JOURNAL OF BACTERIOLOGY, Feb., 1966Copyright @ 1966 American Society for Microbiology

Vol. 91, No. 2Printed in U.S.A.

Physiological Role of Tryptophanase in Control ofTryptophan Biosynthesis in Bacillus alvei

J. A. HOCH AND R. D. DEMOSSDepartment of Microbiology, University of Illinois, Urbana, Illinois

Received for publication 27 September 1965

ABSTRACTHoCH, J. A. (University of Illinois, Urbana), AND R. D. DEMoss. Physiological

role of tryptophanase in control of tryptophan biosynthesis in Bacillus alvei. J. Bac-teriol. 91:667-672. 1966.-Indole excretion occurred early in the exponential growthphase, and derived mainly from biosynthetic intermediates of tryptophan. Tryp-tophan cleavage by tryptophanase contributed about 1.5% of the indole excreted.In the presence of exogenous tryptophan (5 to 10 ,ug/ml), excretion of early indolewas not observed. Experiments with isotopically labeled indole and tryptophanshowed that a low rate of endogenous tryptophan biosynthesis occurred constantlyduring growth. Both exogenously and endogenously supplied tryptophan were de-graded by tryptophanase. As a consequence, the intracellular tryptophan concen-

tration appeared to be maintained at a constant low level. It was suggested that theaction of tryptophanase is an example of an enzymatic mechanism which controlsthe level of a specific metabolite pool.

We previously reported that Bacillus alvei has atryptophanase similar in many respects to thetryptophanase of Escherichia coli (5). The physio-logical state of the enzyme in B. alvei differs fromthat of E. coli in its noninducibility by tryptophanand nonrepressibility by catabolites. The con-stitutive level of the enzyme is considerably lowerthan the fully induced level of the enzyme in E.coli. The seeming paradox of an enzyme quan-titatively similar to the tryptophanase of E. coliwith respect to enzymatic properties, and com-pletely dissimilar with respect to physiologicalstatus, led us to investigate further the role ofthis enzyme in the tryptophan metabolism of B.alvei.

In recent times, a number of investigators havetried to correlate the function of tryptophanaseand tryptophan synthetase enzymes in the con-trol of tryptophan biosynthesis in E. coli. Freund-lich and Lichstein (4) investigated the levels ofthe two enzymes under a number of conditions.They showed that the level of one enzyme isinversely related to the level of the other. Thelevel of tryptophanase increases with higher levelsof tryptophan, while the level of tryptophansynthetase decreases. Newton and Snell (9, 10)showed that tryptophanase can catalyze the syn-thesis of tryptophan from indole and serine.Under certain conditions, tryptophanase canserve in lieu of a normal tryptophan synthetase.

This occurs if the enzyme is constitutive or is in-duced by tryptophan analogues when indole ispresent in the medium. The availability of thismechanism for tryptophan synthesis in the normalnonconstitutive organism is in doubt. Thus,Yanofsky (12) reported that, when tryptophan isthe growth supplement, mutants which lack afunctional B protein and are unable to use indolefor tryptophan biosynthesis reach a lower levelof growth than do tryptophanless mutants whichhave a functional B protein and are able to useindole. This observation was most probably dueto degradation of part of the tryptophan bytryptophanase.When B. alvei is supplied with an exogenous

source of tryptophan, indole appears in the cul-ture medium in response to the exogenous levelof tryptophan (5). Even when low levels of tryp-tophan (e.g., 5 or 10 ,g/ml) are present, indoleappears when the tryptophan is utilized from themedium. Indole is excreted until the exogenoustryptophan is depleted, at which time the indoleis rapidly reutilized by the organism. Reutiliza-tion was shown to be dependent on the functionof a normal tryptophan synthetase. If no trypto-phan is supplied to the medium, indole appearsin the culture medium but at an earlier physio-logical age. Early indole was shown to be a func-tion of tryptophan biosynthesis, generated by thefission of indole-3-glycerol phosphate (IGP).

667

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HOCH AND DEMOSS

To clarify the events leading to indole excre-tion, experiments were undertaken with iso-topically labeled tryptophan. The utilization oftryptophan, and the concomitant indole excre-tion, were investigated, and data are presentedfor a physiological role for tryptophanase in thecontrol of tryptophan biosynthesis. A later com-munication will report investigations on theformation of early indole.

