Enzymes Tryptophan Pathway Three Bacillus Species · complementation observed between the a and 12 components of tryptophan synthetase, andthe dissociation patterns ofthe large and

JOURNAL OF BACTERIOLOGY, Nov. 1973, p. 685-693Copyright 0 1973 American Society for Microbiology

Vol. 116, No. 2Printed in U.S.A.

Enzymes of the Tryptophan Pathway in ThreeBacillus Species

SALLIE O'NEIL HOCH AND IRVING P. CRAWFORDDepartment of Microbiology, Scripps Clinic and Research Foundation, La Jolla, California 92037

Received for publication 29 May 1973

The tryptophan synthetic pathway was characterized in three species ofBacillus, B. subtilis, B. pumilus, and B. alvei. They share the common featuresof a pathway which is subject to tryptophan repression, contains no unexpectedcomplexes among the five enzymes, exhibits dissociable anthranilate synthaseenzymes which do not require phosphoribosyl transferase for amidetransferactivity, contains separate indoleglycerol phosphate synthase and phos-phoribosylanthranilate isomerase enzymes, and contains similar tryptophansynthetase multimers. In looking at these characteristics in detail however,differences among the three species became apparent, as, for example, in thecomplementation observed between the a and 12 components of tryptophansynthetase, and the dissociation patterns of the large and small components ofanthranilate synthase. The results demonstrate some pitfalls in attempting tocompare multimeric enzymes in crude extracts from different organisms.

The tryptophan synthetic pathway in Bacil-lus subtilis has been examined both geneticallyand biochemically in some detail. Six of theseven genes encoding the enzymes of this path-way have been mapped and are contiguous (1,3). When methods for stabilizing these enzymesin cell-free extracts were developed, it becameapparent that they are coordinately regulated(14). B. subtilis was the first gram-positivebacterium in which these enzymes were ex-amined in detail, although the pathway has nowbeen studied in Staphylococcus aureus (21) andthe anaerobe Clostridium butyricum (2) as well.Of course, the genes and enzymes of the trypto-phan pathway have been extensively investi-gated in a number of gram-negative organisms,including Escherichia coli (5) and other entericorganisms, Pseudomonas putida (6),Chromobacterium violaceum (29), and Acineto-bacter calcoaceticus (24, 28).The methods that stabilize and permit reli-

able measurement of the tryptophan enzymes inB. subtilis, gentle lysis for release from the cellsand glycerol or sucrose added to the buffers, canbe applied to other Bacillus species as well, andwe have chosen to examine B. pumilus and B.alvei. Prior to this study, the enzymes of thispathway in B. pumilus had not been studied atall. In extracts of B. alvei, the first enzyme ofthe pathway, anthranilate synthase (AS), hadbeen reported to be quite unstable, and the

second, anthranilate-5-phosphoribosylpyro-phosphate phosphoribosyltransferase (PRT),undetectable (4). It had been reported that thelast enzyme of the pathway, tryptophan synthe-tase (TS), could be stabilized by 30% glycerol(23).

In characterizing the genes and enzymes ofthis pathway in diverse bacteria, some remarka-ble dissimilarities in organization of the geneson the chromosome and in regulatory mech-anisms have been seen (19). However, orga-nisms within one group, such as the entericbacteria (5, 19) or the pseudomonads (8, 22),seem to share one pattern of gene organizationand control. In our study of three Bacillusspecies, of which this paper is the first report,we intended to obtain the same types of infor-mation within a single genus, albeit a large andheterogeneous one, consisting of organisms ableto differentiate into spores. If, in fact, B. subtilisis representative of the members of the genus,this will lend validity to comparisons made withsingle representatives of other genera.

MATERIALS AND METHODSBacterial strains. The mutants of B. subtilis used

in this study were isolated by C. Anagnostopoulos andI. Crawford either by ultraviolet irradiation or bytransformation with the use of nitrous acid-treatedwild-type deoxyribonucleic acid (1; see Table 1). Theanalysis of enzyme levels in such mutants has already

685

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HOCH AND CRAWFORD

TABLE 1. Bacillus strains used in this study

Species and strain Genotype Accumulationadesignation

B. alveiTS2 ......... trpE NoneTS12 ......... trpF AnthranilateTS14 ......... trpF AnthranilateTS15 ......... trpA brady- InG,b indole, an-

troph thranilateTS16. trpF AnthranilateTS19. trpB brady- Indole, InG, an-

troph thranilateTS20 ...... trpC Anthranilate,

CDReTS24 ...... trpF Anthranilate

B. pumilusBpB9 ...... trpD, leu AnthranilateBpB10 ...... trpB Indole, InGBpB124 ...... trpE NoneBpB165 ...... trpB InG, indoleBpB166 ...... trpD AnthranilateBpB176 ...... trpB Indole, InGBpB177 ...... trpD AnthranilateBpB191 ...... trpA brady- InG, indole

trophBpB500 ...... trpE NoneBpB503 ...... trpC Anthranilate,

CDRB. subtilisT15 ...... trpD AnthranilateT16 ...... trpA InGT20 ...... trpB InG, indole

a Measured as described under Materials andMethods. The B. alvei cells were grown overnight at37 C in Spizizen minimal medium (22) containing0.5% glucose plus 0.05% acid-hydrolyzed casein, 10 jigof thiamine and cysteine/ml, and 1 Ag of tryptophan/ml; the B. pumilus cells were grown in the samemanner in the same minimal-glucose medium con-taining 0.05 ;&g of biotin/ml and 0.5 sg of trypto-phan/ml (plus 20 Mug of leucine and histidine/ml forBpB9).

