JOURNAL OF BACTERIOLOGY, June 1971, p. 972-982Copyright i 1971 American Society for Microbiology

Vol. 106, No. 3Printed in U.S.A.

Biochemical Bases for the Antimetabolite Actionof L-Serine Hydroxamate

TETSUYA TOSA' AND LEWIS 1. PIZER

Department of Microbiology, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104

Received for publication 21 December 1970

The amino acid analogue L-serine hydroxamate, which is bacteriostatic for Esch-erichia coli, has been shown to inhibit protein synthesis. The antimetabolite is a

competitive inhibitor of seryl-transfer ribonucleic acid (tRNA) synthetase with a K;value of 30 uM. Mutants resistant to L-serine hydroxamate have been selected, andthree were shown to have seryl-tRNA synthetases with increased K, values. Onemutant contains a 3-phosphoglycerate dehydrogenase which is insensitive to inhibi-tion by L-serine.

In the preceding paper (20), we identified L-serine hydroxamate as a potentially useful serineanalogue for the study of the control of serinemetabolism (20). The suitability of this analoguedepends on the observation that it rapidly in-hibits growth of Escherichia coli and that its in-hibitory action is specifically reversed by lowlevels of exogenous serine. It is expected, there-fore, that a mutant which has the capacity toincrease its intracellular pool of serine wouldhave lower sensitivity to the action of the antime-tabolite. Serine biosynthesis is subject to end-product inhibition (18, 21), a mechanism for in-creasing the pool size would be to lose this con-trol. An alternative possibility would be to en-hance the activity of the biosynthetic enzymes bylowering their Km values, increasing their Vmaxvalues, or elevating their concentration in the cellby derepressing enzyme synthesis. Resistance tothe analogue could also arise from an alterationin the enzyme system(s) which are inhibited byserine hydroxamate. This type of mutation wouldbe expected to occur frequently, not necessarilybe associated with serine biosynthesis, and oftentake the form of a reduced affinity of the sensi-tive enzyme(s) for the analogue.

There are several ways an amino acid ana-logue can function as an antimetabolite (for re-view, see references 14 and 15), and there is prec-edent for an amino acid hydroxamate acting asa feedback inhibitor as well as producing repres-sion ( 1I).To study the regulation of serine synthesis and

to determine the mechanism by which L-serinehydroxamate inhibits growth, we have investi-

' Present address: Chemical Research Laboratory, TanabeSeiyaku Co., Ltd., Osaka, Japan.

gated the effects of the analogue on cellularprocesses and correlated the ability of mutantorganisms to resist the analogue with alterationsin the properties of their enzymes. The results ofthe experiments reported in this paper identifythe primary site of serine hydroxamate action asthe inhibition of the seryl tRNA synthetase andsubstantiate the proposal that some mutants re-sistant to its action would be altered in the regu-lation of serine synthesis.

MATERIALS AND METHODS

Bacteria and selection of mutants. E. coli K-12 (A)was the laboratory stock culture. It lacks auxotrophicmarkers and is identified as carrying the prophage (A)by its nonpermissive behavior towards T4 RII bacterio-phage. This strain plates T4r+ with the same efficiencyas E. coli strain B. This differential phage sensitivitywas used to check the parentage of the mutants iso-lated. The conditions for growing bacteria and themethods for measuring growth were described in theprevious paper (20).The procedure of Adelberg et al. was used for ob-

taining mutants (1). Bacteria in mid-log phase were fil-tered, suspended in tris(hydroxymethyl)aminomethane(Tris)-maleate buffer (pH 6.0), and treated with N-methyl-N'-nitro-N-nitrosoguanadine (100 ,ug/ml) for30 min at 37 C with gentle shaking. This treatmentkilled 99.5% of the cells. The cells were filtered,washed, and suspended in Brain Heart Infusion me-dium. Incubation in this medium was continued for 3hr at 37 C with aeration. The cells were again filteredand washed with M9 medium; a portion was spread ongradient plates which contained DL-serine hydroxa-mate-hydrochloride at a concentration range from 0 toI mg/ml. Discrete colonies that appeared at the highconcentration of hyxroxamate were picked and subse-quently purified by three subcultures on different levelsof serine hydroxamate. Eight clones showing highlevels of resistance to Ser-Hdx were selected for subse-

972

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF L-SERINE HYDROXAMATE

quent studies. The bacterial strains used in this studyare listed in Table 1.

Chemical compounds and procedures. Adenosine tri-phosphate (ATP), nicotinamide adenine dinucleotide(NAD), DL-serine hydroxamate-hydrochloride, L-leu-cine hydroxamate, L-lysine hydroxamate, DL-threoninehydroxamate, glycine hydroxamate were purchasedfrom Sigma Chemical Co. (St. Louis, Mo.). The fol-lowing compounds were purchased from Schwarz Re-search (Orangeburg, N.J.): E. coli transfer ribonucleicacid (tRNA), '4C-L-serine (stan-star grade, 50jiCi/jimole), and '4C-glycine (stan-star grade, 50jiCi/Amole). 14C-phosphoglycerate (PGA), dithiothre-itol (DTT), and unlabeled amino acids were purchasedfrom Calbiochem (Los Angeles, Calif.) 3H-uracil, "C-L-threonine, and "C-L-leucine were purchased fromNew England Nuclear (Boston, Mass.). L-serine hy-droxamate-hydrochloride was synthesized as describedin the accompanying paper (20), and chloramphenicol(CM) was a gift from Parke Davis Co.

