Vol. 6o ITACONIC ACID FORMATION BY A. TERREUS. 1 139

2. Non-proliferating mycelia, grown at pH 2* 1,convert glucose into itaconic acid with about equalyields at pH 6-0 and 2-1. Non-proliferating mycelia,grown at pH 6*0, do not accumulate itaconic acidfrom glucose either at pH 6*0 or 2-1. However,mycelia, grown at pH 6-0 and then transferred to afull growthmedium ofpH 2-1, will graduallydevelopthe ability to produce itaconic acid.

3. The results indicate that an acid environmentis necessary for the formation of an essentialenzyme system operating in the conversion ofglucose into itaconic acid.

The work was supported by grants from the RoyalNorwegian Council for Scientific and Industrial Research.We wish to acknowledge our indebtedness to Dr K. B.Raper for supplying a culture of A. terreus NRRL 1960, toMiss Elizabeth Brigham for assistance in running themanometric analyses and to Dr T. Brun for carrying outsome of the carbon determinations.

REFERENCESBarrenscheen, H. K. & Pany, J. (1941). Die Methoden der

Fermentforschung, 1, 1074, ed. Bamann, E. & Myrback, K.Leipzig: Georg Thieme.

Buch, M. L., Montgomery, R. & Porter, W. L. (1952).Analyt. Chem. 24,489.

Cohen, P. P. (1949). Manometric Technique8 and TissueMetabolism, pp. 166-9, ed. Umbreit, W. W., Burris, R. H.& Stauffer, J. F. Minneapolis: Burgess Publishing Co.

Larsen, H. (1954). Methods in Enzymology, 3, sect. m, ed.Colowick, S. P. & Kaplan, N. 0. New York: AcademicPress. In press.

Lockwood, L. B. & Nelson, G. E. N. (1946). Arch. Biochem.10, 365.

Lockwood, L. B. & Reeves, M. D. (1945). Arch. Biochem. 6,455.

Moyer, A. J. & Coghill, R. D. (1945). Arch. Biochem. 7, 167.Nossal, P. M. (1952). Biochem. J. 60, 349.Pawlowski, F. (1932). Die brautechnischen Untersuchungs-

methoden, pp. 126-8. Mtnchen: R. Oldenbourg.Swim, H. E. & Krampitz, L. 0. (1954). J. Bad. 67, 419.

The Mechanism of Itaconic Acid Formation by Aspergillus terreus2. THE EFFECT OF SUBSTRATES AND INHIBITORS

BY K. E. EIMHJELLEN* AND H. LARSENDepartment of Chemistry, The Technical University of Norway, Trondheim, Norway

(Received 30 July 1954)

Published work on the itaconic acid fermentation isvirtually limited to investigations of the optimumconditions for obtaining high yields of itaconic acidfrom glucose and sucrose. Very little information isavailable concerning the conversion of other carbonsources into itaconic acid by mould fermentation.Calam, Oxford & Raistrick (1939) reported thatglucose was the only compound, among a numbertested, which gave rise to substantial amounts ofitaconic acid in surface culture experiments withAspergilus tereus. Citric, malic, pyruvic and aceticacids, either alone or in mixture, gave no or onlyvery small amounts of itaconic acid.

In the present investigation the ability of A.terreus to form itaconic acid from a variety ofcarbonsources has been investigated, using the methodsdeveloped in our laboratory for the study ofitaconicacid production by cultures growing on a shaker(Larsen & Hovden, in preparation) and by non-proliferating mycelial suspensions (Larsen &Eimhjellen, 1955). The results of these experiments,which have been carried out in both the presence

and absence of metabolic inhibitors, serve as a basisfor a discussion of the mechanism of itaconic acidformation.

EXPERIMENTAL

The experiments were carried out with A. terreus, strainNRRL 1960. Growth experiments were carried out withcultures grown in acid media on a shaking machine asdescribed in the previous paper (Larsen & Eimhjellen,1955); using this technique various compounds were testedas carbon source both for growth and for itaconic acidproduction.To measure the QO, (Ild. 02 consumed/mg. dry wt.

mycelia/hr.) of non-proliferating mycelia the organism wasgrown on the acid glucose medium (Larsen & Eimhjellen,1955) at 30° for 50-60 hr., centrifuged and washed with andsuspended in either tap water acidified to pH 2-1 withH,SO, or in 0*067M sodium-potassium phosphate buffer(pH 7-0). Before use the suspensions were aerated for 2 daysat 300 to reduce the endogenous respiration. Oxygen uptakewas measured in conventional Warburg manometers, andthe CO. was absorbed by 10% (w/v) KOH in the centre well.To test for itaconic acid production from various sub-

strates by suspensions of non-proliferating mycelia, 50-60 hr. cultures grown in the acid glucose medium wereharvested by filtration as described in the previous paper* Fellow of The Technical University.

K. E. EIMHJELLEN AND H. LARSEN(Larsen & Eimhjellen, 1955) and suspended in the acid tap-water solution of pH 2*1. The suspensions were incubatedwith substrate according to Larsen & Eimhjellen (1955).Samples were withdrawn at suitable time intervals andanalysed after removal of the mycelium by centrifuging.

