JOURNAL oF BAcFERIOLOGYVol. 88, No. 3, p. 611-619 September, 1964Copyright @ 1964 American Society for Microbiology

Printed in U.S.A.

NUTRITIONAL AND REGULATORY ASPECTS OF SERINEMETABOLISM IN ESCHERICHIA COLI

LEWIS I. PIZER AND MARY L. POTOCHNY

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

Received for publication 11 April 1964

ABSTRACT

PIZER, LEWIS I. (University of Pennsylvania,Philadelphia), AND MARY L. POTOCHNY. Nutri-tional and regulatory aspects of serine metabolismin Escherichia coli. J. Bacteriol. 88:611-619. 1964.-Growth studies with a serine auxotroph havedemonstrated a relationship between the serinesupplied and the extent of growth, which closelyagrees with the value calculated from the reportedchemical composition of Escherichia coli. Serinecould be replaced by glycine and, to a limitedextent, by L-threonine. Both exogenous serineand glycine regulate their own endogenous synthe-sis from glucose or fructose. The inhibition ofendogenous synthesis of both amino acids byserine was greater than 90% with both carbonsources. Greater amounts of exogenous glycinewere utilized for cell synthesis when fructose wasthe carbon source. Growth conditions affectedthe levels of phosphoglycerate dehydrogenasefound in cell-free extracts. The highest levels werefound in glucose-grown cells and the lowest incells grown in a medium augmented with L-threo-nine, L-methionine, L-leucine, and DL-isoleucine.The levels of serine phosphate phosphatase werenot altered by changes in growth conditions.

Although considerable information has beenobtained on the uptake and metabolism of serinein Escherichia coli (Meinhart and Simmonds,1955; Simmonds and Miller, 1957; Levine andSimmonds, 1960), the major biosynthetic path-way from glucose has only recently been estab-lished (Pizer, 1963; Umbarger, Umbarger, andSiu, 1963). Enzymes have been found in E. coliextracts which catalyse the oxidation of 3-phos-phoglycerate to form hydroxypyruvate-phos-phate, the transamination of hydroxypyruvate-phosphate to form serine-phosphate, and thedephosphorylation of serine-phosphate to yieldserine. These reactions make up the "phospho-rylated pathway" described in rat liver by Ichi-hara and Greenberg (1957). E. coli depends onthis pathway for the biosynthesis of serine, be-

cause mutations which result in the loss of phos-phoglycerate dehydrogenase lead to a growthrequirement for serine.Our experiments were performed in conjunc-

tion with studies on the enzymes of the biosyn-thetic pathway (Pizer, 1963) so that observationsin cell extracts might be related to growing cells.The response of these enzyme levels to altera-tions in cell nutrition was studied to determinewhether nutritional factors regulate serine bio-synthesis.

MATERIALS AND METHODS

Bacteria and growth conditions. E. coli strainsW and Wc- were obtained from S. S. Cohen ofthe University of Pennsylvania, and E. coli strainWs- from S. Simmonds of Yale University. Thebacteria were maintained on nutrient agar slantsand were grown in a synthetic medium whichhad the following composition (in g per liter):NaHPO4, 6; KH2PO4, 3; NH4Cl, 1; MgSO4.-7H20, 0.2; FeSO4*7H20, 0.005. Carbon sourcesand nutrients were sterilized by filtration andsubsequently added to the sterile salts solution.Bacterial growth was measured turbidimetricallyin a Klett colorimeter fitted with a no. 42 filterand was correlated with viable counts made onnutrient agar plates. The cultures were grownaerobically at 37 C.

Overnight cultures in synthetic medium werediluted 1:20 in fresh medium on the morning ofthe growth experiment. When the culture reacheda cell concentration of approximately 2 X 108cells per ml, it was chilled to 12 C and washedtwice by resuspension in cold salts solution andby centrifugation. The cells were then resus-pended in the salts solution to give a final con-centration of 108 cells per ml. To each of a set ofgrowth tubes, 9 ml of the cell suspension wereadded, followed by the carbon and energy sourceand other nutrilites according to the experi-mental design. The growth flasks were of thetype described by Cohen and Barner (1955),

611

PIZER AND POTOCHNY

which allow turbidimetric measurements to bemade in a side arm attached to the flask. Whenthe essential nutrients were present, growth re-sumed without a detectable lag and proceededin an exponential fashion.

