JOURNAL OF BACTERIOLOGY, Mar. 1980, P. 1291-12970021-9193/80/03-1291/07$02.00/0

Vol. 141, No. 3

,8-Alanine Synthesis in Escherichia coliJOHN E. CRONAN, JR.

Department ofMicrobiology, University of Illinois, Urbana, Illinois 61801

The enzyme, aspartate 1-decarboxylase (L-aspartate 1-carboxy-lyase; EC4.1.1.15), that catalyzes the reaction aspartate -- fl-alanine + C02 was found inextracts of Escherichia coli. panD mutants of E. coli are defective in 8-alaninebiosynthesis and lack aspartate 1-decarboxylase. Therefore, the enzyme functionsin the biosynthesis of the ,8-alanine moiety of pantothenate. The genetic lesion inthese mutants is closely linked to the other pantothenate (pan) loci of E. coli K-12.

D-Pantothenate, the condensation product of,8-alanine and D-pantoic acid, is found in the 4'-phosphopantetheine moieties of coenzyme A(CoA) and of the acyl carrier protein of lipidsynthesis (1, 6). Due to our interest in acylcarrier protein, we have attempted to alter theamount of the 4'-phosphopantetheine prostheticgroup of this protein by manipulation of thespecific precursors of the prosthetic group,namely,B-alanine, pantoic acid, and pantothen-ate. In the course of these investigations, werealized that our knowledge of the pantothenatepathway was incomplete in that the pathway offB-alanine synthesis was not yet clear (3, 6).The reactions leading to the fornation of pan-

toic acid, the condensation of pantoate with ,B-alanine to form pantothenate, and the conden-sation of the pantothenate moiety with cysteineto form (ultimately) 4'-phosphopantetheine, arewell established (6). However, the data on ,-alanine synthesis are ofa conflicting and indirectnature.Two pathways of ,B-alanine synthesis have

been proposed for Escherichia coli. The firstpathway is the aspartate 1-decarboxylase path-way discussed in this paper. This pathway wasproposed on the basis of the production of ,B-alanine by intact cells of E. coli incubated in thepresence of aspartate (8, 35) and upon the rever-sal by f,-alanine of growth inhibition by hydrox-yaspartate (28). However, as pointed out byBoeker and Sneli (3) and Brown (6), these datadid not demonstrate that ,B-alanine was formeddirectly from aspartate.The second pathway was proposed by Slot-

nick and Weinfeld (29, 30) from their findingthat growth of a fi-alanine auxotroph of E. coliwas supported by dihydrouracil or ,8-ureidopro-pionate (N-carbamyl,-alanine). These workerssuggested that the de novo synthesis of f)-alanineproceeded from uracil through reduction to di-hydrouracil and by hydrolysis first to ,B-ureido-7)ropionate and then to C02, NH3, and fi-alanine.

In this paper, I demonstrate that the mainpathway for ,8-alanine biosynthesis in E. coli isby the aspartate 1-decarboxylase. After the sub-mission of this paper for publication, Williamsonand Brown (39) reported the purification andproperties of an aspartate 1-decarboxylase fromE. coli W. The present paper has been revisedto include a discussion of that work.

MATERIALS AND METHODSMaterials. Three lots of L-[U-`C]aspartic acid

were purchased from Amersham Corp. or New Eng-land Nuclear Corp. /8-[1-'4C]alanine and ,B-[3-3H]ala-nine were the products ofNew England Nuclear. SilicaGel G thin-layer plates (250,um thick) were purchasedfrom Analtech. The cellulose thin-layer plates werefrom Eastman Chemical Products, Inc. The aminoacids, N-(2,4-dinitrophenol)-amino acids (DNP-aminoacids), glutamic decarboxylase, and vitamins werefrom Sigma Chemical Co. Ketopantoic acid (14) andsuccinic dihydrazine (23) were synthesized by pub-lished procedures.

Preparation of extracts. The various bacterialstrains listed in Table 1 were grown in a rich brothmedium (7) (or in minimal medium as given) to latelog phase at 37°C and collected by centrifugation.After the cells were washed with cold (4°C) distilledwater, they were suspended in cold distilled water andincubated in ice for 20 to 30 min to discharge thecellular amino acid pools (4). After centrifugation thecells were suspended in 0.1 M potassium phosphate(pH 6.8) and disrupted in a French press at 18,000 lb/in2. Large debris and unbroken cells were removed bycentrifugation at 6,000 x g for 10 min. Protein concen-trations were determined by the microbiuret reaction(20) by using bovine serum albumin as the standard.Enzyme assays. Pantothenate synthetase was as-

sayed as described by Miyatake et al. (19). Aspartate1-decarboxylase was generally assayed by the conver-sion of '4C-labeled aspartate to ,B-alanine as assayedby thin-layer chromatography. A standard reactionmixture contained L-[U-_4C]aspartate, 23 ,uM (216mCi/mmol); potassium phosphate, 0.1 M (pH 6.8); andenzyme (0 to 90 ,ug of crude extract) in a final volumeof 50 ,ul. The reaction mixtures (in polypropylenecentrifuge tubes) were incubated at 37°C for 1 to 2 h.