I Indoleglycerol.c 1-(o-Carboxyphenylamino)-1-deoxy ribulose.

been described (14). The mutants of B. alvei wereisolated by J. A. Hoch after ultraviolet irradiation ofthe parent strain B. alvei F (10). The mutants of B.pumilus were isolated by P. Lovett after N-methyl-N'-nitro-N-nitrosoguanidine treatment and/or ul-traviolet irradiation of the parent strain B. pumilusNRRL B-3275 (17). The stocks were maintained onAK agar slants at 4 C or in freezing media (1%peptone, 10% glycerol, and 0.05 M magnesium sul-fate) as a spore suspension at -20 C.

Culture conditions. All strains were grown inSpizizen's minimal salts medium (27) containing0.5% glucose and 0.05% acid-hydrolyzed casein (0.05to 1.0% for B. alvei). The B. subtilis and B. pumilusmedia routinely contained 0.5 to 1.0 Mg oftryptophan/ml; the B. alvei medium, 1 to 10 gg/ml. The B.pumilus medium was also supplemented with 0.1 Mg

of biotin/ml and the B. alvei medium with 10 Mg ofthiamine-hydrochloride/ml. The cultures were grownunder conditions of vigorous aeration at 37 C for 11 to12 h. The cells were harvested by centrifugation at16,000 x g, washed once in 0.1 M potassium phos-phate, pH 75, plus 10% glycerol (vol/vol) if anthrani-late had been accumulated, and resuspended in 0.1 Mpotassium phosphate, pH 7.8, containing 40% glycerol(vol/vol), 0.01 M L-glutamine, and 4 mM MgSO4 orcontaining 0.8 M sucrose (2 ml of buffer/100 ml ofculture media). All buffers containing glycerol orsucrose also contained 1 mM sodium azide. Lysozyme(0.33 mg/ml) and deoxyribonuclease (3.3 Mg/ml) wereadded to the suspension at 37 C for 30 min, followedby centrifugation at 43,000 x g for 30 min. If lysis ofthe B. alvei extracts was poor even after 45 min at37 C, these suspensions were usually disrupted fur-ther by sonic oscillation for 90 s.Enzyme assays. AS was assayed in the amide-

transfer reaction (glutamine as source of the aminogroup) as described for B. subtilis (14). In the amina-tion reaction (ammonia as source of the amino group),the AS reaction mixture contained, in 1 ml, 50 Mmol oftris(hydroxymethyl)aminomethane (Tris)-Cl, pH 8.6,100 ,mol of NH4Cl, 10 ,mol of MgCl1, 190 nmol ofchorismate, and enzyme. PRT, phosphoribosylan-thranilate isomerase (PRAI), indoleglycerol phos-phate synthase (InGPS), and tryptophan synthetaseil2 (TS-f2; EC 4.2.1.20, L-serine hydrolyase [add-ing indole]) subunit activity (TS-B reaction) wereassayed as described by Hoch et al. (14). Tryptophansynthetase a (TS-a; EC 4.1.2.8, indole-3-glycerol-phosphate D-glyceraldehyde-3-phosphate-lyase) sub-unit activity (TS-A reaction) was assayed by followingindole production from indoleglycerol phosphate(InGP). The reaction mixture contained 50 Mmol ofpotassium phosphate, pH 7.8, 225 gmol of KCl, 250Amol of salt-free hydroxylamine, pH 7.8, 0.4 to 2 umolof InGP, excess TS-f2 subunits, and enzyme in areaction volume of 0.5 ml. Indole was determined bythe method of Smith and Yanofsky (25). One unit ofactivity is defined as the appearance of 1 Mmol ofproduct or the disappearance of 1 Mumol of substrate/min. Specific activity is defined as units per milli-gram of protein. Protein was determined colorimetri-cally by the method of Lowry et al. (18), with crystal-line bovine serum albumin as the standard.Accumulation studies. The accumulation of an-

thranilate and 1-(o-carboxyphenylamino)-1-deoxy-ribulose was determined by extracting acidified cul-ture supernatant fluids with ethyl acetate and thenchromatographing the organic layer on Whatman no.1 paper in a solvent system composed of methanol-butanol-benzene-water (2: 1: 1: 1, vol/vol). Indole-glycerol (InG) was determined qualitatively with theuse of ferric chloride reagent (26). Indole accumula-tion was determined by use of the indole reagent ofSmith and Yanofsky (25).

Chemicals. Chorismic acid was prepared by themethod of Gibson (7). InGP was synthesized enzy-matically as described by Wegman and Crawford(29). 5-Phosphoryl-ribose-1-pyrophosphate (PRPP)was purchased as the magnesium salt from SigmaChemical Co. All other chemicals were reagent grade.Sephadex chromatography. Molecular weight de-

686 J. BACTERIOL.

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRYPTOPHAN PATHWAY IN BACILLUS SPECIES

terminations were made with Sephadex G-100 col-umns (2.5 by 26 to 32 cm) equilibrated in 0.1 Mpotassium phosphate, pH 7.5, containing 30% glycerol(vol/vol), and 0.01 M glutamine. The upward flow ofthe column was regulated at 9 to 12 ml per h. The voidvolume was determined with blue dextran. The col-umn was calibrated according to the method ofWhitaker (30) by use of four proteins of knownmolecular weight: bovine serum albumin (68,000),ovalbumin (43,000), trypsin (23,300), and lysozyme(14,300). The cell-free extracts to be applied to thecolumn were first concentrated by the addition ofthree volumes of saturated ammonium sulfate, withstirring over a 20-min period, at 0 to 5 C. Aftercentrifugation at 43,000 x g for 30 min, the precipi-tate was resuspended in a minimal volume of columnbuffer. This procedure generally resulted in a two- tothreefold purification of the tryptophan enzymes.

