Function of the Escherichia coli msbB gene, a multicopy suppressor of htrB knockouts, in the acylation of lipid A. Acylation by MsbB follows laurate incorporation by HtrB.

Overexpression of the Escherichia coli msbB gene on high copy plasmids suppresses the temperature-sensitive growth associated with mutations in the htrB gene. htrB encodes the lauroyl transferase of lipid A biosynthesis that acylates the intermediate (Kdo)2-lipid IVA (Brozek, K. A., and Raetz, C. R. H. (1990) J. Biol. Chem. 265, 15410-15417). Since msbB displays 27.5% identity and 42.2% similarity to htrB, we explored the possibility that msbB encodes a related acyltransferase. In contrast to htrB, extracts of strains with insertion mutations in msbB are not defective in transferring laurate from lauroyl acyl carrier protein to (Kdo)2-lipid IVA. However, extracts of msbB mutants do not efficiently acylate the product formed by HtrB, designated (Kdo)2-(lauroyl)-lipid IVA. Extracts of strains harboring msbB+ bearing plasmids acylate (Kdo)2-(lauroyl)-lipid IVA very rapidly compared with wild type. We solubilized and partially purified MsbB from an overproducing strain, lacking HtrB. MsbB transfers myristate or laurate, activated on ACP, to (Kdo)2-(lauroyl)-lipid IVA. Decanoyl, palmitoyl, palmitoleoyl, and (R)-3-hydroxymyristoyl-ACP are poor acyl donors. MsbB acylates (Kdo)2-(lauroyl)-lipid IVA about 100 times faster than (Kdo)2-lipid IVA. The slow, but measurable, rate whereby MsbB acts on (Kdo)2-lipid IVA may explain why overexpression of MsbB suppresses the temperature-sensitive phenotype of htrB mutations. Presumably, the acyloxyacyl group generated by excess MsbB substitutes for the one normally formed by HtrB.

The htrB gene was first described by Karow and Georgopoulos (1,2) as essential for rapid growth of Escherichia coli on nutrient broth above 33°C. At elevated temperatures, peculiar morphological changes are observed in htrB-deficient strains, including bulging of the cell surface and filamentation (1)(2)(3). At permissive temperatures, htrB mutants display increased resistance to bile salts (3). Based on these phenotypes, Karow and Georgopoulos (1,(3)(4)(5) suggested function(s) for htrB in cell envelope assembly, including possible roles in peptidoglycan, lipopolysaccharide, and fatty acid biosynthesis.
We have recently demonstrated (6) that the htrB gene of E. coli encodes the lauroyl transferase that acylates the key lipid A biosynthesis intermediate (Kdo 1 ) 2 -lipid IV A (7) (Fig. 1). This conclusion is based on assays of extracts prepared from htrB-deficient mutants, in which lauroyl transferase activity is undetectable (6). Conversely, cells overexpressing the htrB gene on hybrid plasmids overproduce the transferase several hundredfold (6). We have purified the overproduced enzyme to homogeneity and have shown by N-terminal sequencing that htrB is indeed the structural gene for the lauroyl transferase. 2 Identification of htrB as the lauroyl transferase is further supported by fatty acid analyses of lipopolysaccharide isolated from htrB-deficient strains of E. coli, in which the amount of laurate is reduced (4). Lipid A from htrB-deficient Hemophilus influenzae is also under-acylated (8).
Karow and Georgopoulos (3,5) identified several genes that, when introduced on hybrid plasmids, suppress the temperature-sensitive growth associated with htrB mutations. The msbA suppressor encodes a putative transport protein with a remarkable similarity to the mammalian mdr genes (5). Most Mdr proteins pump hydrophobic drugs out of animal cells, but some are involved in secretion of phospholipids into bile (9). Although the biochemical function of msbA is unknown, the sequence similarity of msbA to mdr suggests that MsbA might be involved in a translocation process, such as the movement of newly made lipid A from the cytoplasmic surface of the inner membrane to the periplasm (10 -12). Whatever its role, the msbA gene is essential for growth (5).