1291

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

1292 CRONAN

After incubation, 0.2 ml of absolute ethanol and 5 pi ofa mixture of 0.2 mM ,B-alanine-0.2 mM L-aspartatewere added as carrier to each tube, and the resultingmixture was centrifuged to remove precipitated pro-tein. The supernatant and a 100-pl wash (ethanol) ofthe pellet were combined in a second centrifuge tubeand evaporated to dryness under N2 at 800C. Theresidue was dissolved in 50% ethanol and quantita-tively spotted on a silica gel thin-layer plate. The platewas developed once or twice in absolute ethanol-28%ammonia (4:1, vol/vol) and then autoradiographed.The appropriate areas of the silica gel were thenscraped from the plate into a vial containing 0.5 ml ofwater, and 3.5 ml of scintillant (PCS, Amersham) wasadded for scintillation counting.

Alternatively, the purified enzyme could be assayedby 14CO2 release. The reaction was run in rubberseptum-sealed vessels containing a disposable centerwell (Kontes). After incubation, 5 pd of glacial aceticacid was added to the reaction and 200 gu of Hyamine(New England Nuclear) was added to the center well.After the 14C02 had been collected (1 h at 370C), thecenter well was placed into a scintillation vial contain-ing 3 ml of PCS, and the 14C02 was quantitated byscintillation counting.A third assay was sometimes used with the purified

TABLE 1. Bacterial strains

Stan Relevant SucStrain genotypeE. coliaKL16 pan+ K. B. Low (CGSG)AB352 pan+ E. Adelberg (CGSG)AB353 panDI E. Adelberg (CGSG)AB354 panD2 E. Adelberg (CGSG)AB355 panD3 E. Adelberg (CGSG)CY262 pan+ pan+ transductant of

AB354CY263 pan+ pan+ revertant of AB355AB2638 panC5 E. Adelberg (CGSG)AT1371 panC4 A. L. Taylor (CGSG)CS8AspTL aspA22 Marcus and Halpern (18)

(CGSG)CY257 panB6 From strain Hfr 3000

YA139bW pan+ C. WoolfolkM99-1 panC B. Davis (17)M99-2 panD B. Davis (17)

S. tphimurium LT2SA965 pan+ K. Sandersonpan-6 panA6 K. Sanderson (9)pan-4 panB4 K. Sanderson (9)pan-2 panC2 K. Sanderson (9)a All strains are derivatives of E. coli K-12 except

strains W, M99-1, and M99-2, which are E. coli Wstrains. Strains marked CGSG were from the ColiGenetic Stock Center, Yale University, New Haven,Conn.

b Strain CY257 is a leu + panB6 derivative of strainPC0540 (CGSG 5407). Strain PC0540 was transducedfirst to ku+ aceF with P1 phage grown on an aceFstrain and then to ace+ panB with phage grown onstrain Hfr 3000 YA139.

J. BACTERIOL.

enzyme. After incubation, the reaction mixture waspipetted into a column (0.5 by 2.5 cm) ofAG 1-X4 ion-exchange resin (chloride form) in a pipette pluggedwith glass wool. ,-Alanine was eluted with two 0.5-mlwashes with water and the water was counted in 3 mlof PCS. Aspartate remained bound to the column, butcould be quantitatively eluted with 1 ml of 1 N HCI.

Control experiments using i-_[1_-4C]alanine in thefirst and third assays and 14C02 in the second assayshowed that recovery of the compounds was quanti-tative and that no degradation of ,B-alanine occurredin crude extracts. One unit of decarboxylase activity isdefined as 1 pmol of product formed per min. Theassay was linear with time for at least 3 h and withactivity to >1 U per reaction in crude extracts and to>15 U per reaction with the purified enzyme.Enzyme purification. An extract of strain AB352

was prepared as described above and centrifuged at48,000 x g for 30 min. The resulting supernatant washeated in a 670C water bath for 20 min, cooled in ice,and centrifuged at 35,000 x g for 30 min at 40C. Thesupernatant (at 40C) was fractionated with ammoniumsulfate, and the precipitate formed from 42 to 60% ofsaturation was harvested by centrifugation, suspendedin 0.1 M potassium phosphate (pH 6.8), and dialyzedagainst 200 volumes of the same buffer overnight at40C. The enzyme was stored at -20°C and retainedfull activity for at least 2 months. Pantothenate syn-thetase was purified as described previously (24).

