JOUTRNAI OFk BACTERIOLOGY, Oct. 1979, p). 240-24500)21-919;3/79/10-0240/06$02.(X)/O

\ol. 140, No. I

Hydroxy Amino Acid Metabolism in Pseudomonas cepacia:Role of L-Serine Deaminase in Dissimilation of Serine,

Glycine, and ThreonineH. C. WONG AND T. G. LESSIE*

Department of Microbiology, UnWiuersity of Massachusetts, Amherst, Massachusetts 01003

Received for publication 23 JulyN 1979

Growth of Pseudomonas cepacia (P. multitorans) on serine depended uponinduction of a previously undescribed L-serine deaminase distinct from threoninedeaminase. Formation of the enzyme was induced during growth on serine,glycine, or threonine. The induction pattern reflected a role of the enzyme incatabolism of these three amino acids. Both threonine and glycine supportedgrowth of serine auxotrophs and were presumably converted to serine andpyruvate in the course of their degradation. Mutant strains deficient in serinedeaminase, or unable to use pyruvate as a carbon source, failed to utilize serineor glycine and grew poorly with threonine, whereas strains deficient in threoninedehydrogenase or a-amino ,B-ketobutyrate:coenzyme A ligase (which togetherconvert threonine to glycine and acetyl coenzyme A) failed to utilize threonine orderepress serine deaminase in the presence of this amino acid. The results confirmfor the first time the role of a-amino /8-ketobutyrate:coenzyme A ligase inthreonine degradation and indicate that threonine does not mimic serine as aninducer of serine deaminase.

Pseudomonas cepacia (P. multitvorans) is themost nutritionally versatile of the pseudomo-nads (1, 9, 1 1). We have examined the possibilitythat the exceptional versatility of this bacteriummight be related to the evolution of broad-spec-ificity enzymes. Our experiments have focusedon enzymes related to metabolism of branched-chain and hydroxy amino acids (4, 13). In thecase reported here, we have considered whetherthe atypical properties of the isoleucine-sensitivethreonine (serine) deaminase of P. cepacia (4,12, 14) reflect a role of this enzyme in utilizationof hydroxy amino acids as well as in biosynthesisof isoleucine. The results indicate clearly thatthreonine deaminase does not participate in hy-droxy amino acid catabolism. Our experimentsdefine the role of a serine deaminase in degra-dation of both serine and threonine. The presentanalysis of enzymes involved in hydroxy aminoacid utilization and related studies of branched-chain amino acid degradation indicate that thebiochemical diversity of P. cepacia dependsupon its extensive repertoire of enzymes ratherthan the use of broad-specificity enzymes.

MATERIALS AND METHODSBacterial strains. P. cepacia 249 was originally

obtained from R. Y. Stanier (11). All other strainsused in this investigation were derived in our labora-tory from strain 249 and are listed in Table 1. Theindicated auxotrophic derivatives and strains unable

to utilize particular organic compounds as the carbonsource were obtained by the D-cycloserine selectionprocedure described earlier (4).

Preparation of bacterial extracts. Bacteria weregrown in 1-liter flasks which contained 200 ml ofminimal salts medium consisting of 1 mM MgSO., 0.1mM CaCl2, 0.01 mM FeSO1, and pH 6.5 phosphatebuffer (33 mM Na2HPO1, and 67 mM KH2PO.1) sup-plemented with the indicated carbon source. The bac-teria were collected by centrifugation, suspended in 5ml of 20nmM phosphate buffer (pH 6.8) containing 20pg of pyridoxal phosphate per ml and 10 mM 2-mer-captoethanol, and disrupted by sonic treatment asreported earlier (4). I'he disrupted-cell suspensionswere centrifuged for 10 min at 12,000 x g, and thesupernatant fractions were assayed for enzyme activ-itv.Enzyme assays. All enzyme activities are ex-

pressed in terms of nanomoles of product formed perminute per nilligram of protein. Protein was deter-mined by the procedure of Lowry et al. (5).

L-Serine and L-threonine deaminase activities weredetermined at 37°C. The assay mixtures contained ina volume of 0.5 ml: 200 mM TI'ris-hydrochloride buffer(pH 8.5), 0.2 mM pyridoxal phosphate, enzyme prep-aration, and 100 mM L-serine or L-threonine (adjustedto pH 8.5 with NaOH). After 20 to 30 min of incuba-tion, the reactions were terminated bv addition of 0.5ml of 15 (wt/vol) trichloroacetic acid, and pyruvateor a-ketobutyrate formed in the course of the reactionwas determined as the dinitrophenylhydrazone (4).