Protein was determined by the method of Lowry etal. (10). The uptake of 'IC-labeled amino acids and3H-uracil into acid-precipitable material was measuredas described by Smith and Pizer (16). The experimentswere performed in the same manner as the growthstudies described in the previous paper (20). After dilu-tion from a Klett reading of 80 to 40, the culture was

divided into growth tubes; 3H-uracil (10 jCi/,jmole)and "C-leucine (1.5 ,uCi/,umole) at final concentrationsof 0.2 and 0.1 mM were added followed by other com-pounds as required. Samples were removed to filterpaper discs at the desired times, and Klett readingswere recorded. Radioactivity was counted either in a

gas flow counter or in a Packard scintillation spectro-photometer, which had an efficiency of 45% for "C-labeled compounds.

3-Phosphoglycerate dehydrogenase. PGA dehydro-genase was assayed in crude extracts by measuring theconversion of radioactive PGA into serine phosphate(12). Overnight cultures grown in minimal media with2 mg of glucose/ml were collected by centrifugation,

washed with cold medium, suspended in 10 mM Tris-hydrochloride buffer (pH 7.5) containing 5 mM DTT (5ml/g, wet weight, of cells), and disrupted by sonic os-cillation for 1.5 min in a sonic oscillatior (Measuring &Scientific Equipment, Ltd.). After removal of debris bycentrifugation, the cell extract was diluted with bufferuntil serine phosphate synthesis was directly propor-tional to the amount of extract added. The componentsof the assay, incubation conditions, and column frac-tionation were the same as previously reported ( 12).One unit of enzyme activity is I jmole of serine phos-phate formed in 30 min at 37 C.Amino acyl-tRNA synthetase. Amino acyl-tRNA

synthetase was assayed by measuring the conversion of"C-labeled amino acids into an acid-precipitable formas described by Katze and Konigsberg (8). The acti-vating enzymes were partially purified by the followingprocedure. Cells grown in minimal media with 2 mg ofglucose per ml overnight were collected by centrifuga-tion, washed with minimal medium, and disrupted bygrinding with alumina. The disrupted cells were ex-

tracted with 20 mM Tris-hydrochloride buffer (pH 7.5)containing 10 mM MgCl2, I mm DTT, and 0.5 mM

ethylenediaminetetraacetic acid (EDTA). After re-

moval of debris by centrifugation at 12,000 x g for 10min, the supernatant was subjected to ultracentrifuga-tion at 105,000 x g for 2 hr. Streptomycin sulfate (0.3volume, 5% weight by volume) was added to precipi-tate the nucleic acids. After 30 min, the precipitate was

removed by centrifugation, and 2 volumes of saturatedammonium sulfate was added to the supernatant toprecipitate the protein. After 30 min, the protein was

collected by centrifugation, dissolved in 20 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM

MgCI2, I mM DTT, and 0.5 mM EDTA. The proteinwas dialyzed against two changes of this buffer over-

night, and the dialyzed preparation was stored frozenat -20 C.

The assay mixture contained in a final volume of 0.3ml: 100 mM Tris-hydrochloride buffer (pH 7.4), 10 mMMgCI2, 10 mM KCI, 2 mM ATP, I mM DTT, 300,g of

TABLE 1. Bacteria used for this study

Strains Origin Characteristics

K- 1 2 (A) S. S. Cohen Nonpermissive for T4rIIK-12 ser/gly S. Simmonds Requires serine or glycine for growth,

lacks PGA dehydrogenaseK-12 (A) met- J. McKitrick 2-Amino-purine mutagenesis of K-12 (A)

requires methionine for growthHfr-C J. S. Gots Met-, relA-lI Iig Nitrosoguanadine mutagenesis of K-12 (X) Completely resistant to 100 Ag of L-

serine hydroxamate per mlC-1 I Id Nitrosoguanadine mutagenesis of K-12 (A) Completely resistant to 100 ,ug of L-

serine hydroxamate per mlD-1 la Nitrosoguanadine mutagenesis of K-12 (A) Completely resistant to 100 ,ig of L-

serine hydroxamate per mlC-21 Ig Nitrosoguanadine mutagenesis of K-12 (A) Partially resistant to 100 jig of L-serine

hydroxamate per mlA-21 Id Nitrosoguanadine mutagenesis of K-12 (A) Inhibition initiated at 60 minB-i I le Nitrosoguanadine mutagenesis of K-12 (A) Inhibition initiated at 30 minB-21 le Nitrosoguanadine mutagenesis of K-12 (A) Inhibition initiated at 30 minD-22Id Nitrosoguanadine mutagenesis of K-12 (A) Inhibiti-on initiated at 60 min

VOL. 106, 1971 973

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TOSA AND PIZER

E. coli B tRNA, 50 MiM "C-L-serine (25 MCi/Amole),and enzyme. The reaction was initiated with substrate,and incubation was at 37 C. At 5 and 10 min after thestart of the reaction, 50-Mliter portions were removedto filter paper discs, and the discs were immediatelyimmersed in cold 10% trichloroacetic acid. After one

washing with cold 10% trichloroacetic acid, two washingswith cold 5% trichloroacetic acid, and one washingwith acetone, the radioactivity on the discs was deter-mined by use of a scintillation spectrophotometer. Ra-dioactive glycine, threonine, and leucine were substi-tuted for serine where the activating enzymes for theseamino acids were to be assayed. The attachment of Inmole of amino acid to tRNA during a 10-min incuba-tion constitutes a unit of enzyme activity. For the ki-netic studies, the enzyme was diluted so that activitywas proportional to the amount of protein present.