Genetic methods. The genetic methods and themedia used were described previously (7).Chromatography. The solvent systems used in

the identification of the product as ,8-alanine are listedby Niederwieser (21). For chromatography of the freeamino acids on silica gel thin-layer plates, solventsystems 3, 5, 6, and 7 of Niederwieser (21) were usedin various two-dimensional combinations. Either sol-vent system 7 (96% ethanol-28% ammonia, 70:30, vol/vol) or a modification of it (absolute ethanol-28%amumonia, 80:20, vol/vol) was used in one-dimensionalchromatography. For chromatography on thin layersof cellulose (Eastman), the solvents used (in two-di-mensional analysis) were first system 25 and thensystem 27.The DNP-amino acids were synthesized essentially

as described by Niederwieser (21) and extracted intoether. The DNP-amino acids were separated on thinlayers of Silica Gel G in two dimensions by usingvarious combinations of the solvent systems 50, 53,and 54 (21).

Radioactive amino acids were detected by autora-diography. Nonradioactive free amino acids were de-tected by ninhydrin, and the DNP-amino acids weredetected by their characteristic color.

RESULTSIdentification of the reaction products.

Two main products were formed by incubationof L-[U-14C]aspartate with crude extracts of E.coli (Fig. 1). The product of higher chromato-graphic mobility is fumarate formed by the ac-tion of aspartase (13). The identification of thisproduct as fumarate is based on cochromatog-raphy with authentic fumarate and the lack of

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

f-ALANINE SYNTHESIS 1293

this product in a strain deficient in aspartase,strain CS8aspTL (18).The product of lower mobility is ,B-alanine.

The identification of the product as ,8-alanine isbased on (i) Silica Gel G thin-layer chromatog-raphy in four different two-dimensional systems,(ii) on cellulose thin-layer chromatography in atwo-dimensional system, and (iii) Silica Gel Gthin-layer chromatography of its DNP deriva-tive in three different two-dimensional solventsystems. In all cases, the radioactive spot exactlycorresponded with that of f8-alanine or DNP )8-alanine added as an internal standard.The production of ,B-alanine was confirmed by

isotope dilution analysis. The mixture of prod-ucts and substrate resulting from a standardreaction (10-fold increase in volume) was mixedwith 1.2 g of authentic crystalline ,B-alanine,dissolved in water, and crystallized by additionof ethanol. After a second crystallization thespecific activity of the crystals reached the valuefound through four subsequent recrystallizations(106.5 ± 8 cpm/mg).A further confirmation of the identity of the

product was the conversion of the ,8-alanine topantothenate by pantothenate synthetase (D-pantoate:,B-alanine ligase [AMP forming]; EC6.3.2.1) (16, 19, 24). The conversion to panto-thenate required D-pantoate, ATP, and Mg2eand was confirmed by two-dimensional thin-layer chromatography (10).

It should be noted that no synthesis of L-alanine was seen, and therefore no decarboxyl-ation occurred at the C-4 of aspartate. The ab-sence of L-alanine was based on chromato-graphic evidence and on the lack of C02 produc-tion from L-[4-14C]aspartate. This finding is con-sistent with the in vivo evidence of Roberts etal. (26), who reported that aspartate is not aprecursor to L-alanine. Traces of other com-pounds were sometimes formed from L-[U-_C]-aspartate in crude extracts which were tenta-tively identified as lysine, methionine, and thre-onine. These activities were lost upon prolongeddialysis or purification of crude extracts.Partial purification and characterization

of the decarboxylase. Aspartate 1-decarbox-ylase is a soluble enzyme. No activity was foundin a washed membrane fraction. The enzymewas partially purified by heat treatment andammonium sulfate fractionation. Over 94% ofthe initial activity was found in the 42 to 60%ammonium sulfate cut. The resulting enzymepreparations were about 40-fold purified overthe crude extract with a yield of 140%. Theincrease in total activity is attributed to the heatinactivation ofthe competing aspartase reaction.The purified enzyme preparations containedonly traces of aspartase. Similar heat and am-

.:~---:::.solvent

f ront

-<- fumcarate

-- 8- aIaninte

< aspa rt ate

-<- 0rorigin

pan+ pGnDFIG. 1. Autoradiogram of the products of an in-

cubation ofL -[U-14CJaspartate with 90 pg ofproteinfrom strain CY362 (pan) or strain AB354 (panD) asdescribed in the text. The silica gel thin-layer platewas developed in absolute ethanol-28% NH40H (4:1,vol/vol).

