allergenic foods, 5 contact sensitivities, 9-10, 24-26 definitions. 2

INDEX 275

diagnostic tests, sensitivity to flavoring substances, 31-36

assessment strategy, 33 challenge tests, 32, 34-35 patch tests. 36 skin tests, 35

idiosyncratic reactions, 12-13 immediate allergic reactions, 3 intolerances, likelihood, flavoring

substances, 30-31 metabolic food disorders, 11 non-IgE-mediated histamine release,

oral sensitivities, 24-25 proteins as, 6-7, 46, 64

10-1 1

removed or rendered nonallergenic by

published examples, flavoring substances, processing, 6

24-28 Food, Drug and Cosmetic Act, 20-21 Food polysaccharides, see Polysaccharides Food preservation, essential oils role,

249

G

Galactomannans, molecular shapes and

GenPept database, 66 P-Glucan, molecular shapes and

interactions, 175-179 Gluten-sensitive enteropathy, 9 G-protein, involvement in mediation of

cellular CAMP formation, 220 G,-subunit, GTPase activity,

230-231 Gustducin, involvement in sweetener-

induced taste, 227-229.234-235

interactions, 155-156

H

Helix capping, 96 Histamine poisoning. 10 Histamine release reactions, non-IgE-

mediated, 10-1 1 Hydrolyzed proteins, idiosyncratic

reactions, 31 Hydrophobic moment, 101-102

I

Idiosyncratic reactions, 12-13

IgE-mediated allergic reactions, 4-8, 26-27 Intolerance, see Food allergies and

MSG, 30-31

intolerance

K

KOH-amylose complex, molecular shapes and interactions, 165

L

Labeling, flavoring substances, 20-21, 24 Labiatae, 245-267, see also Rosemary

compound activities, 256-261 essential oils as antioxidants, 249-252 evolution as antioxidant sources, 247 health implications, 263-267 plant tissue studies, 247-249 species, active compounds, 246-247

Lactose intolerance, 11 Linalool, 251 Linalyl acetate, 251 Linked-atom least squares analysis, 145

M

Mannan, molecular shapes and interactions,

Marjoram, antioxidant activity, 248 Melittin

amino acid sequences, 107-109 amphipathic structures, 102 emulsion stability, 116, 118

Metabolic food disorders, 11 Methanol extract, rosemary, 253 Milk, as allergen, 26-27 Monosaccharide unit, atom labeling,

Monterpenes, Labiatae, 249-250 MSG

idiosyncratic reactions, 30-31 role in food-borne illness, 13

152-155

146-147

Multipoint attachment theory, sweeteners, 215, 217

INDEX

N Needleman and Wunsch algorithm,

50-51

0

Occupational sensitivities, 31 Oral allergy syndrome, 5 , 7 Oral sensitivities, flavors or flavoring

substances, 24-25 Oregano, health implications, 264-267

P

Papain, as allergen, 7-8 Patch tests, 36 Peanuts, as allergens, 27 Pectic acid, molecular shapes and

Pectinic acid, molecular shapes and

Pectin, molecular shapes and interactions,

interactions, 169, 171

interactions, 169-172

167, 169-172 calcium pectate, 169-172 pectic acid, 169, 171 pectinic acid, 169-172 sodium pectate, 169-170

Penicillin, as food allergen, 8 Peptides, see also Emulsification peptides

amphiphilic design, 99-100 graphical representations, 99-102

computer-aided design, 103-104 secondary structure, 95-99

aliphatic side chains, 98-99 amino acid sequences. 96-97 P-sheet formation, 98 helix formation, 97-99 hydrophobic and hydrophilic

residues, 96 Phenolic acids, cinnamic acid based,

252 Phosphoinositide transduction pathway,

role in taste induced by saccharin and SC45647 sweeteners, 224-227

Phospholipase C, activity, sweetener- induced taste, 224-226

Plasma membrane, sweet taste receptors, 217-219

Poly(cr-~-guluronic acid, molecular shapes and interactions, 173-175

Poly(P-o-mannuronic acid, molecular shapes and interactions, 172-173

Polysaccharides, 131-204 importance, 131-132 mixed, 200, 202 molecular shapes and interactions,

