JOURNAL OF BACrTROLOGY, Apr. 1973, p. 341-350Copyright i 1973 American Society for Microbiology

Vol. 114, No. 1Printed in U.SA.

Catalytic Studies on Tryptophanase fromBacillus alvei

SALLIE O'NEIL HOCH1 AND R. D. DEMOSS

Department of Microbiology, University of Illinois, Urbana, Illinois 61801

Received for publication 30 October 1972

Tryptophanase from Bacillus alvei exhibited the expected spectrum ofpyridoxal-5'-phosphate-dependent reactions. It exhibited L-serine dehydratase,S-alkyl-cysteine lyase, and cysteine desulfhydrase activities, as well as theclassic tryptophanase reactions (all beta elimination reactions). It also acted asa tryptophan synthetase (beta replacement reactions) using indole plus L-serineor L-cysteine or S-methyl-cysteine as substrates. The beta elimination reactionsare simple competitors of the replacement reactions for the same amino acidsubstrates. The kinetics of the reactions were examined in detail using a coupledcontinuous spectrophotometric assay. A product (indole) inhibition study of thebeta elimination reaction with tryptophan showed simple, noncompetitiveinhibition; the same study with allosubstrates showed noncompetitive inhibi-tion by indole. These product studies provided data on the beta replacementreactions as well. The results are discussed in terms of a mechanism for the B.alvei tryptophanase.

Newton et al. (16), using homogeneous tryp-tophanase from Escherichia coli, were able toshow that tryptophanase substrates are notrestricted to tryptophan and its phenyl-sub-stituted derivatives. They demonstrated thatthe enzyme catalyzes a variety of alpha, betaelimination (equation 1) and beta replacement(equation 2) reactions as follows:

RCH2CHNH2COOH + H20-,CH,COCOOH + RH + NH, (1)

RCH2CHNH2COOH + R'H-_R'CH2CHNH2COOH + RH (2)

R represents -SH, -SCH3, -SCH2CH,-OH, -OCH3, the indolyl-radical or its 4-, 5-,or 6-methyl derivative; R'H represents indoleor its 4-, 5- or 6-methyl derivative. Thesereactions are consistent with the general mech-anism for pyridoxal-5'-phosphate (PLP)-dependent enzymes as postulated by Braun-stein (2) and Snell (18).As a part of the comparative studies of

tryptophanase from different physiologicaltypes of bacteria, we examined a purifiedhomogeneous tryptophanase from Bacillusalvei to determine if it also exhibited the broadspecificity seen with many PLP enzymes, todetermine the kinetic patterns from these reac-

' Present addres: Department of Microbiology, ScrippsClinic and Research Foundation, La Jolla, Calif. 92037

tions, and to ascertain whether the B. alveitryptophanase followed the general reactionpostulated for E. coli. There were previousindications (10, 11) that differences in thephysical structures existed between these twotryptophanases.

MATERIALS AND METHODSEnzyme preparation. Tryptophanase was puri-

fied from B. alvei F following the modified procedureof Hoch and DeMoss (11). The procedure for prepar-ing apottyptophanase has also been described in thesame reference.

Tryptophanase assay. The purified tryptopha-nase was assayed using a modification of the methodof Pardee and Prestidge (17). The reaction mixturecontained in a volume of 0.3 ml: potassium phos-phate buffer (pH 8.0), 50 Mmol; PLP, 50 nmol; bovineserum albumin (BSA), 50 ug; and enzyme. Fourdrops of toluene were added from a Pasteur pipette,and the mixture was incubated at 37 C for 5 min. Thereaction was initiated by the addition of 0.2 ml of0.02 M L-tryptophan, and incubation was continuedat 37 C for 10 min. The reaction was terminated bythe addition of 3 ml of tube reagent (14.7 g ofp-dimethylaminobenzaldehyde in 948 ml of 95%ethanol plus 52 ml of concentrated sulfuric acid). Thecolor was read at 568 nm after 20 min. All enzymeactivities are reported as international enzyme units(1 Mmol/min). Protein concentrations were deter-mined by the method of Lowry et al. (13).Where specified, the flask assay for tryptophanase