monium sulfate steps were used by Williamsonand Brown (39).The pH optimum of the aspartate 1-decarbox-

ylase activity was about pH 7.5. The reactionrates at the pH values of 6.8 and 8.0 were about60% of the pH 7.5 rate. A similar broad pHoptimum (pH 6.5 to 7.5) was reported by Wil-liamson and Brown (39).The Michaelis constant of the decarboxylase

for aspartate was about 80 ,uM. Similar valueswere given by crude extracts and by the partiallypurified enzyme preparation. Williamson andBrown (39) reported values of 160 and 300,M,respectively, for the major and minor formsfound in E. coli W. The specific activities re-ported for crude extracts differ between the twolaboratories, but this is due to differing assayconditions. When corrected to the same sub-

VOL. 141, 1980

...

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

1294 CRONAN

strate concentration, assay pH, and assay tem-perature, our results are essentially identical tothose of Williamson and Brown (39). Our assaywas run at a suboptimal pH to minimize com-petition by the aspartase reaction which has ahigher pH optimum (13).The stoichiometry of the reaction was deter-

mined by using the partially purified enzyme(Table 2). For each molecule of aspartate con-sumed, one molecule each of C02 and ,B-alaninewere formed. The same stoichiometry was re-ported by Williamson and Brown (39) for thereaction products of the purified major form ofthe decarboxylase.Evidence for an essential carbonyl

group. Aspartate 1-decarboxylase was inhibitedby a variety of carbonyl reagents. Hydroxyla-mine was the most effective inhibitor (81% in-hibition at 60 uM), but all other carbonyl re-agents tested (NaBH4, D-cycloserine, hydrazine,semicarbazine, and succinic dehydrazine) alsoinhibited aspartate 1-decarboxylase. Williamsonand Brown (39) reported inhibition by hydrox-ylamine and NaBH4 of the major form of thedecarboxylase.Williamson and Brown (39) demonstrated

that the essential carbonyl of the major form ofthe decarboxylase is that of a pyruvoyl residue.However, the prosthetic group ofthe minor formof the decarboxylase was not studied. I foundthat about 10% of the original activity of am-monium sulfate-purified decarboxylase prepa-rations (reportedly [39] a mixture of the twoforms) was inactivated by treatment with cyclo-serine (12), but this could be restored by incu-bation with pyridoxal phosphate. It seems pos-sible therefore that the essential carbonyl groupof the minor decarboxylase form is that of pyri-doxal phosphate.Inhibitor and substrate specificity. In

TABLE 2. Stoichiometry of aspartate 1-decarboxylase reaction'

CO2 released B-AIanine L-AspartateExpt (nmol) formed (nmol) c(1)nume

1 10.5 10.4 9.92 5.8 8.4 5.3

'A standard reaction was run with L-[U-_4C]aspar-tic acid by using 65 ,g of ammonium sulfate-purifiedenzyme in experiment 1 and half that amount inexperiment 2. After incubation, C02 was collected asdescribed in the text, and then carrier /i-alanine andL-aspartic acid were added. ,-[3-3H]alanine (1 ,uCi,37.5 Ci/mmol) was added to monitor the recovery of/i-alanine. The L-[U-`4C]aspartate was uniformly la-beled as determined by ninhydrin treatment and bytreatment with a glutamate-5 (aspartate4) decarbox-ylase preparation from Clostridium welchii (11).

J. BACTERIOL.

agreement with Williamson and Brown (39), Ifound that D-aspartate was neither a substratenor an inhibitor ofthe decarboxylase. We furtheragree that D-serine is a competitive inhibitor andis therefore a plausible target for the knowninhibition of pantothenate synthesis by D-serine(17).Physiological function of aspartate 1-de-

carboxylase. Pantothenate auxotrophs of E.coli and Salmonella typhimurium are of threetypes (Fig. 2). panC mutants respond only topantothenate, whereas panB mutants also re-spond to pantoic or ketopantoic acids (or theirlactones) and panD mutants respond to ,B-ala-nine as well as pantothenate (9, 17). panD mu-tants, therefore, seem specifically defective inthe synthesis of 8l-alanine and thus should bedeficient in aspartate 1-decarboxylase. This isthe case. Four different panD mutants werefound to contain no detectable decarboxylaseactivity (Table 3). The most thoroughly char-acterized mutant, strain AB354, contained<0.3% of the enzyme level found in the wild-typestrain. The other strains, AB353, AB355, andM99-2, contained less than 2% of the normalactivity (Table 3). Mixtures of the inactive ex-tracts with active extracts gave additive activi-ties, indicating the lack of an inhibitor in thepanD extracts. Apan' revertant of strain AB355regained normal decarboxylase activity, and apan' transductant of strain AB354 (the P1phage were grown on strain AB352) also con-tained a nornal level of decarboxylase (Table3). These results indicate that a single mutationis responsible for the lack of decarboxylase inpanD auxotrophs and therefore demonstratethat the major pathway of f8-alanine synthesisin E. coli is by the a-decarboxylation of L-aspar-tate.

