Stringent Control of Fatty Acid Synthesis in Escherichia coli · that st,ringent control is exerted at the level of acetate transport and activation, rather than at the level of fatty

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 245, No. 22, Issue of Nowmher 25, pp. 7957-79130, 1973

Printed in U.S.A.

Stringent Control of Fatty Acid Synthesis in Escherichia coli

POSSIBLE REGULATION OF ACETYL COENZYME A CARBOXYLASE BY ppGpp*

(Received for publication, July 5, 1973)

S. EFTHIMIOS POLAKIS, RAS B. GUCHHAIT, AND M. DANIEL LAKE

From the Department of Physiological Chemistry, The Johns Hopkins Univekty School oj Meclicine, Baltimore, Maryland d1205

SUMMARY

The effect of amino acid starvation on the rate of fatty acid synthesis was examined in stringent (CP 78, ref+) and relaxed (CP 79, rel-) isogenic strains of Escherichia coli (leu-, his-, arg-, fhr-, fhi-). Rates of incorporation of [UJ4C]glucose, [lJ4C]acetate, and 3H,0 into chloroform-methanol-soluble lipids ( >95% phospholipid) were instantly reduced 50 to 60 % by leucine starvation of stringent, but not relaxed, cells. That this depressed rate of lipid labeling was not due to leucine starvation-induced diversion of labeled fatty acyls into chloroform-methanol-insoluble form (e.g. lipopolysaccharide, etc.) is indicated by the fact that the rate of labeling from [lJ4C]acetate of the lipid extract of saponified cells was decreased to the same extent by leucine deprivation. With leucine-deprived rel+ cells, the rate of [1-14C]acetate incorporation into phosphatidylethanolamine and cardiolipin was more drastically curtailed than into phosphatidylglycerol, while the incorporation pattern in reZ- cells (A leucine) was similar to that of nonstarved reZ+ cells. An elevated turnover rate of fatty acyl groups due to amino acid starvation cannot account for the decreased rate of lipid labeling by various precursors, since no detectable loss of 14C-activity from fatty acyls labeled during growth or leucine deprivation occurs during subsequent growth or leucine starvation in the presence of unlabeled precursor. Only minor amounts of labeled lipid are secreted by stringent or relaxed cells grown in the presence of [1-r4C]acetate; leucine deficiency has no significant effect on the rate of labeling of these extracellular lipids. These results lead to the conclusion that fatty acid synthesis per se is subject to stringent control being partially suppressed, i.e. 50 to 60%, by amino acid starvation in rel+ strains of E. coli.

Amino acid starvation of stringent, but not relaxed, strains of E. coli is known to activate ribosomal synthesis of ppGpp and pppGpp. The resultant accumulation of (p)ppGpp appears to mediate stringent control of fatty acid synthesis by blocking the initial committed step of the pathway, i.e. the carboxylation of acetyl-CoA. Of the two catalytic components of the acetyl-CoA carboxylase system, namely biotin carboxylase and carboxyltransferase, only the latter component is inhibited by physiological concentrations (up

* This work was supported by grants from the National Insti- tutes of Health, United States Public Health Service (A&14575) and the American Heart Association, Inc.

to 4 mM) of ppGpp. Two carboxyltransferase-catalyzed reactions-(a) carboxyl transfer from malonyl-CoA to the model acceptor, d-biotin methyl ester, and (b) carboxyl carrier protein-dependent malonyl-CoA-[lJ4C]acetyl-CoA exchange-are inhibited 50 to 60% at saturating concentrations (1.0 to 1.2 mM) of (p)ppGpp. Inhibition is specific in that GTP, GDP, GMP, ATP, ADP, AMP, and cyclic adeno- sine 3’:5’-monophosphate have no effect at comparable concentrations.

The binding of Mn2+ by ppGpp was investigated by electron paramagnetic resonance to evaluate the possible metal scavenging effect of the guanosine nucleotide on the metal- dependent reactions (i.e. biotin and acetyl-CoA carboxylation) catalyzed by the biotin carboxylase component of acetyl- CoA carboxylase. K, values determined at pH 6.8 for the dissociation of Mn2+ from the binary (MneppGpp) and ternary (Mnz*ppGpp) complexes are 1.1 PM and 65 PM, respectively, the former being tighter and the latter an order of magnitude weaker than Mn2+ dissociation from MneATP. Whereas the biotin carboxylase-catalyzed reaction per se (carboxylation of free d-biotin) is not inhibited, 2.5 mu Mn.ppGpp inhibits both acetyl-CoA carboxylation and malonyl-CoA-[1-*%]acetyl-CoA exchange 50 to 60%. These results suggest that stringent control of fatty acid synthesis in E. cofi is mediated through the inhibitory action of (p)ppGpp on the carboxyltransferase component of the acetyl-CoA carboxylase system.

A number of reports (1, 2) have implicated the product of the relaxed’ (Tel) locus in the control of lipid synthesis in Escherichia coli. Evidence supporting this view comes solely from compari- sons of the rates of incorporation of [r4C]acetate or 32Pi into the chloroform-methanol-extractable lipids of reZ+ and rel- strains during growth and amino acid starvation. Since the turnover of the fatty acgl moieties of these lipids was not investigated, it was not established whether fatty acid synthesis per se is affected by amino acid deprivation. Moreover, the possibi1it.y

1 The rel gene of E. coli, sometimes referred to as the KC locus, is responsible for the stringent response, i.e. the curtailment of accumulation of stable RNA species 11pon starvation for a required amino acid (3).

7957

by guest on February 7, 2020http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7958

that st,ringent control is exerted at the level of acetate transport and activation, rather than at the level of fatty acid synthesis, has not been excluded. Although an effect of the stringent control mechanism on acetate transport has not been established, various other transport processes are inhibited (4-S). However, it is known that acetate transport is affected by other factors, e.g., the presence or absence of glucose in the medium (9). In addition, it has been our experience that the rate of [14C]acetate incorporation into the lipids of E. coli, in the presence of glucose as the major carbon source, accounts for <15% of the total rate of lipid synthesis, presumably due to the dilution of the [14C]acety1-CoB pool by unlabeled precursors or rate-limiting acetate t,ransport. Furthermore, it has been reported that the reduction in the rate of [%]acetate incorporation into the lipids of a stringent (Tel+) strain during amino acid starvation is dependent upon the degree of aeration of the culture (2, 10, 11). For these reasons, the relationship of the stringent control mechanism to the regulation of fatty acid synthesis in E. coli was re- examined. Beyond this, we have found that guanosine 5’- diphosphateSdiphosphate and guanosine 5’-triphosphate-3’- diphosphate, which appear to mediate the stringent response (12-14), specifically inhibit acetyl-CoA carboxylase at physiological guanosine nucleotide concentrations. In E. coli the acetyl-CoA carboxylase system is composed of three easily resolved protein components, biotin carboxylase which catalyzes the first half-reaction (Reaction I), carboxyltransferase which catalyzes the second half-reaction (Reaction 2), and carboxyl carrier protein which contains the covalently bound biotinyl prosthetic group (15-19).

Me*+ ATP + HCO,- + CCP-biotin2, BC ’

.CCP-biotin-C02- + ADP +IPi

CCP-biotin-COB- + acetyl-CoA \ CT

.malonyl-CoA + CCP-biotin

Me*+ Net: ATP + HCO,- + acetyl-CoA ~

ICC, CT, CCP-biotin

.malony-CoA + ADP + Pi

This investigation shows that acetyl-CoA carboxylation is

(1)

(2)

(3)

inhibited by physiological concentrations of (p)ppGpp2 and that the site of action of the nucleotides is the carboxyltransferase reacstion. A preliminary report of this work has been presented

(20).