Reagent8. c6s-Aconitic acid was prepared according toMalachowski & Maslowski (1928) and Krebs & Eggleston(1944); DL-citramalic acid according to Michael & Tissot(1892) with the modification introduced by Marckwald &Axelrod (1899); itaconic acid according to Larsen (1954);and DL-isocitric acid according to Fittig & Miller (1889) andOchoa (1948). Oxaloacetic acid was obtained by saponifica-tion of the diethyl ester (Krampitz & Werkman, 1941);commercial pyruvic acid was purified according to Robert-son (1942). Preparations ofL-sorbose, L-rhamnose, L-fucose,L-xylose and D-lyxose were kindly furnished by ProfessorN. A. Sorensen. Other organic reagents were commercialpreparations of highest purity available and were usedwithout further purification.

Analytical methods. Dry matter was determined aspreviously described (Larsen & Eimhjellen, 1955); glucoseand maltose according to Hagedorn-Jensenwith the modifi-cation of Fujita-Iwatake (Barrenscheen & Pany, 1941);total acid by alkali titration; itaconic acid by a semimicromodification of Friedkin's method (Larsen, 1954); citricacid according to Lardy(1949) and cis-aconitic acid accord-ing to Nekhorocheff & Wajzer (1953). Control experimentsshowed that glucose, citric and itaconic acids did not inter-fere with the determination of cu-aconitic acid. Itaconicacid did not interfere with the citric acid determination toany significant extent. Pyruvic acid was determined by themethod ofFriedemann (1950). Acetaldehyde formed by themycelium in the presence of arsenite evaporated from thereaction mixture and was trapped in absorption tubeseither as the bisulphite compound or as the 2:4-dinitro-phenylhydrazone. The amount of acetaldehyde trapped inthe bisulphite solution was determined iodometrically.For the identification of pyruvic acid and acetaldehyde

the 2:4-dinitrophenylhydrazones were prepared and re-crystallized from 96% (v/v) ethanol. The m.p.'s and mixedm.p.'s were determined using authentic samples preparedfrom commercial pyruvic acid and acetaldehyde. The Rpvalues of the 2:4-dinitrophenylhydrazones were determinedby paper chromatography according to Cavallini, Frontali &Toschi (1949).When sodium arsenite was present glucose and itaconic

acid were determined by the methods given above, savethat determinations of blank corrections were made onaqueous solutions of sodium arsenite of the appropriateconcentration instead of on water.The various metabolic reaction mixtures were tested

routinely for the presence of organic acids by paper chro-matography as previously described (Larsen & Eimhjellen,1955).

RESULTS

Growth experiment8. These experiments werecarried out with shake cultures. At the time ofpeakalkali titre the itaconic acid content was deter-mined and the amount of growth judged by visualinspection. The results, which are presented inTable 1, show that A. terreus, strain NRRL 1960,can grow on a large variety of sugars and alcohols;

I955even the methyl pentose L-rhamnose is utilized forgrowth. During growth the pH fell to 2-0-2-2.The ability of the mould to produce itaconic acid

from these compounds presents a somewhatdifferent picture. Some of the substrates gave riseto a fair production of itaconic acid, but the yieldswere variable (Table 1). In other cases no itaconicacid was detected in the fermentation liquor,although the substrate gave rise to good growth.

Table 1. Growth and itaconic acidformationwith sugars and alcohols as substrates

20 ml. medium with 5% (w/v) substrate in 100 ml.flasks. (+), Barely visible growth; +, slight growth;+ +, fair growth; + + +, good growth; -, not tested.

SubstrateInulinStarchGlycogenRaffinoseCellobioseMaltoseLactoseSucroseD-GlucoseD-MannoseD-GalactoseD-FructoseL-SorboseL-RhamnoseL-FucoseD-ArabinoseL-ArabinoseD-XyloseL-XyloseD-RiboseD-LyxosemesoInositolD-MannitolmqsoGalactitol(=dulcitol)

GlycerolEthanol

Growth

+++

+

+

+

+

(+)

Itaconic acidformed in % (w/w)of substrate added

12000

410

575232426

0

1831

0

2311

Growth experiments with citric and acetic acidsshowed that these compounds were effectivelyutilized as sole source of carbon provided that thepH was kept above 4. However, no itaconic acidaccumulated under these conditions, although thepH of the medium, which increased during growth,was frequently readjusted during the experimentto the lower limit at which growth occurred.

Respiration experiments. The results of the QO2measurements are presented in Table 2, and reflectthe ability of A. terreus to attack a large number ofcommon metabolic intermediates. The addition ofmembers of the tricarboxylic acid cycle caused asignificant increase in the rate of oxygen uptake,

140

ITACONIC ACID FORMATION BY A. TERREUS. 2

although the effect ofoxaloacetithe C2 compounds tested, acetgave relatively high QO valuestion of ethylene glycol and glyialmost negligible stimulation oi

Table 2. Oxidation of subst8uwpen,8io at pH 2-1 i

Mycelia suspended in water acipH 2-1 or in 0-067m sodium-potassipH 7 0. To each Warburg vesselmycelial suspension in the main co](20jimoles) of substrate in the sid4added to the mycelial suspension alair. Temp. 30°. O

SubstrateD-GlucosePyruvic acidAcetic acidSuccinic acidDL-Malic acidFumaric acidOxaloacetic acidL-Tartaric acidCitric acidcis-Aconitic acidtrans-Aconitic acidDL-ssoCitric acidItaconic acidac-Oxoglutaric acidEthylene glycolGlycollic acidEthanol