RESULTS

Effect of serine hydroxamate on the growth ofmutant strains. The growth of eight mutantstrains selected for their resistance to serine hy-droxamate was tested in the presence of 1.28 mmDL-serine hydroxamate-hydrochloride. Thegrowth response fit three patterns (Fig. 1). Thefirst group was not inhibited (A-l I 1g, C-I I Id,and D-l 1Ila). The second group (C-21 Ig) was

partially resistant, and the third group (A-211d,Bllle, B-21le, and D-211d) was resistant forperiods of time between 30 and 60 min and then

200

150

' 100aas 80

O 600

40

200 60 120 180

MinutesFIG. 1. Growth of mutants in the presence of L-

serine hydroxamate. The growth of K-12 and serine-hydroxamate-resistant mutant strains was determinedas described in the preceding paper (20). All culturescontained D L-serine hydroxamate-hydrochloride at1.28 mM.

became partially resistant. The generation timesof some of the mutants are presented in Table 2.At higher levels of serine hydroxamate, thegrowth of resistant strains D-1 I la and A-1I1lgwas inhibited. Strain D-I1 la had generationtimes of 150 and 350 min when 1.28 and 2.55mM of L-serine hydroxamate was in the medium.Growth of strain K-12 ser-, which lacks PGA

dehydrognease and consequently has a nutri-tional requirement for serine or glycine, was usedto test the possibility that the analogue wasacting as a false feedback inhibitor shutting offthe supply of serine. The experiment was possiblebecause glycine failed to reverse the serine hy-droxamate inhibition (20). Figure 2 shows that astrain using exogenous glycine for the productionof serine, and not possessing the feedback-sensi-tive enzyme, was still inhibited by serine hydrox-amate in a manner indistinguishable from thewild-type organism. This result indicates that theanalogue does not act on PGA dehydrogenase tocause inhibition of growth. Moreover, the ca-pacity of exogenous glycine to give rise to serinewas below the level necessary to reverse serinehydroxamate inhibition, even when the cells wereutilizing this pathway for serine biosynthesis.

Effect of serine hydroxamate on macromolec-ular synthesis. By using 14C-leucine to measureprotein synthesis and 3H-uracil to measure nu-cleic acid synthesis, it was possible to followsimultaneously the effect of L-serine hydroxa-mate on the two major classes of macromoleculesfound in the cell. Because RNA accumulation isconnected to the ability to make protein, the ef-fect of the antibiotic CM was tested in conjunc-tion with serine hydroxamate. CM allows RNAsynthesis to occur under conditions of aminoacid starvation (9). In the presence of eitherserine hydroxamate, CM, or a combination ofthese two compounds, protein synthesis wascompletely inhibited (Fig. 3A). The addition ofserine hydroxamate alone also inhibited the accu-

TABLE 2. Mass doubling times of E. coli K-12 andserine hydroxamate-resistant mutantsa

L-Serine hv- Bacterial straindroxamate

(MM) K-12 A-lIg C-1iId C-211g D-1Ila

0 70 70 100 70 7064 195128 282191 400320640 x 70 100 100 70

a The mass doubling times were calculated fromplots of Klett readings against time. The time for theKlett reading to increase from 50 to 100 is shown.

974 J. BACTERIOL.

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF L-SERINE HYDROXAMATE

mulation of uracil in nucleic acids (Fig. 3B). Incontrast to the situation with proteiii gynthesis,addition of CM or CM plus serine hydroxamateleads to an appreciable amount of uracil incorpo-ration into nucleic acid. The E. coli K-12 strainused in these experiments was stringent (17) sinceit failed to accumulate RNA in the absence of anamino acid requirement (Table 3). CM had theeffect of allowing RNA synthesis either in theabsence of amino acid requirement or the pres-ence of serine hydroxamate. These results sug-gested that the effect of serine hydroxamate onRNA synthesis was a secondary one related tothe fact that protein synthesis was inhibited bythe serine analogue. This possibility was testeddirectly by measuring the effect of serine hydrox-amate on the accumulation of uracil into nucleicacids in a relaxed strain (Hfr-C; reference 17). Inthis strain, the inhibition of protein synthesis byserine hydroxamate was somewhat slower inappearing and less complete. The effect on uracilincorporation shows that RNA synthesis oc-curred in the presence of L-serine hydroxamatewhen protein synthesis was blocked.The synthesis of nucleic acids in the presence

of serine hydroxamate (Table 3) indicates thatthe analogue does not limit the supply of sefine.Serine is the source of glycine, which is requiredfor purine biosynthesis; had the supply of serinebeen reduced RNA synthesis and protein syn-thesis should have stopped. The results in Table3 indicate that serine hydroxamate must beacting directly on the synthesis of protein.

Inhibition of L-seryl-tRNA synthetase. Sincethe serine analogue appeared to be acting di-rectly on protein synthesis, its effect on the seryl-tRNA synthetase was determined. The activityof seryl-tRNA synthetase was reduced in thepresence of L-serine hydroxamate (Fig. 4A).When the enzyme activity was measured as afunction of L-serine concentration alone and inthe presence of serine hydroxamate, the analogueacted as a competitive inhibitor towards theamino acid (Fig. 4B). The Km for L-serine ob-tained from these studies was 50 gM and the K1for L-serine hydroxamate was 30 ,uM. The en-zyme apparently has a higher affinity for theanalogue than it does for its natural substrate.To test the specificity of the inhibition, seryl-tRNA synthetase activity was measured in thepresence of several amino acid hydroxamates.The results (Table 4) show that 10 mM DL-threo-nine hydroxamate produced 50% inhibition. Thisconcentration of DL-threonine-hydroxamate isapproximately 10 times the concentration thatfailed to effect growth of E. coli K-12 (20).The specificity of inhibition by L-serine hy-

droxamate was also tested. Table 5 demonstrates

200

150

cc

-V

100

80

60

40.