A second suppressor gene, designated msbB, was found to have 27.5% sequence identity and 42.2% similarity to htrB (3) suggesting a biochemical function related to htrB. MsbB reverses the temperature sensitivity associated with htrB mutations when introduced on plasmids that are maintained at high copy number (3). The msbB gene itself is not essential for growth (3). MsbB knockouts greatly reduce the amount of myristate attached lipid A, but they do not affect the laurate content (13). Lipopolysaccharide isolated from msbB mutants contains penta-acylated lipid A (designated (Kdo) 2 -(lauroyl)lipid IV A in Fig. 1) (13). MsbB was discovered independently by Somerville et al. (13), who found that whole E. coli cells harboring msbB insertions are orders of magnitude less immunostimulatory than are wild-type cells.
Using direct enzymatic assays, we now demonstrate that the msbB gene of E. coli encodes a distinct, late-functioning acyltransferase of lipid A assembly (7). MsbB functions optimally after laurate incorporation by HtrB has taken place (Fig. 1). The slow, but significant, rate at which MsbB acylates (Kdo) 2lipid IV A explains why msbB ϩ works only as a high multi-copy suppressor of htrB mutations (3). Our findings support the view that at least one acyloxyacyl residue must be present on a significant fraction of the lipid A moieties of E. coli to allow rapid growth above 33°C. A preliminary abstract of our findings has appeared (14).

EXPERIMENTAL PROCEDURES
Materials-[␥-32 P]ATP was obtained from DuPont NEN. Pyridine, chloroform, methanol, and 88% formic acid were from Fisher. All detergents were of high quality grade (peroxide-and carbonyl-free). Triton X-100 was from Pierce, and Thesit was from Sigma. Acyl carrier protein was purchased from Sigma. Other items were obtained from the following companies: 0.25-mm glass-backed silica gel 60 thin layer chromatography plates (E. Merck), yeast extract and Tryptone (Difco), and DEAE-Sepharose CL-6B (Pharmacia Biotech Inc.).
Bacterial Strains and Plasmids-Strains used in this study are derivatives of E. coli K12, and their genotypes are listed in Table I. Cultures were grown in Luria broth, consisting of 5 g of NaCl, 5 g of yeast extract, and 10 g of Tryptone per liter (15). Antibiotics were added, when required, at 100 g/ml for ampicillin, 12 g/ml for tetracycline, and 10 g/ml for chloramphenicol.
Recombinant DNA Techniques-Plasmid DNAs were isolated using the Wizard miniprep kit (Promega). Other recombinant DNA techniques were performed as described previously (16).  (6). The preparative reaction contained 50 mM HEPES, pH 7.5, 0.2% Triton X-100, 50 M (Kdo) 2 -[4Ј-32 P]lipid IV A (2 ϫ 10 3 dpm/nmol), 50 M lauroyl-ACP, 5 mM MgCl 2 , 50 mM NaCl, and 8 g of lauroyl transferase (specific activity, 2200 nmol ϫ min Ϫ1 ϫ mg Ϫ1 ) per ml. The total reaction volume was 150 l. After incubation at 30°C for 30 min, the reaction mixture was applied as a single line onto a 20 ϫ 20-cm Silica Gel 60 thin layer chromatography plate. After air drying, the plate was developed to the top in the solvent chloroform, pyridine, 88% formic acid, water (30:70:16:10, v/v). The product was detected by a brief autoradiography, and the region of the silica plate containing the desired, more rapidly migrating product was scraped off into a scintered glass funnel. The silica chips were washed once with 4 ml of chloroform. The product, (Kdo) 2 -(lauroyl)-[4Ј-32 P]lipid IV A , was then eluted with three 3.8-ml portions of a single phase, acidic Bligh and Dyer solvent mixture consisting of chloroform, methanol, 0.1 M HCl (1:2:0.8, v/v) (22). The eluted material (12 ml) was converted to a two-phase, acidic Bligh and Dyer system by the addition of 3 ml each of chloroform and 0.1 M HCl. The tube was mixed vigorously, and the phases were separated by a brief centrifugation. The lower phase was transferred to a new glass tube, and 18 drops of pyridine were added before the solvent was evaporated under a stream of nitrogen. The dried lipid was resuspended in 10 mM Tris chloride, pH 7.5, containing 1 mM EDTA. It was stored at Ϫ80°C and was resuspended by brief (1 min) sonic irradiation prior to use.
Acyl-ACPs containing various acyl chains were synthesized from the corresponding free fatty acids and acyl carrier protein using solubilized membranes from the acyl-ACP synthase overproducing strain, E. coli LCH109/pLCH5/pGP1-2, as described previously (6).