L-Threonine dehydrogenase activity was deter-mined at 24°C by monitoring NAD reduction whichoccurre(d during the conversion of I,-threonine to (-

240

on May 31, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. CEPACIA SERINE DEAMINASE 241

TABLE 1. List of strainsStraindeSignation Characteristicsdesignation

P. cepacia 249(ATCC 17616)

249- 13

13-5

13-8

13-5-1

249-16

249- 18

249-23

249-30

Wild type, obtained from R. Y.Stanier (I 1)

Isoleucine auxotroph derived fromstrain 249; deficient in threoninedeaminase (4)

Lysine auxotroph derived from strain13; fails to use diaminopimelic acidin place of L-lysine

Serine auxatroph derived from strain13; deficient in phosphoserineaminotransferase

Serine auxotroph derived from strain13-5, possesses normal levels ofphosphoserine aminotransferaseand phosphatase; presumablyblocked early in serine biosynthesis

Threonine dehydrogenase-deficientderivative of strain 249; unable toutilize threonine as the sole carbonsource

Serine deaminase-deficient derivativeof strain 249; unable to utilize L-serine as the sole carbon source

a-Amino ,3-ketobutyrate:coenzyme Aligase-deficient derivative of strain249; unable to utilize L-threonine asthe sole carbon source

Unable to utilize pyruvate or glucoseas the sole carbon source; growsnormally on acetate or citrate;presumably blocked at level ofpyruvate dehydrogenase

amino ,B-ketobutyrate (4). The assay mixtures con-tained, in 1 ml: 100 mM Tris-hydrochloride buffer (pH8.5), 0.5 mM NAD, enzyme preparation, and 100 mML-threonine (adjusted to pH 8.5 with NaOH).

a-Amino fl-ketobutyrate:coenzyme A (CoA) ligase-dependent condensation of acetyl-CoA and glycine toform a-amino /8-ketobutyrate was determined by mea-suring aminoacetone produced from a-amino ,8-keto-butyrate (4, 7). The a-amino 83-ketobutyrate:CoA li-gase (aminoacetone synthase) assay mixtures con-tained, in 0.5 ml: 200 mM Tris-hydrochloride buffer(pH 8.5), 1 mM acetyl-CoA, enzyme preparation, and300 mM glycine. After 20 min of incubation at 37°C,the reactions were terminated by addition of 0.5 ml of15% (wt/vol) trichloroacetic acid, and aminoacetonewas determined as reported earlier (4).

Resolution of the catabolic serine deaminaseand threonine deaminase. To obtain preparationsof catabolic serine deaminase and threonine deami-nase suitable for comparison of pH optima and kineticand other properties of the two enzymes, bacteria from10 200-ml cultures of strain 249 grown with 0.4% (wt/vol) L-serine were collected by centrifugation, sus-pended in 30 ml of 20 mM phosphate buffer (pH 6.8)containing 20 jg of pyridoxal phosphate per ml and 10

mM 2-mercaptoethanol, and disrupted by sonic treat-ment. Cell debris and unbroken cells were removed bycentrifugation of the disrupted-cell suspension for 15min at 12,000 x g. An equal volume of 2% (wt/vol)streptomycin sulfate dissolved in the above-mentionedbuffer was added to the supernatant to precipitatenucleic acids, and the precipitated material was re-moved by centrifugation. Solid ammonium sulfate (25g/100 ml) was added to the supernatant fraction, andthe precipitate, which contained the bulk of the serineand threonine deaminases present in the crude bac-terial extract, was suspended in ca. 10 ml of pH 6.8phosphate buffer containing pyridoxal phosphate andmercaptoethanol. The preparation was dialyzedagainst 200 volumes of the same buffer and supple-mented with 10% (wt/vol) glycerol, and between 3 and4 ml (100 mg of protein) was applied to a column (2 by20 cm) of hydroxylapatite equilibrated with the samebuffer. The column was eluted with 280 ml of a lineargradient of 20 to 200 mM phosphate buffer (pH 6.8)containing 10% (vol/vol) glycerol, 20,tg of pyridoxalphosphate per ml, and 10 mM 2-mercaptoethanol.Fractions (3.5 ml each) were collected and assayed forL-threonine and L-serine deaminase activities. Thecatabolic serine deaminase eluted first at a phosphateconcentration of 70 mM; L-threonine deaminase elutedat a phosphate concentration of 150 mM. Fractionscontaining the catabolic serine deaminase were pooled,as were those containing L-threonine deaminase activ-ity, and used for studies of kinetic and other propertiesof the two enzymes. The recovery of threonine deam-inase activity was about 90% of the activity present inthe crude bacterial extracts; about 40% of the catabolicserine deaminase activity was recovered. If glycerolwas omitted during chromatographic resolution of thetwo enzymes, the recovery of catabolic serine deami-nase activity was only 10%. The presence of glyceroldid not influence the recovery of threonine deaminase.The extent of purification of the catabolic serine de-aminase and of threonine deaminase was, respectively,28- and 40-fold.The catabolic serine deaminase was sufficiently un-