Glucose or fructose was present in the mediumat 1 mg/ml (5.5 mM), and growth was limited atapproximately 109 cells per ml. Lactate was usedat a final concentration of 11 mm. The enrichedmedium described by Novick and Maas (1961)was used with the following modifications: gly-cine, serine, cysteine, asparagine, biotin, andthe purines were omitted, and arginine wasadded to give a final concentration of 100 ,ug/ml.

Preparation of cell-free extracts. Overnight cul-tures were diluted into fresh medium and wereallowed to grow until they reached the middleof the exponential phase of growth (4 to 5 X 108cells per ml). The cells were chilled and cen-trifuged, and the bacteria were disrupted bygrinding with alumina. The disrupted cells wereextracted with 0.05 M tris(hydroxymethyl)-aminomethane-chloride buffer (pH 7.5); 10 ml ofbuffer were used per gram (wet weight) of cells.Cell debris and alumina were removed by cen-trifugation at 15,000 X g for 15 min. Fraction-ated extracts were prepared from this supernatantfluid by streptomycin and ammonium sulfateprecipitation (Pizer, 1963).

Chemical compounds and procedures. Total andinorganic phosphate were determined by themethod of Bartlet (1959). Protein was measuredby the method of Lowry et al. (1951). Radio-activity measurements were made by countingsamples plated in stainless-steel planchets in awindowless flow counter (Nuclear-Chicago Corp.,Des Plaines, Ill.). All samples were counted to2,000 counts. CI4-phosphoglycerate was a giftfrom N. Pon. L-Serine-phosphate was synthesizedby the method of Neuhous and Korkes (1958).All other materials were purchased from Calbio-chem.Enzyme assays. The capacity to synthesize

serine-phosphate was measured by determiningthe C'4-serine-phosphate formed from C14-phos-phoglycerate. The details of this assay werepreviously published (Pizer, 1963). With theunfractionated extracts used in this study, dupli-cate assays agreed to within 10%, and the en-zymatic activity in different extracts preparedunder similar growth conditions varied from 15to 20%. A unit of enzyme activity is defined as

that amount of enzyme which catalyses the for-mation of 1 ,mole of serine-phosphate per 30min at 37 C. Enzyme assays contained a maxi-mum of 0.3 enzyme units. Specific activities arein terms of enzyme units per milligram of pro-tein, and are based on the average of two to fiveassays. L-Serine-phosphate phosphatase wasassayed as described previously (Pizer, 1963).Specific activities are presented as enzyme unitsper milligram of protein.

Competition between exogenous amino acids(serine or glycine) and glucose and fructose. Themethod of isotopic competition described byRoberts et al. (1955) was utilized with the follow-ing modifications. Specific activities were used tomeasure the quantity of isotope incorporatedinto the compounds of interest. A uracil-requir-ing strain, Wc-, was used instead of the wild-type organism to avoid the incorporation ofglucose carbon into the pyrimidine ring. Theprecision of measuring the radioactivity in themethyl group of thymine was thereby improved.The methyl group of thymine was selected torepresent "one carbon" units because of tech-nical advantages offered in purification andmeasurement of specific activity.The cells were grown as described previously

with C12-glucose or C"2-fructose to a Klett readingof about 60. They were then chilled, centrifuged,and washed with salts solution. After resuspen-sion, samples were added to growth flasks con-taining medium at 37 C. The medium in eachflask contained uracil (0.1 mg/ml) and uniformlylabeled radioactive glucose or fructose (1 mg/ml)with specific activities of 36 X 103 counts permin per ,umole. Flasks A and D had no furtheradditions and served as control cultures. FlasksB and E had 84 mg of L-serine added, and flasksC and F had 60 mg of glycine added. The finalvolume of medium in each of the flasks was 200ml. The initial Klett reading was 60; growth wasstopped by chilling the cultures when the Klettreading reached 130. From the growth studiesdata, it was estimated that the serine added toflasks B and E was six times the quantity usedby the cells during their doubling in mass. Theprocedure used for the fractionation of the cellsinto acid-soluble, lipid, nucleic acids, and proteinhas been described (Roberts et al., 1955).The protein residues were hydrolyzed with 6

N HCl for 16 hr at 105 C. After removal of theHCl by placing the samples in a vacuum desic-

