Isolation and Characterization of Escherichia coli K- 12 Mutants Lacking Both 2-Acylglycerophosphoethanolamine Acyltransferase and Acyl-Acyl Carrier Protein Synthetase Activity*

2-Acyl-glycerophosphoethanolamine (2-acyl-GPE) acyltransferase and acyl-acyl carrier protein (acyl-ACP) synthetase are thought to be dual catalytic ac- tivities of a single inner membrane enzyme. A filter disc replica print method for the detection of acyl-ACP synthetase activity by colony fluorography was used to screen a mutagenized population of cells for acyl-ACP synthetase mutants (aas). All aas mutants lacked both acyl-ACP synthetase and 2-acyl-GPE acyltransferase activities in vitro. There was no detectable acyl- CoA-independent incorporation of exogenous fatty acids into phosphatidylethanolamine or the major outer membrane lipoprotein in aas mutants. Exogenous lysophospholipid uptake and acylation was also lacking in am mutants. Lipoprotein acylation by phospho- lipids synthesized by the de novo biosynthetic pathway was not affected in aas mutants showing that this gene product was not directly involved in lipoprotein bio-genesis. The aas mutants had an altered membrane phospholipid composition and accumulated both 2- acyl-GPE and acylphosphatidylglycerol. Acylphospha-tidylglycerol accumulation was due to the transacylase activity of lysophospholipase


2-Acyl-glycerophosphoethanolamine
acyltransferase and acyl-acyl carrier protein (acyl-ACP) synthetase are thought to be dual catalytic activities of a single inner membrane enzyme. A filter disc replica print method for the detection of acyl-ACP synthetase activity by colony fluorography was used to screen a mutagenized population of cells for acyl-ACP synthetase mutants (aas). All aas mutants lacked both acyl-ACP synthetase and 2-acyl-GPE acyltransferase activities in vitro. There was no detectable acyl-CoA-independent incorporation of exogenous fatty acids into phosphatidylethanolamine or the major outer membrane lipoprotein in aas mutants. Exogenous lysophospholipid uptake and acylation was also lacking in am mutants. Lipoprotein acylation by phospholipids synthesized by the de novo biosynthetic pathway was not affected in aas mutants showing that this gene product was not directly involved in lipoprotein biogenesis. The aas mutants had an altered membrane phospholipid composition and accumulated both 2acyl-GPE and acylphosphatidylglycerol. Acylphosphatidylglycerol accumulation was due to the transacylase activity of lysophospholipase L2 (the pldB gene product) since aaspldB double mutants accumulated 2-acyl-GPE, but not acylphosphatidylglycerol. The aas allele was mapped to 61 min of the Escherichia coli chromosome, and the deduced gene order in this region was thyA-aas-lysA. The biochemical, physiological, and genetic analyses of aas mutants support the conclusion that 2-acyl-GPE acyltransferase and acyl-ACP synthetase are two activities of the same protein and confirm that this enzyme system participates in membrane phospholipid turnover and governs the acyl-CoA independent incorporation of exogenous fatty acids and lysophospholipids into the membrane.