146-201 (1+3)-linked polysaccharides,

175-180 P-glucan, 175-179 scleroglucan, 179-180

(l-+4)-linked polysaccharides, 147-175

alginic acid, 172-175 amylose, 156-168 cellulose, 148-153 galactomannans, 155-156 mannan, 152-155 pectin, 167, 169-172

alternating (1-3) and (1+4)-linked polysaccharides, 180-185

agarose, 184-185 carrageenans, 180-184

arabinan, 198-201 branched polysaccharides, 194-199

Rhizobium trifolii CPS, 194-197 xanthan, 196-199

gellan family, 185-194 potassium native gellan, 186-188 welan, 191-194

potassium gellan, 187-191 morphology to macroscopic properties,

structure-function-property correlations,

x-ray diffraction, 133-146 crystal systems, 136 electron density, 143 equator and meridian, 142 helical, 144 model building and refinement

techniques, 145-146 ordered molecules, 137-144 single crystals, oriented fibers, and

structure factor, 143 unit cell, 135-136, 138

202-203

132-133

powder specimens, 134-137

INDEX 277

Potassium gellan. molecular shapes and

Potassium native gellan, molecular shapes

Proteins, see also Amino acid sequence

interactions, 187-19 1

and interactions, 186-188

alignments, allergenicity potential assessment

amino acid sequences, 47-48 flavoring substance content, 17-20 as food allergens. 6-7, 46, 64 functional domains, 65 hydrophilic side chains, 94 transferred, 45-46

R

Rhizobium trifolii CPS, molecular shapes

Rosemary and interactions, 194-197

antioxidant activity, 248, 258 compounds

isolation and identification, 254-256 relative antioxidant activity levels,

257 extracts

antioxidant activity, 251-254 heat stability, 262-263

health implications, 264-267 monterpenes, 250 oleoresin, 253-254 synergisms and heat stabilities,

258-263 Rosmariquinone. 262

chemoprevention. 266

S

Saccharin, sweetener-induced taste, phosphoinositide transduction pathway, 224-227

Sage antioxidant activity, 248, 258 antioxidant compounds, 257-258 health implications, 264-267 monterpenes, 250

Scleroglucan, molecular shapes and interactions, 179-180

Scombroid fish poisoning, 10

SC45647 sweeteners, sweetener-induced taste, phosphoinositide transduction pathway, 224-227

215 Shallenberger/Acree AH-B model,

Skin pick test, 35 Skin tests, food allergy, 35 Sodium pectate, molecular shapes and

interactions, 169-170 Soy products, allergenicity, 7 Spices

as allergens, 27-28 contact dermatitis, 29

Strawberry, allergy, 11 Structure factor, 143 Sugar, high intake, linkage to metabolic

Sulfites, induced asthma, 12-13 Supercritical carbon dioxide extraction,

Sweetener-induced taste, 211-236

disorders, 212

253

gustducin/transducin involvement,

multiple transduction pathways,

proposed pathways, 234-235 receptor-independent activation,

transduction pathway, downstream

227-229

233-234

230-232

components, 219-227 adenylyl cyclase cascade involvement,

phosphoinositide transduction pathway 219-223

role, 224-221 Sweeteners

amphipathic nonsugar, receptor-

chemical structures, 216 multipoint attachment theory, 215,

Nofre-Tinti model, 217 recognition by taste-receptor cells,

responses of mammalian species to,

ShallenbergerlAcree AH-B model,

independent activation, 230-232

217

214-219

213

215 Sweet molecule, chemical aspects,

215-217

278 INDEX

Sweet-taste transduction, amiloride- sensitive, 229-230

T

Transgenic foods, concerns regarding allergenicity, 64

2,2,2-TrifluoroethanoI, 96 Tumor initiation, inhibition by rosemary,

264-266

Taste bud, 213-214 Taste-receptor cell, recognition stage,

Thyme 21 4-21 9

antioxidant activity, 248-249 health implications, 264-267 monterpenes, 250

Thvmol, 249-250

W

Welan, molecular shapes and interactions,

Wheat gluten, amino acid sequences, 191-194

85-87

X T lymphocytes, activation by allergens, 8 a-Tocopherol, antioxidant effectiveness,

Transducin, involvement in sweetener-

Xanthan, molecular shapes and heat stability, 258-263 interactions, 196-199

X-ray diffraction, polysaccharides, induced taste. 227-229 133-146

I S B N O-L2-OLbY42-b

advances in food and nutrition research volume 42

Documents