341

HOCH AND DEMOSS

described by Hoch et al. (10), with the substitution of200 gmol of potassium phosphate buffer (pH 8.0),was used to assay tryptophanase.Dehydratase (beta elimination) assays. Pyru-

vate formation was measured by a modification ofthe method of Friedemann and Haugen (9). Thereaction mixture contained in a volume of 0.5 ml:potassium phosphate buffer (pH 7.8 or 8.0), 50 umol;PLP, 50 nmol; BSA, 50Mug; the respective amino acid;and enzyme. After 5 min at 37 C, the reaction wasinitiated by the addition of the amino acid andallowed to proceed for 5 to 15 min at 37 C. Thereaction was stopped by the addition of 0.25 ml of0.15% 2,4-dinitrophenylhydrazine in 2 N HCI. After10 min further incubation at 37 C, 1.0 ml of distilledwater was added plus 0.75 ml of 2.5 N NaOH. Thecolor was read after 10 min at 515 nm. Because thesulfhydryl groups of cysteine interfered with thiscolorimetric assay, 0.1 ml of 2.5 N NaOH was used toterminate the reaction when cysteine was the sub-strate, followed by 0.2 ml of 0.1 M iodoacetamide.The mixture was incubated for 5 min before proceed-ing with the addition of the phenylhydrazine reagent.Tryptophan synthetase (beta replacement) as-

says. The disappearance of indole was measured by amodification of the method of Yanofsky (19). Thereaction mixture contained, in a volume of 0.25 ml:potassium phosphate buffer (pH 8.0), 25 nmol; PLP,25 nmol; BSA, 25 Mg; indole, 125 nmol; the respectiveamino acid; and enzyme. After 5 min at 37 C, thereaction was initiated by the addition of the aminoacid and allowed to proceed for 5 to 10 min at 37 C.The reaction was stopped by the addition of 5 ml oftube reagent, and the color was read at 568 nm. Whencysteine was the substrate, the reaction was stoppedwith 0.1 ml of 2.5 N NaOH. Iodoacetamide wasadded as in the dehydratase assay before the additionof the tube reagent.

Continuous spectrophotometric assay system.The rate of formation of pyruvate in the tryptopha-nase and dehydratase reactions was measured in acoupled reaction with lactate dehydrogenase (typeIII, Sigma) and reduced nicotinamide adenine dinu-cleotide (NADH). The assay is a modification of thesystem used by Morino and Snell (14). Initial rateswere measured using a Gilford 2000 or 240 recordingspectrophotometer. The complete reaction mixturecontained, in a volume of 3 ml: potassium phosphatebuffer (pH 8.0), 300 Mmol; PLP, 150 to 300 nmol;BSA, 300 Mg; mercaptoethanol, 3 gmol (150 pmolwhen serine is used as a substrate); NADH, 60 nmol;lactic dehydrogenase, 1 IU; the appropriate aminoacid; and tryptophanase. All components except thelactate dehydrogenase and amino acid substrate werepreincubated for 5 min at room temperature (21-22C). The lactic dehydrogenase was then added. Theamino acid was used to initiate the reaction. Theinitial velocities represent the amount of pyruvateformed in the first minute of the reaction as meas-ured by the decrease in absorbancy of NADH at 340nm. The molar extinction coefficient of NADH is 6.22x 108 M-' cm-' (12).The reaction proceeds at a constant rate until all

the NADH is exhausted except in two instances.

When tryptophan is the substrate, the rate decreasesas the indole formed in the reaction becomes inhibi-tory. At high amino acid and low indole concentra-tions, the rate increases after the indole becomesexhausted and the competing synthetase reaction riolonger inhibits the dehydratase reaction.

Kinetic parameters. Km and vmax values weredetermined on an IBM 7094 computer using aniterative program to fit the data to an hyperbola.Analysis of the product inhibition patterns was madeusing an iterative program obtained from W. W.Cleland (6).

RESULTSTryptophan analogues as tryptophanase

substrates. When tryptophanase was first pu-rified, it was examined for its ability to cata-lyze the degradation of substrates other thantryptophan and the synthesis of tryptophanfrom indole and various amino acids. The onlyactivity observed was that of serine dehydra-tase (10). These results indicated that the B.alvei tryptophanase had a very narrow specific-ity in contrast to the results found with the E.coli tryptophanase (16). To reexamine thisquestion we modified the colorimetric assaysfor the tryptophanase, dehydratase, and syn-thetase reactions to increase their sensitivity.Activity was detected with several tryptophananalogues (Table 1). It is apparent from theseresults that the B. alvei tryptophanase acts onsome indole-substituted tryptophans. It isinactive with side chain-substituted trypto-phan analogues. The activity detected for eachof these analogues relative to tryptophan is notnecessarily indicative of the maximum activityattainable with that substrate. To establishthat value, it would be necessary to estimatethe initial reaction rate as a function of vaiyinganalogue concentration.Subsidiary reactions. Tryptophanase was

also shown to catalyze a variety of subsidiaryreactions, namely beta elimination reactionswith serine, cysteine, and S-methyl-cysteine;and beta replacement reactions using indoleplus serine, cysteine, or S-methyl-cysteine toform tryptophan. The Michaelis-Menten con-stants with standard deviations for the aminoacid substrate in these reactions are shown inTable 2. The relative activities were deter-mined from the vmax values. To avoid confusionin the discussion of these activities, serinedehydratase, S-alkyl-cysteine lyase, and cys-teine desulfhydrase will all be referred to by thecommon title of dehydratase. The Km for theamino acid substrate in the synthetase reactionis an "apparent" value because it was deter-mined only in the presence of a large excess ofindole (0.5 mM). Likewise the cysteine Km for

