An Escherichia coli mutant lacking the cold shock-induced palmitoleoyltransferase of lipid A biosynthesis: absence of unsaturated acyl chains and antibiotic hypersensitivity at 12 degrees C.

An acyltransferase induced by cold shock in Escherichia coli, designated LpxP, incorporates a palmitoleoyl moiety into nascent lipid A in place of the secondary laurate chain normally added by LpxL(HtrB) (Carty, S. M., Sreekumar, K. R., and Raetz, C. R. H. (1999) J. Biol. Chem. 274, 9677-9685). To determine whether the palmitoleoyl residue alters the properties of the outer membrane and imparts physiological benefits at low growth temperatures, we constructed a chromosomal insertion mutation in lpxP, the structural gene for the transferase. Membranes from the lpxP mutant MKV11 grown at 12 degrees C lacked the cold-induced palmitoleoyltransferase present in membranes of cold-shocked wild type cells but retained normal levels of the constitutive lauroyltransferase encoded by lpxL. When examined by mass spectrometry, about two-thirds of the lipid A molecules isolated from wild type E. coli grown at 12 degrees C contained palmitoleate in place of laurate, whereas the lipid A of cold-adapted MKV11 contained only laurate in amounts comparable with those seen in wild type cells grown at 30 degrees C or above. To probe the integrity of the outer membrane, MKV11 and an isogenic wild type strain were grown at 30 or 12 degrees C and then tested for their susceptibility to antibiotics. MKV11 exhibited a 10-fold increase in sensitivity to rifampicin and vancomycin at 12 degrees C compared with wild type cells but showed identical resistance when grown at 30 degrees C. We suggest that the palmitoleoyltransferase may confer a selective advantage upon E. coli cells growing at lower temperatures by making the outer membrane a more effective barrier to harmful chemicals.

A unique glycolipid known as lipopolysaccharide is the major component of the outer surface of the outer membranes of Gram-negative bacteria and forms a barrier around the cell (1)(2)(3). Compounds with molecular weights of less than 600 penetrate the outer membrane via porins, but larger, poten-tially harmful agents, including many antibiotics, are excluded (2). Lipid A is the hydrophobic anchor of the lipopolysaccharide molecule (1, 4 -6). In wild type Escherichia coli grown at 30°C or above, lipid A consists of a ␤,1Ј-6-linked disaccharide of glucosamine that is phosphorylated at the 1-and 4Ј-positions ( Fig. 1) (1, 4 -6). E. coli lipid A usually contains six acyl chains (1,4). (R)-3-Hydroxymyristoyl groups are located at positions 2, 3, 2Ј, and 3Ј of the glucosamine disaccharide, and two short saturated acyl chains are attached to the (R)-3-hydroxymyristoyl groups of the distal unit, forming acyloxyacyl moieties ( Fig.  1, "Normal" lipid A) (1, 4 -6). Laurate is linked to the (R)-3hydroxymyristoyl residue at position 2Ј, and myristate is similarly attached at position 3Ј (1, 4 -6). These so-called "secondary" acyl chains are the same in wild type E. coli and Salmonella.
Recently, a cold shock-induced enzyme that catalyzes the transfer of a palmitoleoyl chain from palmitoleoyl-acyl carrier protein to the lipid A precursor Kdo 2 -lipid IV A was discovered in our laboratory (15). The greater similarity of the palmitoleoyltransferase (LpxP) to the lauroyltransferase (LpxL) than to the myristoyltransferase (LpxM), and the ability of LpxP to use the same Kdo 2 -lipid IV A substrate as LpxL (Fig. 1), led us to suggest that LpxP acylates the 2Ј (R)-3-hydroxymyristoyl chain (15). LpxP activity is induced within 15 min following a shift of exponentially growing cells from 37 to 12°C and peaks after 2 h (15). Appearance of palmitoleoyltransferase activity is correlated with massive accumulation of lpxP mRNA at 12°C (15).
Modification of bacterial cell membranes in response to cold shock is not unprecedented (16,17). Following exposure to low * This work was supported in part by National Institutes of Health Grants GM-51310 (to C. R. H. R.) and GM54882-01 (to R. J. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  temperatures, some Gram-positive bacteria, like Bacillus subtilis, produce a fatty acid desaturase encoded by des, resulting in production of glycerophospholipids with unsaturated acyl chains in addition to the branched chains normally present in Gram-positive bacteria (17,18). The fatty acids of E. coli glycerophospholipids contain some unsaturated acyl chains at all growth temperatures (19), but their composition shifts toward greater unsaturation during cold shock because of increased ␤-ketoacyl-ACP synthase II (FabF) activity (16). Additional acyltransferases that selectively incorporate unsaturated fatty acyl chains into glycerophospholipid are not induced by cold shock in E. coli (16). The constitutive glycerol-3-phosphate acyltransferases of E. coli utilize whatever acyl donors are most abundant in the acyl-ACP and acyl-coenzyme A pools (20 -22). LpxP is the first example of a specific acyltransferase that is induced by cold shock (15).