612 J. BACTERIOL.

SERINE METABOLISM IN E. COLI

cator over KOH, the dinitrophenyl (DNP) deriv-atives of the amino acids were prepared. Samplesof these DNP derivatives were separated by two-dimensional paper chromatography. The solventfor the first dimension, butanol-ammonia (Smith,1958), was allowed to drip off the paper untilthe serine and glycine derivatives had movedabout halfway down the paper. The solvent forthe second dimension was 0.75 M potassium phos-phate, pH 6.0 (Smith, 1958). The concentrationsof the eluted DNP-amino acids were determinedfrom their absorbancies.The bases were liberated from the nucleic

acids by hydrolysis in 6 N HCl for 3 hr at 105 C.These bases were separated and purified by paperchromatography. The spectral properties of thecompounds eluted from the chromatographswere used to determine their purity and con-centration. The chromatography involved thesequential use of the following solvents: iso-propanol-HCl, n-butanol-ammonia, and iso-butyrate-ammonia, for the purification of ade-nine, thymine, and uracil. The specific activitiesof the adenine and thymine determined afterthe isopropanol chromatography did not alterafter chromatography in the n-butanol and theisobutyrate solvents (Smith, 1958). The specificactivity of the uracil decreased after each chro-matographing, indicating a radioactive impurityin this compound. The pyrimidines were thenchromatographed with an ethyl acetate-formatesolvent. The specific activity of the thymine didnot alter, and the radioactivity of the uracilsamples dropped to background.

For purification of guanine, use was made ofits low solubility in water. The regions of theisopropanol-HCl chromatographs which con-tained the guanine were first eluted with a smallvolume of water. These eluates contained radio-active materials which contaminated the gua-nine. The guanine was then eluted from each ofthe papers with a larger volume of 0.01 N HCl,and the specific activities of these solutions weredetermined.

RESULTS

Growth studies. These experiments were de-signed to provide information on the serine re-quirements for growth, and the compoundswhich replace serine. Glucose was used as thecarbon source.The growth of E. coli strain Ws- was directly

0m

Ir-

HOURS

FIG. 1. Growth response of Escherichia coli strainWs- to limiting L-serine. In the lower part of thefigure the change in turbidity with time is shownfor cultures supplied with different amounts of L-serine. The numbers on the curves give the initialL-serine concentration in mg/ml. The upper partof the figure shows a plot of the maximal change inturbidity against initial L-serine concentration.These data were obtained from the curves below.

dependent on the quantity of serine added (0.01to 0.15 mg/ml), although the initial rate ofgrowth was independent of this factor (Fig. 1).From the growth response to measured amountsof serine, it was estimated that 1010 cells in theexponential phase of growth utilized approxi-mately 23 ,umoles of the amino acid.A number of compounds were tested to see

whether they could replace the serine require-ment of strain Ws-. Glycine was active but itwas not used as well as serine, as indicated bythe relative division times. The division timewas approximately 50 min with L-serine as thegrowth factor, and was increased to 120 minwhen glycine was substituted. The glycine waspresent initially at 0.26 mg/ml. Doubling theglycine concentration did not increase the growthrate, but lowering the concentration reduced it.

VOL. 88, 1964 613

PIZER AND POTOCHNY

When the culture was grown in a mixture ofserine and glycine (0.04 and 0.28 mg/ml, re-spectively), the growth rate was initially that ofa control culture containing only serine. When aKlett reading of 80 was reached, the serine wasexhausted and growth stopped in the controlcontaining only serine. In the culture initiallycontaining both amino acids, growth continued,but now at the slower rate characteristic of acontrol culture growing with glycine.These experiments show that L-serine and

glycine can be interconverted, but that serineis the preferred growth factor. This preferencemay be one of entry into the cell or may be dueto limitations in the rates of subsequent enzy-matic reactions. The observation that the rate ofgrowth with L-serine does not depend on highexternal concentrations, as does the rate withglycine, indicates that uptake of glycine could belimiting its ability to promote growth. Levineand Simmonds (1960), studying the serine-gly-cine relationships in an auxotrophic strain ofK-12, showed that exogenous L-serine inhibitedthe uptake of Cl4-glycine, but they did not finda diphasic character to the growth curve ob-tained with a mixture of L-serine and glycine.