Preparation of Cell-free Extracts, Membranes, and Soluble Fractions-Crude cell-free extracts were made from 1-2 liters of logarithmically growing cultures. After harvesting by low speed centrifugation at 2°C, cells were washed once in 30 mM HEPES, pH 7.5, containing 1 mM EDTA and 1 mM EGTA (half the volume of the original culture). The washed cell pellet was resuspended in 30 mM HEPES, pH 7.5, containing 1 mM EDTA and 1 mM EGTA (a volume approximately equal to the volume of the cell pellet). Cells were broken using an ice-cold French pressure cell (SLM Instruments, Urbana, IL) at 20,000 psi. The broken cell suspension was adjusted to 10 mM MgSO 4 , and DNase I was added to 1 g/ml. After a brief sonic irradiation on an ice water bath to decrease the viscosity, the suspension was incubated for 30 min at 30°C. Unbroken cells were removed by centrifugation at 1,000 ϫ g for 10 min. Membranes and soluble fractions were separated by centrifugation at 150,000 ϫ g for 60 min. The supernatant was centrifuged a second time to remove residual contaminating membranes. The membrane pellet was resuspended in 25 ml of 30 mM HEPES, pH 7.5, containing 1 mM EDTA and 1 mM EGTA, and it was centrifuged again as above to generate the final, washed membrane fraction (ϳ40 mg/ml). The membrane suspension was stored at Ϫ80°C.
Protein concentrations were determined with the bicinchoninic assay (Pierce), using bovine serum albumin as the standard (23).
Assay for Acyl-ACP-dependent Acylation of Kdo-containing Precursors-HtrB-catalyzed acylation was assayed using Method I described earlier (6). In most cases, the reaction mixture contained 50 mM HEPES, pH 7.5, 0.1% Triton X-100, 25 M (Kdo) 2 -[4Ј-32 P]lipid IV A (ϳ2 ϫ 10 3 dpm/nmol), 25 M lauroyl-ACP, 0.1 mg/ml bovine serum albumin, and 0.1-1000 g/ml enzyme at 30°C in a final volume of 10 -20 l. With highly purified preparations of HtrB, enzyme stability is increased by including 50 mM NaCl and 5 mM MgCl 2 in the assay mixture (6). The MsbB-catalyzed acylation of the product generated by HtrB, designated (Kdo) 2 -(lauroyl)-[4Ј-32 P]-lipid IV A , was assayed in a FIG. 1. Kdo-dependent "late" acylations of lipid A precursors in extracts of E. coli. The acylation catalyzed by E. coli HtrB is 8-fold selective for laurate over myristate (6), consistent with the composition of the acyloxyacyl residues of lipid A isolated from cells (6). The MsbBcatalyzed acylation of these precursors is 100-fold selective for the penta-acylated (Kdo) 2 -(lauroyl)-lipid IV A (generated by HtrB) over the tetra-acylated precursor, (Kdo) 2 -lipid IV A . In extracts, however, MsbB is only slightly selective for myristate over laurate. The possibility that HtrB can function on the (Kdo) 2 -[myristoyl]-lipid IV A generated at a slow rate by MsbB from myristoyl-ACP and (Kdo) 2 -lipid IV A has not been examined (as indicated by the question mark). The proposed structures of the lipids generated by HtrB and MsbB are surmised based on what is known about the lipid A structure of E. coli (11, [25][26][27][28]. Unequivocal analyses to establish the structures of these acylation products of (Kdo) 2 -lipid IV A generated in vitro have not yet been carried out. Since studies using partially purified MsbB showed a stabilizing effect of sodium chloride, 50 mM NaCl was also included in the reaction in some experiments. MgCl 2 inhibited MsbB slightly. Reactions were stopped by spotting a 4 -5-l sample onto a Silica Gel 60 thin layer chromatography plate. After air drying and developing the thin layer chromatography plate in the solvent chloroform, pyridine, 88% formic acid, water (30:70:16:10, v/v), the plates were exposed to a PhosphorImager screen. The amount of radioactivity present in the labeled spots corresponding to substrate and product were quantitated using a Molecular Dynamics Imager. The specific activity of both HtrB and MsbB was expressed in terms of nmol/min/mg protein.