MATERIALS AND METHODS

Bacteria. The bacteria used in this investigationhave been described earlier (5). All mutants were de-rived from B. alvei ATCC 6348.

Growth. The bacteria were grown overnight in 2%(w/v) Trypticase (BBL). The inoculum for this cul-ture was kept sufficiently small so that the culture wasin exponential growth when transferred. After sedi-mentation at 8,700 X g for 15 min, the cells weretaken up in saline and transferred to a mineral salts

-J00z

0

EE

6

4

2

0 1 2 3 4 5TIME, HRS

6 7 8

FIG. 1. Indole-excretion kinetics at low levels ofexog-enous tryptophan. The wild-type strain was culturedat 37 C in mineral salts-thiamine medium supplementedwith 1% acid-hydrolyzed casein. Samples were with-drawn at 30-min intervals, and the cells were sedimentedby centrifugation at 8,700 X g for 15 min. The super-

natant liquid was extracted with toluene, and indole was

determined in the toluene layer. The initial tryptophanconcentration of the medium for Fig. JA, IB, and ICwas 1, 2, and 3 pg/ml, respectively.

medium (11) supplemented with thiamine hydro-chloride (10 jug/ml) and 1% salt-free acid-hydrolyzedcasein (Nutritional Biochemicals Corp., Cleveland,Ohio).The absorbancy of cultures was measured at

660 m,u, either in a Zeiss PMQ II spectrophotometeror in an Evelyn photoelectric colorimeter.

Determination of tryptophan. Tryptophan in cul-ture supernatant fluids was digested with partiallypurified E. coli tryptophanase by the method of Frankand DeMoss (3). The indole formed was extractedinto toluene, and samples of the toluene layer wereassayed for indole (5). In isotope experiments, asample of the toluene layer was also assayed forradioactivity.

Measurement of radioactivity. Trichloroacetic acid-soluble and -insoluble counts were performed onsamples of the culture as described by Britten andMcClure (1). Indole and radioactivity in culturesupernatant fluids were determined as previouslydescribed (5).

Special chemicals. L-Tryptophan-2-C"4 was pre-pared enzymatically from indole-2-C'4 (Calbiochem)with tryptophan synthetase. L-Tryptophan-UL-IH waspurchased from Volk Radiochemical Co., Burbank,Calif. Indole-UL-H3 was prepared from L-tryptophan-UL-H3 by digestion with tryptophanase.

RESULTS

Indole excretion. A culture growing in a me-dium devoid of exogenous tryptophan excretesindole at an early physiological age (5). Indoleappears at the onset of exponential growth, risesto a maximum, and is completely reutilized be-fore maximal growth is attained. The additionof tryptophan at concentrations of 5 ug/ml orgreater completely represses the formation ofearly indole, but an indole excretion occurs laterin the growth period. Late indole is derived fromthe tryptophan supplied to the medium. Thesource of the early indole was shown by isotopeexperiments to be the fission of IGP (5).

Since 1 ,ug/ml of tryptophan did not repressearly indole formation, and 5 ,ug/ml effectedcomplete repression, it was of interest to ex-amine the indole-excretion kinetics at severallow levels of tryptophan. In Fig. 1, the indole-excretion kinetics for 1, 2, and 3 sg/ml of tryp-tophan are depicted. As the exogenous concen-tration of tryptophan was raised, there was aprogressive decrease in the amount of earlyindole produced and more of a contribution oflate indole. The late indole shoulder at 1 ,g/ml(Fig. 1A) is not usually noticed at this concen-tration, and may be due to a small amount oftryptophan brought in with the inoculum.

If early indole is a manifestation of unre-pressed tryptophan biosynthesis, and late indolethat of the repressed biosynthetic system, raisingthe exogenous tryptophan concentration to inter-

_ c

I I 1 I*

J. BAamEium.668

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TRYPTOPHANASE FUNCTION

S80 _

LUIj 00

z 800z

60:1.E

600 X

40 _ _

0 2 4 6 8TIME, HRS

FIG. 2. indole utilization by the wild-type strain. Thewild-type strain was cultured at 37 C in mineral salts-thiamine medium supplemented with 1% acid-hydro-lyzed casein and 10 jug of indole-2-C'4 per ml. Samplesof the culture were withdrawn at 30-min intervals, andthe cells were sedimented by centrifugation at 8,700 X gfor 15 min. The supernatant liquid was extracted withtoluene, and indole was determined in the toluene layer.Samples of the toluene layer were assayed for radio-activity. Indole, *; counts per minute per millimicro-mole of indole, 0.