stable to preclude its further purification. The prepa-rations obtained after hydroxylapatite chromatogra-phy lost approximately 50% of their activity after 2days in storage at 0'C. Less-purified preparations suchas crude bacterial extracts or the ammonium sulfatefractions used for the chromatographic step were com-pletely stable for at least 2 weeks when stored undersimilar conditions.

Estimation of enzyme molecular weight. Theapparent molecular weights of serine and threoninedeaminases were determined by comparing their sed-imentations in sucrose gradients with that of bovinehemoglobin as described by Martin and Ames (6). Thegradients contained 20 mM phosphate buffer (pH 6.8)and 20 pg of pyridoxal phosphate per ml.

Chemicals. All chemicals were obtained fromSigma Chemical Co., St. Louis, Mo.

RESULTSP. cepacia 249 utilizes the L-forms of serine

and threonine but not the D-forms of theseamino acids as the sole carbon source. In pre-vious experiments, we investigated whether the

VOL. 140, 1979

on May 31, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

242 WONG ANI) LESSIE

atypical threonine deaminase of P. cepaciaplayed a role in threonine catabolism and con-cluded on the basis of mutant studies that it didnot (4). P. cepacia grows more rapidly on L-serine (generation time, 2.5 h) than on L-threo-nine (generation time, 5.5 h). To determinewhether the serine deaminase activity associatedwith threonine deaminase was involved in serinecatabolism, we tested whether P. cepacia 249-13, an isoleucine auxotroph devoid of threoninedeaminase activity (4), could utilize serine. Thisstrain grew normally in medium containing 40pg of L-isoleucine per ml and 0.4% L-serine as thesole carbon source, indicating that threoninedeaminase is not essential for serine utilization.This led us to investigate further the enzymaticbasis of serine utilization.Growth of P. cepacia 249 in medium contain-

ing 0.4%. L-serine as the carbon source resultedin an approximate sevenfold increase in serinedeaminase-specific activity compared to bacteriagrown on alternate carbon sources such as cit-rate or glucose or phthalate (Table 2). The in-crease in specific activity was due to inductionof a previously undescribed catabolic L-serinedeaminase distinct from threonine deaminase.The catabolic enzyme was readily resolved fromthreonine deaminase on the basis of its slowersedimentation in sucrose gradients and its rela-tively weak binding to hydroxylapatite columns(see above). Its molecular weight was estimatedto be 60,000 by comparison of its sedimentationbehavior with that of bovine hemoglobin (64,500molecular weight) and threonine deaminase(190,000 molecular weight). Extracts of the thre-onine deaminase-deficient strain 249-13 (pre-pared from serine-grown bacteria provided withisoleucine) contained only the catabolic serinedeaminase. This enzyme was inactive with L-threonine. (Under our conditions of assay thelimit of detection for activity with L-threonine

was equivalent to about 0.5Zcr of the activityobserved with L-serine.) In this respect, it dif-fered from threonine (serine) deaminase, theactivity of which with L-threonine was sevenfoldgreater than with serine. There was also amarked difference in response of the two en-zymes to the threonine deaminase feedback in-hibitor L-isoleucine. Threonine deaminase ex-hibited no significant serine deaminase activityin the presence of 10-2 M L-isoleucine. The ac-tivity of the catabolic serine deaminase was un-affected by 10-2 M L-isoleucine.The distinguishing features of P. cepacia ser-

ine and threonine deaminases, including kineticproperties, pH optima, and responses to isoleu-cine, are summarized in Table 3. The findingthat the activity of the catabolic serine deami-nase was not influenced by L-isoleucine afforded

T'ABLE 2. Induction of L-serine deaminase in strain249

Enzvyne sp act"

sabonurce Generation -Threo-Carbon so " timne (m) L-Serine de- nine deami-amninase' naseGlucose 0 229 694Citrate 60 235 704Phthalate 65 215 685L-Threonine 300 1,482 512L-Serine 150 2,319 510Glycine 540 1,733 493Casamino acids 60 235 535Ile + Val + Leu 70 202 615

" The bacteria were grown at ;37°C in minimal salts mediumsupplemented with the following amounts of each carbonsource: glucose, citrate, and phthalate, 0.57% (wt/vol); glycine,threonine, and serine, each at 0.4%c (wt/vol); Casamino Acidsat 1%' (wt/vol); and branched-chain anmino acids, each at 0).2%.