The incorporation of an unsaturated fatty acyl chain into lipid A might alter the physical properties of the outer membrane to benefit the cell in cold environments. We have therefore constructed the strain MKV11 with an insertion mutation in the palmitoleoyltransferase structural gene, lpxP. MKV11 lacks detectable cold shock-induced LpxP activity in vitro but retains the constitutive lauroyltransferase (Fig. 1). Palmitoleate-containing species are absent in lipid A isolated from the mutant subjected to cold shock, whereas two-thirds of the lipid A from cold-shocked wild type E. coli contain palmitoleate. Although our lpxP mutant strain exhibits no obvious growth defects at 12°C, the integrity of its outer membrane is compromised. Minimal inhibitory concentrations for rifampicin and vancomycin are 10 times lower in MKV11 than wild type cells at 12°C but are the same at 30°C. Genes encoding cold-induced lipid A acyltransferases are present in many bacteria that survive for long periods outside of animal hosts, including E. coli, all serovars of Salmonella enterica, and Klebsiella pneumoniae, but are absent in pathogens transmitted from animal to animal.

EXPERIMENTAL PROCEDURES
Materials-[␥-32 P]ATP was purchased from PerkinElmer Life Sciences. Surfact-Amps grade Triton X-100 and the bicinchoninic acid (BCA) protein determination kit were obtained from Pierce. All restriction enzymes and the T4 DNA ligase were products of New England Biolabs. Cloned Pfu DNA polymerase and the PCR buffers were obtained from Stratagene. Primers and isopropyl-thio-␤-D-galactoside (IPTG) were purchased from Invitrogen. Deoxyribonucleotides were purchased from Roche Molecular Biochemicals. Sodium chloride, agarose, magnesium chloride, 88% formic acid, and glass-backed Silica Gel 60 thin layer chromatography plates (0.25 mm) were from Merck. Pyridine was from Mallinckrodt Chemical Works. Tryptone, peptone, yeast extract, and other growth media agar were from Difco. All other materials were purchased from Sigma unless otherwise mentioned.
Bacterial Strains and Growth Conditions- Table I lists the E. coli K12 strains and the plasmids used in this study. Cells were cultured at the temperatures noted below in Luria broth (LB), consisting of 10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl/liter (23). Antibiotics were added as necessary at the following concentrations: 30 g/ml for kanamycin, 25 g/ml for chloramphenicol, 25 g/ml for tetracycline, 100 g/ml for ampicillin, and 20 g/ml for streptomycin.
Isolation of DNA for Molecular Biology Applications-All plasmid DNA and DNA fragment isolations were performed following the protocols for the QIAprep Spin Miniprep Kit and the QIAEX II Gel Extraction Kit (Qiagen), respectively. Genomic DNA was isolated using the protocol for bacterial cultures in the Easy-DNA TM kit (Invitrogen). Molecular biology protocols followed are those described by Sambrook et al. (24).
Construction of a Temperature-sensitive Plasmid Carrying an lpxP::kan Insertion-Plasmid pUC4K (Table I)  In cells grown at 30°C or above, the key precursor Kdo 2 -lipid IV A is utilized solely by the lauroyltransferase (8), LpxL. However, in cold-shocked cells, an additional acyltransferase, designated LpxP (15), is induced, which is proposed to incorporate palmitoleate at the same site normally reserved for laurate. In wild type cells, the action of LpxL and LpxP is followed rapidly by the myristoyltransferase, LpxM, generating hexa-acylated lipid A (7,9). About two-thirds of the hexa-acylated lipid A isolated from cells grown overnight at 12°C contains palmitoleate, and the remainder contains laurate. When the PhoP/PhoQ system is activated or when cells are grown on ammonium metavanadate, a portion of the lipid A molecules contain a palmitate residue at position 2, which is incorporated by the outer membrane enzyme PagP using glycerophospholipids as palmitate donors (37). analyzed by agarose gel electrophoresis, and the 1485-bp fragment containing the kan element was purified from the agarose. pSK57 was digested with NotI, which cuts the lpxP gene at position 417 (15,25). The kan cassette was then ligated with T4 DNA ligase into pSK57 at this position using the compatible NotI and EaeI overhanging ends to create plasmid pSMC1. The lpxP::kan fragment was amplified from pSMC1 via PCR using primers to the 5Ј (5Ј-PGCGGGCCATATGTTTC-CACAATGC-3Ј) and 3Ј (5Ј-PGCGGGCCGAATTCAGGAGGGAAAGA-3Ј) ends of lpxP. pMAK705, a vector with a gene for chloramphenicol resistance and a temperature-sensitive origin of replication, was digested with HincII, which yields blunt ends (26). The PCR product containing lpxP::kan was blunt-end ligated into this site, generating the plasmid pSMC2.