S. typhimurium LT2 strains have aspartate 1-decarboxylase levels similar to those of E. coli(Table 3). The enzyme levels of the panD mu-tants of S. typhinurium (22)- were not deter-mined because these strains could not be ob-tained. It should be noted that Williamson andBrown (39) found that E. coli M99-2 had onlyabout 10% of the decarboxylase activity found inE. coli W.The level ofaspartate 1-decarboxylase activity

in strain AB352 was unaffected by Aupplemen-tation of a defined growth medium with f8-ala-nine or a mixture of/-alanine, pantoic acid, andpantothenate (Table 3). Therefore, the decar-boxylase, like other pantothenate biosynthe, i-enzymes (32), is not repressed by end productr.panB and panC auxotrophs of E. coli had nor-mal decarboxylase levels (Table 3).Genetic mapping of the panD locus. The

pan auxotrophs of E. coli K-12 had not been

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 141, 1980

CH3 0 1. (PMnS)I

H-C - C-COOH

CH, THF

a-ketolsovle*ric acid THF

,B-ALANINE SYNTHESIS 1295

HiHi R 2. H JHS H

HO-C -C - C-COOH HO -C-C -c -COOHH CH, NADPH NADP H CH, H

ketopontoic ocid pontoic ocid

3(nC/.> H CHOHO H HHO-C-C C-C- N-C-C-COOH

ATP PPI H CH5H H H H

COOHICH2

HC-NH2

COOH

oasprtic oci

COOH4. (panD CH

COt HC-NH2

H

id

pontothenic acid

.8-alanWoFIG. 2. Pantothenate synthesis in E. coli. The reaction steps are numbered, and the gene loci believed to

code for the enzymes catalyzing the steps are given in parentheses. Step 1 is catalyzed by ketopantoatehydroxymethyl transferase; step 2 could occur on the free acid (as drawn) by using ketopantoate reductase oras the lactone with ketopantoyl lactone reductase. Step 3 is catalyzed bypantothenate synthetase, and step 4is the aspartate 1-decarboxylase reaction. The systematic name for ketopantoic acid is 2-keto-4-hydroxy-3,3-dirmethylbutyric acid. See the text for discussion of this scheme.

classified as thoroughly as the pan auxotrophsof S. typhimurium and E. coli W. We found thatthe Coli Genetic Stock Center collection consistsof one panB auxotroph (responding to pantoicacid and ketopantoic acid) and two panC auxo-trophs in addition to the threepanD auxotrophs.The E. coli K-12 strains were clasified by nu-tritional tests and by cross-feeding experimentswith the E. coli W and S. typhimurium auxo-trophs.Both the panB and panC loci were known to

map between the tonA and dapD loci (min 3)on the E. coli genetic map (2). ThepanA,panB,andpanCloci ofS. typhimurium map in a clusterin the analogous place on the map of that orga-nism (9, 27). However, because the panD locusof S. typhimurium has recently been reported tomap at min 89 (22), we mapped the panD locusof E. coli K-12.

Conjugational crosses of phenocopies of thethreepanD strains with HfrH and HfrC strainsindicated that thepanD locus was closely linkedto thepanB andpanC loci. This was confirmedby P1 cotransduction of the panD, panB, andpanC loci. P1 phage grown on strain AB354(panD) was used to transduce strain CY257(panB) to panB+ in the presence of fl-alanine.All 73 transductants were pan strains whoserequirement was met by,B-alanine. An analogousexperiment gave 95% cotransduction of panBwith panC. Conjugational crosses of the HfrpanD strains AB353, AB354, and AB355 withthe panC strain AT1371 all gave a >98.5% link-age between panD and panC, whereas linkageofpanD with the leu locus at min 1.5 was onlyabout 60%. The results therefore indicate thatthepanB,panC, andpanD loci are tightly linkedand possibly comprise an operon.