Nanomoles of pyruvate or a-ketobutyrate formed perminute per milligram of protein at 37°C.

" Determined in the presence of 10( 2 M L-isoleucine andtherefore does not include serine deaminase activity associatedwith threonine deaminase.

TABLE 3. Properties of L -threonine and L -serine dlearninasesExtenit ot ni

,i-Threonine

Enzrme' Apparent pH optirm biti of inhi- - cK,0 Kornol wt MI-iOCcO MM I - Isoleuc irleeSerinie deaminase 60(),O) Broad region of optimal None 2 x 10 2"

activity between pH 6.5and 9.5'

'I'hreorinle deamitiase 19(,(00 With serine, optimal Complete 6 x 10 2 x 1I) 2activitv was betweenpH 8.5 and 9.5"

T'he preparations used in these determinations were resolvecl on hydroxylapatite columns as described inthe text.

Activity at pH 7 was equivalent to 105%' of that at pH 8.5.T'he K,,, value at pH 7.0 was identical to that at pH 8.5.Acttivity with 1.-serine at )H 7 was olly :3'tof that at pH 8.5.

J. BACRTFRIOL.

on May 31, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. CEPACIA SERINE DEAMINASE 243

a rapid means of determining the amount of thisenzyme in crude bacterial extracts. About 95%of the serine deaminase activity of serine-grownbacteria and 65% of the activity of glucose- orcitrate-grown bacteria was insensitive to inhibi-tion by 10-2 M L-isoleucine. Thus, even the basallevel of catabolic serine deaminase found in cit-rate- or glucose-grown bacteria contributedmore to the total serine deaminase of P. cepaciaextracts than did the serine deaminase associ-ated with threonine deaminase, and in the caseof fully induced cells, the contribution of threo-nine deaminase was relatively insignificant.Induction of L-serine deaminase during

growth on L-threonine or glycine. As notedabove, the catabolic L-serine deaminase was in-active with L-threonine. It was therefore surpris-ing to discover that the enzyme was inducedduring growth on threonine (Table 2). This par-adox was resolved when we were able to confirmby the mutant studies described below that ca-tabolism of threonine proceeds by its stepwiseconversion to a-amino f3-ketobutyrate andthence to acetyl CoA and glycine. Metabolismof the glycine derived by cleavage of a-amino/3-ketobutyrate appears to require its conversionto serine, which is further metabolized to pyru-vate by the action of serine deaminase. Thisexplains the observed induction of serine deam-inase when glycine served as the carbon source(Table 2). The proposed route of threonine ca-tabolism (Fig. 1) is consistent with the observedpattern of induction of serine deaminase duringgrowth on threonine, serine, or glycine. Themutant studies outlined below rule out the pos-sibility that threonine directly mimics serine asan inducer of serine deaminase and support theinterpretation that the physiological inducerduring growth on threonine is serine formed viaglycine.Enzyme deficiencies affecting utilization

of L-serine and L-threonine. We have used aD-cycloserine selection procedure (4) to obtainmutants blocked in the utilization of L-serine orof L-threonine. One such strain, 249-18, whichproved to be deficient in serine deaminase, failedto grow on serine or glycine and grew slowly onthreonine (27-h generation time compared with5.5 h for the wild type). The level of serinedeaminase activity in extracts of bacteria grownon threonine or combinations of threonine andalternate carbon sources such as citrate or glu-cose was less than 2% of that found in the wildtype grown under similar conditions (Table 4).Strain 249-18 had normal levels of threoninedeaminase and of the isoleucine-sensitive serinedeaminase activity associated with this enzyme(data not shown). The results confirm that thecatabolic serine deaminase is essential for serine

L -Thr-L aKB - LL-IIe

AKB

Gly + AcCoA

~4L-Ser Pyr

FIG. 1. Pathways of threonine metabolism in P.cepacia. Threonine deaminase (enzyme 1) and en-zymes required to convert a-ketobutyrate to isoleucinewere formed constitutively. Enzymes of the threoninedegradation pathway, which included L-threoninedehydrogenase (enzyme 2), a-amino 8-ketobutyrate:CoA ligase (enzyme 3), and L-serine deaminase (en-zyme 4), were induced during growth on L -threonine.