EXPERIMENTAL PROCEDURE

Reagents-Enzymes, nucleotides, and other cofactors were purchased from Uoehringer, P-L Riochemicals, and Sigma, and radioactive materials were from New England Nuclear. All other reagents were the best grade available. ppGpp and pppGpp were generously provided by Dr. M. Cashel (National Institute of Child Health and Human Development, National Institutes of Health) and Dr. R. lsyrne (Johns Hopkins Uni-

2 The abbreviations used are : ppGpp, guanosine 5’-diphosphate- 3’.diphosphate; pppGpp, guanosine 5’.triphosphate-3’-diphosphate; 5’.phosphate substitutions are written to the left of the nu- cleoside, and 3’.phosphate substitutions to the right as suggested by Cashel et all (12). (p)ppGpp, ppGpp and pppGpp collec- tivelv: BC. CT, and CCP. the biotin carboxvlase. carboxvl- transfkrase; and carboxyl carrier protein componedts, respectively, of the E. coli acetyl-Cob carboxylase system.

versity School of Medicine) and were purified chromatographi- tally on Whatman DEAE-52 cellulose. The purified guanosine nucleotides gave rise to single ultraviolet light-absorbing spots by polyethyleneimine thin layer chromatography (21) and the RF values were identical with those of authentic ppGpp and

PPPGPP. Bacterial Strains and Growth Conditions-The E. coli K-12 iso-

genie strains CP 78 (Tel+, thi-, arg-, his-, Zeu-, thr-) and CP 79 (Tel-, thi-, arg-, his-, leu-, thr-) were kindly supplied by Dr. Niels Fiil (Universitetets Mikrobiologiske Institut, Copenhagen, Denmark). The Tris-glucose minimal medium of Gallant and Cashel (22) containing 2 mM phosphate, 0.2% glucose, thiamine (50 mg per liter), and required amino acids (1 mM each) was used. Cells were grown in 125-ml cylindrical gas-washing bot- tles at 37” under vigorous aeration; growth was monitored by the increase in the optical density of the culture at 450 nm. Cells were harvested during early logarithmic growth when the optical density of the culture reached a value of 1.0. Under the conditions described in this paper A4s0 nm w?s linearly related to cell dry weight during logarithmic growth.

Conditions for Labeling the Lipid Fraction of E. coli Cells- Cells were harvested by rapid filtration on Millipore filters (HAWP, 0.45 p), washed extensively at 37” with new medium minus labeled precursor. The washed cells were resuspended by vortexing in the appropriate medium (plus or minus leucine) containing labeled precursor to yield a cell density of 1.0 A unit per ml. The total time required for this procedure was 1.5 to 2 min. The following substrates were used: [1-%]acetate (15 &i per pmole) at a concentration of 0.075 or 0.9 InM, [U-W-

glucose (1 &i per pmole) at a concentration of 11.6 InM, and 3Hz0 (10,000 cpm per pmole). After incubation for various periods of time at 37” under constant aeration, l.O-ml aliquots of cell suspension were filtered on Millipore filters, and the cells washed three times with 5 ml of H?O. Filter discs with cells were extracted with chloroform-methanol or were saponified with 2.3 x KOH in 50% ethanol and extracted subsequently as described below. Cell filtrates were collected for the determination of radioactivity incorporated into fatty acyl residues excreted into the medium, as well as for the assay of residual glucose (23) in the medium. At the end of all labeling experiments, both the glucose concentration and the level of labeled precursor were >90% of the initial value.

Extraction of CelMar Lipids with Chloroform-Methanol-The cells and filter were extracted overnight at room temperature with 5 ml of chloroform-methanol (1 :I, v/v) ; the extract and cells were filtered and the filter washed three times with 4 ml of chloroform-methanol (2: 1, v/v). The filtrate and washings were back-washed with 5 ml of water and subsequently with 5 ml of Folch upper phase solution (24). The chloroform layer was transferred into counting vials, the solvent evaporated under a stream of nitrogen, and the labeled lipids counted in a scintillation counter.

Saponification of Cells and Extraction of Fatty Acids-Cells and filter discs were transferred to conical tubes containing 3.5 ml of 2.3 N KOH in 50% ethanol, stoppered, and heated at 85-90” for 6 hours. After evaporation of the alcohol on a boiling water bath, the tubes were cooled to O”, and 3 ml of 6 x HCl were added while vortexing. Free fatty acids were extracted with 6 ml of n-pentane or petroleum ether (b.p. 38.1-47.4”) and the water phase discarded; the upper phase was washed twice with 5 ml of water, the pentane layer t’ransferred to counting vials, the pentane evaporated under a st,ream of nitrogen, and the fatty acids taken up in liquid scint,illator and count,ed.


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7959

Determination of Total Fatty Acyl Groups Excreted into Me- dium--The cell filtrates were concentrated to dryness in a rotary evaporator; when [Wlacetate was present, the residue was resuspended in 5 ml of 0.1 s acetic acid and taken to dryness several times. The residue was resuspended in 2 ml of 2.3 N alcoholic KOH (ethanol-water, l:l, v/v), transferred into a conical tube and the flask was washed twice with 2 ml of alcoholic KOH. The residue and washes were hydrolyzed, and fatty acids were rstracted with pentane or petroleum ether and counted as described above.

Determination of Turnover of Labeled Fatty Acyl Groups during Growth or Leucine Deprivation--To cells entering log phase growth was added 0.017 mM [I-Wlacetate (15 PC1 per pmole). After growing to an optical density of 1.0, the labeled cells were harvested by filtration and washed repeatedly on the filter with the appropriate medium containing unlabeled acetate. The washed cells were resuspended to an optical density of 1.0 in unlabeled complete or leucine-deficient media and the fate of the labeled fatty acyl groups followed during a subsequent l-hour incubation at 37” under vigorous aeration. Aliquots of cells withdrawn at various time intervals were worked up as before, and radioactivity in the chloroform-methanol extract of cells, in the total fatty acyls of cells, and in the total fatty acyls in material excreted into the growth medium was determined as described above.

To label cells during leucinc starvation, growth was allowed to proceed to an optical density of 1 .O unit on complete medium. After harvesting and washing with medium devoid of leucine, the cells were resuspended in the same medium to an optical density of 0.5 unit. [I-W]Acetate was added (0.075 mrvr, 15 &i per pmole), and the cells were incubated for 1 hour at 37” with constant aeration. During this period there was no de- monstrable decrease in viable cell count. After labeling under conditions of leucine starvation, the cells were harvested by filtration, washed with medium devoid of leucine but containing unlabrlcd acetate, and resuspended in unlabeled complete or leucine-deficient media to an optical density of 1.0 unit. The fate of fatty acids labeled in the starvation period was followed during the subsequent growth or leucine starvation as described above.

Distribution of Labeled Patty Acyl Residues in Chlorojorm- dfetlzanol-extractable Lipids-Cells washed with the appropriate medium were resuspended to an optical density of 1.0 in either complete medium or medium lacking leucine. [I-14ClAcetate was added (0.141 mM, 1.6 PCi per ml) and the cells incubated with constant aeration at 37” for 30 min before extraction with chloroform-methanol. The extracts were brought to dryness under reduced pressure, redissolved in chloroform, and chromat- ographed on Whatman silica gel loaded paper SG-81. For two- dimensional chromatography, the first solvent system was petroleum ether-diethyl ether-acetic acid (9O:lO: 1, v/v), and the second solvent was chloroform-methanol-acetic acid (65 : 25: 8, v/v) (25). The latter system was also used for one-dimensional chromatography. Lipid spots were identified by autoradiog- raphy using appropriate standards. The chromatograms were cut into strips, counted, histograms constructed, and total radioactivity in each lipid class calculated.

Measurement of MrP Binding by ppGpp-Measurements were performed at 30” in 48 mM N-tris(hydroxymethyl)methylU aminoethanesulfonic acid buffer, pH 6.8, by electron paramagnetic resonance spectroscopy at 9150 MI-I, essentially according

Manganese ion was titrated with ppGpp, and free manganese was determined at all six peaks of the Mn2f spectrum by com- paring the height of each peak to the height of Mn2+ standards. For each experiment the average free Mn2+ concentration was calculated from the values of fret ;Mn2+ obtained at each ab- sorption maximum, and average values were used to determine the dissociation constants of the ppGpp complexes with lMn2+. Results were fitted by the least squares method.