mediate low levels (e.g., 3 Ag/ml) shows theindole-excretion kinetics for the partially re-pressed state. At the low levels of tryptophanused in this case, it is reasonable to assume thatthe mechanism for the decrease in early indoleproduction is feedback inhibition of tryptophanbiosynthesis by the supplied tryptophan ratherthan repression of the biosynthetic enzymes.The kinetics of indole excretion suggest that aparticular level of intracellular tryptophan mustbe reached before the system is in equilibrium.

Utilization of indole. Figure 2 shows the kinet-ics of indole utilization when B. alvei was trans-ferred to an environment with 10 ,ug/ml of in-dole. The indole was not utilized by the celluntil after an additional amount of biosynthe-sized indole was excreted into the culture me-dium. Furthermore, once indole utilization fromthe medium began, the isotopic specific activityof the exogenous indole decreased. It continuedto decrease during the whole period of growth,even though the indole was being utilized from themedium. Thus, before early indole reached amaximum, indole, at least exogenously suppliedindole, could not be used for tryptophan syn-thesis by the cell. The dilution of the isotopecontent of exogenous indole during utilizationindicates that endogenous indole synthesis wasoccurring during this period. Endogenous indolecould arise either from IGP fission or fromtryptophanase activity on biosynthesized trypto-phan.

10

E 8

z 6-o

2o 4

2 -

0-L4

TIME, HRS

12

FIG. 3. Indole excretion by the wild-type strain in thepresence of radioactive tryptophan. The wild-typestrain was cultured at 37 C in mineral salts-thiaminemedium supplemented with 1% acid-hydrolyzed caseinand 15 ,ug of L-tryptophan-2-CM4 per ml. Samples of theculture were withdrawn at 30-min intervals, and the cellswere sedimented by centrifugation at 8,700 X g for15 min. The supernatant liquid was extracted withtoluene, and samples of the toluene layer were assayedfor radioactivity and for indole. Tryptophan, A; in-dole, 0; toluene-extractable radioactivity per milliliterof supernatant fluid, El; counts per minute per milli-micromole of indole, 0.

Isotopic tryptophan uptake by the wild-typestrain. Figure 3 shows the tryptophan-uptake andindole-excretion kinetics when the wild-typeorganism was grown in the presence of 15 ,ug/mlof L-tryptophan-2-C'4. After about 1.6 genera-tions, detectable tryptophan utilization com-menced; concomitantly, indole appeared in theculture. The amount of indole increased until 6hr, when exogenous tryptophan disappeared fromthe culture. During the period of tryptophanuptake and indole accumulation, the specificactivity of the tryptophan in the medium re-mained at a constant value (the mean of ninedeterminations was 1,551 counts per min permymole), whereas the accumulated indole had aconstant but lower specific activity (the mean ofeight determinations was 1,431 counts per minper m/Amole). Application of t distribution tothese values indicates a high degree of significancebetween the two means (P < 0.001). A similaranalysis for a tryptophan auxotroph grown underthe same conditions (5) shows no significant dif-ference between the specific activity of the indoleaccumulated and of the tryptophan supplied(P = 0.5). The lower specific activity evident

669VOL. 91, 1966

8 XL

on

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HOCH AND DEMOSS

1a HI .e 1o

0~~~~~~~~~~~~~~~~~~~~0E~~~~~~~~~~

00~~~~~~~~~

0 2 4 6 8

TIME, HRS

FIG. 4. Indole and radioactivity in the supernatantliquid during continuous feeding of labeled tryptophan.The wild-type strain was cultured at 37 C in mineralsalts-thiamine medium supplemented with 1% acid-hydrolyzed casein. During growth, randomly labeledDL-tryptophan-HW was fed to the culture at a constantrate by means ofa pump. The tryptophan had a radio-specific activity of 6,800 counts per min per mAmole.The rate offeeding in each case was about 0.01 JUg perml per hr. Samples of the culture were withdrawn at30-min intervals, and the cells were sedimented bycentrifugation at 8,700 X g for 15 min. The super-natant liquid was extracted with toluene, and samples ofthe toluene layer were assayedfor radioactivity andforindole. The feeding period in Fig. 4A was from 2 to6 hr; in Fig. 4B, 0 to 7.5 hr. Indole, 0D; toluene-extract-able counts per minute per milliliter of supernatantfluid, A.