RESULTSEnzyme levels in tryptophan auxotrophs.

Activities observed in crude extracts of 10 B.pumilus auxotrophs grown to a tryptophanlimit are shown in Table 2. None of the mutantsappears to be deficient in PRAI activity. In twoinstances (BpB124 and BpB176), activity wasbelow detectable limits in two reactions of thetryptophan synthetic pathway. In BpB124, thejoint loss of AS and PRT might conceivably beascribed to the existence of a single polypeptidewith both PRT and AS amidotransferase activ-

TABLE 2. Tryptophan biosynthetic enzymesa incell-free extracts of B. pumilus

Specific activity (milliunits/mgClass Mutant of protein)

AS PRT PRAI InGPS TS-A TS-B

1 BpB124 0 0 13 1.4 4.3 19BpB500 0 0.2 24 39 1.5 8.3

2 BpB166 4.0 0 46 81 3.1 13BpB177 4.0 0 45 73 2.6 14BpB9 2.8 0 23 73 1.8 10

4 BpB503 1.7 0.3 20 0 1.1 5.3

5a BpB191 5.7 2.4 49 88 O0 7.3b

5b BpB1O 4.9 1.5 63 127 6.2? 0BpB165 2.4 0.7 32 67 3.0b 0BpB176 4.3 0.6 42 119 Ob Ob

AS, anthranilate synthase; PRT, anthranilate-5-phosphoribosylpyrophosphate phosphoribosyltrans-ferase; PRAI, phosphoribosylanthranilate isomerase;InGPS, indoleglycerol phosphate synthase; TS-A,tryptophan synthetase a; TS-B, tryptophan syn-thetase ,2-

" These TS-A and TS-B activities were assayed inthe presence of exogenous tryptophan synthetase,2 ora subunits, respectively.

ity, as is the case in the enteric bacteria, but thecomplete absence of AS amination activity inthe mutant extract makes this unlikely. In lightof the very low level of activity of InGPS for thismutant, it seems more likely that the primarydefect is a nonsense mutation in the trpE locusexerting polar effects on the next two genes inthe operon. (Very similar results have been seenwith certain trpE nonsense mutants of B. sub-tilis [S. 0. Hoch, unpublished data]. A defini-tive solution to the BpB124 case awaits map-ping of the trp genes and/or isolation of non-

sense suppressors in B. pumilus.) In strainBpB176, the loss of both TS activities due to arevertible point mutation can be hypothesizedto result from a nonsense mutation in the genefor the , chain of this enzyme (31).

Table 3 shows specific activities for the tryp-tophan enzymes found in extracts of eightauxotrophic mutants of B. alvei. No strainslacking PRT were found. There are no examplesof multiple enzyme defects. An observationmade during the course of these studies was

that the specific activities of AS and PRTseemed very dependent on the amount of pro-tein in the crude extract; whenever the proteinconcentration was below 2 mg/ml, these activi-ties were disproportionately diminished, as

though these two enzymes were not releasedfrom treated cells as readily as the others.

Specific activities for four of the tryptophanenzymes in the B. subtilis mutants studiedearlier have proven to be as variable as in B.pumilus and B. alvei, namely, AS, 0.9 to 10.3;PRT, 0.5 to 8.3; PRAI, 29 to 193; and TS-#2, 6.2

TABLE 3. Tryptophan biosynthetic enzymesa incell-free extracts of B. alvei

Specific activity (milliunits/mgClass Mutant of protein)

AS PRT PRAI InGPS TS-A° TS-B"

1 TS2 0 2.6 39 7.0 ND: 3.8

3 TS12 3.4 1.4 0 6.1 3.1 9.8TS14 0.9 2.5 0 3.1 1.3 2.6TS16 2.9 1.9 0 5.4 ND 8.3TS24 0.6 3.9 0 5.9 0.9 2.5

4 TS20 2.0 1.7 45 0 ND 4.8

5a TS15 1.0 0.5 56 5.2 < 0.5 2.4

5b TS19 3.1 7.3 46 6.8 0.6 < 0.5

a Enzymes abbreviated as in Table 2.1 The TS-A and TS-B reactions were assayed with-

out the addition of exogenous tryptophan synthetase#2 and a subunits.

c Not determined.

VOL. 116, 1973 687

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HOCH AND CRAWFORD

to 14.2 (14). The most interesting comparisonconcerns InGPS levels, however. In B. subtilis,the levels in most mutants lie between 0.2 and1.3, except for two strains, T3 and T24, havinglevels of 33 and 66, respectively (14). We stillhave no explanation for these inordinately highInGPS levels. However, in examining the B.alvei and B. pumilus data, it seems that,although this activity shows about the sameintraspecies variability as the other tryptophanenzymes, the median levels are quite differentin the two species. B. alvei mutants haveintermediate levels of InGPS (2.1 to 7.0),whereas B. pumilus mutants have high levelsequivalent to those in B. subtilis T3 and T24 (39to 127).

