Phospholipid Synthesis in Escherichia coli - The Journal of Biological ...

JOURNAL OF BIOLOGICAL CHEMISTRY Val. 256, No. 2. Issue of January 25, pp. 736742,1981 Printed in U.S.A. THE

Phospholipid Synthesis in Escherichia coli CHARACTERISTICS OF FATTY ACID TRANSFER FROM ACYL-ACYL CARRIER PROTEIN TO SN-GLYCEROL 3-PHOSPHATE* (Received for publication,June 11, 1980, and in revisedform, August 26, 1980) Charles 0. Rock$ and Susan E. Goelzg From the Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06510

John E. Cronan, Jr. From the Department of Microbiology, University of Illinois, Urbana, Illinois61801

Two kinetically distinguishable sn-glycerol 3-phosphate (glycerol-P) acyltransferase activities were detected in Escherichia coli inner membranes using acylacyl carrier protein (ACP) substrates. The first system was characterized as having a Michaelis constant (K,) f o r glycerol-P of 90 PM and utilized palmitoyl-ACP to formprimarily 1-acylglycerol-P. Palmitoyl-CoA and cis-vaccenoyl-ACP were also utilized by this system but, with these substrates, significantlymore phosphatidic acid was formed as compared to palmitoyl-ACP. Although palmitoyl-ACPand palmitoyl-CoA had kinetically indistinguishable glycerol-P sites, distinct acyl donor binding sites were inferred from kinetic experim e n t s using acyl carrier protein as an acyltransferase inhibitor. A second enzyme system, characterized as having a K , f o r glycerol-P of 700 PM, was found using palmitoleoyl-ACP as a substrate. This acyltransferase had a slightly higher p H o p t i m u m than the low K , acyltransferase activity, and phosphatidic acid was the major product. Two degradative reactions were identified in this system. O n e reaction yielded diacylglycerol when palmitoyl-ACP was the substrate. The other degradative reaction produced glycerol. Glycerol was formed in all incubations but was most pronounced when palmitoleoyl-ACP was the substrate.

Naturally occurring phospholipids generally have a saturated fatty acid at position 1and an unsaturated fatty acid at position 2 of the sn-glycerol 3-phosphate backbone (1).The glycerol-P' acyltransferase catalyzes the fist committed step in phospholipid biosynthesisacting toesterify one (or perhaps both) fatty acids to glycerol-P in the formation of the intermediate, phosphatidicacid. Since redistributionof acyl moieties following phospholipid synthesis does not occur in Escherichia coli (2), the origin of acyl group asymmetry in this organism must be at the level of the acyltransferase. Acyltransferase activities for two consecutive acylation reactions are located in the inner membrane (3, 4). One enzyme cata-

__

* This research was supported by National Institutes of Health Grant A115650. f Reciuient of National Institutes of Health Postdoctoral Fellowship GM 06258. Present address, Department of Biochemistry, St. Jude Children's Research Hospital,332 North Lauderdale, Memphis, Tenn. 38101. fj Present address, Fredrich Miescher Institute, CH-4002, Basel, Switzerland. The abbreviations used are: glycerol-P, sn-glycerol 3-phosphate, ACPSH, acyl carrier protein; acyl-ACP,acyl-acyl carrier protein. I

.

lyzes the transfer of a fatty acid to position 1 to form 1acylglycerol-P, and a second enzyme catalyzesthe transfer of a second fatty acid to position 2 of the fiist intermediate, thereby forming phosphatidic acid. The questionof the specificity of these membrane-boundenzymes has been the subject of a large and confusing literature (for review, see Ref. 5). Several workers have reported that acyl transfer to glycerolP from saturated acyl-CoA substrates gave 1-acylglycerol-P whereas unsaturated acyl-CoA substrates yielded 2-acylglycerol-P (6-8). Other investigators reported that both saturatedandunsaturatedacyl-CoAs yielded 1-acylglycerol-P and found no specificity in the acyl transfer reaction using membrane preparations (9, 10) or partiallypurified acyltransferase preparations (11, 16). A third laboratory has reported that no significant accumulation of 1-acylglycerol-P occurs, but rather that phosphatidicacid is the major product(9, 10, 14). Acylation specificity has also been studied in the presence of a mixture of saturated and unsaturatedacyl-CoA's (6, 12). These results have shown that the positional distribution of fatty acids inphosphatidic acid (6,13) orphosphatidylglycerol (12) is similar to the distribution in uivo. However, these studies did not delineate the number or specificity of the intermediate reactions conferring this specificity. Okuyama and co-workers (10, 14) have reported that acyltransferase specificity is a function of the glycerol-P concentration. However, these resultswere not confirmed byothers either in vitro (13) or in uiuo,' and since the glycerol-P levels in E . coli are tightly regulated (5, 15), it is difficult to see how this effect would have physiological relevance. Several groups have attempted to purify the acyltransferase in the hope that these conflicts could be resolved by use of more pure preparations (11, 13, 16). Unfortunately, the acyltransferase is inactivated by most detergents used in membrane solubilization (11, 13, 16) and the K,,, for glycerol-P has been found to increase 20fold following solubilization (13).The kinetic properties of the solubilized acyltransferase therefore do not reflect the activity of the protein in its native environment. A major concerninevaluating these data is that most previous investigations have used acyl-CoA substrates as the acyl donor and inspection of the available data indicates that acyl-ACP rather thanacyl-CoA is the physiological acyl donor (for review, see Ref. 5). In light of these data, there is some doubt thatexplanations of the specificity of the acyltransferase system based on invitro assays using acyl-CoA substrates are relevant to thein uiuo situation. An additional complication with acyl-CoA substrates is that they are potent inhibitors Go&, S., and Cronan,J. E., Jr. (1980)J.Bacteriol. 144,462-464.