Creation of the lpxP::kan Insertion Mutant MKV11-MKV11 (MC1061 lpxP::kan) was constructed according to published procedures (26,27). pSMC2 was transformed into wild type MC1061 using 100 mM CaCl 2 treatment, followed by heat shock (24). The selection for bacterial colonies containing pSMC2 was performed at 30°C on LB plates containing chloramphenicol. After growth of a single colony to mid log phase in LB medium at 30°C, selection for recombination of this plasmid into the chromosome was executed at 44°C on LB agar containing chloramphenicol. After three cycles of growth at 30°C to stationary phase with 100-fold back dilution of a colony containing an integrated plasmid, selection for a recombinant that had retained lpxP::kan on the chromosome and had transferred the intact lpxP gene to the temperature-sensitive covering plasmid was performed in the following manner. Single colonies were selected for resistance to chloramphenicol at 30°C but not 44°C by replica plating (27). Mutant MKV11, which was identified in this manner, subsequently was found to grow on LB broth in the presence or absence of kanamycin at 30 and 44°C, indicating that the lpxP covering plasmid could be cured without loss of viability.
A polymerase chain reaction using primers specific to the 5Ј (5Ј-GCAACGGGCCATATGTTTCCACAATGCAAA-3Ј) and 3Ј (5Ј-GCAGTC-CTCGAGTTAAATGTACAACGACG-5Ј) ends of lpxP was performed on genomic DNA isolated from the wild type and strain MKV11, as well as on the excised plasmid pMKV11, isolated from the mutant prior to curing at 44°C. PCR products were run on a 1% agarose gel, and band sizes were compared with molecular weight standards. A band of ϳ1 kb, the size of the intact lpxP gene, was produced from the wild type genomic DNA and from the pMKV11 plasmid, verifying that the intact lpxP gene had been excised from the chromosome with the plasmid. However, a band of ϳ2.4 kb, the size of lpxP::kan, was produced from genomic DNA obtained from the mutant strain, MKV11, confirming that the lpxP gene was disrupted.
Growth Rate Comparisons-Cultures of wild type and mutant strains were grown at 12°C starting at an A 600 of ϳ0.03. The A 600 was measured every 6 -8 h. For prolonged growth rate comparisons, cultures were diluted into fresh medium at regular intervals to maintain cell densities below A 600 ϭ 0.6.
Preparation of Substrates for Enzymatic Assays-Carrier (Kdo) 2lipid IV A was prepared by enzymatic synthesis on a 5-10-mg scale starting with lipid IV A (28). (Kdo) 2 -4Ј-32 P-lipid IV A was made from 4Ј-32 P-lipid IV A by a scaled down version of a similar enzymatic procedure (29) and was purified by thin layer chromatography. Substrates were stored as aqueous dispersions at Ϫ20°C. Lauroyl-and palmitoleoyl-acyl carrier proteins (ACP) were synthesized from commercial ACP using recombinant E. coli acyl-ACP synthase (30). Previous procedures (31) were modified by including an additional purification step (an octyl-Sepharose column) (32) to remove residual non-acylated ACP, which may be present at higher amounts with unsaturated fatty acids because of the reduced activity of acyl-ACP synthase with palmitoleate as the substrate (33).
Cold Shock Conditions and Preparation of Membranes for Use in Enzyme Assays-Shaking cultures (typically 100 ml) were grown on LB broth at 30°C to A 600 of ϳ0.4. The cultures were divided into two equal portions. One was incubated further at 30°C for 2 more hours, and the other was shifted to 12°C for 2 h. Cells were then harvested by centrifugation at 5000 ϫ g in a Beckman JS4.3 swinging bucket rotor, washed in 10 ml of 30 mM Hepes, pH 7.5, centrifuged a second time, resuspended in 5 ml of Hepes buffer, and lysed by passage through a small French pressure cell at 10,000 pounds/square inch. Cell debris was removed by centrifugation at 4000 ϫ g for 10 min, and the resulting cell-free extract was subjected to ultracentrifugation at 100,000 ϫ g for 60 min in a Beckman Ti-70.1 rotor. The supernatant was decanted, and the membranes were homogenized in Hepes buffer. The sample was subjected to a second 60-min ultracentrifugation to remove residual soluble molecules and then resuspended in Hepes buffer to yield final protein concentrations of 1-5 mg/ml, as determined using the BCA assay (34).