TABLE 3. Aspartate 1-decarboxylase levels invarousstrains

Sp actcStrain Charcteristics (U/mg of

protein)

KL16 pan+ 2.9AB352 pan+ 3.1AB354 panD <0.005AB353 panD <0.08AB355 panD <0.06CY266 pan+ tranaductant of 3.2

AB354CY267 pan + revertant of AB355 4.8AB1371 panC 3.4CY257 panB 5.5AB352 Defined medium' 3.6AB352 Defined medium + I- 3.8

alanineAB352 Defined medium + 8- 2,7

alanine, pantoic acid,and pantothenate

AB352 Minimal mediumb 3.3E. coli W pan+ 4.2M99-1 panC of E. coliW 1.2M99-2 panD of E. coliW <0.05SA965 pan+ of S. typhimurium 4.8

LT2

'The defined medium consisted of medium E (36)supplemented with glucose, 0.4%; adenine, 20 ug/ml;thiamine, 1 tg/ml; and vitamin-free casein hydroly-sate, 0.2%. Pantothenate, pantoic acid, and,B-alanine,when present, were added at 0.01%. All other strainswere grown in R broth (7).

bMinimal medium is the defined medium with thecasein hydrolysate replaced with 20 jg each of threo-nine and leucine per mL

'Specific activity of crude extracts; the assay wasperformed at a suboptimal pH and aspartate concen-tration (see text). Units are picomoles minute-' deter-mined at 370C.

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

1296 CRONAN

Lack ofresponse to dihydrouracil. In con-flict with the previous results (29, 30), we findthat none of the E. coli panD auxotrophs re-spond to dihydrouracil. This has also been ob-served for the S. typhimurium strains (22). TheE. coli strains do respond to ,-ureidopropionate(N-carbamyl f,-alanine), but this can be attrib-uted to the spontaneous hydrolysis of the car-bamyl group in the growth medium.

DISCUSSIONThe results presented in this paper and that

of Williamson and Brown (39) demonstrate thatthe synthesis of ,B-alanine in E. coli proceeds bya-decarboxylation of L-aspartate.panD mutantsare deficient in aspartate 1-decarboxylase, andthe enzyme defect seems due to a single muta-tional event. These data, coupled with the factthat dihydrouracil does not support the growthof the E. coli and S. typhimurium panD strains,indicate that the purine degradation pathwayproposed by Slotnick and Weinfeld (29, 30) canbe discarded as a major pathway of ,6-alaninesynthesis.

Aspartate 1-decarboxylase could be a rate-lim-iting step in the synthesis of pantothenate. Theactivity ofthe enzyme should be maximal in vivobecause the intracellular level of pyruvate (400uM) is fivefold greater than the Michaelis con-stant of the decarboxylase for aspartate (15).Under optimal conditions (pH 7.5, saturatingaspartate), a crude extract of 10' E. coli cells hassufficient decarboxylase activity to form about0.8 to 1 ttmol of ,-alanine per h at 370C. This isquite similar to the amount of,B-alanine neededfor 10' cells to double in mass (0.7 to 0.9 junol)(34).These considerations suggest that the forma-

tion of fl-alanine could be the rate-limiting stepin the formation of pantothenate. Powers andSnell (25) have suggested on the basis of endproduct inhibition experiments that ketopantoichydroxymethyl transferase activity could also belimiting, whereas pantothenate synthetase activ-ity seems to be present in a large excess (16, 19,24). The rate-limiting step in pantothenate syn-thesis is probably also the rate-limiting step inCoA synthesis because supplementation of E.coli growth media with pantothenate increasesthe intracellular CoA pool by 10- to 20-fold (1,5).The present work, when coupled with recent

biochemical data, gives a coherent picture of thepantothenate biosynthetic pathway in E. coli(Fig. 2). Apparently the same pathway functionsin S. typhimuriun. However, few of the geneticand biochemical data on this latter organismhave been published. Most of the results areavailable only in summary form (9).

Ketopantoic acid, the first intermediate in thepantoic acid arm of the pantothenate pathway(Fig. 2), is formed from a-ketoisovaleric acid andformyl tetrahydrofolate by the enzyme ketopan-toate hydroxymethyl transferase (25, 32). Thisenzyme is probably coded by the panB genebecause a panB mutant of E. coli W lacks thisenzyme (32). The ketopantoic acid is then re-duced to pantoic acid either directly or afterlactonization to ketopantoyl lactone (37, 38).Reductases for both the acid and the lactonehave been demonstrated in E. coli and, further-more, two forms of the ketopantoyl lactone re-ductase have been reported (37, 38). This mul-tiplicity of enzymes may explain why no panmutants able to use pantoic acid, but not keto-pantoic acid, have been isolated.The pantoic acid formed by reduction is then

condensed with fi-alanine to form pantothenate(16, 19, 24). The condensing enzyme, pantothen-ate synthetase, appears to be the product of thepanC gene because apanC mutant of E. coliWlacks this enzyme (16). The genetic and enzy-mological studies are therefore in good agree-ment except for the panA mutant of S. typhi-murium (9) (no panA mutants are known in E.coli).The panA mutant responds to a-ketoisoval-