Membrane phospholipids are critically important in maintaining the structure of biological membranes, but in addition, portions of these lipids are used in the synthesis of other membrane-associated macromolecules. In Escherichia coli, the observed turnover of the phosphatidylglycerol headgroup is due to its transfer to either the cysteine of the major outer * This research was supported by National Institutes of Health Grant GM-28035, Cancer Center (CORE) Support Grant CA-21765 from the National Cancer Institute, and the American Lebanese Syrian Associated Charities. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should he addressed: Dept. of Biochemistry, St. Jude Children's Research Hospital, Memphis, T N 38105. membrane lipoprotein to form a thioether linkage (1) or to the abundant periplasmic glucose polymers known as membrane-derived oligosaccharides (2). The diacylglycerol remaining in the membrane is phosphorylated by diacylglycerol kinase to yield phosphatidic acid which then re-enters the mainstream of the E. coli phospholipid biosynthetic pathway (3). Acyl moieties at the 1-position of PtdEtn' are also metabolically active, and this process is related to the transacylation reactions that occur during the maturation of the bacterial lipoproteins (4). The resulting 2-acyl-GPE is then recycled to PtdEtn by 2-acyl-GPE acyltransferase (4)(5)(6). This inner-membrane enzyme transfers fatty acids to the l-position either from an acyl-ACP derivative or from nonesterified fatty acids in the presence of ATP and Mg'+ (5,6). Acyltransferase activity is blocked by monospecific ACP antibodies indicating that ACP is required for the ATP-dependent acylation of 2-acyl-GPE ( 5 ) . 2-Acyl-GPE acyltransferase/acyl-ACP synthetase was solubilized and purified from E. coli inner membranes (7). The biochemical data support the concept that both enzymatic reactions are catalyzed by a heterodimer composed of a membrane-bound acyltransferase subunit and ACP (7). The physiological function of 2-acyl-GPE acyltransferase is to regenerate PtdEtn from 2-acyl-GPE formed by transacylation reactions or phospholipase AI action. 2-Acyl-GPE acyltransferase is also thought to be required for the acyl-CoA-independent incorporation of exogenous fatty acids (8) and 2-acyl-lysophospholipids into the cell membrane (9). The goal of the present work was to isolate mutants lacking 2-acyl-GPE acyltransferase and acyl-ACP synthetase activity to corroborate the conclusion that 2-acyl-GPE acyltransferase/acyl-ACP synthetase are dual catalytic activities of the same protein complex, and to confirm the physiological role of the acyltransferase/synthetase in fatty acid and lysophospholipid uptake, and in maintaining membrane phospholipid composition.
Acyl-ACP Synthetase and 2-Acyl-GPE Acyltransferase Assays-Membranes were prepared by differential centrifugation of bacterial cell lysates as described previously (7). The standard acyl-ACP synthetase assay (23) contained 5 mM ATP, 10 mM MgCl,, 2 mM dithiothreitol, 0.4 M LiCl, 60 pM ['Hlpalmitic acid, 15 pM ACP, 2% Triton X-100, and 0.1 M Tris-HC1, pH 8.0, and the indicated amount of membrane protein in a final volume of 40 pl. At the end of 10 min a t 37 "C, 30 p1 of the assay mixture was withdrawn and deposited on a Whatman 3MM filter disc and washed with two changes of chloroform/methanol/acetic acid (3:6:1, v/v) to remove unreacted fatty acid. The filter papers were dried and counted in 3 ml of ACS scintillation solution to determine the amount of [:'H]palmitoyl-ACP formed.
The standard 2-acyl-GPE acyltransferase assay contained 5 mM ATP, 5 mM MgC12, 1 mM dithiothreitol, 50 p~ [''Hlpalmitic acid, 100 p~ 2-acyl-GPE, 10 p~ ACP, 1% Triton X-100, 0.1 M Tris-HC1, pH 8.0, and the indicated amount of membrane protein in a final volume of 40 pl (5). Incubations were terminated after 10 min a t 37 "C by adding 0.2 ml of ethanol. The mixture was evaporated to dryness under a stream of nitrogen, resuspended in chloroform/methanol ( l : l , v/v), and the sample applied to the preabsorbant layer of a Silica Gel G plate. The plate was then developed with chloroform/methanol/ acetic acid (85:15:10, v/v), the PtdEtn area was located with the Bioscan imaging detector, and the amount of ["HIPtdEtn formed was determined by scraping the Silica Gel from the plate and scintillation counting.