342 J. BACTERIOL.

TRYPTOPHANASE FROM B. ALVEI

TABLE 1. Thyptophanase activity relative totryptophan analogues"

Final ActivitySubstrate relative to

(mM) tryptophan

Tryptophan .......... 5.0 1004-Methyl-tryptophan. 5.0 525-Methyl-tryptophan 5.0 226-Methyl-tryptophan 2.5 445-Hydroxy-tryptophan 5.0 6Tryptamine .......... 5.0 0Indole-acetate ....... 5.0 0Indole-acrylate ....... 5.0 0Indole-lactate ........ 5.0 0Indole-pyruvate ...... 2.5 0Indole-propionate .... 2.5 0D-Tryptophan ........ 10.0 0

aConditions: Activities except for D-tryptophanwere measured using the tube assay with 0.1 Mbicine buffer (pH 8.2). The substrates were at ap-proximately pH 7.0. Activity with D-tryptophan (pH8.0) was measured using the tube assay with 0.1 Mpotassium phosphate buffer (pH 8.0).

the dehydratase reaction is an "apparent"value because it was obtained by extrapolation.The Lineweaver-Burk plot for tryptophan syn-thetase with cysteine as substrate is shown inFig. 1. The level of substrate required formaximal activity of this synthetase reaction isin fact inhibitory to the dehydratase reaction asseen in Fig. 1.The Km values are essentially constant for

the same amino acid substrate in both thedehydratase and synthetase reactions underthe assay conditions employed. It must beborne in mind that it cannot be definitelyconcluded from these values that the in vivo

affinities of the enzyme are necessarily con-stant. Km is more properly regarded as anoperational term which is equal to the sub-strate concentration at which the initial rate ishalf-maximal under the specified assay condi-tions. It does not directly reflect the affinity ofthe enzyme for the substrate. We had previ-ously reported (S. 0. Hoch, and R. D. DeMoss,Bacteriol. Proc., p. 118, 1968) an apparentsignificant difference in the Km values for thesame amino acid in the dehydratase and syn-thetase reactions. This difference is more prop-erly attributed to the differences in the buffersystems used, namely potassium phosphate forthe dehydratase reaction and bicine for thesynthetase reaction.The values obtained using apotryptophanase

are included as evidence that the method ofpreparation of the apoenzyme did not appear tomodify the active site affinity for substrate.The following amino acids do not exhibit

detectable activity as substrates for either thedehydratase or synthetase reactions: cystathio-nine, homocysteine, homoserine, methionine,threonine, and tyrosine. Each was tested at afinal concentration of 5 to 10 mM.The Km reported here for tryptophan (0.273

mM), using the more sensitive tube assay,agrees with the value (0.272 mM) reported byHoch et al. (10) using the flask assay.The Michaelis-Menten constants for the

amino acid substrates for the dehydratase andsynthetase reactions appear to be equal withinexperimental error. Indole, the second sub-strate for the synthetase reaction, acts as apotent inhibitor of the dehydratase reaction. Ifthe two reactions share a common catalyticsite, then the apparent effect of indole as an

TABLE 2. Michaelis-Menten constants for the tryptophanase subsidiary reactionsa

Substrate Reaction K. (M) Activity relative totryptophan (%)

L-serine Serine dehydratase (1.29 + 0.11) x 10-1 15Tryptophan synthetase (1.44 + 0.21) x 10- 11

(1.77 X 0.18) x 10-lb

S-methyl-L-cysteine S-alkyl cysteine lyase (1.17 ± 0.10) x 10-2 11Tryptophan synthetase (1.52 + 0.12) x 10-2 16

(1.30 4 0.17) x 10 2b

L-Cysteine Cysteine desulfhydrase (0.90 4 0.12) x 10-3 15Tryptophan synthetase (1.52 + 0.33) x 10-3b 11

L-Tryptophan Tryptophanase (2.73 ± 0.25) x 10-4 100

a Conditions: All assays were in 0.1 M potassium phosphate (pH 7.8-8.0), containing 0.1 mg of BSA per mland 0.1 mM PLP as described under Materials and Methods.

b Apotryptophanase was used in the assay.