736

E. coli Glycerol-3-P Acyltransferases of acyltransferase activity due to formationof detergent micelles (17). This property necessitates theinclusion of serum albumin in the assay to ameliorate the detergent effects of the substrate. The resolution of this contradictory information seems to reside in examining the kinetic mechanism of the acyltransferase using acyl-ACP substrates. Thisconclusion is supported by experiments on unsaturated fatty acid auxotrophs. Significant accumulation of disaturated molecular species is observed in E. coli when unsaturated fattyacid auxotrophs are deprived of exogenous unsaturated fatty acid supplements (18). Restoration of unsaturated fatty acids to the medium results in molecular species possessing typical positional asymmetry (18).These data demonstrate that acylation specificity is not absolute invivo and is controlled in part by the supply of fatty acids. A few studies have appeared in which acylACP's have been used as substratesfor the acyltransferaseof E . coli. Most of thesedata (8, 19) aresuspect since the chemical method used to prepare the ACP thioesters results in a dramatic loss of secondary structure andbiological activity of ACPSH (20). Ray and Cronan (21) have used biosynthetically prepared palmitoyl-ACP as a substrate for acyltransferase, but at the time these studies were done native unsaturated acyl-ACPs could not be prepared. However, these investigators found that acyl-CoA and acyl-ACP do not behave identically in that acyl-CoA acyl transfer is stimulated by Mg2+and inhibitedby guanosine 5"diphosphate-3"diphosphate whereas acyl transfer from acyl-ACP is not affected by addition of either compound. Recently, anenzymatic method to prepare pure native acyl-ACP substrates has been developed (22).Physical studies on these thioesters show that they possess stable native protein structures and do not exhibit detergent properties (23). These developments haveled us to examine acyltransferase specificity using well defined native acyl-ACP substrates. EXPERIMENTALPROCEDURES

Materials-Palmitoyl-CoA was purchased from P-L Biochemicals. Glycerol-P was from Sigma and lipid standards were obtained from Serdary Research Labs. Silica Gel G plates werepurchasedfrom Analtech, Inc. and silica gel-impregnated paper (SG81) was purchased from Whatman. ~n-[U-'~C]Glycerol-3-P (specific activity 165 Ci/mol) and Aquasol were purchased from New England Nuclear. Fatty acid free bovine serum albumin was obtained from Miles Laboratories. ACPSH was purified (24) and acyl-ACP substrates were synthesized (22) as described previously. All acyl-ACP's were pure as judged by native gel electrophoresis (25).All other materials were reagent grade or better. Preparation and Assay of Acyltransferase Activity-An overnight culture of E . coli strain 8 (HfrC, glpR, glpD, phoA, relA1, tonA22, spoT,pit-lO, T2R) was diluted 20- to 40-fold into 200 ml of rich broth (10 g of tryptone, 1g of yeast extract,5 gof NaCl/liter) and grown for approximately 12 h a t 37°C to early stationary phase. T h e cells were harvested by centrifugation, resuspended in1.5 ml of 20mM potassium phosphate buffer (pH 7.5), containing 5 mM EDTA and 50 mM KCl, and disrupted by passage through a French Pressure cell a t 16,000 p.s.i. Unbroken cells and debris were removed by centrifugation a t 10,OOO X g for 10 min, and then inner and outer membranes were resolved in the step gradient system of It0 etal. (26). In this procedure soluble proteins remain above the 15% sucrose layer and the inner membranes form a visible band at the interphase dividing the 15% and 53% sucrose steps. The inner membrane fraction was removed, diluted to 6 ml with 20mM potassium phosphate buffer (pH 7.5), containing 10 m~ MgC12, and sedimented a t 100,000 X g for 1 h. The inner membrane pellet was homogenized in 0.2 to 0.3 ml of the Same buffer and storeda t 4°C a t a protein concentration of 10 to 20 mg/ml. The membranes were generally used within 2 days of their preparation and diluted to a protein concentration of 1 mg/ml immediately before use. Protein was determined using the microbiuret procedure (27) with bovine serum albumin as the standard. Standard incubations for determining acyl transferfrom palmitoylCoA contained 0.1 M Tris-HC1 (pH 8.5); bovine serum albumin, 1 mg/