degradation and that the serine deaminase as-sociated with threonine deaminase cannot sub-stitute for loss of this enzyme.We obtained two classes of mutants unable to

utilize L-threonine as the carbon source. Thefirst was deficient in threonine dehydrogenase,the NAD-specific enzyme which converts thre-onine to a-amino /8-ketobutyrate (2, 4, 7, 8). Thesecond class was deficient in a-amino 13-ketobu-tyrate:CoA ligase, which converts a-amino ,B-ketobutyrate to glycine and acetyl-CoA. Dataindicating the levels of these two enzymes inwild-type and mutant bacteria are presented inTable 4. Strain 249-16 is representative of thefirst class; strain 249-23 is a member of thesecond class. Both the threonine dehydrogenaseand a-amino ,8-ketobutyrate:CoA ligase-defi-cient strains failed to form more than basal levelsof catabolic serine deaminase when grown inmedium containing threonine in combinationwith citrate as a carbon source. Growth of thewild type in this medium resulted in formationof elevated levels of serine deaminase. Membersof both classes of mutants grew normally withL-serine as the carbon source and produced wild-type levels of catabolic serine deaminase underthese conditions. We interpret the results asindicating that threonine degradation must pro-ceed at least to the level of glycine and acetyl-CoA formation for threonine to influence for-mation of serine deaminase. The simplest inter-pretation is that in this case the physiologicalinducer of the catabolic serine deaminase is ser-ine formed from glycine.

It should be noted that the threonine dehy-drogenase-deficient strain 249-16 formed wild-type levels of a-amino 83-ketobutyrate:CoA li-gase when grown on a combination of threonineand citrate (Table 4) and that the a-amino /3-ketobutyrate:CoA ligase-deficient strain 249-23

VOL. 140, 1979

on May 31, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

244 WONG ANI) LESSIE

IABiF: 4. Leuels of enzyvmnes relate(d to serine a(nd threonine degradation in tidd-tvpe and mutant strai'ns ofP. cepacia

Strain Enznme (leficiencv

249 (wild Nonetype)

249-18 Serine deaminase

249-1 6 lThreonine dehydrogenase

249-2; a-Amino fl-ketobutyrate:CoAligase

Enzyme Sp act

Carbonsource-Serine dleamn-Carbon"' source ilise l-Threonine fl-Aninio /1-ke-dehydrogenase tobutvrate:toA ligase

i-Thr 1,489 61 :31,-Thr + Cit 642 29 17i-Ser 2:319 3

VOL. 140, 1979

dicating that it does not play a secondary role inserine catabolism.The inability of strain 249-18 to utilize glycine

or grow normally on L-threonine implicates thecatabolic serine deaminase in the degradation ofboth of these amino acids. The mutant studiesreported here indicate that threonine and gly-cine are readily converted to serine, which ap-pears to be the physiological inducer of serinedeaminase. The results confirm for the first timethe role of a-amino /3-ketobutyrate:CoA ligase inthreonine degradation, and indicate that threo-nine degradation proceeds by cleavage of a-amino /3-ketobutyrate to glycine and acetyl-CoAand not by decarboxylation of this compound toproduce aminoacetone as postulated earlier (2,4, 7). To our knowledge, neither a-amino /8-ke-tobutyrate:CoA ligase nor threonine dehydro-genase-deficient strains have been described forother bacteria presumed to metabolize threoninevia glycine and acetyl-CoA (2, 7, 8).At the onset of these studies, our working

hypothesis was that the use of broad-specificityenzymes such as threonine deaminase andbranched-chain amino acid aminotransferase forboth the biosynthesis and catabolism of aminoacids might underlie the extreme nutritional ver-satility of P. cepacia. Such roles would be en-tirely consistent with the in vitro properties ofthese enzymes (4, 13). Thus, the failure of thre-onine deaminase to substitute for the catabolicserine deaminase in serine degradation or forthreonine dehydrogenase in threonine degrada-tion was unexpected.The apparent inability of threonine deaminase