Other compounds tested which did not supportgrowth included the keto-acids, hydroxypyru-vate, and glyoxylate, either at the concentrationused for their amino acid analogues or at tentimes these concentrations, and D-glycerate atten times the normal serine concentration. E.coli strain Ws- had the same growth character-istics when either fructose or lactate replacedglucose as carbon source. Growth of strain Ws-occurred when the medium was augmented with3 mM threonine. The rate of growth was linearrather than exponential; with either glucose,fructose, or lactate as the carbon source, thetime required for the number of cells to increasefrom 108 to 2 X 108 per ml was approximately190 min. It appears that L-threonine gives riseto L-serine at a rate which limits growth.The rate of growth of the prototrophic strain

of E. coli was measured in the presence and ab-sence of exogenous serine. Glucose was the car-bon source (1 mg/ml), and serine when addedwas at the final concentration of 0.3 mg/ml. Inthe absence of L-serine, the division time was 60min, and in the presence of serine it was 50 min.The amount of growth was approximately 12%greater when L-serine had been added.

Competition experiments. The original experi-ments of Roberts et al. (1955) indicated that, inE. coli strain B, exogenous -serine limited theendogenous synthesis of T-serine and glycinefrom glucose. They also observed that exogenousglycine only limited the incorporation of glucosecarbon into glycine, but, when C'4-fructose orC14-acetate were the carbon sources, glycinelimited the incorporation of isotope into bothT-serine and glycine. It was concluded that inglucose-grown cells L-serine is the precursor ofglycine, and it was postulated that, when thecells are grown in fructose or acetate, metabolicpathways are altered so that -serine is synthe-sized by a novel route. Experiments with cell-free extracts of E. coli strain W defined the serinepathway from glucose and showed that the initialenzyme in the pathway, phosphoglycerate de-hydrogenase, is subject to feedback inhibition(Pizer, 1963). L-Serine is a more effective in-hibitor than glycine by a factor of approximately100. Competition experiments were performedwith E. coli strain W to determine whether thefeedback inhibition observed in cell-free extractsoperates in growing cells, and to ascertain whethergrowth on fructose alters the serine-glycine re-lationship. These experiments were also designedto give quantitative information on the roleplayed by L-serine and glycine in the synthesisof the purine nucleus and the methyl group ofthymine (Table 1).

In the experiment with uniformly labeledCA4-glucose, the supply of exogenous serine re-duced the formation of both serine and glycinefrom glucose to less than 5%. The presence ofexogenous glycine reduced the conversion ofglucose to serine to 48%, while 12% of the gly-cine in protein arose from glucose.The specific activities of the nucleic acid bases

presented in Table 1 provide information on theirorigin when considered in conjunction with aminoacid data. In flask A where no competitor hadbeen added, the specific activities in thymine, inwhich only the methyl carbon was radioactive,and the purines, in which five carbons werelabeled, show that the radioactivity was 2,930counts per min per ,umole of carbon. This figureagrees within experimental error with the aminoacid data. In flask B where nonradioactive serinewas present to compete with the radioactive glu-cose, the specific activity of the thymine indicatesthat not more than 3% of that group was derived

614 J. BACTERIOL.

SERINE METABOLISM IN E. COLI

TABLE 1. Specific activities of nucleic acid bases, serine, and glycine*

Flask Growth conditions Serine Glycine Thymine Adenine Guanine

A C14-glucose 7,794 4,750 2,940 14,900 14,400B C14-glucose plus 329 224 85 2,940 3,360

C'2-serineC C14-glucose plus 4,020 592 1,030 5,620 5,620

C12-glycineD C'4-fructose 7,455 4,310 2,000 12,000E C'4-fructose 166 121 34 2,800

C12-serineF C'4-fructose 2,445 174 335 3,560

C12-glycine

* The specific activities of the amino acids are averages of four independent determinations on thedinitrophenyl derivatives; expressed as counts per minute per micromole.

from glucose. The exogenous serine must supply97% of the "one carbon" fragments which areused for thymine synthesis, with the remaining3% coming most likely from the limited amountof endogenous serine. The specific activities of thepurines allow for one atom of carbon to arise un-diluted from the labeled glucose. Carbon-6 of thepurine ring is derived from CO2 (Roberts et al.,1955), and would be expected to have the samespecific activity as does the glucose carbon. Theremaining purine carbons must be unlabeled andtherefore derived from the unlabeled serine. Posi-tions two and eight are derived from one-carbonfragments, whereas positions three and four arederived from glycine. In flask C where non-radioactive glycine was present, 35% of the thy-mine methyl group was derived from glucose,probably via the carbon of serine. The remainderwas derived from unlabeled glycine. The glycinecontribution to "one carbon" units was greaterthan was expected from its contribution to serine.If the contributions of the purine specific activityexpected from two "one carbon" fragments andone undiluted CO2 molecule are taken into ac-count, the radioactivity in position three and fourshould be 620 counts per min per jAmole. Thisfigure is in agreement with the data obtained forglycine.