VOL. 114, 1973 343

HOCH AND DEMOSS

0 .2 .4 .6 .8 1.0 1.2(CYSTEIWE, MuT4

FIG. 1. Lineweaver-Burk plots for B. alvei trypto-phanase with cysteine as substrate. Initial reactionrates were measured for the subsidiary reactions,tryptophan synthetase (A), and cysteine desulfhy-drase (B). The assays were in 0.1 M potassiumphosphate buffer (pH 8.0) containing 0.1 mg of BSAper ml and 0.1 mM PLP.

inhibitor may be due solely to the diversion ofthe amino acid to the competing synthetasereaction. This postulate was tested by observ-ing the dehydratase reaction in the presence ofvarious levels of indole. The dehydratase reac-

tion was assayed by measuring pyruvate forma-tion. The synthetase reaction was assayed bymeasuring the disappearance of indole as a

measure of tryptophan formation. Tryptophancould not be measured directly by the usualenzymatic assay of Frank and DeMoss (8)because of the presence of saturating levels ofthe respective amino acids for the dehydratasereaction. It is also because of the competitiveeffect of these amino acids that the possible

degradation of the tryptophan formed by thesynthetase reaction can be regarded as negligi-ble. The results for the serine and S-methyl-cysteine dehydratase reactions are shown inTable 3. In each instance, the decrease inpyruvate formation, in response to added in-dole, is approximately equal to the increase intryptophan formation, i.e., the total productsremain constant. Therefore the dehydrataseand synthetase reactions appear to be simplecompetitors for the same amino acid substratesand for the same intermediate enzyme-sub-strate complex.Cation requirement for tryptophanase.

Hoch et al. (10) reported that tryptophanaseshowed an absolute requirement for eitherammonium or potassium ions after the enzymewas desalted and assayed with various mono-valent cations in the presence of Tris-hydro-chloride buffer. Both ions produced the sameVrax, although the ammonium ion was a moreefficient activator with an apparent Km of 3.7mM as opposed to 22 mM for the potassiumion. However, Tris-hydrochloride buffer hasbeen reported to inhibit various univalent ca-tion activated enzymes (1). Therefore the saltrequirement was again examined but this timein the presence of imidazole buffer. Again, bothpotassium and ammonium ions activated theenzyme, as seen in Fig. 2, but the apparent Kmfor the ammonium ions is 0.31 mM as opposed

TABLE 3. Indole inhibition of the dehydratasereactiona

S-methyl-cysteine Senne dehydrataselyase

Indole added(nmol) Pyruvate Indole Pyruvate Indole

formed used formed used(nmol) (nmol) (nmol) (nmol)

0 162 6650 107 46 34 31100 56 93 21 42250 12 160 10 63500 6 163 8 61

a Conditions: The dehydratase reactions were runin 0.1 M potassium phosphate buffer (pH 8.0)containing 0.1 mg of BSA per ml and 0.1 mM PLP,and either 50 Amol of S-methyl-L-cysteine or 100;&mol of L-serine in quadruplicate for each indoleconcentration. After 10 min, two of the tubes wereused to assay the formation of pyruvate in the usualmanner. The reaction in the other two tubes wasterminated by immersion in a boiling water bath for4 min. The unused indole was extracted into 2 ml oftoluene. One milliliter of this layer was assayed forindole in the usual manner. Each value for pyruvateformed and indole used is the mean of the duplicatetube assays.

A 250 0

- ~~~~200_

I-

/50. . . . . . .

1 .2 .4 6 .8 1.0 1.2(CYSTEINE, mMr,

B100-

uj) 80

- ~~~~~Z)60i,<I.- -

>,-4

l'4, /, -, 20

I____ ______ I I II

344 J. BACTFJIOL.

TRYPTOPHANASE FROM B. ALVEI

to 10 mM for the potassium ion. Moreover, theammonium ion produced a higher maximalrate than did the potassium ion.Continuous assays. Because of the limita-

tions inherent in any fixed-time, colorimetricassay for a kinetic study, we examined thereactions in more detail using a continuousspectrophotometric assay. The continuousassay couples pyruvate formation to NADHoxidation by using lactic dehydrogenase.

Indole is an inhibitor of the tryptophanaseand dehydratase reactions because of the com-peting synthetase reactions as shown in aprevious section. Therefore, the concentrationsof both indole and the amino acids were variedin a range encompassing their respective Kmvalues in a product inhibition study to investi-gate the change in catalytic activity connectedwith the binding of indole.The reaction rate of the continuous assay

increased for the first 4 to 5 min in either bicineor potassium phosphate buffers supplementedwith BSA and PLP. The rate could be stabil-ized in potassium phosphate by: (i) adding 1mM mercaptoethanol to the buffer for thetryptophanase and S-methyl-cysteine dehydra-tase reactions, and by adding 50 mM mercap-toethanol for the serine dehydratase reaction;(ii) preincubating the tryptophanase with thecomplete mixture minus substrate and lacticdehydrogenase. The cysteine dehydratase reac-tion could not be assayed because cysteineinhibited lactic dehydrogenase. The reactionsinvolved in this inhibition have been reportedby Dugaiczyk et al. (7).For simplicity, we will use the same ab-

breviations and kinetic constants proposed byMorino and Snell (4) in their analyses of thetryptophanase reaction in E. coli. E is trypto-phanase; T. tryptophan; ET, the enzyme-tryp-tophan complex; EA, the tryptophanase-aminoacrylate-complex; I, indole; C, an al-losubstrate (i.e., an amino acid other thantryptophan such as serine, methyl-cysteine,etc.); Kc, the Michaelis constant for the al-losubstrate.Beta elimination reaction of tryptophan