to participate in threonine or serine degradationmight be due to the operation of feedback con-trols or, in the case of threonine, degradation toan imbalance in branched-chain amino acid bio-synthesis rather than any inherent inability ofthis enzyme to rapidly convert serine and thre-onine to pyruvate and a-ketobutyrate. For ex-ample, the strong inhibition of threonine deam-inase by isoleucine might preclude its partici-pation in amino acid degradation, or the toxicityof a-ketobutyrate might contribute to poorgrowth on threonine. In the presence of excessa-ketobutyrate, the interaction of this com-pound with acetohydroxyacid synthase to pro-duce acetohydroxybutyrate, a precursor of iso-leucine, presumably competes with interactionof the same enzyme with pyruvate to produceacetolactate, a precursor of valine. P. cepaciautilizes a-ketobutyrate as the sole carbon source,if the bacteria are supplemented with valine.However, threonine dehydrogenase-deficientbacteria still fail to utilize threonine even if theyare supplied with valine. We are examiningwhether mutant strains deficient in catabolic

P. CEPACIA SERINE DEAMINASE 245

serine deaminase or in threonine dehydrogenasecan mutate to utilize threonine deaminase as acatabolic enzyme. We expect that selection ofbacteria able to utilize serine or threonine fromserine deaminase- or threonine dehydrogenase-deficient strains might yield mutants uncoupledwith respect to isoleucine inhibition of threoninedeaminase. In the case of growth on threonine,we anticipate a possible requirement for valine.

ACKNOWLEDGMENTSThis work was supported by U.S. Public Health Service

grant GM21198 from the National Institute of General Medi-cal Sciences.

ADDENDUMWe have reexamined the growth behavior of the threonine

dehydrogenase-deficient strain 249-16 and find that it cangrow slowly in medium containing 0.5% L-threonine if themedium is supplemented with 100 pg of L-valine per ml (14-hgeneration time versus 3.5 h for the wild-type strain 249). Theresults indicate that the failure of P. cepacia threonine de-aminase to play a secondary role in threonine degradation isrelated to the toxicity of a-ketobutyrate produced by thisenzyme from threonine.

LITERATURE CITED

1. Ballard, R. W., N. J. Palleroni, M. Doudoroff, and R.Y. Stanier. 1970. Taxonomy of the aerobic pseudomo-nads: Pseudomonas cepacia, P. marginata, P. allicolaand P. caryophylli. .J. Gen. Microbiol. 60:199-214.

2. Bell, S. C., and J. M. Turner. 1976. Bacterial catabolismof threonine: threonine degradation initiated by L-thre-onine-NAD' oxidoreductase. Biochem. .J. 156:449-458.

:3. Isenberg, S., and E. B. Newman. 1974. Studies on L-serine deaminase in Escherichia coli K-12. J. Bacteriol.118:5:1-58.

4. Lessie, T. G., and H. R. Whiteley. 1969. Properties ofthreonine deaminase from a bacterium able to use thre-onine as sole source of carbon .J. Bacteriol. 100:878-889.

5. 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.

6. Martin, R. G., and B. N. Ames. 1961. A method fordetermining the sedimentation behavior of enzymes:application to protein mixtures. J. Biol. Chem. 236:1:372-1379.

McGilvray, D., and J. G. Morris, 1969. Utilization ofL-threonine by a species of Arthrobacter. A novel cat-abolic role for amino acetone synthase. Biochem. .J.112:657--671.

8. Newman, E. B., V. Kapoor, and R. Potter. 1976. Roleof L-threonine dehydrogenase in the catabolism of thre-onine and synthesis of glycine by Escherichia coli. *J.Bacteriol. 126:1245-1249.

9. Palleroni, N. J., and M. Doudoroff. 1972. Some prop-erties and taxonomic division of the genus Pseudomo-na.s. Annu. Rev. Phytopathol. 10:73-100.

10. Pardee, A. B., and L. S. Prestidge. 1955. Induced for-mation of serine and threonine deaminases by Esche-richia coli. J. Bacteriol. 70:667-674.

11. Stanier, R. Y., N. J. Palleroni, and M. Doudoroff.1966. The aerobic pseudomonads: a taxonomic study. J.Gen. Microbiol. 43:149-171.

12. Umbarger, H. E. 1973. Threonine deaminases. Adv. En-zymol. Relat. Subj. Biochem. 37:349-395.

13. Wong, H. C., and T. G. Lessie. 1979. Branched chainamino acid aminotransferase isoenzymes of Pseudom-onas cepacia. Arch. Microbiol. 120:223-229.

14. Wood, W. A. 1969. Allosteric L-threonine dehydratasesof microorganisms. Curr. Top. Cell. Regul. 1:161-182.

on May 31, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/