Enzyme Preparations and Assays-Biotin carboxylase, carboxyltransferase, and carboxyl carrier protein were purified (27, 28)3 from three-fourths log phase E. coli 1% cells grown on enriched medium purchased from Grain Processing Corp. (Mus- catine, Iowa). Riotin carboxylase and carboxyltransferase were homogeneous preparations, and the carboxyl carrier protein was estimated to be >507, pure. All enzyme activities were determined at 30”. Riotin carboxylase was assayed’radiochemically in a total volume of 0.05 ml as described by Dimroth et al. (15) ; however, 1 mM MnClz was used instead of 8 m&t MgCl,. The specific activity of KHWOS was 20 x lo6 cpm per pmole and the reaction was initiated by addition of ATP.‘ The progress of the reaction followed at 1, 2, 3, and 4 min of incubation was linear.

Carboxyltransferase activity was determined either spectro- photometrically in a total volume of 0.2 ml or radiochemically in 0.1 ml, pH 8.0, using biotin methyl ester as model acceptor substrate as described by Guchhait et al. (17, 27), but with the following modifications. The concentration of buffer was reduced to 20 mM, and the coupling enzymes employed in the spectrophotometric assay were freed of ammonium sulfate by dialysis. The progress of the reaction in the radiochemical assay Teas followed kinetically at 1, 2, 3, and 4 min. The reaction m-as initiated by addition of a mixture of substrates.

Malonyl-CoA-[l-i4C]acetyl-CoA exchange activity was meas- ured at pH 6.5 in a volumeof 0.05 ml (27).3 The specific activity of the [l-Wlacetyl-CoA employed was 5 x lo6 cpm per pmole. The reaction initiated with a mixture of malonyl-CoA and [‘“Cl- acetyl-CoA was followed kinetically at 2, 4, 6, and 8 min of incubation.

Acetyl-CoA carboxylase activity was determined at pH 6.8 to 6.9 in a final volume of 0.05 ml containing: imidazole (Cl-) buffer, 50 mM; bovine serum albumin, 0.03 mg; GSH, 3 m&r; KHi4C03, 10 mM (10 to 30 X 10” cpm per pmole) ; MnC12, 1 IIIM; ATE’, 1 mM; and enzyme (25 to 42yc dialyzed ammonium sulfate fraction (as)), 60 to 100 pg. The reaction initiated by the addition of ATP was followed kinetically at 1, 2, 3, and 4 min.

RESULTS

Effect oj Leucine Starvation on Rates of Incorporation 01 [I-‘“Cl- Acetate, [U-W]Glucose, and 3Hz0 into Lipids of E. coli-To as- certain whether the apparent inhibitory effect of amino acid starvation on the labeling of lipids by [i4C]acetate is exerted on fatty acid synthesis per se or upon acetate transport and activation, or both, the rate of [%]acetatc incorporation into the lipids of rel+ cells (CP 78) during leucine starvation was compared with the rates of incorporation of [U-i4C]glucose and 3Hz0. When [W]glucose is used both as labeled precursor and sole carbon source, dilution of the [14C]acetyl-CoA pool by unlabeled precursors is precluded; consequently, the rate of lipid labeling from [‘4C]glucose should more closely approximate the in vivo rate than from [i4C]acetate. Moreover, the use of ‘Hz0 (29-31)

to Cohn and Townsend (26). The volume of the sample was 3 R. B. Guchhait, S. E. Polakis, J. Moss, P. Dimroth, and M. 0.05 ml, and the temperature was controlled to within kO.5”. D. Lane, manuscript in preparation.


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7960

as precursor of fatty acids is not subject to the same limitations as [14C]acetate or glucose and probably provides the best esti- mate of the rate of fatty acid synthesis, since incorporation of tritium from %I120 is independent of the source of acetyl-CoA and unaffected by transport processes. It was found (Fig. 1) that the reduced rate of incorporation of all three radioactive precursors into chloroform-methanol-extractable lipids of stringent cells caused by leucine deprivation is independent of the nature of the precursor. Essentially identical results were obtained using two concentrations of [I-14C]acetate, [U-14C]glucose, or 3Hz0. In all instances a 50 to 60% reduction in rate of lipid labeling was observed when results were normalized to counts per min incorporated per optical density unit of cells. Nor- malization was necessary to account for the fact that cells in complete medium were growing logarithmically while the leu- tine-starved cells were not. During the course of the 60.min

L - 3x10”

Complete

- 2x105

-Iw - IXIO~

[lJ-‘4C]PIUCOs* 11.2mM

/

- 6x10’ Complete

a 2xlo4 20 40 60

MINUTES

I I I

20 IwNu4ToES

60

FIG. 1. Incorporation of labeled precursors into the chloroform- FIG. 2. Incorporation of label from 3Hs0 into the chloroform- methanol-soluble lipids of Escherichia coli (CP 78 Tel+). For methanol-soluble lipids of Escherichia coli (CP 79 rel-). For details see “Experimental Procedure.” l , complete medium; details see “Experimental Procedure.” 0, complete medium; 0, medium minus leucine. 0, medium minus leucine.

incubation in complete medium and leucine-deficient medium optical density at 450 nm increased approximately 100% and <lo’%, respectively. As anticipated (Figs. 2 and 3B), the

rates of incorporation of 3Hz0 and [I-%]acetate into the lipid fraction of relaxed cells (CP 79, Tel-) either in complete medium or in the absence of a required amino acid were essentially identical. These results suggest that the effect of stringent control on the rate of incorporation of [l%]acetate into lipids is not exerted at the level of acetate transport or activation. Hence, [14C]acetate incorporation into the lipid fraction of cells is a valid measure of the stringent control of lipid labeling. However, before it can be concluded that this control mechanism operates at the level of fatty acid synthesis per se, the possibility that amino acid starvation affects fatty acyl turnover or excretion, or both, had to be excluded.

Effect of Leucine Starvation on Rate of [l-lY?]dcetate Incorporic- tion into Total Cellular and Extracellular Fatty Acyl Groups--It

is well documented that E. coli cells excrete into the medium a lipopolysaccharide which may originate from the outer cell membrane (32-38). While the amount excreted into the medium is usually small in relation to total cellular fatty acids, it is in-

Complete \.

. o

/ “\ -leu

A and In

B c A. total fotty acyls B. CHQ-MeOH sol. C.extrocell. fatty acyls

= I I I 8 5xI05r5xI0’

z E 4x13 -4x10’

Complete

2 3XIOJ - 3x103

? 6 2Xld-2x105

~

-leu d lXIOS- 1x10’

&I 0.

20 40 60

Incubation time ,min

FIG. 3. Effect of leucine starvation on the rate of lipid synthe- methanol-soluble lipids of cells. C, extracellular fatty acyl resi- sis in stringent (CP 78 Tel+) and relaxed (CP 79 rel-) strains of dues determined after saponification (as in A) of the residue of Eschem’chia coli. For details see “Experimental Procedure.” the medium after removal of the cells. 0, complete medillm; A, total fatty ncyl residues in cells determined after saponifica- 0, medium minus leucine. Note the difference in the ordinate tion at 90” for 6 hours with 2.3 N alcoholic KOH. B, chloroform- scales of A or B and C.


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7961

Labeling medium-

Incubation medium-

40 60 0 20 40 60 0 20 40 60 Incubation Time , m in.