(equation 1). The reaction is postulated asfollows:

E + T ETETk1 k-2

EA k E + NH3 + pyruvate (3)

The rate equation for the tryptophanase reac-tion derived by Morino and Snell (14) can beexpressed:

E. 1 KT\VI Vmax T,

1 (I 0[.KT+

VmaxK(Ti)k1 KT)

k5 TJ(4)

where E. is the total enzyme concentration; viand vmax are the initial and maximal velocityrates; T and I, the tryptophan and indoleconcentrations; KT, the Michaelis constant fortryptophan; KI, the Michaelis constant forindole as the inhibitor. Vmax, KT and K, aredefined as follows:

k2k5Vmax =

k2 + k.k-l + k2

k2kil(k-l + k)2)

KT = 1+k1(ka + k,)

The initial rate for the tryptophanase reac-tion was measured at varying tryptophan andindole concentrations. Using the data shown inFig. 3, the Michaelis constants were calculatedusing a computer program supplied by W. W.Cleland, In two separate experiments, KT wascalculated to be (4.41 ± 0.61) x 10-4 M and(8.68 1.46) x 10-i M and K, to be (8.62 4

1.33) x 10-1 M and (8.16 4 1.68) x 10-' M.The results indicate that the data fit a pattemof simple noncompetitive inhibition; i.e., thefamily of curves intersect at a common point onthe abscissa in the left quadrant (4, 5). This isin contrast to the pattem seen with the E. colitryptophanase which is that of nonsimple,noncompetitive inhibition, i.e., the family of

C NHCFmM b, NH4CI, mMM0 20 40 60 80 100 120 600

NaCI and KCI, mM

FIG. 2. Cation activation of the tryptophanasereactions. Purified tryptophanase was freed of ca-tions by filtration through a Sephadex G-10 column(1 by 19 cm) equilibrated in 0.01 M imidazole buffer(pH 7.0). The flask reaction mixture contained: 100Mmol of imidazole, (pH 7.8); 200 nmol ofPLP; 200 ,gof BSA; and enzyme. The monovalent cations wereadded as the corresponding chloride. The reactionmixture was incubated for 10 min at 37 C before theaddition of tryptophan to initiate the reaction.

VOL. 114, 1973 345

HOCH AND DzMOSS

(TRYPTOPHAN, mM)P'

'

_600z

X400K

- 200.

LI I

0 .005 .01 015 .02INDOLE, mM

FIG. 3. Pyruvate formation from L-tryptophan bythe tryptophanase reaction and its inhibition byindole. A, The curves are double reciprocal plots ofinitial velocity with respect to tryptophan concentra-tion at the fixed concentrations of indole indicated oneach line. B, The curves are double reciprocal plots ofinitial velocity with respect to indole concentrationat the fixed levels of tryptophan indicated on eachline. C, The curve is a replot ofvmaxI from Fig. 3A asa function of indole concentration.

iurves intersect at a common point above theabscissa in the left quadrant.

If k.1 = k,, equation 4 reduces to:

Et=_(1+ )(1+ ( )) ~(5)VI vma (T) K,

Equation 5 describes the requirements for sim-ple noncompetitive inhibition in accordancewith our experimental results. Equation 4 alsorequires that a plot of Edt', as a function ofindole concentration with tryptophan as theindependent variable yield a second series ofstraight lines with a common intercept in theleft-hand, quadrant located at a distance Kk./k., from the ordinate, as seen in Fig. 3B. Valuesfor K1k/k. were calculated to be 8.63 x 10-6Mand 10.03 x 10-1 M. Thus, the value of k./k1 isapproximately unity, satisfying the require-ments for the validity of equation (11).

It should be noted that the K.m, as alreadydefined,. is only the concentration of substrateat which the reaction rate is half-maximal for aparticular assay system. Thus, since the colori-metric and continuous assay systems use dif-ferent reaction mixtures and reaction tempera-tures, the Km values reported for a particularamino acid substrate from the two differentassay systems will not be necessarily identical.Beta elimination reaction of substrates

other than tryptophan. Unlike the situationfound for E. coli in which indole is also anoncompetitive inhibitor of the allosubstratebeta-elimination reactions, indole is an uncom-petitive inhibitor (4, 5) of these reactions in B.alvei. Again the pattern was determined usingthe computer program of W. W. Cleland. Theresults for the S-methyl-cysteine dehydrataseare shown in Fig. 4. In two separate experi-ments, the Kc for S-methyl-cysteine was cal-culated to be (7.55 ± 1.05) x 10- ' M and (10.40± 0.66) x 10-8 M, and K, to be (7.57 i 0.74) x10- M and (6.98 + 0.34) x 10- M. The resultsfor seine dehydratase are shown in Fig. 5. TheK. for serine was calculated to be (6.12 i 0.41)x 102MandK,tobe(15.48 ±0.34) x 10-1M.