In the experiment with uniformly labeled C14-fructose, the specific activities in the compoundsfrom control flask D were slightly lower thanwhen C14-glucose was used. This reflects lessgrowth in the presence of C14-fructose. WhenC12-serine was present, the endogenous synthesisfrom fructose was reduced in both serine andglycine to approximately 2% of the control value.

When C"2-glycine was present, the endogenousserine synthesis was reduced to 33% of the con-trol and the endogenous glycine synthesis to 4%of the control; 83% of the methyl groups ofthymine were derived from glycine. Exogenousserine efficiently limited the formation of bothserine and glycine from glucose or fructose.Exogenous glycine, while reducing the incorpora-tion of isotope from both carbon sources intoglycine, did not limit serine synthesis from glucoseas well as it did from fructose. The observationthat serine can be synthesized from unlabeledglycine confirms the mutant work and indicatesthat glycine can readily give rise to "one carbon"units in this organism.

Effect of nutrition on enzyme levels. The enzymesresponsible for L-serine biosynthesis were assayedin extracts of cells grown under different nutri-tional conditions. Initially, the effect of an exog-enous supply of L-serine or replacement of glucoseby fructose was tested. The capacity to synthesizeserine-phosphate was lower in extracts from cellsgrown on fructose than in extracts of cells grownon glucose (Table 2). The relationship of this ob-servation to the competition experiments de-scribed above will be discussed later.Growth in the presence of 3 mM L-serine did not

affect the enzyme levels of E. coli strains W, B,or K-12. Because the endogenous serine pool mayrepress the synthesis of these enzymes to a maxi-mal extent, an attempt to reduce the size of thispool was made by accelerating growth in the ab-sence of exogenous serine or its derivatives. Table2 shows that growth rate was increased in theenriched lactate medium used but the enzymaticsynthesis of serine-phosphate was lowered. Al-

N OL. 88, 1964 615

PIZER AND POTOCHNY

TABLE 2. Effect of nutrition on enzyme levels

Specific activity

Growth conditions Division time Serine- Serine-phosphate phosphateformation phosphatase

Mfin

Glucose ........ 55-60 3.04 2.3Fructose ....... 60 1.65 2.1Lactate ........ 80 1.50 1.9Enriched lac-

tate ......... 40 0.50 1.7

TABLE 3. Serine-phosphate synthesis inmixed extracts

Extract from cells grown inC14-serine phosphate*

Glucose Enriched lactate

30t - 1,98060 3,90060 4,20090 - 5,220_ 60 720- 160 1,26060 80 4,38060 160 4,440

* Specific activity: 17,500 counts per min perumole.

t Expressed as micrograms of protein.

though it was not possible to relate enzyme levelsto either serine supply or growth rate, this nutri-tional effect was investigated further. This is nota generalized effect on enzyme synthesis, becausethe last enzyme in the serine pathway, serine-phosphate phosphatase, possesses approximatelythe same specific activity in all the extracts ex-amined. This enzyme was subsequently used as acontrol on the preparation of extracts.Experiments were performed to determine

whether an enzyme inhibitor was present in theextracts prepared from cells grown in enrichedlactate medium. Mixtures of extracts from cellsgrown in glucose and enriched lactate mediumwere assayed for their capacity to synthesizeserine-phosphate. The activity of the mixtureswas compared with that in unmixed extracts(Table 3). Because the addition of an extract ofcells grown in enriched medium did not reducethe enzyme activity in extracts of glucose-growncells, enzyme inhibitors were not present in the

extract of the cells grown in enriched nmedium.Fractionation of extracts from cells grown in theenriched medium failed to increase the specificenzymatic activity. These experiments make itunlikely that the reduced enzyme levels observedwere due to readily dissociable enzyme inhibitorsin the extract.Two enzymes are responsible for the production

of serine-phosphate, phosphoglycerate dehy-drogenase, and serine-phosphate transaminase.Assays of these enzymes in unfractionated ex-tracts are inaccurate and, in view of the pitfallsin comparing enzyme activity in extracts afterfractionation, an indirect approach was made tothe question of which enzyme limits the rate ofserine-phosphate formation in the extracts of cellsgrown in enriched medium. A fractionated ex-tract was prepared from E. coli strain Ws- grownon glucose. This extract lacks the phospho-glycerate dehydrogenase, but has the normalamount of transaminase (Pizer, 1963). Supple-mentation of extracts of cells grown in enrichedmedia with this transaminase preparation failedto increase serine-phosphate formation. It is con-cluded that phosphoglycerate dehydrogenaselimits serine-phosphate formation.To identify the compounds responsible for the