0 60 120 180Minutes

FIG. 2. Inhibition of the growth of E. coli K-12Ser- by serine hydroxamate. The experiment was per-formed as described in the accompanying paper. Thenutritional requirement of the organism was satisfiedby 400 gsg of glycine per ml. The DL-serine hydroxa-mate concentrations were: none, 0; 128 ,Am, A; and1.28 mM, 0.

that, although leucine hydroxamate, threoninehydroxamate, and glycine hydroxamate inhibittheir acyl-tRNA synthetase to an appreciableextent, L-serine hydroxamate at concentrationsthat would completely inhibit the seryl-tRNAsynthetase had no effect on the leucine-activatingenzyme, relatively little effect on the threonineactivating enzyme, but inhibited the glycyl-acti-vating enzyme approximately 60%. The effect ofthe glycyl-activating enzyme may be of interest,considering that glycine hydroxamate was notappreciably better as an inhibitor than the serineanalogue. It is apparent that the inhibition ofseryl-tRNA synthetase by the L-serine hydroxa-mate is sufficiently sensitive and specific to ac-count for the bacteriostatic action of this com-pound. The significance of the inhibition of acyl-tRNA synthetases by other amino acid hydroxa-mates in relation to earlier growth studies isevaluated in the discussion section.Mutant studies. To obtain direct evidence that

the inhibition of seryl-tRNA synthetase by serine

975VOL. 106, 197 1

20 L

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

976

EC.IU 200

TOSA AND PIZER

1None

(L-Ser-Hdx,j CM,(L-Ser-Hdx+CM004*oae

EC.0N0

J. BACTERIOL.

ICM+L-Ser-Hdx

L-Ser-HdxI

0 60 120 180 0 60 120 180Minutes Minutes

FIG. 3. Effect of serine hydroxamate on protein and nucleic acid synthesis in E. coli K-12. After a Klettreading of80 was reached, the culture was diluted with an equal volume of medium and divided into growth tubes.The isotopically labeled compounds and the desired antibiotics were then added. The additions are shown on theappropriate curves. L-serine hydroxamate and CM were present at a final concentration of 100 ,g/ml. The 14C-leucine incorporation in the presence of L-serine hydroxamate, CM, or both compounds was the same and isshown by a single curve (0). The 3H-uracil incorporation in the presence of CM or CM + serine hydroxamatewas the same and is shown by a single curve (A).

TABLE 3. Effect of L-serine hydroxamate (Ser-Hdx) on uracil and leucine incorporation into acid-precipitablematerial

Compound added (counts/min)

Per cent ofStrain None L-Ser-Hdxa L-Ser-Hdxa relative in- -MET

(A) (B) + CM corporation + CM(B/A x 100)'

3H-uracilK-12 680 80 650 650 12K-12 Met- 1,780 216 1,080 1,106 12 112 870Hfr-C3 2,600 1,150 600 600 58

'4C-leucineK-12 2,950 120 20 30 4K-12 Met- 7,020 128 60 36 2 340 154Hfr-C 4,150 300 0 0 8

a L-Ser-Hdx-hydrochloride concentration was 0.64 mM (100 gg/ml).'CM (chloramphenicol) concentration was 100 gg/ml.c Isotope incorporation is presented for the samples removed from the cultures 120 min after initiation of the

experiment. The incorporation kinetics were similar for strains K-12 (X)Met- and Hfr-C to those found for strainK-12 (X) (Fig. 3). In cultures of K-12 Met- and Hfr-C, methionine was present at 40 ,g/ml unless otherwise indi-cated.

hydroxamate was responsible for the bacteria-static properties of this compound, mutant strainswhich had been selected for their ability to growin the presence of the analogue were studied.Assuming that the point of L-serine hydroxamateinhibition was the activating enzyme, there wouldbe two ways to generate a resistant strain. Oneway would be modification of the activating en-zyme itself; another would be to increase the in-,tracellular level of L-serine. Described in the in-troduction are the characteristics in which theactivating enzyme could be altered and the mech-

anisms for increasing the supply of serine.The possibility that the resistance of the eight

mutants which grew in the presence of serinehydroxamate was due to an increased supply ofserine was tested by analyzing the capacity forserine phosphate formation in crude extracts. Itwas observed (Table 6) that the enzyme levelsdid not differ from those found in the wild-typeorganism, and no evidence was obtained for de-repression of enzyme synthesis. Whether the en-zymes were altered in their sensitivity to feed-back control was tested by measuring serine

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 106, 1971 MODE OF ACTION OF L-SERINE HYDROXAMATE

0

0 5 KI

B

L-SER - Km - 50pML-SER-HDX-Ki- 30,aM

0 2 3 5 IOM

Reaction Time (min.) ([L-rX)

FIG. 4. Effect of serine hydroxamate on L-seryl-tRNA synthetase. (A) The assay was performed with 0.1 unitof enzyme activity in the incubations, a serine concentration of 50 ,uM, and the L-serine hydroxamate concentra-tions indicated. (B) The standard assay was used with 0.16 units ofenzyme, and the serine concentration varied as

shown. The concentration of L-serine is in moles per liter, and the Ser-Hdx concentration was 50 Mm (0). The ra-

dioactivity (counts per minute) present in the sample removed at 5 min was used as the measure of velocity, andthe reciprical of counts/min is plotted on the I/v axis. In both experiments, serine had a specific activity of 25ACiIAmole.eTABLE 4. Effect of amino acid hydroxamates on L-

seryl-tRNA synthetasea

Amino acid hydroxamates Activity Inhibition(5 mM) (counts/min) (%)

None 676 0L-Leucine hydroxamate 658 3DL-Threonine hydroxamate 328 50Glycine hydroxamate 640 5L-Lysine hydroxamate 744 0L-Sefine hydroxamate 0 100

a Standard assay was performed with 0.09 units ofenzyme, 14C-L-serine (50 gCi/,gmole), and the L-aminoacid hydroxamates at 5 mM. The concentration of DL-threonine-hydroxamate was 10 mM. The activity isshown for the 10-min time point.

