The ATP Binding Cassette Multidrug Transporter LmrA and Lipid Transporter MsbA Have Overlapping Substrate Specificities*

LmrA is an ATP binding cassette (ABC) multidrug transporter in Lactococcus lactis that is a structural and functional homologue of the human multidrug resistance P-glycoprotein MDR1 (ABCB1). LmrA is also homologous to MsbA, an essential ABC transporter in Escherichia coli involved in the trafficking of lipids, including Lipid A. We have compared the substrate specificities of LmrA and MsbA in detail. Surprisingly, LmrA was able to functionally substitute for a tempera-ture-sensitive mutant MsbA in E. coli WD2 at non-per-missive temperatures, suggesting that LmrA could transport Lipid A. LmrA also exhibited a Lipid A-stimu-lated, vanadate-sensitive ATPase activity. Reciprocally, the expression of MsbA conferred multidrug resistance on E. coli . Similar to LmrA, MsbA interacted with pho-toactivatable substrate [ 3 H]azidopine, displayed a daunomycin, vinblastine, and Hoechst 33342-stimulated vanadate-sensitive ATPase activity, and mediated the transport of ethidium from cells and Hoechst 33342 in proteoliposomes containing purified and functionally reconstituted protein. Taken together, these data demonstrate that MsbA and LmrA have overlapping substrate specificities. Our observations imply the presence of structural elements in the recently published crystal structures of MsbA in E. coli and Vibrio cholera (Chang,

Multidrug transporters interfere with the drug-based control of cancer and infectious diseases by mediating the extrusion of cytotoxic drugs from the cell (1,2). Whereas ion-coupled transporters have a dominant role in efflux-based multidrug resistance in prokaryotic organisms, ATP binding cassette (ABC) 1 transporters have such a role in eukaryotic organisms (3). Certain human ABC transporters are well conserved in bacteria. For example, LmrA in the Gram-positive bacterium Lacto-coccus lactis is homologous to the human multidrug resistance P-glycoprotein MDR1 (ABCB1) (4,5), overexpression of which confers resistance on human cancer cells to chemotherapy (6). LmrA and P-glycoprotein MDR1 have very similar specificities for chemotherapeutic drugs and modulators, and surprisingly, LmrA can functionally substitute for P-glycoprotein MDR1 in human lung fibroblast cells (7).
One of the controversial issues regarding multidrug transporters is their role in cell physiology. Multidrug transporters may have a purely protective function, but they may also be involved in the transport of substrates (e.g. lipids and lipid soluble metabolites) that share physico-chemical properties with drugs (8,9). Interestingly, LmrA and P-glycoprotein MDR1 share a significant sequence similarity with MsbA, an essential ABC transporter in Gram-negative Escherichia coli that is involved in the biogenesis of the outer membrane (10). The outer membrane is an asymmetric bilayer composed of glycerophospholipids on its inner leaflet and Lipid A, the hydrophobic anchor of lipopolysaccharides, on the outer leaflet. The Lipid A moiety is a hexa-acylated disaccharide of glucosamine and is a potent activator of innate immunity in mammals via the Toll-like receptor 4 (11,12). The enzymes that synthesize phospholipids and Lipid A are well characterized and are present in the cytoplasm and cytoplasmic membrane (13). Although the precise mechanisms by which phospholipids and Lipid A are targeted to the outer membrane are unknown, recent observations demonstrate that MsbA is required for the export of all newly made lipopolysaccharides and phospholipids to this compartment of the cell envelope (10,14).
MsbA was the first ABC transporter to be crystallized and analyzed by x-ray crystallography, allowing the determination of its structure at 4.5-Å resolution for MsbA in E. coli (15) and at 3.8-Å resolution for MsbA in Vibrio cholera (16). The implications of the MsbA structure for the structure and function of LmrA and P-glycoprotein MDR1 make a comparison of the substrate specificities of these transporters highly relevant. Here, we report that the lmrA gene can complement the msbA gene in E. coli and that MsbA and LmrA have overlapping specificities for Lipid A and chemotherapeutic and cytotoxic drugs.
pNHLmrA (pNZ8048 containing lmrA with a coding region for an amino-terminal hexa-histidine tag) was grown at 30°C in M17 medium (Difco) supplemented with 20 mM glucose and 5 g/ml chloramphenicol (18).