VF

* Endogenous QO = 10 is subtra

ic acid was small. Of Formation of itaconic acid by non-proliferatingjic acid and ethanol mycelia. The extent of itaconic acid formation byi, whereas the addi- non-proliferating mycelia in the presence of acolic acid caused an variety of substrates is shown in Table 3. Allfoxygen uptake. attempts to reduce the considerable endogenous

activity were invariably accompanied by an even'rates by mnycelial stronger reduction of the capacity of the suspensionand pH 7m0 to form itaconic acid from substrates. The rate ofand pH 7.0 endogenous itaconic acid production was particu-dpfiedhohat HbSO4 tof larly rapid at the start ofthe experiment and usuallywere added 2b0mle of fell to about zero after 2-4 hr. (cf. Larsen &mpartment and 0*2 ml. Eimhjellen, 1955, Fig. 1). For the evaluation of suche bulb. Substrate was experiments it was thus made the rule to startt zero time. Gas phase measuring the changes taking place 2-3 hr. after the*(./mg. d t/h) suspension had been transferred to the aerateda______________ tubes in the water bath, and to subtract the amountH 2-1 pH 7 0 of endogenous itaconic acid production from that-17 -22 formed in the presence of the substrate.- 6 -11 By determining the amount of substrate utilized-6 -17 it was possible to calculate the 'yield' of itaconic-14 - 8 acid from the substrate. Though the yield was-8 _11 identical in duplicate experiments with the same-14 -12 substrate, mycelia harvested at different times

-1 showed large variations. This effect has been-12 -11 attributed largely to the extraordinary sensitivity of-14 -8 the mycelium to slight variations in the working

- conditions during harvesting. In experiments on- 4 itaconic acid production from glucose by mycelial

-10 - suspensions, yields of 0-65% (mole for mole) have0 been obtained; usually the molar yield was about2 40% (i.e. about 29% by weight). In calculating

-34 these figures itaconic acid formed by endogenouscted from all figures. activity was subtracted.

Table 3. Formation of itaconic acid by non-proliferating mycelia

Mycelia suspended in tap water solution of pH 2-1. Expt. no. 1 lasted for 10 hr., Expt. no. 2 for 9 hr. and Expt. no. 3for 10 hr. ____! _ _- r s I

SubstrateNoneD-GlucoseGlycerolGlyceric acidGlycollic acidEthanolCitric acidcis-Aconitic acidDL-Malic acidNoneD-GlucosePyruvic acidOxaloacetic acidPyruvic acid +oxaloacetic acid

DL-Citramalic acidNoneD-GlucoseMaltoseAcetic acid

Substrateadded(mg.)

140415340400375140140280

145190140190140280

14014050

Itaconic acid formed (mg.)t A-

Total minusTotal endogenous1852 3433 1530 120 -1833 1543 2518 034 161451799

132759474

37-7-5-5

1

3220

-23

Expt.no.1

2

3

Vol. 6o 141

K. E. EIMHJELLEN AND H. LARSENMaltose, which gave rise to good growth but no

itaconic acid in the growth experiments, was alsoreadily utilized by non-proliferating mycelia grownin an acid glucose medium. Concomitantly with theutilization of maltose itaconic acid was produced ata rate well above that found in the absence of sub-strate (Table 3). It thus seems probable thatmaltose can be converted into itaconic acid by A.terreue.

molar yield of itaconic acid than did glucose, butoccasionally the yield was as high or even higher.Table 4 shows the results of an experiment in whichcitric acid and glucose were tested separately and ina mixture. In this particular case the yield ofitaconic acid from citric acid was higher than thatfrom glucose. It should be noted that when citricacid and glucose were tested together, no citric acidwas consumed by the mycelium, whereas glucose

Table 4. Forrnation of itaconic acid in the pre8ence of gluco8e and citric acidNon-proliferating mycelia suspended in tap-water solution of pH 2-1. Added at start: tube 1: no addition; tube 2:

705,umoles glucose; tube 3: 775Amoles citric acid; tube 4: 705!Lmoles glucose +775jumoles citric acid. Figures belowrefer to changes taking place in the tubes between fourth and tenth hr. after addition of substrates.

Glucose Citric acid Itaconic acid MolarTube utilized utilized formed yield*no. Substrate (I&moles) (Itmoles) (gsmoles) (%)1 None - 102 Glucose 390 165 403 Citric acid 435 225 494 Glucose + citric acid 395 0 160 38

* Assuming that one mole of substrate can theoretically give rise to one mole of itaconic acid. Itaconic acid formed byendogenous activity has been subtracted.

Table 5. Formation of itaconic acid in the presence of gluco8e and cis-aconitic acidNon-proliferating mycelia suspended in tap-water solution of pH 2-1. Added at start: tube 1: no addition; tube 2:

775!tmoles glucose; tube 3: 805 pmoles cu-aconitic acid; tube 4: 775 jLmoles glucose +805 .&moles c8-aconitic acid.Figures below refer to changes taking place in the tubes between fourth and ninth hr. after addition of substrates.

Glucose c68-Aconitic Itaconic acid MolarTube utilized acid utilized formed yield*no. Substrate (I&moles) (!moles) (jmoles) (%1 None 252 Glucose 435 130 243 c68-Aconitic acid 0 254 Glucose +cis-aconitic acid 465 60 250 43

* Assuming that one mole of substrate can theoretically give rise to one mole of itaconic acid. Itaconic acid formed byendogenous activity has been subtracted.