Assay for Palmitoleoyltransferase Activity-Assay mixtures (15 l final volume), containing 50 mM Hepes, pH 7.5, 250 mM NaCl, 5 mM MgCl 2 , 0.1% Triton X-100, 6 M (Kdo) 2 -4Ј-32 P-lipid IV A (80,000 cpm/ nmol) and 25.0 M palmitoleoyl-ACP, as indicated, were preincubated at 12°C for 10 min. Cell membranes, diluted in 30 mM Hepes, pH 7.5, as appropriate, were then added to give a final protein concentration of 0.1 mg/ml. The reactions were incubated at 12°C for 8 h (15), and then 5-l portions were spotted onto a Silica Gel 60 TLC plate. Substrate and products were resolved by chromatography in the system chloroform, pyridine, 88% formic acid, water (30:70:16:10, v/v). The plate was dried and exposed to a Molecular Dynamics PhosphorImager screen, which was analyzed using ImageQuant version 1.2 software for the Macintosh.
Assay for Lauroyltransferase Activity-Assay mixtures (15 l final volume) (8), containing 50 mM Hepes, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 , 0.1% Triton X-100, 6 M (Kdo) 2 -4Ј-32 P-lipid IV A (80,000 cpm/nmol) and 12.5 M lauroyl-ACP, as indicated, were preincubated at 30°C for 10 min. Cell membranes were diluted and added to the reaction mixtures to a final protein concentration of 0.1 mg/ml. The reactions were incubated at 30°C for 10 min at which point 5-l portions were spotted onto a Silica Gel 60 TLC plate, which was developed and analyzed as described above for the palmitoleoyltransferase.
Cloning and Expression of the lpxP Gene Behind a T7lac Promoter-The lpxP structural gene was obtained by PCR using pSK57 (15) as the template. Primers specific for lpxP containing an NdeI restriction site at the 5Ј end (5Ј-GCAACGGGCCATATGTTTCCACAATGCAAA-3Ј) and an XhoI restriction site at the 3Ј end (5Ј-GCAGTCCTCGAGTTAAAT-GTACAACGACG-5Ј) were used. The PCR product was digested with NdeI and XhoI and ligated into the similarly digested vector, pET 21a ϩ (Novagen). The resulting hybrid plasmid, pMKV2, was transformed into BLR(DE3)pLysS cells for controlled protein expression from the T7lac promoter, which is activated following T7 polymerase expression upon induction with IPTG. Overexpression and optimal enzymatic activity of LpxP was achieved by growing the plasmid-containing cells in LB broth at 37°C to A 600 of ϳ0.5, followed by induction with 1 mM IPTG and further incubation at 12°C for an additional 10 h or at 37°C for 3 h. Cell membranes from BLR(DE3)pLysS cells containing either pMKV2 or the pET21a ϩ vector control, grown with and without IPTG, were isolated and washed as above. The palmitoleoyltransferase assay was performed on these membranes as described above, except that the protein concentration was 0.005-0.01 mg/ml in the assay mixture, and the incubation time was reduced to 5 min at 12°C, as indicated.
Antibiotic Minimal Inhibitory Concentrations-Cultures were grown from an initial A 600 of 0.01 in the presence of a range of antibiotic concentrations for 24 h at 30°C or for 72 h at 12°C. The minimal inhibitory concentration was defined as the lowest antibiotic concentration at which no measurable bacterial growth was observed. Experiments were performed in triplicate.
Isolation and Mass Spectrometry of Lipid A Samples-Typically, 200-ml cultures were grown at 12°C from A 600 of ϳ0.02 to A 600 ϭ 0.5. Cells were collected by centrifugation at 5000 ϫ g in a Beckman JA14 rotor at 4°C and resuspended in 40 ml of phosphate buffered saline (35). Lipid A was isolated from chloroform/methanol-extracted cells by hydrolysis of the residue at 100°C in 12.5 mM sodium acetate, pH 4.5, and 1% SDS, as described previously (36). Lipid A 1,4Ј-bisphosphate species were resolved on a 1-ml DEAE-cellulose column equilibrated in chloroform/methanol/water (2:3:1, v/v), which was eluted stepwise with increasing concentrations of ammonium acetate (36).