Analysis of the Lipid A 4Ј-Monophosphates Extracted from 32 P i -labeled Cells-Cells were grown and labeled by the method of Galloway and Raetz (24), with several modifications. Briefly, bacteria from 30°C overnight cultures were inoculated into 20 ml of fresh LB broth and grown for several hours at 30°C until A 600 had reached ϳ0.1. Next, two 5.0-ml portions of each culture were transferred into two new culture tubes, and 32 P i (5 Ci/ml) was added to both. One was grown for 3 h at 30°C, and the other was grown for 3 h at 42°C. 32 P-Labeled cells were washed twice with 5.0 ml of phosphate-buffered saline, pH 7.4, and resuspended in 0.8 ml of the same buffer. Washed cells were extracted at room temperature for 60 min with a single phase Bligh and Dyer mixture, formed by the addition of 2 ml of methanol and 1 ml of chloroform. After centrifugation in a clinical centrifuge at top speed for 20 min, the pellets were recovered and washed once with 5.0 ml of a single phase of Bligh and Dyer mixture, consisting of chloroform/methanol/water (1:2:0.8, v/v). The pellets were resuspended in 3.6 ml of 0.2 M HCl by sonic irradiation in a bath. The resuspended pellets were incubated at 100°C for 90 min. The acid-hydrolyzed material was converted to a two-phase Bligh and Dyer system by addition of 4 ml of chloroform and 4 ml of methanol. After centrifugation at low speed, the upper phases were removed, and the lower phases were washed twice with 4.0 ml of pre-equilibrated acidic upper phase of a two-phase Bligh and Dyer system, generated by mixing chloroform, methanol, 0.2 M HCl (2:2:1.8, v/v). The washed lower phases were dried under a nitrogen stream. The lipid A 4Ј-monophosphates were redissolved in 100 l of chloroform/methanol (4:1, v/v), and approximately 1000 cpm of the samples was applied to each lane of a Silica Gel 60 thin layer chromatography plate. The plate was developed in the solvent of chloroform, pyridine, 88% formic acid, water (50:50:16:5, v/v). The plate was dried and exposed overnight to a PhosphorImager screen.

Activity of the htrB Encoded Lauroyl Transferase of E. coli Is
Not Reduced by msbB Insertion Mutations-Extracts of strains bearing htrB mutations do not catalyze lauroyl-ACP-dependent acylation of (Kdo) 2 -[4Ј-32 P]lipid IV A , as shown in Fig. 2 by the absence of (Kdo) 2 -(lauroyl)-lipid IV A formation (product a in lane 2). The transfer of laurate from lauroyl-ACP to (Kdo) 2 -[4Ј-32 P]lipid IV A was efficient in extracts of strains harboring an insertion in msbB (Fig. 2, lane 6), as in the wild type (Fig. 2,  lane 4). A similar pattern was observed when myristoyl-ACP was substituted for lauroyl-ACP (not shown), except that the initial rates of (Kdo) 2 -[4Ј-32 P]lipid IV A acylation were 5-10-fold slower with myristoyl-ACP, given the selectivity of HtrB for laurate (6).
To approach this problem, we initially incubated membranes of wild-type cells that overexpress either htrB ϩ or msbB ϩ with (Kdo) 2 -[4Ј-32 P]lipid IV A and lauroyl-ACP. As shown in Fig. 3, panel B, overexpression of htrB ϩ in the presence of msbB ϩ on the chromosome resulted in very rapid formation and accumulation of a monoacylated product at the relatively low extract concentrations employed (4 g/ml). Alternatively, overexpression of msbB ϩ in the presence of htrB ϩ on the chromosome (Fig.  3, panel A) did not increase the overall extent of (Kdo) 2 -[4Ј-32 P]lipid IV A acylation compared with wild-type (hence the need to use 100 g/ml membranes), but it did result in the accumulation of the diacylated material (product b) at the expense of the monoacylated product a. The difference between HtrB and MsbB overproduction in a wild-type background is especially apparent when comparing the product distribution at the 40-min time point in panel A of Fig. 3 (excess MsbB) to the 10-min time point in panel B of Fig. 3 (excess HtrB). In both cases the conversion of substrate to product is similar (ϳ30%), but diacylated material is generated in strains overproducing MsbB, whereas monoacylated product is formed when HtrB is overproduced. The results of Fig. 3 provide support for the idea that MsbB catalyzes the second acylation and can use lauroyl-ACP as the donor under the in vitro assay conditions employed.