phosphate production in the presence of 40 IAM L-serine. One strain (C-l lId) failed to show anyinhibition, whereas the wild-type organism andremaining seven mutants were inhibited by ap-proximately 70% (Table 6). A detailed study ofthe feedback inhibition of this mutant strain(Fig. 5) showed that the enzyme in this mutantwas highly resistant to serine inhibition. Mixingexperiments and experiments in which the serinewas preincubated with the mutant extract une-quivocally showed that the failure to observe inhi-bition was not due to destruction of L-serine. We

TABLE 5. Effect ofamino acid hydroxamates onamino acyl-t-RNA synthetasesa

Amino acid Activity Inhibi-Substrates Aminoxacid (counts/ tionhydroxamates min) (%)

4C-L-leucine None 127 0(40 ,Ci/,mole) L-Leucine hydroxa- 59 54

mateL-Serine hydroxa- 127 0mate

'4C-L-threonine None 877 0(50 uCi/gmole) DL-Threonine hy- 41 95

droxamateL-Serine hydroxa- 685 21

mate

4C-glycine None 50 0(50 gCi/Mmole) Glycine hydroxa- 16 78

mateL-Serine hydroxa- 20 60

mate

The standard assay was performed with the activity shownfor the 10-min time point. The substrates were at 50 gM, andthe L-amino acid hydroxamates were at 5 mM. The DL-threo-nine-hydroxamate was at 10 mm. The amount of enzyme addedwas equivalent to 0.09 units of seryl-tRNA synthetase.

conclude that in strain C- Ild the mutation al-ters the PGA dehydrogenase, making the enzymeless sensitive to inhibition by serine, and inferthat, as a consequence of this alteration the pool

977

E

z

a)tt:

L0

Cc

60.O-

E0Ua.

010

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

978 TOSA A

of serine in the cell increases to the level whereserine hydroxamate no longer acts as an effectiveinhibitor of seryl-tRNA synthetase.An analysis of the amino acyl-tRNA synthe-

tase activity in three other mutants was thenundertaken. One of these mutants (A-l1lg) ex-hibited a four-fold increase in the level of seryl-tRNA synthetase (Table 7). One of the mutants,(C-211g), contained 15% of the activity of thewild-type, the third (D-111a) slightly higher ac-tivity than the wild-type organism. The strain A-lllg which had the greater activity for serineactivation had approximately wild-type levels ofthreonyl- and leucyl-activating enzymes (Table 7).The lowered level of enzyme activity detected instrain C-21 Ig suggested that the structure of thisenzyme had been altered. This was verified by

TABLE 6. Serine-phosphate formation inhibition by L-serinea

Serine- Ihbto

Bacterial phosphate Relative Inhibitionformation activit bstan (units/mg of acivt L-eil

protein)

K-12 1.32 100 65A-l1lg 1.60 122 80A-21 Id 1.55 120 76B-ll le 1.86 141 77B-211 e 1.97 149 76C-lIld 1.65 125 0C-21 Ig 1.01 83 76D-lIla 2.07 157 77D-22 Id 1.34 102 72

a The specific activity of the substrate (phosphoglyc-erate) was 20,000 counts per min per umole.

100~

80

O, 60

.E 40

.a_

E

lu20

C-Ilid

0 5 10 20 30 40 50x 10-5MConc. of L-Ser

FIG. 5. Inhibition ofserine-P formation by L-serine.Standard assay conditions were used with the finalconcentration of L-serine in the assay indicated.

4D PIZER J. BACTERIOL.

measuring the temperature stability of this en-zyme. The data obtained by using the procedureof Sundharadas et al. (19) showed that the mu-tant enzyme was 50% inactivated by 5 min oftreatment at 57 C, whereas the wild-type orga-nism was considerably more stable (Fig. 6). Theinstability of the activating enzyme from C-21lgdid not affect bacterial growth at 42 C. ThetRNA synthetase from mutant strains A-l I lgand D-l1lg showed the same temperature sta-bility as the enzyme from K-12. A detailed studyof the kinetic properties of the activating en-zymes from the mutant strains was undertaken.It was observed (Fig. 7) that the wild-type andmutant enzymes had approximately the same Kmvalues for serine and that, in all cases, serinehydroxamate acted as a competitive inhibitor. K,values were obtained for the mutant strains thatwere three to five times higher than those foundwith the parental type K-12. These results aresummarized in Table 8.

DISCUSSIONThe primary observation described in this

paper is that L-serine hydroxamate inhibits theactivity of seryl-tRNA synthetase. The inhibitionis competitive with L-serine. Since the hydroxa-mate group eliminates activation of the ana-logue, the inhibition presumably results fromexclusion of the amino acid substrate from theenzyme. The in vivo consequences of reducedseryl-tRNA synthetase activity are inhibition ofprotein synthesis and growth. In strains thatexert stringent control over RNA synthesis, thereis, together with the inhibition of protein syn-thesis, a marked reduction in the accumulationof uracil into RNA. It appears that, even in thepresence of an adequate supply of serine, RNAsynthesis does not proceed normally if the at-tachment of the amino acid to the seryl-tRNA isinhibited. Serine behaves like other amino acids

TABLE 7. Comparison of aminoacyl-tRNA synthetaseactivitiesa

Aminoacyl-tRNA synthetase activitiesBacterial (units/mg of protein)strains

L-Serine L-Threonine L-Leucine

K-12 11.6 1.5 11.3A-11 Ig 51.3 1.6 6.9C-211g 1.8D-1 I la 16.6

a Standard assay conditions were used with L-serinepresent at 25 AM (specific activity, 50 MCi/gmole), andthe enzyme was diluted to give proportionately withadded protein. The activity of the sample taken at 5min was used to calculate the specific activity shown.