Spectra were acquired in the negative ion linear mode using a Kratos Analytical (Manchester, UK) matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometer equipped with a 337 nm nitrogen laser, a 20-kV extraction voltage, and time delayed extraction. Each spectrum was the average of 50 shots. The lipid A 1,4Ј-bisphosphate samples were prepared for MALDI/TOF analysis by depositing 0.3 l of the sample dissolved in chloroform/methanol (4:1 v/v), followed by 0.3 l of the matrix, which was a mixture of saturated 6-aza-2-thiothymine in 50% acetonitrile and 10% tribasic ammonium citrate (9:1 v/v). The sample mixtures were allowed to dry at room temperature prior to mass analysis. Hexa-acylated lipid A 1,4Ј-bisphosphate from wild type E. coli (Sigma) was used as an external standard for calibration.

RESULTS
The Effect of an lpxP Insertion Mutation on E. coli Growth at 12°C-MKV11 is a mutant of E. coli MC1061 in which a kanamycin gene cassette is inserted into the lpxP gene (Table  I). Because palmitoleoyltransferase activity is not present in MC1061 at 30°C (15), the insertion mutation should have no effect on viability or growth at 30°C and above. This was verified by our ability to cure the lpxP mutant of its lpxP covering plasmid at 44°C. However, given the fact that lpxP is induced by cold shock, it was possible that lpxP might be required for growth below 30°C. Consequently, the growth of MKV11 at 12°C was compared with that of MC1061. As shown in Fig. 2, the lpxP::kan strain grows at approximately the same rate as wild type E. coli at low temperatures.
Absence of Palmitoleoyltransferase Activity in the lpxP Mutant-Multiple copies of the plasmid pSK57 (15), which contains a 5.6-kb segment of the E. coli genome including the lpxP gene, direct the overproduction of palmitoleoyltransferase activity (15). It was therefore anticipated that a disruption of the lpxP gene would completely eliminate palmitoleoyltransferase activity from the cell. Cultures of MC1061 and MKV11 were grown to early log phase, and a portion of each culture was shifted to 12°C for 2 h. Membranes were then isolated and assayed for palmitoleoyltransferase activity. As shown in Fig.  3, conversion of the acceptor substrate, (Kdo) 2 -4Ј-32 P-lipid IV A , to a more hydrophobic species representing (Kdo) 2 -4Ј-32 Ppalmitoleoyl-lipid IV A only occurred in the cold-shocked, wild type cells and was absolutely dependent upon the inclusion of palmitoleoyl-ACP in the reaction system. No residual palmitoleoyltransferase activity was detected in MKV11. This finding is consistent with the proposal that lpxP is the structural gene for the palmitoleoyltransferase. A small amount of a more rapidly migrating radioactive species (Fig. 3, lane 6) likely arises from the action of the outer membrane enzyme PagP (37), which incorporates a palmitate chain on the hydroxymyristate residue at the 2-position of the proximal unit (Fig. 1), using endogenous glycerophospholipids as donors (38). PagP activity is greatly induced under conditions that activate the PhoP/PhoQ two-component signaling system (37) but not under the growth conditions employed in the present study.
Normal Levels of Lauroyltransferase Activity in Membranes of MKV11-Lauroyltransferase activity (7,8) was assayed in cell membranes isolated from both MC1061 and MKV11, with and without exposure to cold shock. As shown in Fig. 4, normal amounts of lauroyltransferase activity are present in the membranes of both strains, irrespective of cold shock treatment. The specific activity for the transferase in all membrane preparations is approximately the same (1.04 Ϯ 0.11 nmol/min/mg). In addition to confirming the idea of distinct genes encoding for the lauroyltransferases and palmitoleoyltransferases (Fig. 1), this result also shows that expression of the lauroyltransferase is not affected by cold shock or by the absence of the palmitoleoyltransferase.