Incorporation of Exogenous Fatty Acids and Lysophospholipids-Strains were grown to the mid-logarithmic phase (approximately 5 X 10' cells/ml) in glycerol minimal medium containing 0.5% Brij-58 and labeled with 25 pCi/ml of [''Hlpalmitic acid. Culture samples (1 ml) were placed on ice, and the cells were harvested by centrifugation a t 12,000 X g for 15 min at 4 "C. The cell pellet was washed with icecold unlabeled medium. The lipids were extracted by the method of Bligh and Dyer (24) and were separated on Silica Gel G thin-layer chromatography plates developed with chloroform/methanol/acetic acid (55:20:5, v/v). The location of radioactivity on the thin-layer plate was determined with the Bioscan imaging detector. The labeled phospholipids were identified by their migration with standards and then scraped from the plate and counted in 3 ml of scintillation fluid. The same experimental approach was used to assess the ability of strains to incorporate exogenous lysophospholipids except that 100 /IM 2-acyl-GPC was added to the culture at the same time as the [''HIpalmitate.
Phospholipid Composition-The phospholipid composition of aas mutants was determined by steady-state [:"P]orthophosphate labeling as described by Nishijima and Raetz (25). Strains UB1005 and LCHl ( a a s -I ) were continuously labeled with 50 +Ci/ml of '"Pi for five generations of logarithmic growth a t 37 "C in rich medium. Cells were harvested, and the phospholipids were extracted (24) and analyzed by two-dimensional thin-layer chromatography on Silica Gel G layers developed first with chloroform/methanol/water (65:25:4) followed by chloroform/methanol/acetic acid (65:25:10). Phospholipids were visualized by autoradiography, and the radioactive areas were scraped from the plate and counted in 3 ml of scintillation solution. Phospholipids were identified by comparison with standards.
Lipoprotein Acylation-Strains LCH22, LCH31, LCH40, LCH41, SJ126, and SJ127 were grown in glycerol minimal medium containing 0.5% Brij-58 to a density of 3.2 X lo8 cells/ml. The cultures were labeled with 25 pCi/ml of ['Hlpalmitic acid for 1 h a t 37 "C. The cells were harvested, extracted, and the phospholipid labeling pattern determined as described above. Lipoprotein acylation was determined by boiling the cell pellets for 3 min in SDS-gel electrophoresis sample buffer and separating the acylated proteins on a 15% acrylamide/ 0.4% bisacrylamide gel prepared in 0.375 M Tris-HC1, pH 8.8, containing 0.1% SDS and 7.55 M urea according to Tian et al. (26). Fluorography was performed as described (4).
Genetic Analysis-The aas allele was first localized to a region of the chromosome by mating strain LCHl ( a a s -I ) with the collection of Hfr::TnIO mapping strains (14). Tet' exconjugants were selected on rich agar containing 10 pg/ml of tetracycline hydrochloride and either nalidixic acid or streptomycin to counterselect against the Hfr. The exconjugants were scored for acyl-ACP synthetase activity by replica printing. Ply,,-mediated transduction of aas with other markers in the 60-65 min region located aas near 61 min. Significant cotransduction between aas, argA, thyA, and lysA was found, and these markers were used to refine the map location. The location of aas with respect to thyA was determined from a three-factor cross.
Cotransduction frequencies were converted to map distance using the formula of Wu (27).

Isolation of Acyl-ACP Synthetase Mutants-A
colony autoradiography method was developed to screen a mutagenized population of strain UB1005 for the presence of acyl-ACP synthetase activity. Acyl-ACP formation (a dark colony) was completely dependent on ATP and the addition of exogenous ACP to the assay mixture increased the intensity of colony labeling. However, there was sufficient endogenous ACP present to produce adequate colony labeling for routine screening. Two strains, LCHl ( a m -I ) and LCH2 (ass-2), identified by this procedure had identical biochemical and physiological properties although strain LCHl ( a m -I ) was used for most of the experiments in this report. Strain LCHl did not have a growth phenotype under any of the standard laboratory culture conditions we employed. Filter disc screening of mutant colonies and assays of cell extracts were performed at 30 and 42 "C in order to recognize temperature-sensitive mutations in acyl-ACP synthetase; however, extracts from strains LCHl and LCH2 lacked acyl-ACP synthetase activity at both 30 and 42 "C, and no temperature-sensitive mutants were found.