It should be noted here that in the B. pumilusand B. alvei strains we have studied the TS-Areaction is easily demonstrable in derepressedcrude extracts, whereas in B. subtilis additionall2 subunits must be added.Tryptophan repression. Exogenously added

tryptophan has been shown to repress the syn-thesis of TS in B. subtilis (14). The effect ofexogenous tryptophan on this enzyme in strainB. pumilus BpB166 is presented in Table 4.Clearly, the activity decreased proportionatelyfor the TS-A and TS-B reactions with increasingtryptophan concentrations. In a similar experi-ment, we showed that there was no further in-crease in specific activity when cells grown at1 ,gg/ml were tested. A similar point of optimalderepression (near 1.5 gg/ml) has been shown forstrain T20 of B. subtilis.Tryptophan repression experiments in B.

TABLE 4. Tryptophan repression of tryptophansynthetase in B. pumilusa

Tryptophan Tryptophan synthetase (mU/mg)(Ug/mO) 2a

2 14.9 5.335 8.44 2.8010 0.79 0.2

a B. pumilus BpB166 was grown for 12 h in Spizi-zen's minimal medium containing 0.5% glucose,0.05% acid-hydrolyzed casein, 0.1 jg of biotin/ml, andvarious amounts of tryptophan. The cells were har-vested and resuspended in 3 ml of 0.1 M potassiumphosphate, pH 7.8, containing 15% glycerol (vol/vol),5 mM magnesium sulfate, 0.1 mM pyridoxal-5'-phos-phate, and 1 mM sodium azide. Lysozyme (1 mg),ribonuclease (20 Mg), and deoxyribonuclease (20 Mg)were added to the cell suspension which was incu-bated at 37 C for 30 min, following by centrifugationfor 30 min at 34,800 x g. TS-02 subunit was assayed inthe TS-B reaction, and TS-a subunit in the TS-Areaction.

alvei are complicated by the fact that thisspecies contains the enzyme tryptophanase,which degrades tryptophan to indole, pyruvate,and ammonia. However, Catena and DeMosswere able to show that, for at least 80 min afterthe addition of tryptophan to a growing culture,the synthesis of AS decreased markedly (4); ithad previously been shown that tryptophandegradation does not begin for at least 90 minunder similar culture conditions, though shortlythereafter tryptophan begins to decrease and istotally absent from the medium by 6 h (11, 23).Under our culture conditions, after a 12-hgrowth period there was no evidence of repres-sion of the tryptophan biosynthetic enzymeswith initial tryptophan levels from 1 to 10jig/ml. In effect, these results corroborate thoseof Roth et al. (23), who found no difference inTS-B activity between wild type and variousauxotrophs grown under conditions similar toours. We suggest that the study of repression ofthe tryptophan enzymes in a strain like B. alveihaving a constitutive tryptophanase can only beaccomplished in short-term experiments or withtryptophanase-less mutants.

It should be noted here that considerablymore tryptophan had to be added to cultures ofB. alvei tryptophan auxotrophs to obtaingrowth yields comparable to those of the otherspecies (Fig. 1). Moreover, B. alvei cultures hadto be inoculated with 50 to 100 times as manycells as the other strains (a 1.7% [vol/vol]inoculum of cells growing exponentially in L-broth) to avoid the selection of prototrophicrevertants through tryptophan exhaustion be-fore carbon source depletion. Even under theseconditions, B. alvei auxotrophs growing in 5 Mgof tryptophan/ml only reached densities equiva-lent to cultures of the other two species at 1jsg/ml. No such marked tryptophan require-ment was seen with the wild type or tryptophanbradytrophs of B. alvei.Molecular weights of TS components. The

TS components of B. subtilis have been purifiedto homogeneity and have molecular weights of82,000 and 26,000 for the f2 and a subunits,respectively (12, 13). When an ammonium sul-fate concentrated extract of a trpE mutant,BS73, was applied to a Sephadex G-100 columnequilibrated with 0.1 M potassium phosphate,pH 6.6, containing 15% glycerol, 1 mM mercap-toethanol, and 0.1 mM pyridoxal-5'-phosphate,there was an almost complete dissociation intof2 and a components (13). In contrast, when asimilar preparation from B. subtilis T15 wasapplied to a Sephadex G-100 column equili-brated as described in Materials and Methodswith glutamine-containing buffer to stabilize

688 J. BACTrERIOL.

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRYPTOPHAN PATHWAY IN BACILLUS SPECIES

450

400

350F

300k

D 250

4-,- 200[

150k

loo1

2 4 6Tryptophan (jg

FIG. 1. Effect of tryptophan on finaiof B. alvei tryptophan auxotrophs. B. c

was grown for 16 h at 37 C in Spizizmedium containing 0.5% glucose, 1.09lyzed casein, 10 pg of thiamine/ml,amounts of tryptophan. B. subtilis T2pumilus BpB166 (0) were grown for 12 tin the same minimal-glucose mediu0.05% acid-hydrolyzed casein and varioitryptophan, plus 0.1 pg of biotin/ml for