reduction in enzyme activity, the enriched me-dium was modified by omitting comlponents.Initially, several components, grouped togetheron the bases of chemical similarity, were omittedfrom the medium (Table 4). From these data, itis possible to eliminate the compounds in four ofthe six groups from a role in the regulation ofenzyme levels because their absence from themedium did not increase enzyme activity. Wheneither the group of aliphatic amino acids (leucine,isoleucine, valine, alanine) or methionine andthreonine were omitted, the enzyme level in-creased, indicating that at least one amino acidfrom each group was responsible for the originalreduction in enzyme activity. To determinespecifically which of these six amino acids wasinvolved, each was omitted in turn from a growthmedium which contained the remaining five. Theenzymatic activity in the extracts prepared fromcells grown in these media was measured (Table5). Alanine and valine were not involved in en-zyme regulation, but the other four amino acidswere required to minimize the enzyme level. Whenthe medium was enriched with L-threonine,L-methionine, L-leucine, and DL-isoleucine, the

616 J. BACTERIOL.

SERINE METABOLISM IN E. COLI

specific activity of serine-phosphate synthesis was0.52, whereas it was 1.67 in the lactate-growncontrol. It appears that four amino acids are

involved in the regulation of the biosynthesis ofphosphoglycerate dehydrogenase. The lowestconcentration of these amino acids which givesthis effect has not been determined, but doublingthe amino acid concentration failed to furtherreduce the enzyme levels. E. coli strains B andK-12 also showed this reduction in enzyme levelwhen grown with these four amino acids.

DISCUSSION

Growth experiments with strain Ws- estab-lished that the major biosynthetic pathway fromglucose to L-serine is blocked as a result of theloss of phosphoglycerate dehydrogenase (Pizer,1963). The slow growth which occurred whenexogenous L-threonine was provided probablyresults from the conversion of threonine to glycine(Roberts et al., 1955). Endogenous threonine doesnot satisfy the serine requirement of strain Ws-,since growth did not occur when the exogenous

supply of serine was exhausted. The failure ofendogenous threonine to be converted to glycineprobably reflects the maintenance of the threoninepool at a level too low for the functioning of theenzyme responsible. A major factor in regulatingthe size of the L-threonine pool is the feedbackinhibition described by Wormser and Pardee(1958). Other strains would be expected to havedifferent degrees of control and different threo-nine supplies, and could make more iserine fromthis source. Simmonds and Miller (1957) reportedthat a serine-requiring strain of E. coli K-12 can

obtain 14% of its glycine from glucose. Thisstrain has the same enzymatic deficiency as doesstrain Ws- (Pizer, unpublished data). Whenexogenous threonine is available, it can be con-

centrated in the cell to the point where itsenzymatic conversion to glycine occurs.

In the absence of a secondary source of L-

serine, the growth found with limited quantitiesof L-serine should reflect the minimal require-ments for cell synthesis. Roberts et al. (1955) pro-vided a table of the quantitative composition ofE. coli strain B, and information on what com-

pounds can derive their carbon from serine.Inspection of their data shows that approximately12 ,tmoles of L-serine or compounds directly de-rived from it, approximately 9.4 jAmoles ofglycine or compounds derived directly from it,

TABLE 4. Serine-phosphate synthesis in extracts ofcells grown on modified enriched media*

Serine-Omitted compounds phosphate

synthesist

Leucine, DL- isoleucine, L-valine,DL-alanine ........................ 0.92

L-Methionine, L-threonine .......... 0.93L-Histidine, L-arginine, L-lysine 0.52L-Glutamate, L-aspartate ........... 0.50L-Tyrosine, L-phenylalanine, L-tryp-tophan, L-proline ................. 0.53

Vitamins, uracil .................... 0.54None ........................... 0.50

* The enriched media described in Materialsand Methods was modified by omitting the com-pounds indicated.

t Expressed as micromoles per milligram per30 min.