Mass Spectrometry of Lipid A from Wild type E. coli and MKV11 Grown at 12°C-To demonstrate the importance of the palmitoleoyltransferase during cold shock in living cells, lipid A 1,4Ј-bisphosphate species were isolated from MC1061 and MKV11 grown continuously at 12°C to mid-log phase (A 600 ϳ 0.5). The hexa-acylated lipid A 1,4Ј-bisphosphates, purified by chromatography on DEAE-cellulose (36), were analyzed by MALDI/TOF mass spectrometry in the negative mode. Major peaks for lipid A from wild type cells were observed at m/z 1851.5 and 1797.4 (Fig. 5). The larger peak at m/z 1851.5, representing about two-thirds of the total, is attributed to [M Ϫ H] Ϫ of a cold-adapted lipid A molecule containing a palmitoleate residue in place of laurate (Fig. 1). The smaller peak at m/z 1797.4 represents the [M Ϫ H] Ϫ ion derived from "normal" E. coli lipid A (Fig. 1), which contains laurate in acyloxyacyl linkage at position-2Ј. The presence of the latter species in cells at 12°C is consistent with the in vitro assay data (Fig. 4), demonstrating that the lauroyltransferase activity is intact in membranes isolated from cold-shocked cells. Additional minor peaks at m/z 1769.4 and 1824.0 (Fig. 5) likely represent lipid A molecules lacking two methylene units, when compared with the two major species; these probably arise from the slow incorporation of shorter acyl chains into the acceptor (Kdo) 2lipid IV A by the late acyltransferases (7-9). Lipid A was also isolated from cold-shocked MKV11 to verify that the palmitoleoyltransferase was not functional in vivo and to ensure that no other modifications of the lipid A molecule were occurring in its absence. The mass spectrometry in the negative mode (Fig. 6) shows only a single major peak at m/z 1797.4, which is indistinguishable from the signal obtained with lipid A of wild type cells grown at 30°C or above (Fig. 1). In the absence of the palmitoleoyltransferase, the lauroyltransferase is fully functional at 12°C, indicating that the two enzymes must compete for the precursor (Kdo) 2 -4Ј-32 P-lipid IV A in wild type cells at 12°C (Figs. 1 and 5). The absence of other major lipid A species also confirms that no alternative modifications are induced in the absence of LpxP. The minor peaks at m/z 1769.5 and 1825.6 ( Fig. 6) are again interpreted as the loss or gain of two methylene units due to slow incorporation of shorter or longer acyl chains into the common precursor (Kdo) 2 -lipid IV A by the late acyltransferases. The absence of the peak at m/z 1851.5 in Fig. 6 proves that the lpxP gene product is solely responsible for palmitoleate incorporation into lipid A in E. coli cells at low temperatures.
Massive Overproduction of Palmitoleoyltransferase in Cells Expressing lpxP Behind a T7lac Promoter-The lpxP gene was cloned into a T7 polymerase expression vector, pET21a ϩ , to confirm that controlled overproduction of the LpxP protein from a non-native promoter results in increased in vitro palmitoleoyltransferase activity. As shown in Table II, massive overproduction of palmitoleoyltransferase activity is observed only in membranes of strain BLR(DE3)pLysS cells containing the lpxP gene behind the T7lac promoter on pMKV2. The membranes from cells induced with IPTG at 12°C have a specific activity of 44.3 ϫ 10 3 pmol/min/mg, nearly 6000-fold greater than wild type, cold-shocked cells (Table II). This establishes that LpxP is necessary and sufficient for the conversion of (Kdo) 2 -4Ј-32 P-lipid IV A to a more hydrophobic product in a reaction that is absolutely dependent upon the inclusion of palmitoleoyl-ACP. Interestingly, membranes from BLR(DE3)pLysS/pMKV2 that were induced with IPTG at 37°C also overexpress LpxP activity but at a lower specific activity (Table II). These findings indicate that LpxP may not be stable when overproduced at 37°C.
Antibiotic Hypersensitivity of MKV11 during Cold Shock-As lpxP gene expression is not required for growth in liquid medium (Fig. 2) or on agar plates at 12°C, the effect of lpxP inactivation on antibiotic resistance was determined at cold temperatures. Five antibiotics from different major structural classes were tested (Table III). Erythromycin had no effect on either strain at both temperatures at concentrations below 500 g/ml. However, the minimal inhibitory concentrations for rifampicin and vancomycin are 10-fold less for the mutant than for the wild type at 12°C (Table III). Bacitracin and fusidic acid also inhibited the growth of the mutant at lower concentrations than wild type during cold shock. Therefore, it is necessary for E. coli to incorporate an unsaturated fatty acid into its lipid A at low temperatures to maintain on optimal barrier for protection against harmful agents in the environment. DISCUSSION We have constructed a mutant of E. coli (MKV11) containing a kanamycin resistance cassette insertion in the structural gene of the palmitoleoyltransferase, lpxP, involved in the biosynthesis of cold-adapted lipid A (15). In contrast to its parental strain, MC1061, membranes isolated from this mutant do not convert the lipid substrate for LpxP, Kdo 2 -lipid IV A , to Kdo 2 -(palmitoleoyl)-lipid IV A (Figs. 1 and 3) and cannot incorporate any palmitoleate into its lipid A in vivo during cold shock.  1-4) or the wild type control (lanes 5-8) grown at 12 (lanes 1, 2, 5, and 6) or at 30°C (lanes 3, 4, 7, and 8). Lanes 9 and 10 are no protein controls. Conversion of Kdo 2 -4Ј-32 P-lipid IV A to Kdo 2 -4Ј-32 P-palmitoleoyl-lipid IV A occurred only in the presence of palmitoleoyl-ACP with membranes from wild type cold-shocked cells and not with membranes from mutant MKV11 grown at 12°C. Small amounts of a more rapidly migrating diacylated species (upper arrow) likely arise by action of the outer membrane acyltransferase PagP (37), but the possibility that LpxP incorporates a second palmitoleoyl chain at a slow rate cannot be excluded. The identity of a slowly migrating product (lower arrow) present in all lanes is unknown, but its formation is independent of LpxP or palmitoleoyl-ACP.