AS, the two components eluted togeestimated molecular weight of 77,0A portion of the preparation apjcolumn was held at 4 C for the sai

time and then assayed for TS airecovery of TS-B activity in the c

tions was complete relative to thsample, though the latter had only R

activity after 24 h. There was 61%the TS-A activity relative to thsample, which had 78% residualappears, therefore, that the prepcspecies in glutamine-containing bthose supplemented with 30% glyca# composition, as though absencetor diminishes #-# affinity without i

interactions.A concentrated extract of B. pum

under these conditions showed a pactivity at a molecular weight of

TS-A activity was detected in the fractions,however, and no a component could be locatedby stimulation of added f2 component in theTS-B reaction. The undiluted sample held at4 C retained only 16% of its original TS-Aactivity. The experiment was repeated with amore concentrated preparation from BpB177,with the use of the same buffer minus glutamine(Fig. 2B). Here, both the TS-A and TS-Bactivities cochromatographed at a molecularweight of 61,600, with distribution slightlyskewed to the low side. There was 68% recoveryof TS-B activity in the fractions (the undilutedcontrol lost no activity in 24 h), but the TS-Aactivity found was only 42% of the control,which had itself only 84% residual activity.Both TS activities in a preparation from B.

alvei TS12 co-chromatographed at an averagemolecular weight of 86,500 i 5,300 (four deter-minations; Fig. 2C is one example). The shoul-der seen with the TS-A activity was not alwayspresent. Recoveries were generally good, 65% forTS-B (100% residual activity in the control after2 days) and 78% for TS-A (84% residual in the

8 10 control).Molecular weights of PRT, PRAI, and

/ml) InGPS. When estimated by Sephadex G-100growth yield gel filtration, PRT had a molecular weight of

alvei TS2 (0) 45,300 i 2,900 in two separate experiments with,en's minimal B. pumilus mutants BpB10 and BpB191 (Fig.Yo acid-hydro- 3A is an example). PRAI and InGPS elutedand various close together, though there was always a slight

to (A) and B. skewing of one or both peaks; molecular weightsm containing for these two enzymes with strains BpB10,us amounts of BpB133, and BpB191 were 24,300 i 1,300 forB. pumilus. PRAI and 25,800 ±+ 1,600 for InGPS.

PRT from B. alvei was calculated to be53,600 2,500 with mutants TS12 and TS19.

ther with an The results with TS19 are shown in Fig. 3B.00 (Fig. 2A). InGPS was calculated to have a weight ofplied to the 26,200 in TS12. In a separate experiment withme length of TS19, both PRAI and InGPS were calculated toctivity. The be 26,400 in molecular weight.olumn frac- Molecular weight of AS and its com-e undiluted ponents. In our initial characterization of the57% residual tryptophan enzymes in B. subtilis, we did notD recovery of observe dissociation of AS on columns equili-e undiluted brated with buffers containing 30%o glycerol andactivity. It 10 mM glutamine. Kane et al. have since

)nderant TS partially purified the enzyme, reporting that ituffers, even contains two subunits, a large component of,erol, has an approximately 84,000 molecular weight and aof the cofac- small component of 16,000 molecular weight,affecting a-# and that glutamine promotes association be-

tween these subunits (15).ilus BpB124 As with B. subtilis, gel filtration of B. pumi-eak of TS-B lus AS in the presence of 10 mM glutamine andf 83,000. No 30% glycerol shows a single peak of amidetrans-

689VOL. 116, 1973

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HOCH AND CRAWFORD J. BACrERIOL.

Fraction NumberFIG. 2. Sephadex G-100 chromatography of tryptophan synthetase. The extracts were prepared and the

columns were run as described under Materials and Methods. The activities assayed were f,, subunit in theTS-B reaction (0) and a subunit in the TS-A reaction (0). The recovery of activity off the column was calcu-lated in reference to a portion of the extract held at 4 C for the time required to run the column. (A) B. subtilisT15 examined on a column (2.5 by 28 cm); each component assayed in the presence of an excess of the othersubunit; (B) B. pumilus BpBJ77, on a column (2.5 by 31.4 cm); (C) B. alvei, on a column (2.5 by 32 cm); the f12component assayed in the presence of exogenous TS-a from TS19, and the a subunit assayed in the presence ofexogenous TS-,h, from TS15.

25 35 45

E

(A

0t1-

Ln

0-

cr

U)4-,

E

0-

25 35

U)-0

a._

20 EU)I._

1.5 .?

E1.0 Ul)

CL

c05

Fraction NumberFIG. 3. Sephadex G-100 chromatography of phosphoribosyltransferase (PRT), phosphoribosylanthranilate

isomerase (PRAI), and indoleglycerolphosphate synthase (InGPS). The extracts were prepared and the columnswere run as described under Materials and Methods. The activities assayed were PRT (A), PRAI (0), andInGPS (0). (A) B. pumilus BpB1O, examined on a column (2.5 by 26 cm); (B) B. alvei TS19, on a column (2.5by 28 cm).

fer activity (Fig. 4A). The fractions contained84% of the activity of the undiluted controlsample. There was no increase in activity whensamples from fractions in the molecular weightrange of 15,000 to 30,000 were added to a sample

from the peak (fraction 30). Low activity in theamination reaction was observed in a distribu-tion identical to the amidetransfer activity.When a similar preparation was examinedunder the same conditions but in the absence of

690

I)

IE

th

a-E

CL

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRYPTOPHAN PATHWAY IN BACILLUS SPECIES

glutamine, both activities were again detectedin the 86,000 molecular weight region, but inbarely measurable amounts. Again there was nostimulation of the amidetransfer activity byadmixture of fractions from the molecularweight region of 16,000 to 22,000.