.N

n

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 106, 1971 MODE OF ACTION OF L-SERINE HYDROXAMATE

30

(K-12=5 x 0-5M25 A-lllg 2 5x 105MKm C-211g= 6.6 x 105M

D-lal - 8.3x 105M

1004

0-1. 50 ~

25C-211g

E

I0

10 , I I I0 5 lo Is 20

Minutes at 570CFIG. 6. Thermal inactivation of seryl-tRNA synthe-

tase. The procedure of Sundharadas et al. (19) for de-termining the rate of heat inactivation was modified bythe substituting DTT for f3-mercaptoethanol. Afterremoval of enzyme samples to chilled tubes, the en-zyme was diluted 10 times with reaction buffer; tRNA,A TP, and I4C-serine were then added to initiate theassay. The amount of enzyme added was adjusted togive the same amount of activity before heating.

0 1 2 3 5I ) x 104

FIG. 7. Kinetics of L-seryl-tRNA synthetase. Thestandard assay was employed with between 0.05 to 0.2units of enzyme purified from the mutant strains. Theserne (25 gCiI,umole) concentration was varied asshown, and the reciprocal of the molarities was multi-plied by 104 plotted on the I /s axis. The velocity of thereaction was measured by determining the radioactivityin the sample taken at 5 min. The reciprocals of thenumber of counts per minute in these samples weremultiplied by 10o and have been plotted on the l/vaxis.

TABLE 8. Summary of the properties of L-seryl-tRNA synthetases from mutant strains

Thermal inactivationat 57 C relative Kinetic parameters

Growth inhibition Relative activity (%) re- (#M)Bacterial by 1.28 mM L-seryl-tRNA maining after Characteristics ofstrains DL-serine h synthetase mutant seryl-

droxamate activity K5 for tRNA synthetasedroxamate (%)5 min 10 min K,,,a for L-serine

L-serine hydroxa-mate

K-12 Complete in- 100 70-90 48-67 50 30hibition

A-11 lg Complete re- 450 91 53 50 167 Excess enzyme,sistance K, alteration

C-21 lg Partial resist- 15 33 13 66 85 Heat sensitive,ance Ki alteration

D-11 la Complete re- 140 78 47 83 167 Ki alterationsistance

aReported Km 7 X 10-5 M (8).

(e.g., valine and glycine) in this respect (4, 5).Because serine hydroxamate is a competitive in-hibitor, its effectiveness in stopping growth de-pends on the concentration of serine hydroxa-mate and serine in the cell and the relative affini-ties of these two compounds for the activatingenzyme. The second parameter can be expressedas the ratio of Ki (Ser-Hdx) to Km (Ser). Thisratio is approximately 1; therefore, the sensitivityof the bacteria to low concentrations of analogue

must be due to low intracellular concentrationsof serine.

Evidence that the bacteriostatic action ofserine hydroxamate is due to the inhibition of theactivating enzyme and that the in vitro measure-ments of enzyme activity are pertinent to theconditions in the cell comes from an analysis ofthe properties of the mutants resistant to theanalogue. Three of the resistant mutants haveactivating enzymes with increased K1 (Ser-Hdx)

979

in0

K-12

A- IlgC-211gD-lila

10 M

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TOSA AND PIZER

values and, in keeping with the competitive char-acter of the inhibition, higher levels of serine hy-droxamate are required to inhibit their growth.In addition to this kinetic evidence that muta-

tions to resistance are accompanied by changesin seryl-tRNA synthetase, one resistant strainpossesses an enzyme with an altered structurewhich leads to increased lability at 57 C. Thesedata confirm that the activating enzyme is thepoint of inhibition and that the Ki (Ser-HDX) toKm (Ser) ratio is critical in establishing the levelof sensitivity.The second factor which should effect sensi-

tivity is the intracellular concentration of serine.The normal endogenous serine concentrationappears to be maintained at a low level. Additionof serine to the medium relieves the inhibition ofgrowth by increasing the intracellular concentra-tion. A mutation which relieves end-product inhi-bition confers resistance to serine hydroxamate,presumably by allowing expansion of the endoge-nous pool. It appears that serine production istightly regulated by the end-product inhibition ofPGA dehydrogenase. The addition of exoge-

nous glycine did not overcome serine hydroxa-mate inhibition; we conclude that, althoughthe conversion of glycine to serine was adequatefor growth, it did not lead to an elevated concen-

tration of serine in the cell. The possibility thatthe conversion of serine to glycine is regulatedwarrants further investigation.The properties of the mutants support our

ideas on the mechanism of action of the ana-

logue. By making some assumptions we can fur-ther test the relationship between the inhibitionof seryl-tRNA synthetase and the inhibition ofgrowth. If we equate the rate of growth (recip-rocal of generation times given in Table 2) inthe absence of serine hydroxamate with Vmax(l/VmaX, time in minutes for the Klett readingto go from 50 to 100 in the uninhibited culture)of the activating enzyme and the rate of growthin the presence of different concentrations ofserine hydroxamate to v, the inhibited level ofenzyme activity, we can attempt to fit the equa-tion