eric acid and valine as well as ketopantoic acid,pantoic acid, and pantothenate (9). Demerecand co-workers, therefore, proposed that thismutant is defective in the synthesis of a-ketoiso-valeric acid (9). This proposal does not seemlikely, as such a block would engender require-ments for valine and leucine as well as panto-thenate (33). It seems more likely that panAmutants produce an altered ketopantoate hy-droxymethyl transferase with a decreased affin-ity for a-ketoisovalerate and thus the panA mu-tant might be a type ofpanB auxotroph. Studiesto test this hypothesis are in progress.Williamson and Brown (39) reported two dis-

tinct forms of aspartate 1-decarboxyle thatdiffered greatly in both size and charge. How-ever,panD mutants lacked both forms (Table 3;39), and thus the two forms are related. Perhapsthey share a common subunit coded by thepanD gene. The major form of decarboxylaseappears to be composed of three dissimlar sub-units (39), and the large size of the minor com-ponent suggests a subunit structure.

Circumstantial evidence presented in this pa-per suggests that the minor form may be apyridoxal phosphate enzyme. It would be mostinteresting if E. coli does contain two enzymeswith different prosthetic groups, particularly be-cause the relative amounts of the two decarbox-ylase forms depend on the composition of themedium in which the cells are cultured (39).

J. BACTERIOL.

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

V,-ALANINESYNTHESIS 1297

ACKNOWLEDGMENTSThis work was supported by Public Health Service grant

AI 16650 from the National Institute of Allergy and InfectiousDiseases.

I thank Barbara Bachmann for her advice and the gifts ofstrains.

LITERATURE CITED1. Alberts, A. W., and P. RI Vagelos. 1966. Acyl carrier

protein. VIII. Studies of acyl carrier protein and coen-zyme A in Eschenichia coi pantothenate or ,-alanineauxotrophs. J. Biol. Chem. 241:5201-5204.

2. Bachmsnn, B. J., K. B. Low, and A. L Taylor. 1976.Recalibrated linkage map of Eswherichia coli K-12.Bacteriol. Rev. 40:116-167.

3. Boeker, E. A., and E. S. Snell. 1972. Amino acid decar-boxylases, p. 217-263. In P. Boyer (ed.), The enzymes,vol. VI, Academic Pres Inc., New York.

4. Britten, R. G., and F. T. McClure 1962. The amino acidpool in Ewcherichia coli. Bacteriol. Rev. 26:292-335.

5. Brown, G. M. 1959. The metabolism of pantothenic acid.J. Biol. Chem. 234:370-380.

6. Brown, G. M. 1970. Biosynthesis of pantothenic acid andcoenzyme A, p. 73-80. In M. Florkin and E. H. Stotz(ed.), Comprehensive biochemistry, vol. 21. Elsevier,Amsterdam.

7. Cronan, J. E., Jr., D. F. Silbert, and D. L Wulff. 1972.Mapping of the fabA locus for unsaturated fatty acidsynthesis in Escherichia coli. J. Bacteriol. 112:206-211.

8. David, W. E., and H. C. LUchatein 1950. Aspartic aciddecarboxylase in bacteria. Proc. Soc. Exp. Biol. Med.73:216-218.

9. Demerec, M. E., E. L. Lahr, E. Balbinder, T. Miyake,C. Mack, D. Mackay, and J. Iahidu. 1959. Bacterialgenetics. Carnegie Inst. Washington Yearb. 58:433440.

10. Frank, M. J., P. G. Trager, J. B. Johnson, and S. H.Rubin. 1973. Characterization of crystalline calciumpantothenate by physiochemical asay methods. J.Pharm. Sci. 62:108-111.

11. Greenberg, D. M., and K. Rothateln. 1957. Methods forchemical synthesis, isolation, and degradation oflabeledcompounds. Methods Enzymol. 4:662-731.

12. Klosterman, H. J. 1979. Vitamin B, antagonists of nat-ural origin. Methods Enzymol. 62:483-496.

13. Liesa, K., D. Mecke, and H. Holzer. 1966. Zur regulationdu synthese von aspartase in E. coli. Biochem. Z. 346:244-251.

14. UApton, A. H., and F. M. Strong. 1949. Synthesis ofcompounds related to pantothenic acid. J. Am. Chem.Soc. 71:2364-2367.

15. Lowry, 0. H., J. Carter, J. B. Ward, and L Glaser.1971. The effect of carbon and nitrogen sources on thelevel of metabolic intermediates in Ewcherichia coli. J.Biol. Chem. 246:6511-6521.