In contrast, B. alvei AS readily dissociatesunder these conditions. An experiment with anextract of TS12 in 30% glycerol and 10 mMglutamine is shown in Fig. 4B. A peak ofamination activity was seen at a molecularweight of 55,500, associated with amidetransferactivity. The recovery of amination activity was28%. When fractions from this peak were com-bined and mixed with later fractions, a peak ofamidetransferase-stimulating material at an es-timated molecular weight of 21,800 was found.After mixing fractions, a total recovery of 25% ofthe amidetransfer activity was calculated.Catena and DeMoss have reported no apparentdissociation of the B. alvei enzyme run over

A B 8

1.2

1.0 3 3

1Z1_ 2-~~~~~~.61 20.22

E<0.4-

0.2-

25 35 30 40 50 60 70

Fraction NumberFIG. 4. Sephadex G-100 chromatography of an-

thranilate synthase (AS). The extracts were preparedand the columns (2.5 by 31.4 cm) were run asdescribed under Materials and Methods. The activi-ties assayed were AS amidetransfer reaction (0) andAS amination reaction (0). The recovery of activityfrom the column was calculated in reference to aportion of the extract held at 4 C for the time requiredto run the column. (A) B. pumilus BpB177; (B) B.alvei TS12. Tubes 36 and 37 were pooled and assayedfor glutamine-dependent activity. The peak desig-nated by the closed circles represents the increase inthis activity when the smaller molecular weightsamples were mixed with the pool. Inset: Whitakerplot of molecular weight versus the ratio of elutionvolume to void volume. The void volume was deter-mined with blue dextran; the standards used werebovine serum albumin (68,000), ovalbumin (43,000),trypsin (23,300), and lysozyme (14,300).

Bio-Gel P-200 in the presence of 100 mMglutamine and 20 mM Tris-Cl, pH 7.2; themolecular weight was estimated at 90,000 (4).Interspecific complementation between B. sub-tilis and B. alvei AS components was notattempted, but is an experimental approachopened by these results.

Interspecies complementation of TS com-ponents. We have shown previously that B.subtilis 2 component could be complementedwith a components from E. coli and P. putida,resulting in a hybrid molecule having about 30%of the activity of the homologous complex in theTS-B reaction (12); we observed no heterolo-gous stimulation of B. subtilis a component inthe TS-A reaction using 2 components from thesame organisms (13). Using a crude extractfrom B. subtilis T16 as a source of fl2 subunits.we have now sought similar heterologous com-plementation with a subunits from the othertwo Bacillus species. The B. pumilus and B.alvei extracts were concentrated by the additionof 3 volumes of saturated ammonium sulfate, asin the sample preparation for Sephadex col-umns. Saturating levels of B. pumilus BpB1OTS-a produce the same specific activity in theTS-B reaction as the B. subtilis T20 subunit(data not shown). However, little or no activitywas evoked in this reaction in repeated at-tempts with preparations from B. alvei TS12and TS19.

DISCUSSIONExtracts of B. pumilus and B. alvei mutants

were prepared by the procedure involving gentlelysis in the presence of glycerol or sucrose wedevised for characterizing the tryptophan bio-synthetic enzymes of B. subtilis. All six activi-ties in the pathway can be assayed in suchextracts. No inexplicable examples of multipleenzyme defects were noted among 10 B. pumi-lus and 8 B. alvei tryptophan auxotrophs; thesingle instances of B. pumilus mutants lackingboth TS activities or both AS and PRT areprobably attributable to nonsense mutations.Moreover, in Sephadex gel filtration experi-ments we found no evidence for enzyme aggre-gates of the five tryptophan enzymes whichmight result in genetic pleiotropy. There weregood indications of a multimeric structure forTS (B. pumilus) and AS (B. alvei), however.The dissociable TS subunits of B. subtilis

have been purified to homogeneity and theirmolecular weights have been determined to be82,000 (TS-,62) and 26,000 (TS-a). Yet, underthe conditions used in this study (a pH of 7.5, 10mM glutamine, and 30% glycerol in the buff-ers), the TS did not elute in the position of an

VOL. 116, 1973 691

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HOCH AND CRAWFORD

a202 or a#2 complex, nor did the a and 02components separate. The co-elution of TS-Aand TS-B activities at a molecular weight ofabout 77,000 suggests the prevalence of an a#species. A similar pattern was found for B.pumilus and B. aluei, though these enzymes areless well characterized and there were someindications of slight dissociation of free achains. The molecular weight of the major B.pumilus species varied from 61,000 to 83,000 inseveral experiments with different mutants,and the B. alvei enzyme was found between80,000 and 94,000. The presence of high concen-trations of glutamine in these experiments,added to encourage the association and stabilityof AS, undoubtably affected the pyridoxal-5'-phosphate cofactor of the 2 subunit. Loss ofthis cofactor markedly weakens ,-, interactionsin E. coli (9). We assume it may have a similar,perhaps even more drastic, effect in these Bacil-lus enzymes, and that the sizes we have mea-sured by Sephadex gel filtration are not indica-tive of the molecular weight of the TS enzymecomplex under more physiological conditions.Thus, it seems unwise to use size estimation ofdissociable enzyme complexes in crude extractsfor comparative purposes. Nevertheless, thegeneral pattern of the three Bacillus TS en-zymes seems similar.