737

ml; palmitoyl-CoA, 12.5 pM; MgC12, 5 mM; [U-"C]glycerol-P (l0,OOO cpm/nmol), 250 p ~ and ; 4 to 16 pg of inner membrane protein in a final volume of40 pl. For assaying acyl transferfrom acyl-ACP, standard incubations contained 0.1 M Tris-HC1 (pH 8.5); bovine serum albumin, 1 mg/ml; acyl-ACP 25 pM; [U-'4C]glycerol-P (10,000 cpm/ ; 4 to 16 pgof inner membrane protein in a final nmol), 200 p ~ and volume of 40 pl. In some experiments ( a s indicated), 5 mM MgClz was . included and the glycerol-P concentration was 250 p ~ Incubations were terminated after5 min a t 21OC by pipetting 35 pl of the reaction mixture onto a Whatman No. 3MM fdter disc. Filter discs were processed by a modification of the procedure of Goldfine (28) by washing (20 ml/fdter) twice with ice-cold 5% trichloroaceticacid followed by ice-cold 1%trichloroacetic acid, drying under a hot air stream, and counting in Aquasol. The assay was found to be linear from 1 to 12 min and from 4 to 16 pg ofprotein with a5-min incubation time. A unit of acyltransferase activity is defined as a nanomole of product formed/min. Chromatographic Analysis-The samples analyzed were obtained by adding 0.2 ml of CHCI&HaOH (1:2; v/v) to an acyltransferase reaction mixture. The samples were then evaporated to dryness under N2 and suspended in 0.03 ml of CHCla/CHZOH (l:l, v/v), and the entire mixture spotted on the chromatogram. Glycerol, lysophosphatidic acid, phosphatidic acid, and neutral lipids were separated on silicic acid-impregnated chromatographic papers (Whatman SG81). Immediately before use the paperswere dipped in 0.01 M Na2C03 and dried a t 100°C for a t least 3 h. Samples of acyltransferase reaction mixtures were then spotted on the paper strips (23 X 1.75 cm) which were first developed for 14.5 cm in an ascending mode in CHCL/ CHzOH/CH&OOH/H? (7020:12:4, by volume). After drying, the paper strips were developed in the same direction (descending mode) for 20 cm in petroleum ether/ether/CHKOOH (23:21:0.5, byvolume). The papers were dried and cut into0.5 cm sections, and each section was counted in Aquasol scintillation solution. Neutral lipids were identified by chromatography on boric acidimpregnated Silica Gel G thin layer plates. The plates were obtained from Analtech and were impregnated by dipping in 0.4 M boric acid in methanol and drying for at least 3 h a t 100OC. The plates were developed in chloroform/acetone (95:5, v/v) to 15 cm from the origin. Chromatographic standards were run in parallel. Conversion ofMonoacylglycero1-P to Monoglycerides-Acyltransferase incubations were terminated by addition of 0.2 ml of CHCla/ CHJOH (1:2, v/v). The mixture was evaporated to dryness under N2 and beef liver alkaline phosphatase (0.2 mg) (33) in 0.5 ml of 1 M TrisHCI buffer (pH 6.8) was added. After incubation for 2.5 h a t 21°C, the lipids were extracted (32) following acidification with HCl. T h e chloroform phase was washed with 0.01 N HCl, evaporated under nitrogen, and chromatographed on boric acid-impregnated plates as described above. About 20%of authentic 2-mono-oleoin added as an internal standard isomerized to the 1-isomer during the dephosphorylation and chromatographic procedures, and thus the observed amounts of the 2-isomer are probablyunderestimations of thetrueamounts formed. RESULTS

General Properties of the Acyltransferase Reaction-In agreement with previous investigators (21, 29), we observed that acyl transfer from palmitoyl-CoA to glycerol-P was enhanced 1.4- to 2-fold by the presence of 4 mM Mg2' in the assay system, but that acyl transfer from acyl-ACP was unaffected by the presence of Mg2+ (21). Bovineserum albumin has beenfound to markedly stimulate acyl transfer from palmitoyl-CoA, presumably due to its abilitybind to acyl-CoA and thus protect the acyltransferase from inactivation due to the detergent properties of this substrate (16, 17, 21, 29). Bovine serum albumin had onlyaminimal effect on acyl transfer from palmitoyl-ACP (Fig. 1); however, a small stimulation was consistently observed (Fig. 1) and thus, bovine serum albumin was included in all assays with acyl-ACP. It should be noted that no interactionof bovine serum albumin with acyl-ACP can be detected by gel filtration experiments (data not shown). In contrast, acyl transfer from palmitoylCoA was stimulated by an increasein the bovine serum albumin concentration in the assay, although very high albumin concentrations inhibited activity (Fig. 1).We added bo-

E. coli Glycerol-3-P Acyltransferases

738

TABLE I Kinetic constants forvarious acyltransferase substrates Initial velocities were determined in 5-min assays as described under “Experimental Procedures” and the kinetic constants were obtained by Lineweaver-Burk analysis of the data (31). MyristoylACP and oleoyl-ACP were poor substrates for the reaction, and due to thelimited availability of these substrates thekinetic constants for these two substrates were extrapolated from a limited data set. Our uncertainty in the exact V,, value for these substrates is reflected by listing them as