FIG. 4. Normal levels of lauroyltransferase in membranes of
lpxP mutant MKV11. The same membranes assayed for LpxP activity at 12°C in Fig. 3 were also assayed for lauroyltransferase at 30°C (8), using 0.1 mg/ml membrane protein with lauroyl-ACP as the acyl donor, as described under "Experimental Procedures." Lanes 1-4 are assays with membranes isolated from the lpxP mutant MKV11 grown either at 12 (lanes 1 and 2) or at 30°C (lanes 3 and 4).  represent the corresponding assays with membranes isolated from the wild type control MC1061. Lanes 9 and 10 are no protein controls. The evennumbered assays were done in the presence and the odd-numbered assays in the absence of lauroyl-ACP. Conversion of Kdo 2 -4Ј-32 P-lipid IV A to Kdo 2 -4Ј-32 P-(lauroyl)-lipid IV A occurred only in the presence of lauroyl-ACP. The lauroyltransferase activity was not affected by the growth temperature or by the lpxP insertion mutation.
The absence of the lpxP gene product does not confer any obvious growth phenotype in nutrient broth at low temperatures (Fig. 2). Furthermore, the inner and outer membranes of wild type and mutant cells do not appear to differ appreciably in their protein or lipopolysaccharide core composition, as judged by gel electrophoresis (data not shown). The ratio of glycerophospholipids to lipid A (normally about 10:1) is also not affected (data not shown). The absence of a strong phenotype associated with lpxP inactivation is a little surprising, considering that two-thirds of the lipid A species isolated from wild type E. coli cultured under cold shock conditions contain a palmitoleoyl residue (Fig. 5). It is possible that the up-regulation of unsaturated fatty acid biosynthesis and the increased incorporation of these fatty acids into the glycerophospholipids (19) during growth at low temperatures provide sufficient fluidity to maintain cell viability.
Interestingly, however, the integrity of the mutant outer membrane is altered during cold shock, as determined by antibiotic resistance studies. A 10-fold greater sensitivity to rifampicin and vancomycin verifies that incorporation of an unsaturated fatty acid during cold shock, although not required for cell survival, is essential for the maintenance of an effective outer membrane barrier. Although we favor the view that the outer membrane is leaky when the palmitoleoyl moiety is not incorporated into lipid A, we cannot entirely exclude the possibility that export of these antibiotics is compromised.
The activity of the lauroyltransferase (LpxL) (8) is not affected in mutant cells grown at both 12 and 30°C, confirming that the lauroyltransferase and palmitoleoyltransferase are distinct enzymes (Fig. 1). LpxL activity does not decrease in wild type cells upon cold shock indicating that there must be competition between LpxL and LpxP activity during growth at low temperatures. The composition of the acyl-ACP pool clearly must influence the ability of these enzymes to catalyze their respective acylation reactions at 12°C. The increase in activity of ␤-ketoacyl-ACP synthase II at cold temperatures (a direct effect on the enzyme) (16) may result indirectly in the diminished availability of lauroyl-ACP relative to palmitoleoyl-ACP. In vitro, the specific activity of LpxP is at least an order of magnitude lower than that of LpxL (15), but to date, the opti- FIG. 5. Mass spectrometry of lipid A species from wild type E. coli grown at 12°C. Lipid A species were isolated from the wild type strain, MC1061, after 32 h at 12°C to A 600 ϳ0.5. MALDI/TOF mass spectrometry in the negative mode reveals a peak at m/z 1851.5, consistent with [M Ϫ H] Ϫ of a hexa-acylated lipid A species in which a palmitoleoyl moiety has replaced laurate (Fig. 1). The peak at m/z 1797.4 is interpreted as [M Ϫ H] Ϫ of normal hexa-acylated lipid A containing laurate, as seen in wild type cells grown at 30°C or above (Fig. 1). The minor peaks at m/z 1824.0 and 1769.4 are consistent with structural variants lacking two methylene units, when compared with the major lipid A species at m/z 1851.5 and m/z 1797.4, respectively. These variants may arise because of the lack of absolute chain length specificity of LpxM (9). Lipid A isolated from wild type cells grown at 30°C or above contains only the species with m/z 1797.4 (data not shown).