Table 3 shows that the substrates tested can bedivided into three classes: (1) Substrates causing aclear-cut increase in itaconic acid production, viz.glucose, maltose, glycerol, ethanol, glyceric, citricand malic acids. (2) Substrates which neither causean increase in production of itaconic acid nor aninhibition of its endogenous formation, viz. ci8-aconitic and citramalic acids. (3) Substrates in-hibiting the endogenous formation of itaconicacid, viz. pyruvic, acetic, oxaloacetic and glycollicacids.

Formation of itaconic acid in the presence of citricacid. The addition of citric acid resulted in a sub-stantial accumulation of itaconic acid by non-proliferating mycelia (Table 3). Citric acid wasutilized at a rate about equal to that of glucose. Inseveral experiments with citric acid as substrateyields of itaconic acid ranging from nil to about50% (mole for mole) have been found. When testedsimultaneously citric acid generally gave a lower

was converted into itaconic acid as if no citric acidwas present. Other experiments ofsimilar type sub-stantiated this latter finding and showed that theattack on citric acid did not start until all glucosehad been consumed.

Formation of itaconic acid in the presence of cis-aconitic acid. Non-proliferating mycelia, preparedso as to test their itaconic acid-producing activity,attacked cw-aconitic acid either very slowly or notat all. Although in some experiments a significantamount of cis-aconitic acid was metabolized, noitaconic acid was produced over and above thatformed by endogenous activity. But when amixture of ci8-aconitic acid and glucose was added,an amount of itaconic acid accumulated which wasfar greater than that accumulating in the presenceofglucose alone (Table 5). However, the glucose wasutilized at the same rate in the two cases, and onlyvery little cis-aconitic acid disappeared from themixture.

142 I955

ITACONIC ACID FORMATION BY A. TERREUS. 2

Formation ofitaconic acid in the presence ofpyruvicacid. Although pyruvic acid was rapidly meta-bolized by non-proliferating mycelia, the amount ofitaconic acid formed was less than that produced bythe control without substrate (Table 3). When amixture of glucose and pyruvic acid was added, noglucose was consumed until all the pyruvic acid hadbeen utilized. The amount of itaconic acid formed inthe presence of such a mixture was not only smallerthan that formed in the presence of glucose alone,but considerably smaller than that formed endo-genously (Table 6).

Table 6. Formation of itaconic acid in the pre8enceof gluco8e and pyruvic acid

Non-proliferating mycelia suspended in tap-watersolution of pH 2-1. Added at start: tube 1: no addition;tube 2: 780,umoles glucose; tube 3: 2270 zmoles pyruvicacid; tube 4: 780pmoles glucose +2270 tmoles pyruvicacid. Figures below refer to changes taking place in thetubes during the first 9 hr. of incubation.

Tubeno.1234

SubstrateNoneGlucosePyruvic acidGlycose +pyruvic acid

Glucoseutilized(Imoles)

725

0

Pyruvicacid

utilized(jAmoles)

11701010

Itaconicacid

formed(jumoles)

1352657042

Inhibition of metaboli8m by fluoroacetate. Table 7shows the results of an experiment on the influenceof sodium monofluoroacetate on itaconic acidformation by non-proliferating mycelia in thepresence of glucose. In the presence of a concentra-tion of fluoroacetate of 10-"M or higher, the yield ofitaconic acid was greater than that found in theabsence of fluoroacetate.The values for the yield of itaconic acid (Table 7,

column 5) have been calculated on the assumptionthat fluoroacetate has no significant influence onthe rate of endogenous itaconic acid formation.This is only true in part, since separate experimentshave shown that at fluoroacetate concentrationsgreater than 10-8M, the rate of endogenous itaconicacid formation is lowered; consequently the corre-sponding values for '% molar yield' in Table 7 areprobably too low. On the other hand, separateexperiments on the rate of endogenous itaconic acidformation in the presence of 10-4M and 2 x 10-4Mfluoroacetate have occasionally shown a smallincrease in the rate over and above that in theabsence of fluoroacetate. However, in the lattercase the effect could only be shown during the first2-3 hr. of the experiment, when the endogenousactivity was greatest. At concentrations below10-8M it therefore seemed permissible to disregard

the influence of fluoroacetate on the endogenousactivity in experiments of the type reported inTable 7, in which the conversions taking placebetween the second and the sixth hour have beenused as a basis for the calculations.

Table 7. Formation of itaconic acidfrom glucosein the presence offluoroacetate

Non-proliferating mycelia suspended in tap-watersolution of pH 2-1. Added at start 460umoles glucose andvarying amounts of sodium monofluoroacetate. Figuresbelow refer to changes taking place between second andsixth hr. Endogenous production of itaconic acid in theabsence of fluoroacetate amounted to 30jumoles, and hasbeen subtracted from the values for itaconic acid producedin the presence of glucose given below.

Fluoro-acetate

(M)0

2 x 10-510-'

2 x 10-45 x 1o-,2 x 10-85 x 10-s

10-2

Glucoseutilized(I.moles)23023520518516516012075

ItaconicInhibition acid pro-of glucose duced fromutilization* glucose

0) (mol5)0 1050 105

11 12520 11028 10530 10048 7067 40

Molaryieldt(%)464561596463t58153$

* Inhibition of utilization rate in % of rate in theabsence of fluoroacetate.

t Assuming that one mole of glucose can theoreticallygive rise to one mole of itaconic acid.

$ Values possibly too low, see text.