Acylation by MsbB Follows HtrB-catalyzed Acylation of (Kdo) 2 -[4Ј-32 P]Lipid IV A -To evaluate further the function of MsbB and its relationship to HtrB, we prepared membranes from strains of msbB or htrB-deficient cells overexpressing htrB and msbB, respectively (6). In this way, we could study cell extracts that contained only one or the other enzyme.
As shown in Fig. 4, panel A, membranes of strain MLK53/ pBS233 (containing only msbB ϩ ) did not catalyze any measurable acylation of (Kdo) 2 -[4Ј-32 P]lipid IV A under the conditions employed. Membranes of cells containing only HtrB (Fig. 4,  panel B) catalyzed mainly the formation of product a. In Fig. 4, panel A, the (Kdo) 2 -[4Ј-32 P]lipid IV A and acyl-ACP were first preincubated for 60 min with 10 g/ml membranes of MLK53/ pBS233. At time 0, a second 60-min incubation was started. Samples were then withdrawn during the second incubation Specific Activity of MsbB in Extracts of Mutants and Plasmid-bearing Strains-A direct assay for MsbB was developed using 25 M (Kdo) 2 -(lauroyl)-[4Ј-32 P]lipid IV A as the acceptor and 25 M acyl-ACP as the donor (see "Experimental Procedures"). Acylation was detected by thin layer chromatography and PhosphorImager analysis as in Fig. 4. With the direct assay, the specific activity of MsbB in extracts of wild-type cells was ϳ0.3 nmol/min/mg, but in extracts of msbB-deficient mutants, the activity was below the limits of detection (ϳ0.03 nmol/min/mg). Overexpression of msbB ϩ on multi-copy hybrid plasmids resulted in ϳ100-fold higher specific activity of MsbB in cell extracts or membranes compared with wild-type, as shown in Table II. The large size of the observed effects on MsbB-specific activity supports the proposal that msbB is the structural gene encoding the MsbB acyltransferase.
Solubilization and Partial Purification of MsbB-Washed membranes (5 ml) from E. coli MLK53/pBS233 were diluted in 20 mM Tris chloride, pH 7.8, containing 1 mM EDTA and 1 mM EGTA to a final protein concentration of ϳ17 mg/ml protein (208 mg of protein in a total volume of 12 ml). The membrane suspension was mixed with an equal volume of solubilization buffer, giving final concentrations of 2.5% Triton X-100, 20 mM Tris chloride, pH 7.8, 1 mM EDTA, 1 mM EGTA, 100 mM sodium phosphate, 150 mM NaCl, and 10% glycerol. The solubilization mixture was incubated with slow stirring at 5°C for 2 h. Insoluble material was removed by centrifugation for 60 min at 170,000 ϫ g at 5°C. The supernatant was transferred to a new tube and diluted to a final volume of 72 ml with 20 mM Tris chloride, pH 7.8, 1 mM EDTA, and 1 mM EGTA to reduce the NaCl concentration. The solubilized proteins were loaded onto a DEAE-Sepharose CL-6B column (45-ml packed bed-volume) equilibrated in 0.2% Thesit, 20 mM Tris chloride, pH 7.8, 1 mM EDTA, 1 mM EGTA, 50 mM NaCl, and 10% glycerol. After washing the column with 140 ml of equilibration buffer, bound proteins were eluted with a 400-ml linear gradient of 50 -500 mM NaCl in 0.2% Thesit, 20 mM Tris chloride, pH 7.8, 1 mM EDTA, 1 mM EGTA, and 10% glycerol. The absorbance at 280 nm of each fraction was measured (Fig. 5, open circles), and the four peak fractions (Fig. 5, shaded area) of MsbB activity (filled circles) were pooled. The combined fractions (20 ml) were diluted 4-fold in 20 mM Tris chloride, pH 7.8, 1 mM EDTA, and 1 mM EGTA, and the material was loaded onto a 1.8-ml DEAE-Sepharose CL-6B column, equilibrated as above, to concentrate the sample. The total protein was then eluted into a total volume of 1.5 ml of 0.05% Triton X-100, 20 mM Tris chloride, pH 7.8, 1 mM EDTA, 1 mM EGTA, 200 mM NaCl, and 10% glycerol. The fractionated MsbB was stored in aliquots at Ϫ80°C.