in the indole accumulating in the culture mediumis expected if a constant low rate of endogenousindole synthesis occurred during this period.Once the tryptophan disappeared from the

medium, the accumulated indole was reutilizedfor tryptophan synthesis. During this time, thespecific activity of the indole fell drastically,indicating the endogenous synthesis of indole at asignificantly higher rate. Thus, even though suffi-cient tryptophan is present to supply the needsof the cell, endogenous synthesis of indole ortryptophan, or both, proceeds constantly.

Contribution of tryptophanase to early indole.We have previously shown that the addition of I

pg/ml of tryptophan to the culture medium does

not affect early indole formation, whereas higheramounts (5 ,ug/ml) repress formation. Further-more, if isotopically labeled tryptophan is addedto the medium at a level of 1 ,lg/ml, none of thelabel appears in the early indole formed (5).These data ruled out the cleavage of tryptophanas a source of early indole, and suggested thatthe indole arises from fission of IGP. The ex-periment, however, did not test the possibilitythat tryptophan cleavage occurs after the levelof indole has reached a maximum, because theisotope was completely incorporated into pro-tein by that time.To assess any contribution of tryptophan

cleavage, it was necessary to keep labeled trypto-phan in the internal tryptophan pool during theperiod in question. This prerequisite was met bycontinuously feeding to the culture low concen-trations (ca. 0.01 ,Ag per ml per hr) of typtophanof high specific activity. This concentration oftryptophan would not alter the kinetics of earlyindole formation, and would allow detection ofany label in the indole. Figure 4 shows the kinet-ics of the appearance of labeled indole in thisexperiment. As was previously shown, tryptophancleavage does not contribute to indole formationduring the rising slope of indole appearance,giving strength to the argument that IGP fissionis the predominant source of early indole. Oncethe indole reached a maximum and reutilizationbegan, the cleavage of tryptophan was moreapparent. The fact that the specific activity ofthe indole increased even though it was beingreutiiized suggests that the indole was undergoingcyclic formation and reutilization.The assumption that the indole is undergoing

cyclic formation and utilization is based on thecleavage of a small percentage of the inputtryptophan to indole. The input tryptophan con-tained 6,800 counts per min per mnmole. Sincethe tryptophan was randomly labeled withtritium, the specific activity of the indole portionof the molecule is unknown. If one assumes that50% of the counts are in the indole ring, themaximum indole accumulating from the radio-active tryptophan (Fig. 4A) is 0.25 nw,nole; i.e.,exogenous tryptophan provides about 1.5%of the total indole at 6 hr.

DIscussioN

The constitutive production of tryptophanaseby B. alvei is manifest in its effect on tryptophanutilization. The addition of tryptophan, even insmall amounts, results in a subsequent excretionof indole into the culture medium. The amountof indole excreted depends on the oriinaltryptophan concentration supplied. The concen-

670 J. BAc:rEiuoL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TRYPTOPHANASE FUNCTION

tration of excreted indole represents, at most,15 to 20% of the original tryptophan suppliedwhen intermediate levels (5 to 15 ,ug/ml) areused. The addition of higher levels of tryptophanresults in indole accumulation and incompletereutilization by the cell. When low levels (1 to3 ,ug/ml) of tryptophan were supplied, the indolethat was formed was reutilized by the cell beforemaximal growth was attained.

It was desirable to identify the immediateprecursor of early indole, since indole can ariseas a consequence of IGP fission as well as oftryptophanase activity. The addition of 1 ,g/mlof labeled tryptophan to the medium did not re-

press early indole synthesis, and none of the labelappeared in the indole; this suggests that theearly indole was due to the fission of IGP, or to adifferent, unknown, pathway for indole synthe-sis. The feeding experiments, involving continu-ous addition of labeled tryptophan, pointed toIGP as the immediate precursor of indole, sincethe early indole was not significantly labeled.The feeding experiments revealed another

phenomenon, which was not recognized earlier.Even though early indole had reached a maximalvalue and was declining, the isotopic specificactivity of the indole was rising. This observationindicated that tryptophan fission was occurringsimultaneously. The fission of IGP as the mecha-nism for early indole synthesis was now furthercomplicated by the fact that indole was beingproduced by tryptophan fission when the indolewas being reutilized. The total early indole mustthen be a composite of IGP fission and trypto-phan fission, though the latter contributes only aminute fraction of the total indole.