B. pumilus and B. alvei TS enzymes differfrom the B. subtilis enzyme by not requiringadditional ,2 subunits for the assay of a sub-units in the TS-A reaction in vitro. This differ-ence may be due to affinity differences betweenhomologous a and ,B chains under the nonphysi-ological conditions of the TS-A assay. Whencomplementation between heterospecific TSsubunits was assayed, B. pumilus a. subunitproved equal to the homologous one in stimulat-ing B. subtilis 02 subunit. We were unable todemonstrate an active complex between B.alvei a and B. subtilis f2 subunits, though theexperiment did not establish whether or not thehomologous B. alvei enzyme dissociated enoughto permit a hybrid enzyme to form. This resultdeserves further investigation, in light of thefact that both E. coli and P. putida a subunitsprovide 30% of the maximal stimulation of B.subtilis f2 subunit (12).The PRT of both B. pumilus and B. alvei is

comparable in size to that of B. subtilis. Al-though no PRT-deficient mutants of B. alveiwere found, fractionation of the AS from thisspecies on a G-100 column showed that theamidotransferase-stimulating component has amolecular weight near 22,000, clearly differentfrom PRT and similar to the small, glutamineamidotransferase subunit demonstrated in B.subtilis (15), P. putida (22), and other orga-

nisms. As in B. subtilis, both the PRAI andInGPS enzymes of B. pumilus and B. alvei arebelow 31,000 in molecular weight. The standarddeviations of the determinations of these twoenzymes in all three species overlap, so that wecannot recognize any definite differences in size.By comparison with the molecular weights ofthe fused PRAI-InGPS enzymes in the entericbacteria, which are approximately twice aslarge, these two enzymes in the Bacillus specieswe have studied appear to be separate entities,as they are in other nonenteric bacteria (19).The first enzyme of the pathway, AS, is

multimeric in all Bacillus species studied, as itis in all organisms studied to date (19). In B.subtilis it is composed of a large subunit E(84,000 molecular weight) which by itself cancatalyze only the amination reaction, and asmall subunit X (16,000) which when complexedwith subunit E permits catalysis of the amido-transferase reaction. When examined on aSephadex G-100 column in the presence ofbuffer containing 10 mM glutamine, subunits Eand X form a complex with a second peak ofdissociated subunit X (16). When the B. pumi-lus AS was examined on a Sephadex G-100column in the presence of glutamine, there wasno apparent dissociation of a subunit X; itspresence was inferred by the loss of amide-transfer activity when the column was run inthe absence of glutamine. Thus, the B. pumilusAS dissociates less easily than the B. subtilisenzyme. On the other hand, the B. alvei ASfollowed the B. subtilis pattern with the appear-ance of an apparent subunit X of 22,000 molecu-lar weight. However, the large enzyme speciescatalyzing the amination and some amidetrans-fer AS activity was calculated to be only 55,500in molecular weight, approximately half the sizeof the B. subtilis complex and of that reportedby Catena and DeMoss (4) for B. alvei AS in thepresence of 100 mM glutamine. This fact, inconjunction with the low recovery of activityfrom the column, indicates that the B. alvei ASmay actually be an E2X2 tetramer that requiresa high glutamine concentration to maintain thisstructure under conditions of gel filtration.Thus, B. subtilis, B. pumilus, and B. alvei

share the common features of a tryptophanbiosynthetic pathway which is subject to tryp-tophan repression, contains no unexpected com-plexes among the five enzymes, exhibits disso-ciable AS enzymes which do not require PRTfor amidetransfer activity, contains separateInGPS and PRAI enzymes, and contains similarTS multimers. Overall, there is a consistencyamong the three bacilli equal to or greater thanthat among the genera of the enterobacteria.But, in looking at these characteristics in detail,

692 J. BACTERIOL.

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TRYPTOPHAN PATHWAY IN BACILLUS SPECIES

differences among the three species becomemore apparent. The interaction between thehomologous a and 02 components in crudeextracts of B. pumilus and B. alvei is such thatexogenous a is not required to assay the TS-Areaction as in B. subtilis. Yet when heterologousa and 2 subunits are mixed to assay the TS-Breaction, only B. pumilus a complements the B.subtilis subunit 02. The AS complex of B.pumilus is more stable to gel filtration than thatof either B. subtilis or B. alvei; the latterorganism presents an elution pattern similar toB. subtilis, but the complex of subunits E and Xappears much smaller. These differences do notcorrelate precisely with the relatedness of thethree species as estimated by their guanine pluscytosine content, which is 32.8 (B. alvei), 40 (B.pumilus) and 43.8 (B. subtilis; 20). Our dataemphasize the dangers inherent in attemptingto draw too many conclusions from size estima-tions or subunit complementation results ob-tained with crude extracts.

ACKNOWLEDGMENTS

This investigation was supported by National ScienceFoundation grant GB6841 and by Public Health Service grantGM-18627 from the National Institute of General MedicalSciences. Part of this work was carried out while S.O.H. was a

recipient of an American Cancer Society postdoctoral fellow-ship (PF 489).

The technical assistance of Nancy Vaughan and HelenCoukoulis is gratefully acknowledged.

The B. pumilis strains were generously provided by PaulLovett and the B. alvei strains by Ralph DeMoss.

LITERATURE CITED

1. Anagnostopoulos, C., and I. P. Crawford. 1967. Le groupedes genes regissant la biosynthese du tryptophane chezBacillus subtilis. C. R. Acad. Sci. (Paris) 265:93-96.

2. Baskerville, E. N., and R. Twarog. 1972. Regulation ofthe tryptophan synthetic enzymes in Clostridium buty-ricum. J. Bacteriol. 112:304-314.

3. Carlton, B. C., and D. D. Whitt. 1969. The isolation andgenetic characterization of mutants of the tryptophansystem of Bacillus subtilis. Genetics 62:445-460.

4. Catena, A., and R. D. DeMoss. 1970. Physiological andkinetic studies with anthranilate synthetase of Bacillusalvei. J. Bacteriol. 101:476-482.