Stability and Assay of Purified MsbB-Using DEAE-Sepharose-purified MsbB, we examined the stability and time dependence of product formation, using enzymatically synthesized (Kdo) 2 -(lauroyl)-[4Ј-32 P]lipid IV A and myristoyl-ACP as At the times indicated, 5 l of each reaction mixture was spotted onto a Silica Gel 60 thin layer chromatography plate. The plate was developed, and the radioactive spots were detected as in Fig. 2. Arrows indicate the substrate (Kdo) 2 -[4Ј-32 P]lipid IV A and the two predominant acylation products (a and b) that are observed in unfractionated systems.

FIG. 4. MsbB-catalyzed acylation of (Kdo) 2 -[4-32 P]lipid IV A is accelerated by preincubation with HtrB.
Membranes (10 g/ml) of pKS12/MLK1067 (an HtrB over-producer lacking MsbB) were incubated with (Kdo) 2 -[4Ј-32 P]lipid IV A as described in Fig. 3 but with both 25 M lauroyl-ACP and 25 M myristoyl-ACP as acyl donors in a total volume of 80 l. After 60 min of preincubation to allow the HtrBcatalyzed synthesis of (Kdo) 2 -(lauroyl)-[4Ј-32 P]lipid IV A (product a), the reaction was split into two equal portions. To one portion (panel C) membranes (10 g/ml final concentration) of pBS233/MLK53 (an MsbB overproducer lacking HtrB) were added in 4 l. To the second portion (panel B) a corresponding volume of buffer was added. The incubation was continued for another 60 min at 37°C. A third reaction mixture (panel A) with identical concentrations of substrates and buffers, but containing only membranes (10 g/ml) of pBS233/MLK53 (an MsbB overproducer lacking HtrB), was preincubated and then incubated for another 60 min in parallel. At the times indicated following the preincubation, 5-l portions of each reaction were spotted onto Silica Gel 60 thin layer chromatography plates. The plates were developed, and radioactive spots were detected as in Fig. 2. Arrows indicate the locations of the substrate, (Kdo) 2 -[4Ј-32 P]lipid IV A , and the two predominant acylation products (a and b).
the substrates (Fig. 6). In contrast to the significant activation previously observed with HtrB (6), we found that MsbB was slightly inhibited by the presence of 5 mM MgCl 2 (Fig. 6). However, 50 mM NaCl was beneficial (Fig. 6), as in the case of HtrB (6), and MsbB was stable to preincubation under assay conditions for 30 min in the presence of NaCl (not shown). The MsbB assay carried out in the presence of 50 mM NaCl displayed excellent linearity with respect to time (Fig. 6) and protein concentration (not shown).
Using the improved conditions, the fractions generated in the course of solubilizing and purifying MsbB were reassayed. Table II shows that the final MsbB preparation is purified about 24-fold relative to washed membranes from the overproducing strain employed, and about 2400-fold relative to crude extracts of wild-type cells.
Selectivity of MsbB for Acyl Chain Length-As shown in Fig.  7, DEAE-Sepharose-purified MsbB displays a slight kinetic preference for myristoyl-ACP over lauroyl-ACP. Other acyl donors, including decanoyl-, palmitoyl-, palmitoleoyl-, and (R)-3hydroxymyristoyl-ACP are inactive as acyl donors when (Kdo) 2 -(lauroyl)-[4Ј-32 P]lipid IV A is employed as the acceptor (Fig. 7). The virtual absence of activity with decanoyl-ACP is interesting. HtrB does utilize decanoyl-ACP as a donor in vitro at about one-third of the rate of lauroyl-ACP. Given the fact that we have tested the same acyl-ACP preparations with both HtrB (6) and MsbB (Fig. 7), we conclude that the optimal activity of MsbB with various acyl-ACPs under our assay conditions is set approximately two carbons longer than that of HtrB. In living cells, however, the selectivity of HtrB for lau-rate and of MsbB for myristate must be higher. No decanoate is found in the position normally occupied by laurate on lipid A under ordinary growth conditions in living cells (Fig. 1), and very little laurate is incorporated in place of myristate (11, [25][26][27][28].
State of Lipid A Acylation in msbB-and htrB-deficient Mutants and the Effect of Plasmids Bearing msbB ϩ -In previous studies (24), we described a simple, radiochemical method for  determining the lipid A content of E. coli. Our procedure involves the labeling of cells with 32 P i , followed by Bligh-Dyer extraction of the glycerophospholipids (24). Next, the cell resi-due is treated with O.2 M HCl at 100°C to release all the lipid A moieties of lipopolysaccharide as a series of 4Ј-monophosphates (24). These compounds are separated into hexa-acylated, penta-acylated, and tetra-acylated species, using silica gel thin layer chromatography (24), and they can be detected by autoradiography or PhosphorImager analysis.