[(l/V)/(l /Vmax)] I

= lKm - (Km[I]/Kjj1/[S] (I)

to the growth data obtained in the presence ofdifferent levels of serine hydroxamate. EquationI is a convenient rearrangement of the standardMichaelis equation for a competitive system (2):v = Vmax [S]/fKm (I + [I]l/Ki) + [S]f. EquationI describes the competitive inhibition of an en-

zyme reaction (2); to use it, we make the addi-tional assumption that [I], the inhibitor concen-

tration in the cell, is equal to the concentration

of serine hydroxamate added to the culture me-dium. This is probably justified since we havedemonstrated the passive entry of serine hydrox-amate into the cell (20). If our assumptions arevalid and the reduction in growth is solely due toa drop in the rate of charging seryl-tRNA, wepredict that the serine concentration in the cell[S] should remain relatively constant at the dif-ferent levels of growth inhibition. We could thenequate the computed value of [S] to the in vivoserine concentration maintained by end-productinhibition. Substituting the numerical values forgrowth rates (Table 2) and the average values forKm (ser) and Ki (Ser-Hdx) (Table 7) into equa-tion 1, the intracellular concentration of serinewas computed to be 88, 87, and 79 uM for cellsof strain K-12 growing in the presence of 64,128, and 192 uM L-serine hydroxamate, respec-tively. The relative constancy of the calculatedserine concentrations support the validity of ourassumptions and strongly suggests (i) that thepool of serine is not expandable by endogenoussynthesis, (ii) that the inhibition of seryl-tRNAsynthetase reduces growth to a level, v, (iii) thatthe uninhibited rate of growth represents Vmax ofseryl-tRNA synthetase.

Points i and iii above require further amplifi-cation. If the serine pool was expandable by en-dogenous synthesis, it should increase when theactivating enzyme is inhibited and, in the case ofa competitive inhibitor, thereby overcome theinhibition. More than a transient inhibition ofgrowth by serine hydroxamate depends, there-fore, on the constancy of the intracellular serineconcentration. The constant serine concentrationcould result from limitations on substrate supplyfor the biosynthetic pathway, e.g., PGA, NADor glutamate, or end-product inhibition of PGAdehydrogenase. The second explanation appearslikely because the calculated intracellular serinepool of approximately 90 AiM would markedlyinhibit the PGA dehydrogenase (Fig. 5). Also,substrate limitations cannot be restricting serinesynthesis, since a mutant without this feedbackcontrol increases the serine pool, as judged by itsresistance to the antibiotic. It appears that serinesynthesis is proceeding in the cell at a greatlyinhibited rate. Derepression of the biosyntheticenzymes did not occur during growth in the pres-ence of the analogue, and resistant mutants withelevated enzyme levels were not found.

By equating the velocity of the seryl-tRNAsynthetase with the rate of growth, we assumethat the activity of this enzyme is the pacemakerfor bacterial growth. This assumption is plausiblewhen the culture is growing in the presence ofthe antibiotic (i.e., the determination of v), but isless likely in the absence of serine hydroxamate.

980 J. BACTERIOL.

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MODE OF ACTION OF L-SERINE HYDROXAMATE

We have tested the consequences of inaccuracyin the growth rate measurements by substitutingalternative values for v and Vmax into equation 1.Recalculations of [SI with changes in v and Vmaxof 4 10 min have shown that [S] remained rela-tively constant if the same Vmax value was usedfor a series of cultures and that this analysis isnot sensitive to small alterations in growth rate.

In view of these calculations and the fact thatan intracellular serine concentration of 90 uMwould not saturate the activating enzyme with aKm of 50 ym, we conclude that we are not usingthe true Vmax value in equation I and seryl-tRNA synthetase alone does not limit normalgrowth. However, it does appear that, unlike thePGA dehydrogenase, the seryl-charging enzymeis not regulated by a feedback control and isfunctioning in the cell at the maximum rate al-lowed by the availability of its substrate. Directmeasurements of [SI are being undertaken todetermine whether the computed value of 90 Mmis in agreement with experimentally determinedpools.When growth data and the K,(Ser-Hdx) ob-

tained for strain C-1 I Id was put into equation 1,the calculated serine concentrations were 380 and202 Mm for growth in the presence of 1.25 and2.5 mM serine hydroxamate. The variation andthe magnitude of the [S] values suggest that athigh serine hydroxamate concentrations equationI does not hold and the slow growth is not solelydue to a reduction in seryl-tRNA synthetase ac-tivity.The inhibition of the leucyl-, threonyl- and

glycyl-tRNA synthetases by their correspondingamino acid hydroxamates shows that the inhibi-tion of the activating enzymes is not restricted toserine hydroxamate and raises the question whythese analogues fail to inhibit bacterial growth.The concentration of analogues, 5 mM, that gavemarked inhibition of the activating enzymes wasapproximately 10 times higher than that used inthe growth studies; although growth might havebeen inhibited by higher concentrations, the largequantitative difference in the effectiveness of theamino acid hydroxamates is noteworthy. Fromthe few preliminary measurements of the inhibi-tion of the activating enzymes, it appears thatthe K1 values for the other analogues are higherthan the K1 for serine hydroxamate. The differ-ences in K; could account to a large extent forthe absence of an observed effect on growth. Inaddition, the extent of enzyme inhibition at agiven analogue concentration depends on the Kmfor the amino acid and its concentration in thecell. For the amino acid hydroxamate to be apotent antimetabolite, the pool size of the aminoacid must be regulated to avoid an increase when

the hydroxamate inhibits the charging of thetRNA. The serine concentration is maintained ata low level by end-product inhibition. Otheramino acids might be maintained at a higherconcentration if their biosynthetic pathways lackthis type of control or had a less sensitive feed-back enzyme. Derepression of enzymes that arerepressed by amino acyl tRNA would also effectthe pool size. Different species of bacteria mightvary considerably in the intracellular concentra-tions of the amino acids and the mechanisms ofregulating these concentrations. The suscepti-bility to the bacteriostatic action of amino acidhydroxamates might provide a simple method forfinding similarities in regulatory processes andidentifying differences among bacterial species.The inhibition of the activating enzyme by the

amino acid hydroxamates should be consideredwhen using hydroxamate formation to assay en-zyme activity. The degree to which inhibitionoccurs will vary from amino acid to amino acidand possibly account for the observation byHoagland, Keller, and Zamecnic (7) and byHirsh and Lippman (6) that there is a lack ofcorrespondence when some amino acyl-tRNAsynthetases are assayed by hydroxamate forma-tion, pyrophosphate exchange, and amino acyl-tRNA formation. Although the concentration ofamino acid hydroxamate produced would beinitially low, its synthesis would occur in the ac-tive site, the ideal location for it to act as an in-hibitor.