Several experiments have been carried out to seewhether citric acid accumulated when suspensionsof non-proliferating mycelia metabolized glucose inthe presence of fluoroacetate. The results of theseexperiments were equivocal. In three out of a totalof five experiments, in which concentrations offluoroacetate of 2 x 10-4M and 5 x 1O-'M were used,no citric acid accumulated over and above the smallamount formed in the absence of fluoroacetate. Theamount detected accounted for about 1% (mole formole) of the glucose consumed. However, thepresence of fluoroacetate caused a clear-cut increasein the yield of itaconic acid in each case. In two ofthe five experiments a significant amount of citricacid accumulated over and above that found in theabsence of fluoroacetate. In the presence of5 x 10-4'M fluoroacetate the amount of citric acidformed accounted for 3-1 % (mole for mole) of theglucose utilized; in the absence of fluoroacetatecitric acid similarly accounted for 1-5%. Thecorresponding figures for itaconic acid formationwere 46% (mole for mole) and 34% (mole for mole),respectively. The fifth experiment, in which a con-centration of 2 x 1O-"M fluoroacetate was used, gavesimilar values.

Vol. 6o 143

K. E. EIMHJELLEN AND H. LARSEN '

Table 8 shows the results of an experiment on theinfluence of fluoroacetate on itaconic acid formationwith citric acid as substrate. Fluoroacetate causeda pronounced increase in the yield of itaconic acidfrom citrate. It should also be noted that fluoro-acetate was a much more powerful inhibitor of citricacid utilization (Table 8, column 3) than of glucoseutilization (Table 7, column 3).

In quantitative experiments with non-prolifer-ating mycelia, the amounts of glucose utilized anditaconic acid, pyruvic acid and acetaldehyde pro-duced in both the presence and absence of sodiumarsenite were estimated. In the absence of sodiumarsenite glucose was converted into itaconic acidwith a yield of 40% (mole for mole). No acetalde-hyde and only a faint trace of pyruvic acid was

Table 8. Formation of itaconic acidfrom citric acid in the presence offluoroacetateNon-proliferating mycelia suspended in tap-water solution of pH 2-1. Added at start 730 jmoles citric acid and

varying amounts of sodium monofluoroacetate. Figures below refer to changes taking place between fourth and eighth hr.Itaconic acid produced (,moles)

Inhibition of , Apparent itaconicFluoro- Citric acid citric acid (a) In the (b) In the acid production Molaracetate utilized utilization* absence of presence of from citric acid, yieldt

(M) (,umoles) (%) citric acid citric acid b - a (j.moles) (%)0 400 0 20 90 70 185 x10-5 195 51 20 90 70 362x 10-4 95 76 20 80 60 63

10-8 0 100 10 10 0* Inhibition of utilization rate in % of rate in the absence of fluoroacetate.t Assuming that one mole of citric acid can theoretically give rise to one mole of itaconic acid.

Table 9. Metabolism of glucose in the presence of arseniteNon-proliferating mycelia suspended in tap-water solution of pH 2-1. Added at start 800,umoles glucose and sodium

arsenite to a final concentration of 10-2m. Figures below refer to changes taking place between third and twelfth hr.,during which 230jumoles glucose were consumed.

ProductsPyruvic acidAcetaldehydeTheoretical recovery

Amount products formed(1,&moles)

, ~~~A(a) In the (b) In theabsence of presence ofglucose glucose

80 42530 120

Apparentproduction

from glucose,b - a (jmoles)

34590

Productionin % of

theoretical*752095

* Assuming that one moleof glucose can theoreticallygive rise to two molesof pyruvic acid or two moles ofacetaldehyde.

Inhibition of metabolism by arsenite. Preliminaryexperiments on the effect of sodium arsenite on theutilization of glucose by non-proliferating myceliashowed that substantial amounts of pyruvic acidaccumlated in the reaction tubes. The pyruvic acid2:4-dinitrophenylhydrazone had m.p. 2120; themixed m.p. with an authentic sample (m.p. 2110)was 211-5' (all m.p.'s are uncorrected). The RFvalues of the new and the authentic samples of the2:4-dinitrophenylhydrazones were identical.

It was also noticed that the air coming from thetubes to which sodium arsenite had been addedsmelt of acetaldehyde. On passing the air throughan acid solution of 2:4-dinitrophenylhydrazinea copious precipitate was formed, which, afterrecrystallization, had m.p. 1500. The mixed m.p.with an authentic sample of acetaldehyde 2:4-dinitrophenylhydrazone (m.p. 1520) was 1490. TheR. values of the new and the authentic sampleswere identical.

detected; the latter compound accounted for lessthan 1 % (w/w) of the glucose consumed. Althoughin the presence of 10-2M sodium arsenite the rate ofglucose utilization was lowered to 30% ofthe rate inthe absence of arsenite, it remained constant duringthe 12 hr. of the experiment. No itaconic acidaccumulated in the reaction mixture. However,practically all glucose consumed could be accountedfor, on a molar basis, as pyruvic acid and acetalde-hyde. The results of the experiment are summarizedin Table 9. The figures show that substantialamounts of pyruvic acid and acetaldehyde alsoaccumulated in the absence of glucose (column 2),apparently owing to an inhibition of the endogenousmetabolism. On the assumption that the endo-genous metabolism is equally inhibited in thepresence of glucose, the figures for endogenousmetabolism have been subtracted from those ob-tained in the presence of glucose to give apparentproduction of pyruvic acid and acetaldehyde from

144 I955

ITACONIC ACID FORMATION BY A. TERREUS. 2

glucose (column 4). Since one mole of glucose can,

on the basis of the Embden-Meyerhof scheme, giverise to two moles of pyruvic acid in an aerobic fer-mentation, 75 % of the glucose utilized are ac-

counted for as pyruvic acid. Similarly, assuming byanalogy with well-known biochemical reactions thatone mole of glucose can, theoretically, give rise totwo moles of acetaldehyde, 20% of the glucoseutilized is accounted for in the latter product.Thus the amounts ofpyruvic acid and acetaldehydefound are 95% of that expected on a theoreticalbasis.