FIG. 4. Turnover of fatty acyl groups during growth and leucine denrivation in strinnent (CP 78 rel+) and relaxed (CP 79 ret?-1

activity per ml of medium was determined at different times during incubation. 0, total fatty acyl residues determined as free fatty acids after saponification of the cells; q , chloroform- methanol-soluble lipids of cells; A, extracellular fatty acyl residues determined as free fatty acids after saponification of the dried medium.

strains of Escherich& coli For details see “Expermrental Pro- cedure.” Cells labeled with [1-l%]acetate during growth (Panel a) or during leucine starvation (Panel B) were harvested and washed to remove the labeled substrate and then transferred to the incubation medium (complete or minus leucine) and the radio-

creased during amino acid starvation (38). Furthermore, a fraction of the fatty acyl groups synthesized by the cell is incorporated into cell wall components not extractable by the chloroform-methanol procedure. Consequently, it could be argued that the differences observed during amino acid starvation of stringent strains were not due to decreased fatty acid synthesis, but to increased incorporation of synthesized fatty acids into cell wall components which were either not readily es&actable or were extruded into the medium. This possibility was tested by determining the rate of incorporation of [r4C]- acet,ate into the petroleum ether-soluble fraction after comp1et.e saponification of the harvested cells and the growth medium. Comparison of A and B in Fig. 3 reveals that a significant fraction (about 2Oojg) of the cellular labeled fatty acyls became extractable by organic solvent only after saponification. However, [i4C]acetate incorporation into total cellular fatty acyl groups was inhibited by amino acid starvation to the same extent (50 to 607,) as [i4C]acetate incorporation into the chloroform- methanol extract. In growing cells, 83% of the total label in fatty acids was extractable with chloroform-methanol, while during starvation t)he amount extractable was reduced to 78%.

The amount of W activity incorporated into fatty acyls of material excreted into the medium was insignificant in all cases (Fig. 3C) and not materially altered by leucine starvation. It can be concluded, therefore, that amino acid deprivation causes a marked reduction in the rate of labeling of fatty acyls per se in the stringent strain (Tel+) of E. coli.

E.ffect oj Leucine Starvation on Turnover4 of Fatty Acids Xyn- fhesized during Growth or starvation-Since the reduction in rate of incorporation of radioactive precursors into fatty acyls of the st,ringent ,strain (CP 78) during leucine starvation (Figs. 1 and 3) could hare resulted from accelerated turnover of fatty acyl moieties rather than from a reduced rate of synthesis, the fate of

d “Turnover” as used in this paper is defined as the conversion of fatty acyl groups into a form not extractable with petroleum ether after saponification and acidification.

fatty acids synthesized in stringent and relaxed cells during growth and starvation was examined. It is evident from Fig. 4A that fatty acids of relf cells, labeled with [14C]acetate during growth on complete medium, showed no loss of label during subsequent incubation in unlabeled medium either in the presence or absence of leucine. Not only did the 14C activity in total fatty acids remain constant, but label in the chloroform-methanol extract did as well. This result argues against transfer of fatty acyls from structures extractable with chloroform-methanol to substances (e.g. lipopolysaccharide) insoluble in this solvent. Moreover, only an insignificant amount of radioactivity was detected under any circumstance in fatty acyls of substances excreted into the medium (<500 cpm of a total of 33,000 cpm in lipids). Similarly, there was essentially no turnover of W- fatty acyls in the lipids of the rel- strain (CP 79) labeled during growth (Fig. 4A, lower panels). Virtually identical results were obtained in correlative experiments (Fig. 4B) in which the lipids of both rel+ and rek cells were labeled with [i4C]acetate during leucine starvation; no detectable loss of i4C label in fatty acyl groups or the chloroform-methanol extract occurred during subsequent incubation of the cells in complete or leucine-deficient medium. Again insignificant amounts of radioactive fatty acyl groups were excreted into the medium (Fig. 4B).

While E. coli possesses an inducible system for /? oxidation of fatty acids (39, 40), this pathway is strongly repressed by glucose (39), a component of the medium employed here; hence, this degradative system would not be expressed and, therefore, could not account for the reduced rate of labeling of fatty acyl groups observed during amino acid starvation. Moreover, Crowfoot et al. (41) have described a mutant of E. coli (fao-6) which is p oxidation- and yet it exhibits stringent control of fatty acid synthesis. This observation argues against the possibility that a small pool of fatty acyl CoA’s is rapidly degraded, resulting in turnover of fatt.y acids before incorporation into phospholipids. Since the r t a e of incorporation of labeled precursors into total cellular and extracellular fatty acyl groups is


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7962

curtailed by amino acid deficiency and yet fatty acyl turnover is not altered, one is led to the conclusion that fatty acid synthesis per se is blocked by amino acid starvation.

In contrast to t.he stability of the fatty acyl residues of E. cobi phospholipids (this work, 42), the phosphoryl group turns over at an appreciable rate and, in addition, the extent of turnover varies among different classes of phospholipids. Thus, substantial turnover of phosphatidylglycerol and cardiolipin has been observed during growth and starvation of both stringent and relaxed strains (2, 43, 44). On the other hand, phosphatidylethanolamine has been shown to be exceedingly stable under most conditions. Turnover of the phosphoryl group of the latter phospholipid has only been observed during amino acid starvation of relaxed strains (2) and in shifting from growth at 37” to growth at 10” (42). It is interesting that in neither case has turnover of fatty acids been demonstrated (this work, 42).

E$ect of Leucine Starvation on Distribution OJ Radioactivity from [I-14C]Acetate in Lipids of Xtringent and Relaxed Strains of E. coli-An investigation of [PC]acetate incorporation into the chloroform-methanol extractable lipids of rel+ or rel- strains of E. coli grown on complete or leucine-deficient media showed that >957; of the label appeared in phospholipids (phosphatidyl- ethanolaminc, phosphatidylglycerol, and cardiolipin). Two- dimensional thin layer chromatography and radioautography revealed virtually no free fatty acids, in agreement with the findings of Mindich (45). The results summarized in Table I show no substantial difference in the distribution of 1% activity incorporated into the three major classes of phospholipids in rel+ or rel- cells growing in complete medium or rel- cells deprived of leucine. The possible exception to this may be cardiolipin which comprised ~37~ of the label in reZ+ and rek cells grown on complete medium, but 6% in leucine-starved rek cells. Aside from the latter difference, the relative insensitivity of relaxed cells to amino acid starvation is similar both with respect to labeling pattern and rate of fatty acid synthesis. However, leucine deprivation of rel+ cells caused a large relative decrease in the percentage of label incorporated into phosphatidyletha- nolaminc and a corresponding increase in labeling of phosphatidylglycerol; the synthesis of cardiolipin under these circum- stances appears to be abolished. Considering the fact that total fatty acyl, and presumably phospholipitl, synthesis was reduced

TIHLE I

DisLribuLio?~ of label front [l J4C]acelale into the major phospholipid classes of E. coli

Exponentially growing cells (A430 mn = 1) were harvested and resllspended in new medium to a cell density of 1.0 A lmit. [l- 14C]Acetate was added, and the cells were incubated with constant aeration at, 37” for 30 min before they were harvested and extracted. Chromatography of lipids was carried ollt on silica gel-loaded paper (Whatman SG 81) with chloroform-methanol- acetic acid (G5:25:8, v/v) solvent. For details see“Experimenta1 Procedllre.” PE, phosphatidylethanolamine; PG, phosphatidylglycerol; CL, cardiolipin.

Strain and condition

[WIAcetate incorporation into

PI? I ICL

PC

I % lOlO1 ‘C aciisil>

CP78 (/Xl+) complctc / 59.2 I x.0 3.3 CP78 (~,el+) minIIs lerlcine. . 48.4 <0.5 CPi9 (I,&) c-ompletr 37.7 3.3 Cl’79 (ret) minIIs lellcine 36.3 (i .9

by about 507. by leucine starvation, net pl~osphaticl~lethanol~ amine synthesis must have been curtailed to a much greatel estcnt (70 to 75yJ than net phosphatidylglyccrol synthesis (25 to 30%). Moreover, net de novo cardiolipin synthesis appears to be completely blocked by leucinc deprivation of relf cells.