TABLE 5. Serine-phosphate synthesis in extractsof cells grown in amino acid-enriched media*

Serine-Omitted amino acids phosphate

synthesizedt

L-Leucine . ......................... 1.07DL-Isoleucine .......... ............. 1.07L-Methionine .......... ............. 0.77L-Threonine ........... ............. 0.70DL-Alanine . ........................ 0.50L-Valine .................0............40L-Valine, DL-alanine................ 0.52All............................ 1.67

* The medium was supplemented with thefollowing six amino acids: L-leucine, DL-isoleucine,L-methionine, L-threonine, DL-alanine, and L-valine, at the concentrations used in the completeenriched medium.

t Expressed as micromoles per milligram per30 min.

and approximately 11 /Amoles of compounds de-rived from "one carbon" fragments are present in1010 cells during the exponential phase of growth.Because each L-serine molecule can give rise to aglycine and a "one carbon" fragment, the totalquantity of L-serine required is approximately 22Amoles. The balance between the serine growthrequirement, 23 Amoles per 1010 cells, and thequantity of compounds derived from it in thecell, indicates an efficient utilization of theexogenous supply of serine. Endogenously pro-duced serine would be expected to be used with

VOL. 88, 1964 617

PIZER AND POTOCHNY

equal efficiency. This amino acid is of major im-portance as a biosynthetic intermediate, sincethe quantity of glucose carbon which passesthrough serine prior to incorporation into biosyn-thetic end products is approximately 15% of thetotal assimilated carbon. These calculations pre-dict the absence of degradative reactions actingon L-serine. The reported failure of L-serine toinduce L-serine deaminase (Pardee and Prestidge,1955) is in agreement with this prediction.The results of the competition experiments

clearly show that the feedback inhibition of phos-phoglycerate dehydrogenase by serine and glycineobserved in cell-free extracts functions in the in-tact cell. In our experiments, specific activitiesof the compounds were used to measure theamount of isotope incorporated. This methodoffers advantages over the methods used inprevious studies, because quantitative recoveriesare not required and multiple determinations oneach sample can be made. Previously, reliance wasplaced on the quantitative measurement of thematerial in the test sample. The use of specificactivities appears particularly warranted with acompound like serine, which undergoes degrada-tion during hydrolysis. When serine was the com-petitor, our results agreed with those of Robertset al. (1955). However, when glycine was thecompetitor, our results differed from those re-ported. It was reported that when glucose was thecarbon source glycine did not reduce L-serinesynthesis, but did with 100% efficiency whenfructose was the carbon source. Our resultsshowed that glycine functioned as a competitorwith both carbon sources, was more efficient whenfructose was used, but still did not attain 100%efficiency. The ability of glycine to compete todifferent extents with different carbon sourcesmay have several alternative explanations.Growth on fructose may initiate a novel biosyn-thetic pathway which is regulated by glycine; theenzymes made during growth on fructose may bemore sensitive to feedback inhibition by glycine;or growth on fructose may lower the endogenousL-serine supply, making glycine a better competi-tor.The initiation of a new pathway appears un-

likely, because strain Ws-, which is blocked inthe pathway from phosphoglycerate to L-serine,maintains its nutritional requirement whengrown on fructose. Measurements of the inhibi-tion of the enzymes in the biosynthetic pathway

by glycine showed that the phosphoglycerate de-hydrogenase was inhibited to the same extent inextracts from glucose- and fructose-grown cells.The reduced levels of phosphoglycerate dehy-drogenase in extracts of cells grown in fructosesupports the hypothesis that growth on fructoselowers the endogenous L-serine synthesis. Otherfactors, such as reduction in phosphoglycerateconcentration, might contribute to lowered en-dogenous L-serine synthesis. A direct measure-ment of the endogenous serine pool in cells grownunder these different conditions would helpanswer this question.The results of the competition experiments

show that L-serine provides essentially 100% ofboth the glycine for purine synthesis and the"one carbon" fragments for thymine and purinesynthesis. Studies with a serine-glycine auxotrophby Pitts, Stewart, and Crosbie (1961) showedthat 2-C14-glycine was used for thymine andpurine synthesis. The specific activities obtainedindicate that one-carbon fragments were also ob-tained from glucose; but whether endogenousserine was their immediate precursor cannot bedetermined, since the specific activity of serineitself was not reported.The observation that four amino acids not di-