In conclusion, there is the possibility thatamino acid hydroxamates could be a useful classof chemotherapeutic agents affective against viraland bacterial infections or malignant conditions.Whether these compounds are effective in a par-ticular situation will depend on the same factorsthat distinguish serine hydroxamate from otherhydroxamates in inhibiting the growth of E. coli.Cultured human cells do not possess tight feed-back controls on the synthesis of the nonessentialamino acids (3). In the case of serine, there is areport that blood cells, unlike other tissues, re-quire exogenous serine (13). These types of dif-ferences from organism to organism or tissue totissue might provide a basis for selective hydrox-amate toxicity. An experimental program basedon the considerations presented in this paperappears warranted.

ACKNOWLEDGMENTS

This research was supported by Public Health Service grantAl-03957 from the National Institute of Allergy and InfectiousDiseases.

LITERATURE CITED

1. Adelberg, E. A., M. Mandel, and G. C. C. Chen. 1965.Optimal conditions for mutagenesis by N-methyl-N'-Nitro-N-Nitrosoguanidine in Escherichia coli K12.

VOL. 106, 1971 981

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TOSA AND PIZER

Biochem. Biophys. Res. Commun. 18:788-795.2. Dixon, M. 1953. The determination of enzyme inhibitor

constants. Biochem. J. 55:170-171.3. Eagle, H., C. L. Washington, and M. Levy. 1965. End

product control of amino acid synthesis by culturedhuman cells. J. Biol. Chem. 240:3944-3950.

4. Eidlic, L., and F. C. Neidhardt. 1965. Protein and nucleicacid synthesis in two mutants of Escherichia coli withtemperature-sensitive amino acyl ribonucleic acid syn-

thetases. J. Bacteriol. 89:706-711.5. Folk, W. R. and P. Berg. 1970. Isolation and partial char-

acterization of Escherichia coli mutants with alteredglycyl transfer ribonucleic acid synthetases. J. Bacteriol.102:193-203.

6. Hirsh, D. I., and F. Lipmann. 1968. The divergence inreactivity of aminoacyl transfer ribonucleic acid synthe-tases of Escherichia coli with hydroxylamine. J. Biol.Chem. 243:5724-5730.

7. Hoagland, M. B., E. B. Keller, and P. C. Zamecnick.1956. Enzymatic carboxyl activation of amino acids. J.Biol. Chem. 218:345-358.

8 Katze, J. R., and W. Konigsberg. 1970. Purification andproperties of seryl transfer ribonucleic acid synthetasefrom Escherichia coli. J. Biol. Chem. 245:923-930.

9. Kurland, C. G., and 0. Maaloe. 1962. Regulation of ribo-somal and transfer RNA synthesis. J. Mol. Biol. 4:193-210.

10. Lowry, H., N. J. Rosebrough, A. L. Farr, and R. J. Ran-dall. 1951. Protein measurement with the Folin phenolreagent. J. Biol. Chem. 193:265-275.

11. Norton, S. J., and Y. T. Chen. 1969. ,B-Aspartylhy-droxamic acid: its action as a feedback inhibitor and a

repressor of asparagine synthetase in Lactobacillus ara-

binosus. Arch. of Biochem. Biophys. 129:560-566.12. Pizer, L. 1. 1963. The pathway and control of serine bio-

synthesis in Escherichia coli. J. Biol. Chem. 28:3934-3944.

13. Regan, J. D., H. Vodopick, S. Takeda, W. H. Lee, and F.M. Faulcon. 1969. Serine requirement in leukemia andnormal blood cells. Science 163:1452-1453.

14. Richmond, M. H. 1962. The effect of amino acid-ana-logues on growth and protein synthesis in microorga-nisms. Bacteriol. Rev. 26:398-420.

15. Shive, W., and C. G. Skinner. 1963. Amino acid analoguesin metabolic inhibitors, vol. 1., p. 2. In R. M. Hochsterand J. L. Quastel (ed.). Academic Press Inc., New York.

16. Smith, H. S., and L. 1. Pizer. 1958. Abortive infection ofEscherichia coli strain W by T2 bacteriophage. J. Mol.Biol. 37:131-149.

17. Stent, G. S., and S. Brenner. 1961. A genetic locus for theregulation of ribonucleic acid synthesis. Proc. Nat.Acad. Sci. U.S.A. 47:2005-2014.

18. Sugimoto, E., and L. 1. Pizer. 1968. The mechanism of endproduct inhibition of serine biosynthesis. J. Biol. Chem243:2081-2089.

19. Sundharadas, G., J. R. Katze, D. Soil, W. Konigsberg,and P. Lengyel. 1968. On -the recognition of serinetransfer RNA's specific for unrelated codons by thesame seryl-transfer RNA synthetase. Proc. Nat. Acad.Sci. U.S.A. 61:693-700.

20. Tosa, T., and L. 1. Pizer. 1971. Effect of serine hydroxa-mate on the growth of Escherichia coli. J. Bacteriol.106:000-000.

21. Umbarger, H. E., M. A. Umbarger, and P. M. L. Siu.1963. Biosynthesis of serine in Escherichia coli and Sal-monella typhimurium. J. Bacteriol. 85:1431-1439.

982 J. BACTERIOL.

on January 14, 2020 by guesthttp://jb.asm

.org/D

ownloaded from