FIG. 6. Mass spectrometry of lipid A species from mutant MKV11 grown at 12°C. Lipid A was isolated from the lpxP mutant strain MKV11 after growth at 12°C to A 600 ϳ0.5. The peak at m/z 1797.4 is interpreted as [M Ϫ H] Ϫ of hexa-acylated lipid A, containing a laurate moiety (Fig. 1). The prominent species at m/z 1851.5 seen in Fig. 5 is completely absent, confirming the function of LpxP as the palmitoleoyltransferase in living cells. The minor peaks at m/z 1769.5 and 1825.8 represent variants containing two fewer or two additional methylene units than the species at m/z 1797.4, possibly reflecting relaxed substrate specificity of LpxM in cells at 12°C (9). When MKV11 is grown at 30°C or higher, the lipid A species are the same as in 12°C grown cells (data not shown). a MC1061 and MKV11 were assayed for 8 h at 12°C with 0.1 mg/ml membrane protein in the final assay mixture. The short times (5 min) for which the membranes of the plasmid-containing constructs were assayed at 12°C were not sufficient to detect the wild type chromosomal levels of LpxP of the host strain. Membranes from all plasmidcontaining strains were assayed at 0.01 mg/ml in the final assay, except membranes of BLR(DE3)pLysS/pMKV2, which were at 0.005 mg/ml. mal conditions for assaying these enzymes have not been studied in depth.
Palmitoleate is preferentially incorporated into glycerophospholipids by 1-acyl-glycerol-3-phosphate acyltransferase (encoded by plsC) (39), and palmitoleoyl-ACP makes up a large fraction of the acyl-ACP pool isolated from a mutant defective in glycerol 3-phosphate formation (40). Therefore, the incorporation of palmitoleate into lipid A is not limited by substrate availability at normal growth temperatures (30°C and above). The LpxP reaction must be limited by the absence of enzyme protein and/or enzyme activity at higher temperatures. It has been determined previously that the level of lpxP mRNA at 30°C is barely measurable but increases by several orders of magnitude following cold shock (15). However, the induction of a T7lac promoter-driven construct of lpxP with IPTG at 37°C yields more LpxP protein than is seen at 12°C (Fig. 7), but it is less active (Table II). These findings indicate that LpxP activity may be regulated at several levels.
Multiple mechanisms of regulation are seen with other cold shock proteins, including the well characterized CspA protein (41)(42)(43). The latter is an RNA-binding protein, and its expression is regulated at the level of transcription by an AT-rich upstream element (41)(42)(43). In addition CspA self-regulates at the level of translation and mRNA stability via the "cold box," a sequence upstream of the ribosome-binding site, as well as a "downstream box" within the structural gene (41)(42)(43). The 5Ј-untranslated regions of CspA and other cold shock proteins are unusually long, often around 150 bases, within which important sequences for regulation appear to be located (41)(42)(43). It has been determined recently that the 5Ј-untranslated region of the lpxP gene is 152 bases long, 2 and it follows that this segment may be important for regulation. Further studies of lpxP transcription and translation are in progress.
All E. coli cold shock proteins discovered to date, with the exception of LpxP, are involved with the maintenance of intracellular processes required for macromolecular synthesis, including replication, transcription, translation, and protein folding (41)(42)(43). The palmitoleoyltransferase is the first protein in E. coli specifically induced by cold shock whose activity has effects on the outer membrane that are beneficial to the cell. The increased incorporation of unsaturated fatty acids into a membrane structure is termed "homeoviscous" adaptation (44,45), and results in a membrane containing more fatty acyl chains with lower melting temperatures in order to maintain the appropriate fluidity. Further characterization of the palmitoleoyltransferase, its regulation, and the phenotype of mutant MKV11 should result in a better understanding of this adaptation process and its importance to Gram-negative bacteria.