Guanosine 5’-Dipllospkate-S’-diplaospkate (ppCpp) os Possible Mediator of Stringent Control of Fatty Acid Synthesis at the Level of Acetyl-CoA Carbozylase-The in vivo labcling experiments with rel+ and rel- strains dtscribed in the preceding sections indicate that stringent control is exercised at the level of fatty acid sy11- thesis iu E. coli. These findings raise two important questions. (a) At what site(s) in the fatty acid synthetic pathway is strill- gent control exerted? (b) What factor(s) mediate this control? Since acetyl-CoA carboxylase catalyzes the committed strp of fatty acid synthesis in E. coli and since there is prrccdent for its regulation in animal cells, this enzyme was considered a likely candidate as a target of stringent control. In this connection, Gallant and Cashel (12) have demonstrated the appearance of two unusual nucleotides-ppGpp and pppGpp-in stringent, but not relaxed, strains of E. coli during amino acid starvation. The kinetics of formation of these nucleotides upou removal of a required amino acid and of their disappearance upon resupple- mentation (13, 14) is compatible with their suggested role as mediators of the stringent response. Furthermore, these nucleotides have been shown to inhibit key enzymatic activities of various pathways subject to stringent control (46-51). Hence, we considered the possibility that ppGpp or pppGpp, or both, might act as negative effecters on the catalytic components of E. coli acetyl-CoA carboxylase.

Interestingly, it was found that the carboxyltransferasr component, one of the two resolved catalytic elements of the E. coli acetyl-Cob carboxylase system, is inhibited by ppGpp and pppGpp. As illust,rated in Fig. 5, both carbosyltransferas;c- catalyzed transcarbosylation (Reaction 4) and malonyl-Co& [l-14C]acetyl-Cod exchange, which measures the second half- reaction (Reaction 2) in acetyl-CoA carboxylation, are inhibited by ppGpp. pppGpp, which also blocks the model caarbosyl transfer reaction (Fig. 5), was not available in sufficient quantity to test its effect on the cxchango reaction.

Malonyl-CoA + cl-biotin methyl ester F 1’.Y- (1)

carboxy-c-biotin methyl ester + ace@-CoA

Malonyl CoA-[‘“Clacetyl-CoA exchange is inhibitc~tl to the same extent by ppGpp whether tested with crude unresolved acetyl- CoA carboxylase preparations or with a reconstituted system of homogeneous carbosyltransferase and carboxyl carrier protein (Fig. 5, lower panel). Importantly, maximal inhibition of both carboxyltransferase-catalyzed reactions, i.e. 50 to 6O’i;, is achieved at concentrations of ppGpp which ~orrcspolld to those generated in vivo (up to 4 mtix) during amino acid starvation of rel+ cells (50). Mort’over, 50 to 60% inhibition at saturating ppGpp concentration agrees well with the extent to which fat’ty acid synthesis is cturtailcd in vivo by amino a(%1 deprivation (Figs. 1 to 3).

It should also be noted that inhibition of the carbosyltransferase reaction by ppGpp is specific; none of the other nucleotides tested, including 5’.GTI’, 5’-GDP, 5’-Gi\Il’, 5’ATI’, 5’.-iDI’, 5’.AMP, and cycBlic atlcnosine 3’ : 5’.moaoy)~losphate, wcrc inhibitory at a concentration where the ppGpp effect i:: maximal (set Table II). Furthermore, it appears that the two inhibitory nuclcotides, ppGpp and pppGpp, bind at a common site on the


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7963

pppGpp(CT-catalyzed carboxyl

ppGpp (CT-catalyzed carbo~yl L!gQ;::)

O=crude enz

[(P)PPGPP] .mM

overoll+Mn2+ q ‘exchanw +Mn2+

\ eicha&Mn2+

FIG. 5 (left). Inhibition of the carboxyltransferase (CT)- catalyzed reactions by pppGpp and ppGpp. For details of the enzymatic assays see “Experimental Procedure.” l’op panel, inhibition by pppGpp of carboxyl transfer from malonyl-CoA to free d-biotin methyl ester catalyzed by homogeneous carboxyltransferase (4.8 milliunits). Middle panel, inhibition by ppGpp of the same reaction catalyzed by homogeneous carboxyltransferase (4.8 milliunits). Bottom panel, inhibition by ppGpp of malonyl-CoA-[lJ4C]acetyl-CoA exchange catalyzed by crude acetyl-CoA carboxylase (70 pg of dialyzed 25 to 427, satrlrated ammonium sulfate fraction equivalent to an exchange activity of 0.12 nmole per min) (0) or by a system reconstituted from 5.3 milliunits of homogeneous carboxyltransferase and 0.11 nmole of carboxyl carrier protein (CCP) which catalyzes the exchange at a rate of 0.27 nmole per min (0 ).

FIG. 6 (right). Inhibit,ion of the acetyl-CoA carboxylation and malonyl-CoA-[l-~4C]acetyl-CoA exchange reaction by ppGpp. For details of the enzymatic assays see “Experimental Pro- cedure.” q , inhibition by Mn’ppGpp of acetyl-CoA carboxylation catalyzed by a crude acetyl-CoA preparation (82 rg of dialyzed 25 to 42% ammonium sulfate fraction, equivalent to a carboxylase activity of 0.3 nmole per min.); 0, inhibition by Mn.ppGpp of the malonyl-CoA-[l-14C]acetyl-CoA exchange catalyzed by a crude acetyl-CoA carboxylase preparation (70 ,.~g of dialyzed 25 to 427, ammonium sulfate fraction equivalent to an exchange activity of 0.12 nmole per min); 0, inhibition by Mn.ppGpp of the malonyl-CoA-[l-~4C]acetyl-CoA exchange in a system reconstituted from 5.3 millilmits of pure carboxyltransferase and 0.11 nmole of carboxyl carrier protein (equivalent to an exchange activity of 0.27 nmole per min) ; A, inhibition by ppGpp of the malonyl-CoA-[I-‘%]acetyl-CoA exchange in the same reconstituted system.

carboxyltransferase. At a concentration of 0.4 mM, each nucleotide inhibits the model reaction by about 20%, and their effect, in a reaction mixture containing 0.4 mM of each nucleotide, is additive (Table III). However, in a reaction mixture containing a saturating concentration of either pppGpp (1.0 mM) or ppGpp (1.24 mM), the effect is not additive; in fact, the addition of either nucleotide, in a mixture already containing the other at saturating concentration, is virtually without effect (Table III).

Finally, the effect of ppGpp on the reactions dependent upon the biotin carboxylase component, including the carboxylation of free d-biotin and the carboxylation of acetyl-CoA, was tested. Uoth of these reactions are quite sensitive to divalent cation to ATI’ ratios; maximal activities at pH 6.8 to 6.9 are achieved

TanLlc II

Speci,/kity of inhibitiotL by (p)ppGpp oj reaclions calal!yzed by CarOoxyltransferase

For details of the assay methods see “I~Zxperimental Proce- dure.” The amounts of pure enzymes used in the assays were: 5.3 millilmits for carboxyltransferase in the spectrophotometric assay of the carboxyl transfer reaction; 5.3 millirmits of carboxyltransferase plus 0.11 nmole of carboxyl carrier protein in the exchange reaction; 0.15 milliunit of biotin carboxylase in the radiochemical assay for biotin carboxylase. The numbers in parentheses are the rates of the control assays in nanomoles per min.

Additions

None (control). 1.2m~ppGpp..... 1.2 rnM pppapp. 1.2 mM GTP. 1.2 mM GDP., 1.2 mM GMP.. ..__ 1.2 mM ATP . 1.2 mM ADP. 1.2 mM AMP.. 1.2 mM CAMP. . 2.5 mM ppGpp.Mn. 2.5 rnnr lj:I>TA.Mn 2.5 mM GTP.Mn.. 2.5 mM GDP.Mn.. 2.5 mM ATP.Mn..

Carboxyltransferase-catalyzed

Carboxyl transfer from Malonyl-CoA- ~olnnn;;~;A [‘4ClscetyLCoA

methyl ester exchange

% control

100 (5.3) 48 51 92

115 08

100 111 101 113

100 (0.2i) 49

90

98

9G 98

101

-B

-

,iotin carboxylase- catalyzed

carboxylation of d-biotin

100 (0.15)

131 120

TABLE III

Adtlitivity of inhibition of curboxyltramferase

by PPGPP and PPPGPP

For details of the assay method see “Experimental Procedllre.” Homogeneous carboxylt,ransferase was assayed spectrophoto- metrically.