DISCUSSION

The growth experiments with various sugars andalcohols indicate that closely related compoundsdiffer markedly in their ability to act as sources ofitaconic acid (Table 1). It is especially striking thatno itaconic acid was detected in the fermentationliquor in a number of cases where good growth was

observed and where the pH was found to be withinthe range of 2-0-2-2 (Larsen & Eimhjellen, 1955).No general relation has been found between thechemical structure of the substrates and itaconicacid production, thus indicating that a cautiousinterpretation ofthe data is desirable. This is furthervalidated by the fact that maltose, which gave riseto good growth but no itaconic acid in the growthexperiments (Table 1), is readily converted intoitaconic acid by non-proliferating mycelia (Table 3).Previous investigations (Larsen & Hovden, inpreparation) have shown that the itaconic acid-producing mechanism is extremely sensitive toexternal influences, thus suggesting that the greatvariations in the yields of itaconic acid from thecompounds listed in Table 1 might in part be due totoxic trace contaminants in the preparations. Asimilar argument might be used in the cases ofno or

only slight growth, although our experience indi-cates that growth is far less sensitive to externalinfluences than itaconic acid production. On theother hand, the slight growth observed in some ofthecases might be due to contamination of the sub-strates with utilizable carbon compounds.The fact that citric and acetic acids are not con-

verted into itaconic acid, although they supportgood growth, is in agreement with the finding ofCalam et al. (1939). This phenomenon might beexplained on the basis of the finding reported in theprevious paper (Larsen & Eimhjellen, 1955) that theenzyme system producing itaconic acid is formedonly at a low pH. It seemed possible, therefore,that although the pH in the present experiment wasas low as could be tolerated by the organism whencitric or acetic acids were substrates for growth, itwas none the less not low enough to allow the forma-tion of the itaconic acid-producing system. Thishypothesis was tested by allowing non-proliferating

10

mycelia, from cultures grown in glucose medium atpH 2-1, to act on citric and acetic acids. Tables 3and 4 show that under these conditions citric acid,though apparently not acetic acid, does give rise toitaconic acid.

Kinoshita (1931) suggested that itaconic acidmight be formed from glucose by A. itaconicu8 viacitric and aconitic acids as intermediate steps, thelatter giving rise to itaconic acid by a simple de-carboxylation. The evidence behind the suggestionwas his observation that citric acid accumulated incalcium carbonate-buffered cultures, whereas ita-conic acid was the main product in acid cultures.

Corzo & Tatum (1953) have recently reportedthat in experiments with [methyl-14C]acetate, [carb-oXy-14C]acetate and asymmetrically 14C-labelledcitrate they have obtained evidence for the relation-ship of itaconic acid to the tricarboxylic acid cycleintermediates. These authors used the same strainof A. terreus as that used in the present investiga-tion. They found that itaconic acid is formed fromcitric acid by a pathway which involves loss of thecarbon atom derived from the acetate carboxyl,and the conversion ofthe carbon atom derived fromthe methyl group of acetate into the methylenecarbon of itaconic acid. It should be noted that ifcw-aconitic acid were an immediate precursor ofitaconic acid, the findings of Corzo & Tatum areincompatible with the present knowledge of theasymmetric behaviour of citric acid (Potter &Heidelberger, 1949).The results of the present investigation support

the finding of Corzo & Tatum (1953) that itaconicacid is formed from citric acid and suggest that someactivated form of citric acid is an intermediate in theconversion of glucose into itaconic acid by A.terreu8. This conclusion is based upon the followingobservations with non-proliferating mycelia: (a)citric acid can be converted into itaconic acid withan efficiency which is at least as great as that in theconversion of glucose (Table 4); (b) citric acid canaccumulate when glucose is the substrate, particu-larly in the presence of fluoroacetate; (c) in thepresence of both glucose and citric acid no citricacid isutilizeduntil all glucose is consumed (Table 4).The latter observation indicates that, when meta-bolized, these two substrates form the same inter-mediate, and that this intermediate, which ispossibly an activated form of citric acid, is morereadily formed from glucose than from citric aciditself.The experiments reported above have shown that

fluoroacetate stimulates the formation of itaconicacid both from glucose (Table 7) and, in particular,from citric acid (Table 8). In the latter case it waspossible to raise the molar yield of itaconic acid from18 to 63 %. According to present knowledge thetoxic action of fluoroacetate is due to its conversion