5. Creighton, T. E., and C. Yanofsky. 1970. Chorismate totryptophan (Escherichia coli)-anthranilate synthetase,PR transferase, PRA isomerase, InGP synthetase, tryp-tophan synthetase, p. 365-380. In H. Tabor and C. W.Tabor (ed.), Methods in enzymology, vol. 17A. Aca-demic Press Inc., New York.

6. Enatsu, T., and I. P. Crawford. 1968. Enzymes of thetryptophan synthetic pathway in Pseudomonas putida.J. Bacteriol. 95:107-112.

7. Gibson, T. 1964. Chorismic acid purification and somechemical and physical studies. Biochem. J. 90:256-261.

8. Gunsalus, I. C., C. F. Gunsalus, A. M. Chakrabarty, S.Sikes, and I. P. Crawford. 1968. Fine structure map-ping of the tryptophan genes in Pseudomonas putida.Genetics 60:419-435.

9. Hathaway, G. M., and I. P. Crawford. 1970. Studies on

the association of #-chain monomers of Escherichia coli

tryptophan synthetase. Biochemistry 9:1801-1808.10. Hoch, J. A., and R. D. DeMoss. 1965. Physiological

effects of a constitutive tryptophanase in Bacillus alvei.J. Bacteriol. 90:604-610.

11. Hoch, J. A., and R. D. DeMoss. 1966. Physiological role oftryptophanase in control of tryptophan biosynthesis inBacillus alvei. J. Bacteriol. 91:667-672.

12. Hoch, S. 0. 1973. Tryptophan synthetase of Bacillussubtilis. Purification and characterization of the #2component. J. Biol. Chem. 248:2992-2998.

13. Hoch, S. 0. 1973. Tryptophan synthetase of Bacillussubtilis. Purification and characterization of the a

component. J. Biol. Chem. 248:2999-3003.14. Hoch, S. O., C. Anagnostopoulos, and I. P. Crawford.

1969. Enzymes of the tryptophan operon of Bacillussubtilis. Biochem. Biophys. Res. Commun. 35:838-844.

15. Kane, J. F., W. M. Holmes, and R. A. Jensen. 1972.Metabolic interlock. The dual function of a folatepathway gene as an extra-operonic gene of tryptophanbiosynthesis. J. Biol. Chem. 247:1587-1596.

16. Kane, J. F., and R. A. Jensen. 1970. The molecularaggregation of anthranilate synthase in Bacillus sub-tilis. Biochem. Biophys. Res. Commun. 41:328-333.

17. Lovett, P. S., and F. E. Young. 1970. Genetic analysis inBacillus pumilus by PBS1-mediated transduction. J.Bacteriol. 101:603-608.

18. 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.

19. Margolin, P. 1971. Regulation of tryptophan synthesis, p.389-446. In H. J. Vogel (ed.), Metabolic regulation, vol.5. Academic Press Inc., New York.

20. Normore, W. M., and J. R. Brown. 1970. Guanine pluscystosine (G + C) composition of bacteria, p.H-24-H-74. In H. A. Sober (ed.), Handbook of Bio-chemistry. The Chemical Rubber Co., Cleveland, Ohio.

21. Proctor, A. R., and W. E. Kloos. 1973. Tryptophanbiosynthetic enzymes of Staphylococcus aureus. J.Bacteriol. 114:169-177.

22. Queener, S. F., and I. C. Gunsalus. 1970. Anthranilatesynthase enzyme system and complementation inPseudomonas species. Proc. Nat. Acad. Sci. U.S.A.67: 1225-1232.

23. Roth, C. W., J. A. Hoch, and R. D. DeMoss. 1971.Physiological studies of biosynthetic indole excretion inBacillus alvei. J. Bacteriol. 106:97-106.

24. Sawula, R. V., and I. P. Crawford. 1972. Mapping of thetryptophan genes of Acinetobacter calcoaceticus bytransformation. J. Bacteriol. 112:797-805.

25. Smith, 0. H., and C. Yanofsky. 1962. Enzymes involvedin the biosynthesis of tryptophan, p. 794-806. In S. P.Colowick and N. 0. Kaplan (ed.), Methods in En-zymology, vol. 5. Academic Press Inc., New York.

26. Smith, 0. H., and C. Yanofsky. 1963. Intermediates in

the biosynthesis of tryptophan, p. 590-597. In S. P.Colowick and N. 0. Kaplan (ed.), Methods in Enzy-mology, vol. 6. Academic Press Inc., New York.

27. Spizizen, J. 1958. Transformation of biochemically defi-cient strains of Bacillus subtilis by deoxyribonucleate.Proc. Nat. Acad. Sci. U.S.A. 44:1072-1078.

28. Twarog, R., and G. L. Liggins. 1970. Enzymes of thetryptophan pathway in Acinetogacter calcoaceticus. J.Bacteriol. 104:254-263.

29. Wegman, J., and I. P. Crawford. 1968. Tryptophansynthetic pathway regulation in Chromobacterium vi-olaceum. J. Bacteriol. 95:2325-2335.

30. Whitaker, J. R. 1963. Determination of molecular weightsof proteins by gel filtration on Sephadex. Anal. Chem.35:1950-1953.

31. Yanofsky, C., and I. P. Crawford. 1959. The effects of

deletions, point mutations, reversions and suppressormutations on the two components of the tryptophansynthetase of Escherichia coli. Proc. Nat. Acad. Sci.U.S.A. 45:1016-1026.

VOL. 116, 1973 693

on February 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from