rectly involved in the biosynthesis of L-serineregulate the level of phosphoglycerate dehy-drogenase was unexpected and difficult to inter-pret in terms of a control mechanism beneficialto the cell. Possibly, this unusual regulation isconnected with the role serine plays as a precursorto a large number of metabolites. The capacityfor serine synthesis is reduced only when severalamino acids are available to the cell and the me-dium is relatively rich. The other amino acidswhich play a major role in the synthesis of met-abolic intermediates are glutamate and aspar-tate. Halpern and Umbarger (1960) showed thatgrowth of E. coli on glutamate or casein hydroly-sate reduced the level of glutamic dehydrogenaseto 25 and 4%, respectively, of that found inglucose-grown cells. Whether specific componentsof casein hydrolysate were responsible for thereduction in the enzyme level was not investi-gated. It appears that the transaminase responsi-ble for aspartate biosynthesis is also involved inthe production of other amino acids (Rudman andMeister, 1953). If the transaminase levels areregulated by the cell, it is probable that severalcompounds are involved.

618 J. BACTERIOL.

SERINE METABOLISM IN E. COLI

ACKNOWLEDGMENTS

We wish to thank N. Pon for the gift of Cl4-phosphoglycerate, and S. S. Cohen and S. Sim-monds for cultures of bacteria.

This investigation was supported by PublicHealth Service grant AI-03957.

LITERATURE CITED

BARTLETT, G. R. 1959. Phosphorous assay incolumn chromatography. J. Biol. Chem.234:466-468.

COHEN, S. S., AND H. BARNER. 1955. Enzymaticadaptation in a thymine requiring strain ofEscherichia coli. J. Bacteriol. 69:59-66.

HALPERN, Y. S., AND H. E. UMBARGER. 1960.Conversion of ammonia to amino groups inEscherichia coli. J. Bacteriol. 80:285-288.

ICHICHARA, A., AND D. M. GREENBERG. 1957.Further studies on the pathway of serineformation from carbohydrate. J. Biol. Chem.224:331-340.

LEVINE, E. M., AND S. SIMMONDS. 1960. Metaboliteuptake by serine-glycine auxotrophs of Esch-erichia coli. J. Biol. Chem. 235:2902-2909.

LOWRY, 0. H., N. J. ROSEBROUGH, A. L. FARR,AND R. J. RANDALL. 1951. Protein measure-ment with the Folin phenol reagent. J. Biol.Chem. 193:265-275.

MEINHART, J. O., AND S. SIMMONDS. 1955. Serinemetabolism in a mutant strain of Escherichiacoli strain K-12. J. Biol. Chem. 213:329-341.

NEUHOUS, F. C., AND S. KORKES. 1958. Phospho-serine, p. 75-79. In C. S. Vestling [ed.], Bio-

chemical preparations, vol. 6. John Wiley &Sons, Inc., New York.

NOVICK, R. P., AND W. K. MAAS. 1961. Control byendogenously synthesized arginine of theformation of ornithine transcarbamylase inEscherichia coli. J. Bacteriol. 81:236-240.

PARDEE, A. B., AND L. S. PRESTIDGE. 1955. In-duced formation of serine and threoninedeaminases by Escherichia coli. J. Bacteriol.70:667-674.

PITTS, J. A., J. A. STEWART, AND G. W. CROSBIE.1961. Observations on glycine metabolism inEscherichia coli. Biochim. Biophys. Acta50:361-363.

PIZER, L. I. 1963. The pathway and control ofserine synthesis in Escherichia coli. J. Biol.Chem. 238:3934-3944.

ROBERTS, R. B., P. H. ABELSON, D. B. COWIE,E. T. BOLTON, AND R. J. BRITTEN. 1955.Carnegie Inst. Wash. Publ. 607.

RUDMAN, D., AND A. MEISTER. 1953. Transamina-tion in Escherichia coli. J. Biol. Chem. 200:591-603.

SIMMONDS, S., AND D. A. MILLER. 1957. Metab-olism of glycine and serine in Escherichia coli.J. Bacteriol. 74:775-783.

SMITH, I. 1958. Chromatographic and electro-phoretic techniques. Interscience Publishers,Inc., New York.

UMBARGER, H. E., M. A. UMBARGER, AND P. M. L.SIu. 1963. Biosynthesis of serine in Escherichiacoli and Salmonella typhimurium. J. Bacteriol.85:1431-1439.

WORMSER, E. H., AND A. B. PARDEE. 1958. Regula-tion of threonine biosynthesis in Escherichiacoli. Arch. Biochem. Biophys. 78:416 432.

VOL. 88, 1964 619