In any case where indole was being reutilizedby the cell, the indole was diluted by endog-enously synthesized indole. While the indole was

undergoing reutilization, the isotopic specificactivity of the indole was continually undergoingdilution. Simple cyclic excretion and reutilizationof indole would not be expected to alter the spe-cific activity of the indole. Intracellular mixing oftryptophan, formed endogenously and fromexogenous indole, would effect dilution of theisotope in indole. Thus, we suggest that the ex-

'ernal indole and the internal tryptophan are

part of a cycle mediated by tryptophanase andtryptophan synthetase.The physiological significance of such a cycle

is apparent if one considers the effect of thiscycle on the internal metabolic pool of trypto-phan. The concerted action of the two enzymeswould be expected to keep the metabolic poolof tryptophan at a nearly constant size, by virtueof the affinity of the enzymes for their respectivesubstrates and the equilibrium tryptophan value

for the dual enzyme system. If the size of themetabolic pool determines the degree of inhibi-tion of tryptophan biosynthesis by feedbackinhibition, the action of this enzyme systemwould keep the rate of endogenous tryptophanbiosynthesis at a low but finite level.Tryptophanase is postulated to act in the role

of an overflow valve to keep the size of the meta-bolic pool below a certain level. The indoleformed by tryptophanase activity can be easilyshunted back to tryptophan via tryptophan syn-thetase. The concerted action of degradative andbiosynthetic enzymes in the maintenance of aspecific level of a particular metabolite may bedesignated as a "specific metabolite pool con-trol" mechanism. The general applicability of thistype of control mechanism can be envisionedwhere the product of a degradative enzyme (orenzymes) of a metabolite can be reutilized forbiosynthesis of the metabolite.The pool size maintained by this mechanism

is not sufficiently large to eliminate completelythe biosynthesis of a small amount of tryptophan.That endogenous synthesis is maintained at aconstant rate is evident from the constant butlower specific activity of the indole excreted dur-ing tryptophan uptake. A number of observationsindicate that the lower specific activity of indoleis significant. If a mutant unable to synthesizetryptophan is subjected to the same experiment,the indole that appears in the culture mediumhas the same isotopic specific activity, withinexperimental error, as that of the tryptophansupplied (5). Furthermore, an analysis of trypto-phan incorporation from the growth medium inwild-type and mutant organisms indicates thatthe wild-type organism incorporates tryptophanat a lower rate than does the mutant (Hoch andDeMoss, unpublished data). The present data indi-cate that the endogenous rate of synthesis is sig-nificant even under conditions of tryptophanexcess. After exhaustion of tryptophan, feedbackinhibition is released and endogenously synthe-sized indole appears, even though indole uptakeis occurring at a rapid rate. This situation is prob-ably due to a period of adjustment when feedbackrelease overcompensates for the decreasedtryptophan level, and the safety valve (trypto-phanase) corrects the overcompensation by ex-creting indole. When exogenous tryptophan isexhausted, the internal tryptophan pool size ismaintained at a new lower level so that endog-enous synthesis can supply the needs of the cell.The new level of tryptophan is not high enoughfor tryptophanase appreciably to effect cleavageto indole.

It is conceivably advantageous for the cell tocontinue tryptophan synthesis at a low level

671VOL. 91, 1966

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HOCH AND DEMOSS

under conditions of excess exogenous tryptophan,especially if part of the tryptophan pathwaywere used for the biosynthesis of other smallmolecules. The general aromatic pathway syn-thesizes phenylalanine, tyrosine, p-aminobenzoicacid, and p-hydroxybenzoic acid, in addition totryptophan; several of the enzymes involvedappear to be common to all products. In additionto these products, the synthesis of vitamin Kand ubiquinone has been shown to occur viathis pathway, although the precursors have notbeen established (2). The possibility exists that anintermediate of tryptophan biosynthesis is alsoan immediate precursor of one of the aromaticproducts or some other small molecule neededin small amounts. Morgan et al. (8) showed thattryptophan exerts feedback inhibition on thefirst enzyme specific to anthranilate synthesis.If the precursor for the hypothetical small mole-cule occurred subsequent to this reaction, itwould be advantageous for the cell to avoid totalfeedback inhibition. The maintenance of a smalltryptophan pool would satisfy the requirement,and can be achieved by degradation of excesstryptophan by the action of tryptophanase.The tryptophan-indole cycle of B. alvei ap-