Carboxyl transfer’from malonyl- CoA to d-biotin$xthyl ester

Additions I Nanomoles per

min Percentage of

inhibitlon

None. .............................. 0.4 mM ppGpp. .................. 0.4 mM pppGpp .................... 0.4 mM ppGpp + 0.4 mM pppGpp. ... 1.2 mM ppGpp. .................... 1.0 rnM pppGpp. .................. 1.2mMppGpp+ l.OmMpppGpp .... 2.2 mM ppGpp. ..................

4.9 0 4.0 17 3.7 23 3.0 39 2.3 52 2.5 49 2.1 58 2.2 5G

with Mn2+:ATP = l:l, 1Mg2+:ATP = 2.5:1, and Co~f:ATP = 2.5:1, the V,,, o btained with any of thcsc divalent cations being the same. However, activity declines, although not steeply, on either side of the optimal metal to ATP ratio. Since free ATP is also inhibitory in both the biotin carboxylasc and in the acctyl-CoA carboxylation reactions, meaningful inhibition studies with ppGpp could only be conducted with levels of ATP, Mn2+, and ppGpp such that the concentration of free -Yl’l’ did not increase significantly and the ratio of &In2+:Al’P did not


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7964

deviate substantially from 1: 1. This proviso necessitated the determination of the affinity of ppGpp for Mn2+. The dissociation constants for Mn2+ from the ternary (Mn2+)2,ppGpp complex were assessed by titrating MnClz with ppGpp and measur- ing the electron paramagnctic resonance of free Mn2+ at pH 6.8; the values obtained were 1.1 pM and ~65 pM. Since the KI, for the Mn2+.ATP complex, 5.3 pM, determined under the same conditions, is more than an order of magnitude lower than that for the Mn2+ bound most weakly by ppGpp, in the inhibition studies sufficient Mn2+ was added to saturate ATP and the tight binding site of ppGpp. Thus in all experiments with the biotin carboxylase and acetyl-CoA carboxylase systems, the concentration of Mn2+ used was equal to the concentration of ATP plus ppGpp.

As illustrated in Table II, at a Mr?+.ppGpp concentration of 2.5 mnl, which maximally inhibits malonyl-CoA-[l-W]acetyl- CoA exchange and acetyl-CoA carboxylation (Fig. 6), biotin carboxylase activity is not inhibited or is slightly increased. Consistent with the finding that free ppGpp at saturating concentration inhibits the carboxyltransferase-catalyzed model and exchange reactions 50 to 60% (Fig. 5 and Table II), saturating Mn2+.ppGpp blocks malonyl-CoA-[l-14C]acetyl-CoA exchange and acetyl-CoA carboxylation to the same extent. However, the concentration of Mn2+ ppGpp required for maximal effect on the latter-mentioned reactions, about 2 mM and still within the physiological range, was twice as high as the concentration of free ppGpp needed for 507, inhibition of carboxyltransferase. This difference is most likely due to a lower affinity of the enzyme for -\W+.ppGpp. The inhibition cannot be attributed to the free ppGpp in the mixture because its calculated concentration is miniscule. Direct measurement by electron paramagnetic resonance of free Mn2+ . m reaction mixtures containing 2.5 mM ppGpp shows its concentration to be <O.l InM.

because of the differences in the concentration of nucleotide required for inhibition of acetyl-CoA carboxylation by Mn2+. ppGpp and of carboxyl transfer by free ppGpp, the effects of Mn2+.ppGpp on both reactions were compared in the same experiment. As shown in Fig. 6, the pattern of inhibition by Mn2f.ppGpp of malonyl-CoA-[l-i4C]acetyl-CoA exchange, catalyzed by either crude acetyl-CoA carboxylase preparations or by purified enzymes in a reconstituted system, was quantitatively similar to the pattern observed for acetyl-CoA carboxylation. Moreover, 507, inhibition of either reaction was achieved with the same concentration of Mr?+.ppGpp (about 2 mM). Fur- thermore, as shown in Table II, the inhibition of the exchange react,ion is specific. All of the nucleotide chelates of Mn2+, as well as RIn+.EDTA, tested at a concentration of 2.5 mM were not inhibitory. It is interesting that biotin carboxylase is ac- tivated to a small extent by Mn2f.ppGpp (Table II). This effect is not specific, however, because, as shown in the same table, Mn2+.EDTA exhibited the same degree of activation. In addition, acctyl-CoA carboxylase is also influenced to the same extent by the chelate (data not shown). The activation phenomenon is most likely of no physiological significance and may simply constitute a manifestation of the degree of accessi- bility of the metal to the enzyme, the former being more acces- sible when present in the less ionic chelated form.

On the basis of the results presented in this paper, it appears that the inhibition of the over-all reaction by ppGpp is the result of the specific inhibition of the carboxyltransferase component of t,he acetyl-CoA carboxylase. Unfortunately, it is not possible to test the specificity of the inhibition of the over-all carboxylase reaction directly, because various nncleotides are known to be

competitive inhibitors of the biotin carboxylase component of the complex. Thus, YW+.ADP, Mn2+.GDP, and Mn2+ .GTP are all competitive inhibitors of biotin carboxylase with respect to Mn2+ .ATP.

DISCUSSION

The results presented in this paper provide definitive proof that stringent control in E. coli is exercised on fatty acid synthesis rather than on precursor transport (acetate and glucose) or fatty acyl turnover. In addition, our results suggest that this control is mediated by (p)ppGpp and is exerted at the level of acetyl-CoA carboxylase. The intracellular concentration of ppGpp plus pppGpp rises following amino acid starvation from 0.1 to 0.4 mM to 2 to 4 mM (50). Our results indicate (Fig. 6) that maximal inhibition of acetyl-CoA carboxylase is manifested at these physiological (p)ppGpp concentrations and, in particular, that the focus of action is the carboxyltransferase reaction (Fig. 5, Tables II and III). Furthermore, we have shown (Table II) that the inhibition is specific, as other nucleotides at comparable concentrations do not inhibit the carboxyltransferase-catalyzed reactions.

In the discussion to follow, attention will be focused on the broader aspects of this control mechanism. Fine control of biosynthetic pathways, i.e. control at the level of catalysis per se, is generally exerted at an early step in the metabolic sequence. In most cases this step is relatively irreversible, thus committing the product of the reaction to the pathway. Biosynthetic processes leading to the storage of reserve sources of energy are fre- quently regulated by feed-forward activation. The fatty acid biosynthetic pathway in animal t.issues fulfills such a role, and existing evidence (52, 53) strongly suggests that this pathway is regulated by citrate activation of the committed step catalyzed by acetyl-CoA carboxylase. The rationale for this mechanism suggests that the accumulation of citrate in the cytoplasmic compartment serves to indicate that acetyl-CoA, reducing power, and energy in the form of ATP are all in abundance for the biosynthesis of fatty acids.

In contrast to animal systems, where fatty acids are used principally as a reserve source of energy, in E. coli fatty acids are primarily incorporated into phospholipids, which in turn are used for structural purposes in membranes. Hence, regulation of fatty acid synthesis in this organism would be expected to reflect the requirement for phospholipids in growth and cell proliferation. The elegant work of Maaltie and Kjeldgaard (54) on the control of macromolecular biosynthesis has established fundamental quantitative relationships between the rate of growth and such parameters as cell mass, number of nuclei, DNA, RNA, and protein content per cell. Their results suggest that, in a state of balanced growth, the rates of synthesis of the various cell components should be constant. Furthermore, during shift-up or shift-down experiments, it appears (54) that a new steady state is established and that this new state is also characterized by constant, but different, rates of synthesis of macromolecular cell constituents. Consequently, it is likely that a control mechanism is operative which coordinates the rates of synthesis of the macromolecules and components of the cell membrane to the rate of growth. This, in turn, implies the existence of one or more growth modulator(s), and (p)ppGpp may be one of these. The kinetics of appearance and disappearance of this nucleotide during amino acid starvation and refeed- ing (12-14), its accumulation during shift-down experiments or carbon-source starvation (13, 14, 55), as well as the fact that (p)ppGpp concentration is inversely proportional to the RNA


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

content of the cell (14) and growth rate (54) lend support to this assumption. Finally, the recent interesting experiments of Haseltine et al. (56) and Haseltine and Rlock (57) have established that (p)ppGpp is formed on the ribosome by pyrophos- phoryl transfer from ATP to GDP or GTP as an idling step in protein synthesis. The reaction requires both the 30 S and 50 S ribosomal subunits, the G translocation factor, the stringent factor-found only in stringent cells-messenger RNA, and an uncharged transfer RNA in the acceptor site of the ribosome. These observations also suggest that the mode of formation of (p)ppGpp is intimately connected to the cessation of protein synthesis and, thus, to inhibition of growth. Inhibition by (p)ppGpp of acetyl-CoA carboxylase, an enzyme which should supply malonyl-CoA at a rate proportional to cell proliferation, is a rational mechanism for controlling fatty acid synthesis.