Biochemical Characterization of Strain LCHl -Membranes were isolated from strains UB1005 and LCHl (aas-1) and the specific activities of 2-acyl-GPE acyltransferase and acyl-ACP synthetase were compared. Membranes from strain UB1005 had acyltransferase activity or 0.21 nmol/min/mg and synthetase activity of 20 pmol/min/mg. In contrast, Strain LCHl ( a m -1 ) had c0.01 nmol/min/mg of acyltransferase activity and c0.1 pmol/min/mg of synthetase activity. The same results were obtained with strain LCH2 (am-2) (not presented). The acyltransferase/synthetase was postulated to be responsible for the uptake and incorporation of fatty acids into PtdEtn in strains (fadD) lacking acyl-CoA synthetase activity (8). In a control experiment, strain LCH29 ( f a d D ) incorporated exogenous ['Hlpalmitate exclusively into PtdEtn as anticipated (8). In contrast, strain LCH30 (aas-I fadD) did not incorporate exogenous ["Hlpalmitate into PtdEtn (Fig. 1) or any other phospholipid class (not shown) illustrating that a m mutants lacked the acyl-CoA-independ- 2-Acyl-GPE Acyltransferase/Acyl-ACP Synthetase M u t a n t s ent pathway for incorporation of fatty acids into phospholipid. A second physiological process ascribed to this enzyme system was the uptake and acylation of exogenous 2-acyl-lysophospholipids (9). This parameter was measured in aas mutants using exogenous 2-acyl-GPC as the lysophospholipid and ["HI -palmitic acid (9). 2-Acyl-GPC uptake and acylation was easily detected in fadD mutants, hut was absent in aas fadD double mutants (Tahle 11). Phosphatidylcholine formation was also hlocked in strain LCHl ( a a s -I ) which was capable of assimilating large amounts of exogenous fatty acid for phospholipid synthesis via the acyl-CoA-dependent ( f a d D ) pathway (Tahle II), showing that the ACP-dependent acyltransferase/synthetase was the only mechanism for the uptake and acylation of 2-acyl-lysophopholipids in E. coli. These data illustrate that t.he a m mutation results in the concomit.ant loss of hoth 2-acyl-GPE acyltransferase and acyl-ACP synthetase activities in oitro and blocks the acyl-CoA-independent incorporation of exogenous fatty acids and lysophospholipids into the membrane i n uioo.
Alterations in Phospholipid Composition Due to the aas Mutation-The presence of t h e aas mutants did not cause a major change in the amounts of the major phospholipid species, hut two minor phospholipid components accumulated in aas mutants. First, lysophosphatidylethanolamine content increased (Fig. 2) as was anticipated since this phospholipid is a suhstrate for the acyltransferase/synthetase. However, the largest difference between wild type and aas mutants was the accumulation of another phospholipid identified as acyl-PtdGro (Fig. 2). Acyl-PtdCro is formed hy the action of lysophospholipase L, which transfers the fatty acid from the 2-acyl-GPE to the headgroup glycerol of PtdGro (28,29). Hydrolysis of 2-acyl-GPE to fatty acid and GPE is a second reaction catalyzed by lysophosphlipase LL' (28). T o confirm the identity of the phospholipid that accumulates in a m mutants, we examined the phospholipid composition of pldR mutants that lack lysophospholipase L, activity (see Fig. 5). Strain LK29 (pldR12) did not contain detectahle levels of acyl-PtdGro. Strain LCH66 (aas-I pldRI2) possessed elevated levels of lysophosphatidylethanolamine, hut did not accumulate acyl-PtdGro (Fig. 3).