As shown in Fig. 9 (lanes 2 and 3), the lipid A-derived 4Ј-monophosphates of wild-type cells, labeled for 3 h at either 30 or 42°C in LB broth (15) with 32 P i , consist mostly of hexaacylated, some penta-acylated, and a few tetra-acylated forms. The relative abundance of the hexa-acylated species may be underestimated slightly by our procedure because of partial deacylation caused by the acid hydrolysis (24). However, strains defective in msbB yield mostly penta-acylated 4Ј-monophosphates with only minor amounts of the tetra-acylated species at both 30 or 42°C (Fig. 9, lanes 4 and 5), consistent with the proposed function of MsbB (Fig. 1). Under nonpermissive conditions (42°C), strains defective in htrB or in both htrB and msbB (Fig. 9, lanes 7 and 9, respectively) yield mostly tetraacylated lipid A 4Ј-monophosphates. Interestingly, when grown at 30°C, these mutants do generate significant amounts of penta-acylated and even some hexa-acylated forms (Fig. 9,  lanes 6 and 8), the origin of which is not entirely clear (see "Discussion"). However, when msbB ϩ is overexpressed in htrBdeficient cells, growth and considerable amounts of additional penta-acylated 4Ј-monophosphates are restored at 42°C (Fig.  9, lane 11 versus lane 7). These observations support the view Each of the cultures was divided into two 5-ml portions, and 32 P i (5 Ci/ml) was added to both. One culture was grown for 3 h at 30°C, and the other was grown for 3 h at 42°C. After this period to allow for uniform labeling of the cellular lipids, the lipid A 4Ј-monophosphate fractions were isolated by a modification of the procedure of Galloway and Raetz (24). Approximately 1000 cpm of isolated lipid A [ 32 P-4Ј]monophosphates was applied to each lane of a Silica Gel 60 thin layer chromatography plate. The plate was developed in the solvent chloroform, pyridine, 88% formic acid, water (50:50:16:5, v/v). The plate was then dried and exposed overnight to a PhosphorImager screen. The positions of the hexa-acylated, penta-acylated, and tetra-acylated [ 32 P-4Ј]monophosphates, as well as the origin, are indicated. Lanes 1 and 12 show the position of a tetra-acylated [ 32 P-4Ј]monophosphate standard that was generated by acid hydrolysis of [4Ј-32 P]lipid IV A . Even lanes are 4Ј-monophosphates from cells grown at 30°C and odd lanes from cells grown at 42°C. The strains are as follows: lanes 2 and 3, W3110; lanes 4 and 5, MLK1067; lanes 6 and 7, MLK53; lanes 8 and 9, MLK986; lanes 10 and 11, MLK53/pBS233. that high level overexpression of MsbB bypasses the need for HtrB action prior to MsbB-catalyzed acylation. Rapid cell growth apparently requires that a critical fraction of the lipid A moieties of lipopolysaccharide be at least penta-acylated. The enzymatic source of these acylations seems not to be as important. DISCUSSION As shown by the results of Figs. 2-4, 8, and 9, msbB encodes a distinct acyltransferase of lipid A assembly that functions optimally after laurate incorporation by HtrB has occurred. The proposed intermediates generated by MsbB and HtrB are shown in Fig. 1. These structural proposals are based on what is known about lipid A released from E. coli lipopolysaccharide by acid hydrolysis (11, [25][26][27][28]. The proposed structures also are in accord with the molecular weights of (Kdo) 2 -lipid IV A acylation products generated previously with crude cell extracts (7). The lauroyl and myristoyl groups are attached by HtrB and MsbB to the lipid IV A moiety of (Kdo) 2 -lipid IV A (7). However, the structures of the acylated derivatives of (Kdo) 2 -lipid IV A generated using homogeneous preparations of HtrB and MsbB remain to be established unequivocally.