In addition to an inhibition of fatty acid synthesis by amino acid starvation of stringent cells, the results presented in Table I suggest a preferential incorporation of radioactivity into the phosphatidylglycerol fraction. As pointed out under “Results,” it appears that net phosphatidylethanolamine synthesis is curtailed to a greater extent than phosphatidylglycerol synthesis. This apparent blockade of phosphatidylethanolamine synthesis may be related to the yet unexplained observation of Raetz and Kennedy (58) that the “committed” step of phosphatidyleth- anolaminc synthesis is catalyzed by an enzyme, phosphatidyl- serine synthase, which is tightly bound to ribosomes. In this connection, it should be re-emphasized that the synthesis of (p)ppGpp, the apparent mediator(s) of the stringent control of lipid synthesis, is also synthesized on the ribosome.

Available data on the stringent control of fatty acid synthesis (this work, 1, 2) indicate that, per cell, fatty acid synthesis is reduced by only 50 to 60%, while protein synthesis ceases completely and RNA synthesis appears to be inhibited to a greater extent (13, 14). The reason for this continued, although reduced, level of fatty acid synthesis in the absence of growth is not known; however, it is conceivable that successful transitions during alterations in nutritional state may require changes in the membrane composition and, therefore, continued net phospholipid synthesis. Examples of such changes are the increased formation of cyclopropane fatty acids at the end of log phase growth (59, 60), although this change is not vital to cell survival, and the changes in the fatty acid composition of cell membranes when cells are shifted from high to low temperature (42). That changes in the membrane composition may be important factors for cell survival is suggested by the fact that viability decreases soon after cessation of growth of unsaturated fatty acid auxo- trophs of E. coli deprived of oleate (61, 62), after transfer to nonpermissive temperatures of a mutant with a temperature- sensitive glycerol 3-phosphate acyltransferase (63), and after transfer of an unsaturated fatty acid auxotroph grown at 37” with elaidate to 27” in the absence of oleatc (64-66). In addition, it is possible that amino acid starvation induces the synthesis of certain membrane enzymes and proteins and that this induction requires net phospholipid synthesis. Both, induction of enzymes during starvation (67, 68) and a requirement for net phospholipid synthesis for induction (69%72), have been observed experimentally.

The results of Silbert et al. (73, 74) and Esfahani et al. (64, 65) suggest that supplementation with fatty acids results, at least in certain strains, is a curtailment of the synthesis of saturated and unsaturated fatty acids. Supplementation with unsaturated fatty acids (64, 65, 73) preferentially decreased the incorporation of [Tlacetate into unsaturated fatty acids; however, supple-

7965

mentation with saturated fatty acids (74), while reducing incorporation of [i4C]acetate into saturated fatty acids preferentially, also resulted in reduced incorporation of label into the unsaturated fraction. The authors suggest that a primary control mechanism is operative at the level of glycerol 3-phosphate acyltransferase which determines the fatty acyl composition of the phospholipid. In addition, the results of Wakil and his col- leagues (65) suggest that cells unable to degrade long chain (CL,) unsaturated fatty acids compensate by adjusting both the chain length and the quantity of the saturated fatty acids synthesized. It is not yet clear whether the over-all curtailment of fatty acid biosynthesis in the presence of exogeneous fatty acids is the result of feedback inhibition of the biosynthetic pathway by accumulat- ing products, e.g., acyl-CoA, or of repression of synthesis of the enzymes involved.

It is evident that the phospholipid-synthesizing systems of E. coli exhibit extreme versatility and adaptability and this, no doubt, is of survival value to the organism. The formation of a functionally efficient and stable membrane for a given set of conditions appears to be vital, and a multiregulatory system both at the level of synthesis and degradation is probably required to achieve this. The observations described in this paper suggest that the inhibition of acetyl-CoA carboxylase by ppGpp is one of the regulatory mechanisms operating at the biosynthetic level, by which the cell regulates the supply of malonyl-Coh and the rate of phospholipid synthesis for growth.

Acknowledgments-We thank Mr. Eberhard Zwergel for out- standing technical assistance, Dr. Albert Mildvan for carrying out the electron paramagnetic resonance measurements, Drs. M. Cashel and R. Ryrne for generously providing us mibh ppGpp and pppGpp, and Dr. N. Fiil for supplying the strains of E. coli used in these experiments.

REFERENCES

1.

2.

3.

4.

5. 6. 7.

8.

9.

10.

11.

12. 13. 14.

15.

16.

17.

18.

19.

SOKAWA, Y., NAKAO, E., AND KAZIRO, Y. (1968) Biochem. Bio- phys. Res. Commun. 33, 108-112

GOLDEN, N. G., AND POWELL, G. L. (1972) J. Biol. Chem. 24’7, 6651-6658

STENT, G., AND BRENNER, S. (1961) Pyoc. ATat. Acud. Sci. U.S.A. 47, 2005-2014

SOKA~A, Y., AND KAZIRO, Y. (1969) Iliochem. Biophys. Res. Commun. 34, 99-103

EDLIN, G., AND NEUHARD, J. (19G7) J. fiIoZ. Biol. 24, 225-230 EDLIN, G., AND MUL$E, 0. (1966) J. IQ’oZ. Biol. 16, 428-434 NIERLICH, D. P. (1968) Proc. Nat. Acad. Xci. U.S.il. 60, 131-k

1352 EDLIN, ct., END STEXT, G. S. (1969) Proc. byat. dcnd. Sci.

U.S.A. 62, 475-482 WAGNIGR, C., ODOM, R., ASD BRIGGS, W. T. (1972) Biochem.

Biophys. Res. Commun. 47, 1036-1043 TROPP, B. E., MEADE, L. C., AND THOM.~S, P. J. (1970) J. Biol.

Chem. 246, 855-858 NUNN, W. D., AND TROPP, B. E. (1972) J. Bacterial. 109, 162-

168 CASHXL, M., >IND GALLANT, J. (1969) iVature 221, 838-841 CASHEL, M..(1969) J. Viol. Chem. 244,3133-3141 LAZURINI. It. A.. CASHEL, M.. AND GALI~ANT. J. (19il) 1.

Niol. Chkm. 246; 43814385 ’ DIMR~~PH, P., GUCIIIIAIT, It. B., STOLL, E., AND LASE, AI. D.

(1970) Proc. Arat. Acad. Sci. U.S.A. 67, 1353-1360 DIMROTH, P., GUCHHAIT, R. B., AND La~c, M. D. (1071) 2.

Physiol. Chem. 362, 351-354 GUCHHAI~~, II. B., Moss, J., SOI~OI,SI<I, W., >ISD Lasr:, 31. D.

(1971) Proc. iVat. Acad. Sci. U.S.A. 68, 653-657 ALB~RTS, A. W., ;IND VAGI”LOS, P. R. (1968) Proc. ,\-nt. Acad.

Sci. U.S.A. 69, 561-568 AL~~RTS, A. S., NERVI, A. M., ~WD V~~J~I,OS, P. 1:. (1969)

Proc. Nut. Acatl. Sci. U.S.A. 63, 1310-1320


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7966

20.