Protein Acylation i n aus Mutants-In strains ( f a d D ) lacking acyl-CoA synthetase activit,y, exogenous fatty acids were transported into the cell, esterified to PtdEtn, and then transferred to the amino terminus of lipoprotein (4,5,8). To clarify t h e role of aas in these steps, protein acylation by exogenous ["Hlpalmitate was examined in aas mutants carrying either the pKEN126 or pJG310 expression vectors (Fig. 4). In the control strains, SJ126 (fad[) pKEN126) and SJ127 (fadD pJG310), fatty acids were incorporated into PtdEtn and efficiently transferred to hoth the lipoprotein (Lpp) and the lipoprotein- [ "l']orthophosphate during logarithmic Krowth in rich mrtlittm. Thr cells were harvested at a density of 1 X lo'' rells per ml. and thr phospholipids were extracted anti analyzed hy two-dimensional thinlaver chromatographv as descrihetl undrr "FCxprrimental 1'rnrr.dtrrrs." A total of 8.7 X IO'' cpm was spotted onto each thin-layrr plate and hoth autoradiographs were expnsed lor 1 6 h.
pJG310) failed to incorporate exogenous [ 'Hlpalmitate into either phospholipids or lipoproteins. Strains IXH.10 (nay fndK pKEN12fi) and LCH41 ( a m fad!; pJC.310) were capahle of uptake and esterification of exogenous fatty acids hv the acvl-CoA-dependent pathway, and in these two strains hoth phospholipids and lipoproteins were laheled hy exogenous [ ' H I palmitate. These data show that the acyltransferase/synthetase was required for incorporation of exogenous fatty acids into PtdEtn, hut this enzyme system was not involved in the transacylation of fatty acids from PtdEtn to the lipoproteins.
Genetic Idcation of aas on thP Rnctprial Chromosome-The replica print method was used to localize the ans gene on the E. coli chromosome. Strain LCHl was mated with a series of Hfr::TnIO strains (14). and Tet" recomhinants were selected and assayed for acyl-ACP synthetase activitv. These experiments localized the aas gene between 60 and fi5 min of the chromosome. Next, a series of PI,,,-mediated transduction experiments was employed using selectahle markers in this region of the chromosome, and the gene was localized to 61 min (Fig. 5 ) . To confirm the gene order in this region, a threepoint transduction experiment was performed hv infecting strain LCHl ( a a s -I ) with Plv,r phage grown on strain LCH.17 (ar,qA::TnIO th-yA). Tet" recomhinants (242) were selected and scored for thy4 on plates and nas using the colony fluorography method. The results were: t h y K nos*, 19; thy.4 aas-, 25; thyA' ads+, 8; and thy.4' nus-, 190; confirming the gene order as ar,qA-thyA-aas. We did not find a recombinant that was defective in acyl-ACP synthetase hut not ?-acyl-GPE acyltransferase or o i c~ mv"ya supporting the concept that a single gene encodes hoth activities.

DISCC!SSlON
Our results provide additional support for the conclusion that. 2-acyl-GPE acyltransferase and acvl-ACP synthetase are I' NI), not detected. were contintrously laheled with ["~l'lr)rthophosphate during logarithmic growth in rich medium. ('ells were harvested at a density o f 1 X 10'' cells/ml, and the phospholipids were extracted and analyzed by twodimensional thin-layer chromatography as descrihed under "Experimental I'rocedures." A total of 7.2 X 10' cpm was spotted onto each thin-layer plate, and hoth autoradiographs were exposed for 72 h. intermediate is then transferred t o 7-acyl-GI'E t o fnrm I'trlEtn and regenerate enzyme-hound ACI'. Rot h enzyme art ivit ies copurify and are associated with a single prc~tcin species ( 7 ) . T h e ans mutants were selected solely on the basis of defective acyl-ACP synthetase activity; however, the mutants ohtained lacked hoth acyltransferase and synthetase activities in [Ytro. T h e nus allele hehaves as a single genetic locus in P I , , ?mediated transductions and is located at min 61 o f the chromosome (Fig. 5 ) . Cloning and DNA sequencing o f the nos locus will be the next step in verifying that a single structural gene encodes both activities.