DISCUSSIONA reexamination of the subsidiary reactions

catalyzed by B. alvei tryptophanase was under-taken using more sensitive colorimetric assays.Tryptophanase follows the classic pattern es-tablished in studies with the enzyme from E.coli and is active with tryptophan analoguessubstituted on the indole moiety, but not withanalogues substituted in the three carbon sidechains. Tryptophanase also catalyzes the beta

J. BAennoL.346

TRYPTOPHANASE FROM B. ALVEI

elimination reactions of serine dehydratase,S-alkyl-cysteine lyase and cysteine desulfhy-drase, and the beta replacement reaction oftryptophan synthetase using indole plus serine,S-methyl-cysteine, or cysteine as substrates.The Michaelis-Menten constants for eachamino acid substrate were equivalent for thedehydratase and synthetase reactions. More-over, the two reactions appeared to be simplecompetitors for an intermediate enzyme-sub-strate complex and, from the calculation ofv..., also appeared to operate at approxi-mately the same rates.More detailed kinetic experiments were car-

ried out to elucidate the mechanism of thetryptophanase-catalyzed reactions. The majortryptophanase reaction is the beta eliminationreaction in which tryptophan is the substrate.Indole is a simple, noncompetitive inhibitor ofthis reaction as shown by a product inhibitionstudy where the inhibition pattern was deter-mined by computer analysis. Equation 5, aspostulated in Results, satisfies the experimen-tal results, namely that:

(5)

This equation is of course based on the assump-tion that k., = k5. Although the E. coli workersalso concluded that k-l = k5, they in factreported a pattern of nonsimple, noncompeti-tive inhibition which does not fit equation 5.Thus it is here that we see the first importantdifference between the B. alvei and E. colienzymes. Although we conclude that the "gen-eralized" scheme for the tryptophanase reac-tions as postulated for E. coli holds for B. alvei,the basic equation describing these reactions isnot the same for the two organisms. And it isthis equation, rather than the scheme, which ismore important mechanistically in describingthe reactions involved.The subsidiary beta elimination reactions

using serine or methyl-cysteine as substratewere also examined by an indole inhibitionstudy; the pattern here was that of uncompeti-tive inhibition. Despite the disparity in inhibi-tion patterns for the beta elimination reactions,the replots of slopes and intercepts are alllinear, indicating that amino acid substratesbind in each case to just one enzyme form. Thetwo inhibition patterns indicate, however, thatthe amino acid substrate and indole react withdifferent enzyme forms. The noncompetitiveinhibition pattern indicates that the additionof tryptophan and indole to the enzyme is

separated by at least one reversible step. Theuncompetitive inhibition pattern indicatesthat the addition of serine or S-methyl-cysteineand indole to the enzyme is separated by anirreversible step. Cleland (5) defines an irre-versible step as the release of a product at zeroconcentration (not at finite concentration), oraddition of a substrate at infinite concentration(saturation). Such results are not incompatible

-A 600INDOLE, mM

L 0Ko ioK02

01

_/ A<X~~~~~~~0 10 .20 .30 40 50

(S-METHYL-CYSTEINE, mM) -'

400-

200

O .005 01 as5 02INDOLE, m M

FIG. 4. Pyruvate formation from S-methyl- L-cys-teine by the dehydratase reaction and its inhibitionby indole. A, the curves are double reciprocal plots ofinitial velocity with respect to S-methyl-cysteineconcentration at the fixed concentrations of indoleindicated on each line. B, The curves are doublereciprocal plots of initial velocity with respect toindole at the fixed levels of S-methyl cysteine in-dicated on each line. C, The curve is a replot ofVmax-1 from Fig. 4A against indole concentration.