Bioch. 1955, 60

Vol. 6o 145

146 K. E. EIMHJELLEN AND H. LARSEN I955into fluorocitrate, which compound is a specificinhibitor of aconitase (Peters, 1952; Morrison &Peters, 1954). There has been some confusion as towhether aconitase is one or two enzymes, but mostinvestigators now seem to agree that only onecatalytic centre of the aconitase protein is con-cerned in the enzymic conversion of citrate intoiocitrate, possibly via c8-aconitate (Buchanan &Anfinsen, 1949; Martius & Lynen, 1950; Morrison,1954). If this is the case, the consequence of thepresent observations would be that citric acid is amore immediate precursor of itaconic acid than isci8-aconitic acid. Besides the experiments withfluoroacetate further support for this viewpoint isfurnished by the observation that non-proliferatingmycelia are unable to convert cw-aconitic acid intoitaconic acid (Table 3), and by the observation thatwhen a mixture of glucose and cw-aconitic acid ismetabolized the amount of itaconic acid formed isappreciably greater than the sum of the amounts ofitaconic acid formed from each ofthe two substrates(Table 5).In interpreting the latter observation the possi-

bility was considered that the presence of glucosemight stimulate the conversion of cu-aconitic acidinto itaconic acid. However, even if this were thecase, the amounts of itaconic acid formed underthese conditions were so great that by far the largerpart of it must have been formed from glucose, andwith a greater efficiency than the correspondingconversion in the absence of ci8-aconitic acid. Itcan therefore be concluded that ci8-aconitic acidaffects the metabolism of the mycelium in such away that glucose is converted into itaconic acidwith a significantly higher yield than in the absenceof c8-aconitic acid. This conclusion suggests thatc8-aconitic acid inhibits the tricarboxylic acidcycle at the stage of ci8-aconitic acid itself, thusdiverting a greater proportion of the metabolicintermediates to the pathway leading to the forma-tion of itaconic acid. Integrating the observationspresented it thus seems possible that the presence ofcw-aconitic acid promotes the conversion of acti-vated citrie acid into itaconic acid.

Thus postulating activated citric acid as a closeprecursor of itaconic acid, one might envisage itsdecarboxylation to citramalic acid followed by adehydration to form itaconic acid. The experimentsshowed, however, that DL-citramalic acid was notconverted into itaconic acid under the conditionstested (Table 3).No satisfactory explanation can be formulated at

present for the fact that pyruvic, acetic and oxalo-acetic acids, and a mixture of pyruvic and oxalo-acetic acids are not, apparently, converted intoitaconic acid and even inhibit its endogenous forma-tion (Table 3). It should be noted in this connexionthat in a mixture of glucose and pyruvic acid, the

utilization ofglucose is completely inhibited whereaspyruvic acid is consumed as if no glucose werepresent (Table 6). It might therefore be thatpyruvic, acetic and oxaloacetic acids completelyinhibit the endogenous formation of itaconic acid,and that the small but definite production ofitaconic acid observed in the presence of thesecompounds is due to their partial conversion intoitaconic acid.The fact that non-proliferating mycelia had very

little or no ability to convert pyruvic acid intoitaconic acid (Table 3), not even in the presence ofoxaloacetic acid, might also be taken as an indica-tion that pyruvic acid is not an intermediate in theconversion of glucose into itaconic acid. However,the experiments with arsenite strongly suggest thatglucose is converted into a C3 compound before theformation ofitaconic acid, and that the C. compoundis pyruvic acid or a compound closely related to it(Table 9). Arsenite is reported to inhibit oxidativedecarboxylation of a-oxo acids (Walker, 1949), thusfurther suggesting that the Ca compound is normallydecarboxylated to form 'active' acetic acid.

Acetaldehyde, the other product which accumu-lated when glucose was metabolized in the presenceof arsenite (Table 9), may be formed by the de-carboxylation ofpyruvic acid as in yeast. The latterreaction is known to be little affected by arsenite(Szent-Gyorgyi, 1930).

SUMMARY

1. Experiments with growing and non-prolifer-ating mycelia of Aspergiu8s terreu4s, strain NRRL1960, have shown that this organism is able to con-vert a number of different compounds into itaconicacid.

2. Experiments with non-proliferating mycelia,to which were added metabolic inhibitors and sub-strates singly or in mixtures, have led to thefollowing postulate: glucose is converted intoitaconic acid via a C3 stage and activated citric acid.The Cs stage is pyruvic acid or a compound closelyrelated to pyruvic acid. The results indicate thatactivated citric acid is a closer precursor to itaconicacid than is cis-aconitic acid.

The work was supported by grants from the RoyalNorwegian Councb for Scientific and Industrial Research.We are indebted to Professor N. A. Sorensen for a gift ofrare sugars and to Dr A. Nygaard for a gift of sodiumfluoroacetate.

REFERENCESBarrenscheen, H. K. & Pany, J. (1941). Die Methoden der

Fermentfor8chung, ed. Bamann, E. & Myrback, K. 1,1074, Leipzig: Georg Thieme.

Buchanan, J. M. & Anfinsen, C. B. (1949). J. biol. Chem.180, 47.

Vol. 6o ITACONIC ACID FORMATION BY A. TERREUS. 2 147

Calam, C. T., Oxford, A. E. & Raistrick, H. (1939).Biochem. J. 33, 1488.

Cavallini, D., Frontali, N. & Toschi, G. (1949). Nature,Loud., 163, 568.

Corzo, R. H. & Tatum, E. L. (1953). Fed. Proc. 12, 470.Fittig, R. & Miller, H. E. (1889). Liebigs Ann. 255, 43.Friedemann, T. E. (1950). Circular No. 16, Dept. Nutrition& Metabolism, Northwestern University Medical School,Chicago.