Besides HtrB and MsbB, there are at least two additional enzymes that may be able to acylate (Kdo) 2 -lipid IV A in E. coli extracts. One of these transfers palmitate from the 1-position of a glycerophospholipid to the (R)-3-hydroxy moiety of the Nlinked (R)-3-hydroxymyristate of the proximal glucosamine unit (20). The palmitoyltransferase has not been purified, and the gene encoding it is unknown. In addition, we have recently obtained evidence for a cold shock-induced enzyme that transfers palmitoleate from palmitoleoyl-ACP to (Kdo) 2 -lipid IV A . 3 In living cells of E. coli and Salmonella typhimurium, cold shock causes the incorporation of an equivalent of palmitoleate into lipid A at the expense of laurate (29,30). Palmitoleate is not incorporated into (Kdo) 2 -lipid IV A by purified HtrB (6).
As shown in Fig. 9, E. coli mutants defective in htrB or in both htrB and msbB contain lipid A moieties that are mainly tetra-acylated when cells are grown at 42°C for 3 h. The presence of excess msbB ϩ on hybrid plasmids maintained at high copy number restores some penta-acylated lipid A species (Fig. 9) and suppresses the temperature-sensitive growth phenotype associated with htrB mutations (3). These findings suggest that the slow rate of acylation of (Kdo) 2 -lipid IV A catalyzed by MsbB (Fig. 8, product a) may be sufficient to generate a significant number of acyloxyacyl moieties when MsbB is overproduced in htrB-deficient cells. We presume that (Kdo) 2 -(myristoyl)-lipid IV A , having the proposed structure shown at the bottom of Fig. 1, is being generated when the htrB mutation is suppressed by overexpression of msbB ϩ .
E. coli mutants defective in htrB or in both htrB and msbB do contain some penta-and hexa-acylated lipid A species when grown at 30°C (Fig. 9). These species might arise by action of the palmitoyltransferase (20) or the cold-induced palmitoleoyltransferase noted above. To determine the origin of these more highly acylated lipid A species present in mutants defective in both htrB and msbB, it will be necessary to carry out additional chemical analyses. Preliminary results with laser desorption mass spectrometry 4 indicate that lipid A isolated from mutants defective in both htrB and msbB grown at 30°C consists of a mixture of tetra-acylated and penta-acylated forms. Both forms contain four hydroxymyristoyl chains, but the penta-acylated species contains an additional palmitate or palmitoleate.
The effects of the msbA multi-copy suppressor (5) on the extent of acylation of the lipid A moieties present in mutants defective in both htrB and msbB remain to be examined. Recently, Polissi and Georgopoulos (31) have found that lipopolysaccharide accumulates in the inner membranes of htrB Ϫ deficient mutants. Introduction of both msbA ϩ and orfE ϩ (an essential downstream gene of unknown function) on hybrid plasmids partially restored-deficient translocation of lipopolysaccharide to the outer membrane in htrB mutants (31). Insertion mutations that conditionally inactivate both msbA and orfE also appear to accumulate lipopolysaccharide in their inner membranes (31). If msbA is really involved in lipid A export, as anticipated because of its sequence similarity to mammalian Mdr proteins, one should see no restoration of lipid A acylation upon suppression of htrB Ϫ msbB Ϫ mutants by msbA ϩ . If msbA, like msbB, encodes a distinct acyltransferase, however, then the extent of lipid A acylation would be increased when mutants defective in both htrB and msbB are complemented by msbA.
A search of genome data bases indicates that H. influenzae contains htrB and msbB homologues (8). Based on mass spectrometry of lipid A isolated from htrB mutants of H. influenzae (8), it is likely that HtrB function is conserved in this organism. All known HtrB and MsbB sequences contain the peptide LCFP in the vicinity of residue 70.
The identification of the genes that are responsible for the early and late acylations of lipid A should facilitate the construction of E. coli strains possessing greatly reduced or altered endotoxin activities. The msbB insertion mutations described by Somerville et al. (13) reduce the ability of intact E. coli to activate macrophages by several orders of magnitude. By further modifying the structure and composition of E. coli lipid A, it may be possible to create strains with virtually no endotoxin activity. For instance, substitution of E. coli lpxA with Pseudomonas lpxA (32), coupled with deletion of msbB, would result in the generation of a lipid A with considerable similarity to that of Rhodobacter sphaeroides (33,34). Lipid A of R. sphaeroides actually antagonizes the activities of wild-type E. coli endotoxin (33,34). Such E. coli mutants, if viable, might be very useful for the expression and preparation of proteins intended for injection into animals, since quantitative removal of lipopolysaccharide would not be as critical. Modification of the structure of lipid A in living bacterial cells might also provide new insights into the functions of lipid A in outer membrane biogenesis and during infection of animals.