600

S-ME-CYS, mM7' ~~~~~~3

- ~~~~~'-400-M ~~~~~6

[ 2~~~~00

INDOLE, mM

VOL. 114, 1973 347

& 1 KT IVI Vmax T K,

HOCH AND DEMOSS

(SERINE mM) 1

B400

.) SER, mM

.~~~~~~ 0

~200

0 005 01 015 02INDOLE, mM

C ~~~~~~400C

0 005 01 015 02NDOLE mM

FIG. 5. Pyruvate formation from L-serine by thedehydratase reaction and its inhibition by indole. A,The curves are double reciprocal plots of initialvelocity with respect to serine concentration at thefixed concentrations of indole indicated on each line.B, The curves are double reciprocal plots of initialvelocity with respect to indole concentration at thefixed concentrations of serine indicated on each line.C, The curve is a replot of vm.x. from Fig. 5A againstindole concentration.

for the two reactions. In the tryptophanasereaction the beta moiety which is removed isthe indolyl radical; this step is presumablyreversible because of the tryptophan synthe-tase reaction. In the dehydratase reaction, thebeta moiety is either the hydroxide or methylmercaptan radical; it would be feasible topostulate this elimination step as an irreversi-ble sequence.The complexity of the subsidiary reactions

was further emphasized when we examined thekinetics in terms of the beta replacementreaction. Tryptophan synthetase catalyzes abisubstrate reaction. According to Cleland (3),reciprocal plots of initial rate versus substrateconcentration at fixed concentrations of the

second substrate will indicate whether thereaction is of the "sequential" or of the "pingpong" type. The mechanism is sequential if allsubstrates must add to the enzyme before anyproducts are released. In this case, the family ofcurves from the double reciprocal plot willintersect to the left of the ordinate. The mecha-nism is ping pong if one or more products arereleased before all substrates have added to theenzyme. Such a mechanism could result in aseries of parallel curves similar to the pattemobserved in uncompetitive inhibition. Based onthe assumption that the beta elimination andbeta replacement reactions are simple competi-tors of each other, the inhibition by indole ofthe beta elimination reaction at a given sub-strate concentration is equivalent to the rate ofthe beta replacement reaction at that particu-lar substrate and indole concentrations. Thedata from Fig. 4 and 5 were plotted in termsof the tryptophan synthetase reaction. Whenthe allosubstrate, not indole, is the fixed sub-strate, the family of lines intersect on the leftordinate indicating a "sequential" mecha-nism. But, when indole is the fixed substrateand the allosubstrate is varied, one obtains afamily of concave curves. According to Cleland(2), these results indicate that there is analtemative reaction sequence for the indole, orthe indole combines more than once during thereaction sequence. The latter hypothesis couldbe interpreted in terms of a cooperative effectin the binding of the indole molecules beforethe initiation of the reaction.

It is always hazardous to postulate a mecha-nism or reaction sequence solely from one typeof data, i.e., kinetic parameters, binding con-stants, etc. This is especially relevant for thesesubsidiary reactions where the kinetics indicaterather complex relationships. We in fact madekinetic derivations for simple variations of thebasic tryptophanase reaction (e.g., the forma-tion of a dead end enzyme-allosubstrate-indolecomplex) to explain the subsidiary beta elimi-nation reactions. We concluded that we couldnot set forth a mechanism with reactions tosubstantiate it that would satisfy the kineticdata and be explained solely by the parametersexplored in the kinetic experiments. That therecould be a conformational change involvedwith the two types of reactions was suggestedby the comparison of the dehydratase andsynthetase reaction with cysteine as the al-losubstrate using the fixed time, colorimetricassay (the cysteine inhibited the lactic dehy-drogenase used in the continuous, spectropho-tometric assay). The level of substrate required

348 J. BAcrmo1L.

TRYPTOPHANASE FROM B. ALVEI

for maximal activity of the synthetase reactionis in fact inhibitory to the dehydratase reac-tion. We know that the tetramer is the activespecies for all reactions with the B. alvei trypto-phanase (D. D. Whitt and R. D. DeMoss,Biochim. Biophys. Acta, in press). We havealready shown that PLP appears vital to thestructural integrity of the B. alvei enzyme (11).Thus cysteine may be inhibitory to the dehy-dratase reaction because the enzyme is in aconformation such that the PLP is more acces-sible to the cysteine, and upon formation ofthe cysteine-PLP complex, the enzyme startsto dissociate.These kinetic experiments with the subsidi-