Kinoshita, K. (1931). Acta Phytochim., Tokyo, 5, 271.Krampitz, L. 0. & Werkman, C. H. (1941). Biochem. J. 35,

595.Krebs, H. A. & Eggleston, L. V. (1944). Biochem. J. 38,426.Lardy, H. A. (1949). Manometric Techniques and Tissue

Metabolism, pp. 163-4, ed. Umbreit, W. W., Burris, R. H.& Stauffer, J. F. Minneapolis: Burgess Publishing Co.

Larsen, H. (1954). Methods in Enzymology, 3, sect. m, ed.Colowick, S. P. & Kaplan, N. 0. New York: AcademicPress.

Larsen, H. & Eimhjellen, K. E. (1955). Biochem. J. 60,135.Malachowski, R. & Maslowski, M. (1928). Ber. dtsch. chem.

Ge8. 61, 2521.Marokwald, W. & Axelrod, S. (1899). Ber. dt8ch. chem. Ge8.

82, 713.Martius, C. & Lynen, F. (1950). Advanc. Enzymol. 10, 167.Michael, A. & Tissot, G. (1892). J. prakt. Chem. 46, 285.Morrison, J. F. (1954). Biochem. J. 56, xxxvi.Morrison, J. F. & Peters, R. A. (1954). Biochem. J. 56,

xxxvi.Nekhorocheff, J. & Wajzer, J. (1953). Bull. Soc. Chim. biol.,

Pari8, 85, 695.Ochoa, S. (1948). J. biol. Chem. 174, 133.Peters, R. A. (1952). Proc. Boy. Soc. B, 139, 143.Potter, V. R. & Heidelberger, C. (1949). Nature, Lond., 164,

180.Robertson, W. van B. (1942). Science, 96, 93.Szent-Gyorgyi, A. (1930). Biochem. J. 24, 1723.Walker, T. K. (1949). Advanc. Enzyrnol. 9, 537.

Biosynthesis of Fatty Acids in Cell-free Preparations2. SYNTHESIS OF FATTY ACIDS FROM ACETATE BY A SOLUBLE ENZYME SYSTEM

PREPARED FROM RAT MAMMARY GLAND

BY G. POPJAK* A ALISA TIETZtThe National In8titute for Medical Research, MiU Hill, London, N. W. 7

(Received 15 November 1954)

We have reported previously the synthesis of short-and long-chain fatty acids from acetate by cell-freesuspensions (homogenates) of mammary gland oflactating rats and sheep (Popj6k & Tietz, 1954a).The following conditions were required for syn-thesis in the preparations of rat mammary gland:(a) aerobic incubation, (b) the co-oxidation of anyof the three keto acids: pyruvate, oxaloacetate orx-oxoglutarate, and (c) the addition of adenosinetriphosphate (ATP), although the last was notstrictly required with pyruvate and a-oxoglutarate.The maximum synthesis of fatty acids from acetateoccurred in the presence of acetate (0-02M), oxalo-acetate (002M) and of ATP (O-O1M) and when thegas phase was air instead of pure 02-

In this article the preparation and some of theproperties of a soluble enzyme system from themammary gland of lactating rats are described. Theresults have already been presented in a preliminaryform (Popjak & Tietz, 1953, 1954b).

MATERIAL AND METHODSPreparation of the soluble enzyme system. Homogenates of

the mammary gland of lactating rats, 10-14 days afterparturition, were prepared first as described previously(Popjak & Tietz, 1954a) and were then fractionated byhigh-speed centrifuging. The homogenate, from whichwhole cells, cell debris, nuclei, etc., have been removed bypreliminary centrifuging at 400g for 10 min., was centri-fuged first at 25 000g for 30 min. and at 00. A clear,transparent pink supernatant (Sp. I) was taken off andfiltered through a small pad of cotton wool to remove a thinfilm of fat from the top. The sediment, designated as'mitochondria', was washed twice by dispersion in fresh,ice-cold buffer and centrifuging at 25 000g for 15 min.A sample of Sp. I was centrifuged further at 2-4° and at104000g for 30 min. in a Spinco preparative centrifuge. Thesupernatant (Sp. II), which in appearance was similar toSp. I, was taken off and filtered as described for Sp. I. Thesediment, a pinkish brown translucent pellet, designated as'microsomes', was washed once with fresh buffer andsedimented at 104000g for 10 min.The 'mitochondria' obtained from 5 ml. of homogenate

were suspended with the aid of a small glass homogenizer in2-5 ml. of buffer or in the same volume of Sp. I, to provide2 ml. for incubation and 0-5 ml. for determination of dryweight. The 'microsomes' obtained from 10 ml. of Sp. Iwere treated in the same way with buffer or Sp. II.

After it was found that all the enzymic activity requiredfor fatty acid synthesis was present in Sp. II, the procedurewas shortened. Homogenates were prepared in eitherphosphate buffer (0-154m-KCI, 100 parts; 0-154m-MgCl2,

* Present address: M.R.C. Experimental Radiopath-ology Research Unit, Hammersmith Hospital, DucaneRoad, London, W. 12.

t Holder of the Glaxo Laboratories and Friends of theHebrew University of Jerusalem Scholarship while workingat the National Institute for Medical Research.Paper 1 of this series: PopjAk & Tietz (1954a). Present

address: Department of Biochemistry, Hadassah MedicalSchool, Jerusalem, Israel.

10-2