21.

22. 23.

24. 25.

26. 27.

28.

29. 30.

31. 32.

33. 34.

35.

36.

37.

38.

39.

40.

41.

42. 43.

44.

45. 46.

47.

POL.~J;IS, S. E., GUCHX~IT, R. B., ANI) LANE, M. D. (1973) Abslrucls of the 16th International Conference on the Biochem- istq of Lipicls, Nottingham, England

CASHEL, M., LAZZARINI, It., ANI) KALIIACHEI~, B. (1970) J. Chromatogr. 40, 103-109

GALL.~KT, J., llND CASHYL, M. (1967) J. Mol. Biol. 26, 545-553 &XIX, M. W. (1963) in Methods of Enzymatic Analysis (BERG-

>~JSEYIX. H. U.. ed) x)u. 117-123. Academic Press. New York R~DIS, k. S. (1969)‘Methods Enzymol. 14, 245-254' SKIPSKI, V. P., AND BIRCL~Y, M. (1969) Methods Enzymol.

14, 530-598 COHN, M., AND TOI~NSEND, J. (1954) l\rature 173, 1090-1091 G~CHII~IT, It. B., POLAKIS, S. E., AND LAN]<, M. D. Methods

Enzymol., in press GuCHHBIT, It. B., POLAKIS, S. E.. AND LANE, M. D. Methods

Emymoi., in prbss JUNGAS. 1%. L. (19681 Biochemistru 7. 3708-3717 W~TSO;, J. A., ‘AND ‘LOX-ENST~IN~ J.‘M. (1970) J. Biol. Chem.

245,5993-6002 LO\~ENSTEIN, J. M. (1971) J. Biol. Chem. 246,629*32 MUNICIO, A. M., DIAZ, T., AND MAHTINEZ, A. (1963) Biochem.

Biophys. Res. Commun. 11, 195-200 BISHOP, D. G., AND WORI~, E. (1965) Biochem. J. 96, 567-576 TAYLOR, A., Knox, K. W., AND WORK, E. (1966) Biochem. J.

99,53-61 KNOX, K. W., VI!XIC, M., AND WORK, E. (1966) J. Bacterial. 92,

1206-1217 WORK, F:., KNOX, K. W., AND VESK, M. (1966) Ann. N.Y. Acad.

Sci.133,438449 KNOX, K. W., CULLICN, J., AND WORI~, E. (1967) Biochem. J.

103,192-201 ROTHFIE:LD, L., AXD PEARLMAN-KOTHENCZ, M. (1969) J. Mol.

Biol. 44, 477-492

WEEKS, G., SHAPIRO, M., BURNS, R. O., AND WAKIL, S. J. (1969) J. Bacterial. 97, 827-836

CROXT-FORT, P. D., ESFAHANI, M., AND WAICIL, S. J. (1972) J. Bacterial. 112,1408-1415

OVEIZATH, P., PAULI, G., AND SCHAIHER, II. U. (1969) Eur. J. Biochem. 7,559-574

OKUYAMA, H. (1969) Biochim. Biophys. Acta 179, 125-134 KANFER, J., AND KMNNBDY, E. P. (1963) J. Biol. Chem. 238,

2919-2922 K~NEMASA, Y., AKAMATSU, Y., AND NOJIMA, S. (1967) Bio-

chim. Biophys. Acta 144, 382-390 MINDICH, L. (1972) J. Bacterial. 110, 96-102 GALLANT, J., ERLICH, II., HALL, B., AND LAFFLER, T. (1970)

Cold Spring Harbor Symp. Quant. Biol. 35,397405 CASIIEL, M. (1970) Cold Spring Harbor Symp. Quant. Biol.

35, 407-413

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60. 61.

62.

TRAVERS, A., K~MEN, R., AND CASHEL, M. (1970) Cold Spring Harbor Symp. Quant. Biol. 35, 415-418

GALLANT, J., IRR, J., ANI) CIISHEL, M. (1971) J. Biol. Chem. 246, 5812-5816

HOCHSTADT-OZER, J., AND CASHEL, M. (1972) J. Biol. Chem. 247, 7067-7072

ARAI, K., ARM, N., KAU-AKITA, M., AND KAZIRO, Y. (1972) Biochem. Biophys. Res. Commun. 48, 190-196

LANE, M. D., Moss, J., RYDXR, E., AND STOLL, E. (1971) Adv. Enzyme Regul. 9, 237-251

LANE, M. D., AND Moss, J. (1971) in Metabolic Pathways (VOGEL, H. J., ed) Vol. 5, p. 23, Academic Press, New York

MAAL$E, O., AND KJELDGAARD, N. 0. (1966) Control of Macro- molecular Synthesis, W. A. Benjamin, Inc. New York

HARSHMAN, It. B., AND YAMAZAKI, H. (1971) Biochemistry 10, 398G3982

HASELTINE, W. A., BLOCK, R.; GILBERT, W.; AND WEBER, K. (1972) Nature 238, 381-384

HASELTINE, W. A.. AND BLOCK, R. (1973) Proc. Nat. Acad. Sci. U.S.A. 76, 1564-1568

R.AETZ. C. R. H.. AND KENNEDY. E. P. (1972) J. Biol. Chem. 247,iOO8-2014 '

\ ,

KNIVETT, V. A., AND CULLEN, J. (1967) Biochem. J. 103, 299- 306

CRONAN, J. E., JR. (1968) J. Bacterial. 95, 2054-2061 HENNING, U., DENNERT, G., REHN, K., AND DEPPE, G. (1969)

J. Bacterial. 98, 784-796 SCHAIRER, H. U., AND OVER~TH, P. (1969) J. Mol. Biol. 44,

209-214 63. GI~ASER, M., BAYI~R, W. H., BE:LL, R. M., AND VAGELOS, P. R.

66. CROWFOOT, P. D., OICA, T., ESFAHANI, M., AND WAKIL, S. J.

(1973) Proc. Aiat. Acad. Sci. U.S.A. 70,385389 64. ESFAHANI, M., BARNES, E. M., AND WAICIL, S. J. (1969) Proc.

Nat. Acad. Sci. U.S.A. 64, 1057-1064

(1972) J. Bacterial. 112, 1396-1407

65. ESFAHANI, M., IONED~, T., AND WAIIIL, S. J. (1971) J. Biol.

67. SUSSMAN, A. J., AND GILVARG, C. (1969) J. Biol. Chem. 244.

Chem. 246, 50-56

6304-6306 68. MANDELSTAM, J. (1960) Bacterial. Rev. 24, 289-308 69. Fox, C. F. (1969) Proc. Nat. Acad. Sci. U.S.A. 63,850-855 70. Hsu, C. C., AND Fox, C. F. (1970) J. Bacterial. 103, 410-416 71. IZUI, K. (1971) Biochem. Biophys. Res. Commun. 45, 1506-1512 72. ROBBINS, A. R., AND TOTMAN, B. (1972) Proc. Nat. Acad. Sci.

U.S.A. 69, 2125-2129 73. SILBEILT, D. F., COHEN, M., AND HARDER, M. E. (1972) J.

Biol. Chem. 247, 1699-1707 74. SILBERT, D. F., ULRRIGHT, T. M., AND HONEGGER, J. L.

(1973) Biochemistry 12, 164-171


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

S. Efthimios Polakis, Ras B. Guchhait and M. Daniel LaneREGULATION OF ACETYL COENZYME A CARBOXYLASE BY ppGpp

: POSSIBLEEscherichia coliStringent Control of Fatty Acid Synthesis in

1973, 248:7957-7966.J. Biol. Chem.

http://www.jbc.org/content/248/22/7957Access the most updated version of this article at

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/248/22/7957.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/content/248/22/7957

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;248/22/7957&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/248/22/7957

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=248/22/7957&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/248/22/7957

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/248/22/7957.full.html#ref-list-1

http://www.jbc.org/

Stringent Control of Fatty Acid Synthesis in Escherichia coli · that st,ringent control is exerted at the level of acetate transport and activation, rather than at the level of fatty

Documents