ary dehydratase and synthetase reactions rein-force the conclusion from the tryptophanasekinetics that the complete mechanism for thetryptophanase enzyme in B. alvei is dissimilarto the one postulated for E. coli. Morino andSnell proposed for E. coli that the subsidiarybeta elimination with the allosubstrates fol-lowed the same steps as the beta eliminationwith tryptophan with the exception that theformation of the enzyme-aminoacrylic complexwas irreversible. The uncompetitive inhibitionpattern observed in B. alvei does not fit thishypothesis. For the beta replacement reac-tions, Morino and Snell proposed that indole asthe second substrate participates in just onereversible step. This is not the case for the B.alvei enzyme where we obtained a family ofconcave curves in the Lineweaver-Burk plotswhen indole was the fixed substrate and theallosubstrate was varied. These differences arenot unexpected in light of the structural differ-ences between the two enzymes which could beexpected to influence the catalytic functions(11, 15). Both enzymes are tetramers heldtogether by noncovalent bonds; both enzymeshave four PLP cofactors per molecule. How-ever, the PLP appears to be required for thestructural integrity of the B. alvei tryptopha-nase. When apotryptophanase is produced byincubation with cysteine, the tetramer reversi-bly dissociates at a pH above 6 producingsignificant amounts of monomer at pH 8, evenat 20 C and in the presence of potassium. TheE. coli tryptophanase only becomes less com-pact upon removal of the PLP; the tetramerdissociates to a dimer only when the pH israised above 8 or the temperature is lowered to5 C in the presence of potassium. Dissociationto the monomer is seen upon the addition of alow concentration of sodium dodecyl sulfate orthe elevation of the pH to 8.8 or higher.We will need a more complete understanding

of the subunit interactions of the B. alveitryptophanase before we can postulate a mech-anism for the subsidiary reactions catalyzed bythis enzyme. We are able to postulate anequation that satisfies the kinetics of the basictryptophanase reaction, namely the degrada-tion of tryptophan.

ACKNOWLEDGMENTSWe thank Burt Zemer for suggestions regarding the

kinetic arguments.This work was supported in part by Public Health Service

research grants No. Al 2971 from the National Institute ofAllergy and Infectious Diseases and No. AM11696 from theNational Institute of Arthritis and Metabolic Diseases. S. 0.Hoch was supported by Public Health Service predoctoralfellowship No. GM32724 from the National Institute ofGeneral Medical Sciences.

LITERATURE CITED1. Betts, G. F., and H. J. Evans. 1968. The inhibition of

univalent cation activated enzymes by tris (hydrox-ymethyl) aminomethane. Biochim. Biophys. Acta167:193-196.

2. Braunstein, A. E. 1960. Pyridoxal phosphate, p. 113-184.In P. D. Boyer, H. Lardy, and K. Myrback (ed.), Theenzymes. Academic Press Inc., New York.

3. Cleland, W. W. 1963. The kinetics of enzyme-catalyzedreactions with two or more substrates or products. I.Nomenclature and rate equations. Biochim. Biophys.Acta 67:104-137.

4. Cleland, W. W. 1963. The kinetics of enzyme-catalyzedreactions with two or more substrates or products. II.Inhibition: nomenclature and theory. Biochim. Bio-phys. Acta 67:173-187.

5. Cleland, W. W. 1963. The kinetics of enzyme-catalyzedreactions with two or more substrates or products. III.Prediction of initial velocity and inhibition patternsby inspection. Biochim. Biophys. Acta 67:188-196.

6. Cleland, W. W. 1963. Computer programmes for proc-essing enzyme kinetic data. Nature (London) 198:463-465.

7. Dugaiczyk, A., M. Malecki, and J. Eiler. 1968. Effect ofcysteine on L-a-glycerophosphate and lactic dehy-drogenase reactions. J. Biol. Chem. 243:2236-2240.

8. Frank, L. H., and R. D. DeMoss. 1957. Specific enzymicmethod for the estimation of L-tryptophan. Arch.Biochem. Biophys. 67:387-397.

9. Friedemann, T. E., and G. E. Haugen. 1943. Pyruvicacid. II. The determination of keto acids in blood andurine. J. Biol. Chem. 147:415-442.

10. Hoch, J. A., F. J. Simpson, and R. D. DeMoss. 1966.Purification and some properties of tryptophanasefrom Bacillus alvei. Biochemistry 5:2229-2237.

11. Hoch, S. O., and R. D. DeMoss. 1972. Tryptophanasefrom Bacillus alvei. I. Subunit structure. J. Biol.Chem. 247:1750-1756.

12. Horecker, B. L., and A. Kornberg. 1948. The extinctioncoefficients of the reduced band of pyridine nucleo-tides. J. Biol. Chem. 175:385-390.

13. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

14. Morino, Y., and E. E. Snell. 1967. A kinetic study of thereaction mechanism of tryptophanase-catalyzed reac-tions. J. Biol. Chem. 242:2793-2799.

15. Morino, Y., and E. E. Snell. 1967. The subunit structure

VoL. 114, 1973 349

HOCH AND DEMOSS

of tryptophanase. I. The effect of pyridoxal phosphateon the subunit structure and physical properties oftryptophanase. J. Biol. Chem. 242:5591-5601.

16. Newton, W. A., Y. Morino, and E. E. Snell. 1965.Properties of crystalline tryptophanase. J. Biol. Chem.240:1211-1218.

17. Pardee, A. B., and L. S. Prestidge. 1961. The initialkinetics of enzyme induction. Biochim. Biophys. Acta49:77-88.

18. Snell, E. E. 1958. Chemical structure in relation tobiological activities of vitamin B., p. 78-126. In R. S.Harris, G. F. Morrian, and K. V. Thimann (ed.),Vitamins and hormones, vol. XVI. Academic PressInc., New York.

19. Yanofsky, C. 1955. Tryptophan synthetase from Neuro-spora, p. 233-238. In S. P. Colowick and N. 0. Kaplan(ed.), Methods in enzymology, vol. II, Academic PressInc., New York.

350 J. BACTERIOL.