16. Mass, W. K. 1952. Pantothenate studies. III. Descriptionof the extracted pantothenate-synthesizing enzyme ofEacherichia coli. J. Biol. Chem. 198:23-32.

17. Mass, W. K., and B. D. Davis. 1950. Pantothenatestudies. I. Interference by D-serine and L-aspartic acidwith pantothenate synthesis in Escherichia coli. J.Bacteriol 60:733-745.

18. Marcus, KL, and Y. S. Halpern. 1969. Mapping of theaspartase gene in Escherichia coli K-12. Isr. J. Med.Sci. 5:413-415.

19. Mlyatake, K., Y. Nakano, and S. Kitaoka. 1979. Pan-

tothenate synthetase from Escherichia coli. MethodsEnzymol. 62:215-219.

20. Munkres, K. D., and F. KL Richards. 1965. The purifi-cation and properties of Neurospora maleate dehydro-genase. Arch. Biochem. Biophys. 109:466-479.

21. Niederwieser, A. 1972. Thin-layer chromatography ofamino acids and derivatives. Methods Enzymol. 25:60-99.

22. Ortega, K. V., A. Cardena, and D. UbLiera. 1975.panD, a new chromosomal locus of SahnoneUa typhi-murium for the biosynthesis of ,B-alanine. Mol. Gen.Genet. 140:159-164.

23. Parikh, I., S. March, and P. Cuatreca . 1974. Topicsin the methodology of substitution reaction with aga-rose. Methods Enzymol. 34B:77-102.

24. Pflelderer, G., A. Krelllng, and T. Wielad. 1960. tYberPantothen-sauresynthetase aus E. coil. Biochem. Z.333:302-307.

25. Powers, S. G., and E. E. Sneil. 1976. Ketopantoatehydroxymethyl-transferase. II. Physical, catalytic, andregulatory properties. J. BioL Chem. 251:3786-3793.

26. Roberts, R. B., P. H. Abelson, P. B. Cowle, E. T.Bolton, and R. J. Britten. 1963. Studies of biosyn-thesis in Escherichia coli. Carnegie Inst. Washington.Publ. 607:247-272.

27. Sanderson, K. E., and P. E. Hartman. 1978. Linkagemap of SalmoneUa typhinwrium, edition V. Microbiol.Rev. 42:471-519.

28. Shlve, W., and J. Macow. 1946. Biochemical tranfor-mation as determined by competitive analogue-me-tabolite growth inhibition. L. Some transformation in-volving aspartic acid. J. Biol. Chem. 162:451-462.

29. Slotnick, I. J. 1956. Dihydrouracil as a growth factor fora mutant strain of Escherichia coli. J. Bacteriol. 72:276-277.

30. Slotick, I. J., and H. Weinfeld. 1967. Dihydrouracil asa growth factor for mutant strains of Escherichia coli.J. Bacteriol. 74:122-126.

31. Snell, E. E. 1977. Pyruvate-containing enzymes. TrendsBiochem. Sci 2:131-136.

32. Teller, J. H., S. G. Powers, and E. E. Snell. 1976.Ketopantoate hydroxymethyltransferase. I. Purifica-tion and role in pantothenate biosynthesis. J. Biol.Chem. 251:3780-3785.

33. Umbarger, H. E. 1975. Regulation of amino acid biosyn-thesis in microorganisms, p. 1-56. In H. R V. Cunstein(ed.), MTP international review of science vol. 7. But-terworths, London.

34. Van den Bosch, H., J. R. Williamson, and P. R.Vagelos. 1970. Localization of acyl carrier protein inEscherichia coli. Nature (London) 228:338-341.

35. Virtanen, A I., and T. Laine. 1937. The decarboxylationof D-lysine and L-aspartic acid. Enzymologia 8:266-270.

36. Vogel, H. J., and D. K. Bonner. 1956. Acetylorithinaseof Escherichia coil: partial purification and some prop-erties. J. Biol. Chem. 218:97-106.

37. Wilken, D. R., H. L King, and RI E. Dyar. 1975.Ketopantoic acid and ketopantoyl lactone reductases.J. Biol. Chem. 250:2311-2314.

38. Wilken, D. R., H. L King, and R. E. Dyar. 1979.Ketopantoyl lactone reductases. Methods Enzymol. 62:209-215.

39. Williamson, J. M., and G. M. Brown. 1979. Purificationand properties of L-aspartate-a-decarboxylase, an en-zyme that catalyses the formation of,B-alanine in Esch-erichia coli. J. Biol. Chem. 254:8074-8082.

VOL. 141, 1980

on March 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from