Physiological analysis of nns mutants confirms the role o f 2-acyl-GI'E acyltransferase/acyl-A('I' synthetase in memhrane phospholipid metaholism. First, this enzyme system is responsihle for the uptake and incorporat ion of exogcnow fatty acids into the 1-position of I'tdEtn (Fig. 6). 'There is nn detectahle phospholipid synthesis from exogenous fatty acids in nos fndD douhle mutants (Fig. I ) , illrlstrnting that the acyl-CoA synthetase rfndll) and acyl-ACI' synthetase ( n n s ) are components of the only two pathways for the uptake and incorporation of extracellular fatty acids in I-,*. r d i . Second. the inability ofnnsfndll and nns mutants t o acylatr exogenous lysophospholipids ( Table 11) shows that the acyltransferase/ synthetase pathway is the only route for the esterification of extracellular lysophospholipids. Third, the fact that acyl-GPE and acyl-PtdGro accumulate in am mutants (Figs. 2 and 3) indicates that the resynthesis of PtdEtn is the primary metabolic fate of 2-acyl-GPE. The possibility that the acyltransferase may also be directly involved in transferring fatty acids to the amino terminus of the lipoprotein is ruled out by our experiments (Fig. 4). Although lipoprotein acylation by exogenous fatty acids is blocked in am fadD double mutants, lipoprotein acylation does occur in am mutants via the acyl-CoA synthetase (fadD)-dependent incorporation of exogenous fatty acids into PtdEtn which are then transferred to the lipoprotein (Fig. 4).
The am mutation does not have a major effect on membrane phospholipid composition, but the minor alterations observed point to the function of the acyltransferae/synthetase system and lysophospholipase L2 in membrane phospholipid metabolism (Fig. 7). 2-Acyl-GPE acyltransferase appears to be the most economical mechanism for scavenging 2-acyl-GPE generated from phospholipid turnover, and the accumulation of acyl-GPE in am mutants indicates that the acyltransferase/ synthetase is the most active pathway under normal growth conditions. However, E. coli inner membranes also contain a acyltransferase/acyl-ACP synthetase (the aas gene product) or can be either hydrolyzed to fatty acid plus G P E or the acyl moiety transferred to the headgroup of PtdGro by the action of lysophospholipase L2 (the pldB gene product). specific phospholipase that either degrades 2-acyl-lysophospholipids to fatty acid and GPE or catalyzes the transfer of fatty acids to PtdGro to form acyl-PtdGro (15, 28,29). This route for the catabolism of 2-acyl-GPE provides an alternate mechanism to prevent 2-acyl-GPE from accumulating. The observation that acyl-PtdGro accumulates in am mutants demonstrates that the lysophospholipase L2 pathway is significant in the absence of acyltranferase/synthetase activity.
Blocking the lysophospholipase Lz route inpldB mutants does not lead to a large increase in acyl-GPE accumulation (Fig.  3) indicating that there are other enzymes that catabolize 2acyl-GPE. A membrane-associated transacylase activity has been characterized that converts two 2-acyl-GPEs to PtdEtn and GPE ( l l ) , and there is a second lysophospholipase located in the soluble fraction (30). Both of these activities are distinct from thepldB gene product (14), and the role of these enzymes in controlling 2-acyl-GPE levels is not clear. Although lysophospholipids are potent detergents and would be expected to disrupt membrane structure if present in high concentrations, the lack of a growth phenotype in am mutants illustrates that there are mutiple mechanims to prevent the accumulation of lysophospholipids and that the elevated levels of acyl-GPE observed in these strains are not deleterious.