pears to be similar to the anthranilic acid cycle inNeurospora described by Matchett and DeMoss(6, 7). The anthranilic cycle, however, is inducibleand may serve primarily in a catabolic role.Matchett (5a) found another route of tryptophandegradation in this organism, in addition to theanthranilate cycle. A constitutive tryptophantransaminase is present, which converts trypto-phan to indole-pyruvic acid. The striking simi-larity of this system to the tryptophanase systemof B. alvei is in the kinetics of indole-pyruvic acidaccumulation during tryptophan utilization.When tryptophan was utilized from the medium,the extracellular level of indole-pyruvic acid in-creased, reaching a maximal value just before allof the tryptophan has disappeared from themedium. Indole-pyruvic acid was promptly re-utilized. These kinetics correspond exactly tothose of late indole formation in B. alvei; in bothorganisms, the respective enzymes are constitu-tive.

ACKNOWLEDGMENTSThis investigation was supported by Public Health

Service research grant E-2971 from the National

Institute of Allergy and Infectious Diseases. J. A.Hoch was a trainee, supported by Public HealthService Microbiology Training Grant GM-510 fromthe Division of General Medical Sciences.

LITERATURE CrrED1. BRITTEN, R. J., AND F. T. McCLuRE. 1962. The

amino acid pool in Escherichia coli. Bacteriol.Rev. 26:292-335.

2. Cox, G. B., AND F. GIBSON. 1964. Biosynthesisof vitamin K and ubiquinone, relation to theshikimic acid pathway in Escherichia coli.Biochim. Biophys. Acta 93:204-206.-

3. FRANK, L. H., AND R. D. DEMoss. 1957. Specificenzymic method for the estimation of L-trypto-phan. Arch. Biochem. Biophys. 67:387-397.

4. FREUNDLICH, M., AND H. C. LICHSTIN. 1962.Tryptophanase-tryptophan synthetase systemsin Escherichia coli. I. Effect of tryptophan andrelated compounds. J. Bacteriol. 84:979-987.

5. HOCH, J. A., AND R. D. DEMoss. 1965. Physio-logical effects of a constitutive tryptophanasein Bacillus alvei. J. Bacteriol. 90:604-610.

5a. MATCHErT, W. H. 1965. The utilization of tryp-tophan by Neurospora. Biochim. Biophys.Acta 107:222-231.

6. MATCHETT, W. H., AND J. A. DEMoss. 1963.Direct evidence for a tryptophan-anthranilicacid cycle in Neurospora. Biochim. Biophys.Acta 71:632-642.

7. MATCHETIT, W. H., AND J. A. DEMOSS. 1964.Physiological channeling of tryptophan inNeurospora crassa. Biochim. Biophys. Acta86:91-99.

8. MORGAN, P. N., M. I. GIBSON, AND F. GIBSON.1963. The conversion of shikimic acid intocertain aromatic compounds by cell-free ex-tracts of Aerobacter aerogenes and Escherichiacoli. Biochem. J. 89:229-239.

9. NEWTON, W. A., AND E. E. SNELL. 1964. Cata-lytic properties of tryptophanase, a multifunc-tional pyridoxal phosphate enzyme. Proc. Natl.Acad. Sci. U.S. 51:382-389.

10. NEWTON, W. A., AND E. E. SNELL. 1965. Forma-tion and interrelationships of tryptophanaseand tryptophan synthetases in Escherichia coli.J. Bacteriol. 89:355-364.

11. SMITH, 0. H., AND C. YANOFSKY. 1962. Enzymesinvolved in the biosynthesis of tryptophan,pp. 794-806. In S. P. Colowick and N. 0.Kaplan [ed.], Methods in enzymology, vol. 5.Academic Press, Inc., New York.

12. YANOFSKY, C. 1960. The tryptophan synthetasesystem. Bacteriol. Rev. 24:221-245.

672 J. BACrERioL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from