Preparation of Inside-out Membrane Vesicles-For protein expression in L. lactis, cells were grown to an A 660 of 0.3 and incubated for 2 h at 30°C in the presence of 40 pg/ml nisin. Protein expression in the E. coli DE3 cells was induced at an A 660 of 0.5 through the addition of 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside (Melford Laboratories Ltd., Suffolk, UK) followed by incubation for 2 h at 30°C. Cells were harvested by centrifugation at 13,000 ϫ g for 15 min. The cell pellet was washed at 4°C in 50 mM potassium HEPES (pH 7.4), resuspended to an A 660 of 40 in HEPES buffer containing 5 mM ␤-mercaptoethanol (VWR International, Dorset, UK), 2 mg/ml lysosyme (Sigma), 10 g/ml DNase (Sigma), 10 mM MgSO 4 , and Complete protease inhibitor mixture (Roche Applied Science), and incubated for 30 min at 30°C. Cells were broken by passage twice through a Basic Z 0.75-kilowatt Benchtop Cell Disruptor (Constant Systems, Northants, UK) at 20,000 p.s.i. Potassium EDTA (pH 7.4) was then added to a final concentration of 15 mM. Unbroken cells and cell debris were removed by centrifugation at 13,000 ϫ g for 15 min at 4°C. Inside-out membrane vesicles were harvested by centrifugation at 125,000 ϫ g for 30 min at 4°C. The membrane vesicles were resuspended to a protein concentration of 30 mg of membrane protein/ml in 20 mM potassium HEPES (pH 7.4) containing 5 mM ␤-mercaptoethanol and stored in 150-l aliquots in liquid N 2 .
Liposomes-Lipids were extracted from 10 liters of L. lactis culture as described (18), with modifications. L. lactis NZ9000 cells harboring pNZ8048 (vector) were grown to an A 660 of about 1 and harvested by centrifugation at 13,000 ϫ g for 15 min. Cells (50 g wet weight) were resuspended in 1 ml of potassium phosphate (pH 7.0) and frozen and thawed. Subsequently, 4 mg/ml lysosyme, 10 g/ml DNase, and 10 mM MgSO 4 were added, and the cell suspension was incubated for 30 min at 30°C. All subsequent steps took place in glass containers in an N 2 atmosphere in the dark. Cells were stirred overnight in 200 ml of chloroform and 400 ml of methanol at 4°C. Undissolved particles were removed by centrifugation at 1,000 ϫ g for 10 min at 4°C. Water (100 ml) and chloroform (100 ml) were then added, and the mixture was stirred for 3 h at 20°C. Phases were separated overnight in a separation funnel, and the bottom layer containing the lipids was collected. Lipids were dried in a rotory evaporator and dissolved in chloroform to a final lipid concentration of 0.1 g/ml. One ml of the mixture was dripped into 10 ml of ice-cold acetone containing 5 mM ␤-mercaptoethanol, stirred overnight at 4°C, and centrifuged for 15 min at 1,000 ϫ g at 20°C to remove insoluble particles. The lipid solution was dried under N 2 gas and dissolved in 15 ml of diethyl ether containing 5 mM ␤-mercaptoethanol. Subsequently, the lipid mixture was stirred for 1 h at 20°C and centrifuged at 1000 ϫ g at 20°C. The lipid solution was dried in a rotory evaporator, rehydrated in 20 mM potassium phosphate (pH 7.4) containing 5 mM ␤-mercaptoethanol by vortexing for 30 min, and resuspended by sonication in the same buffer to 20 mg lipid/ml. Liposomes were stored in liquid N 2 .
Reconstitution of Purified MsbA in Proteoliposomes-The purification of amino-terminal His 6 -tagged MsbA was performed according to previously described methods (17,18) with modifications. Inside-out membrane vesicles were solubilized on ice for 1 h in Buffer A (50 mM potassium HEPES (pH 7.4), 10% (v/v) glycerol, 0.5 M NaCl, 5 mM ␤-mercaptoethanol, 5 mM MgSO 4 ) containing 2% (v/v) lauryl dimethylamine N-oxide (Fluka Biochemika, Buchs, Switzerland). The mixture was centrifuged at 125,000 ϫ g for 30 min, after which the solubilized protein in the supernatant was mixed with Ni 2ϩ -nitrilotriacetic acid resin suspension (Qiagen, West Sussex, UK) (10 mg of protein/ml suspension), preequilibrated with Buffer B (Buffer A supplemented with 10 mM imidazole and 0.1% (w/v) n-dodecyl ␤-D-maltoside (Sigma)). After washing with 20 volumes of Buffer B and 40 volumes of Buffer A containing 50 mM imidazole and 0.1% n-dodecyl ␤-D-maltoside, the protein was eluted with 5 volumes of Buffer A containing 200 mM imidazole and 0.1% n-dodecyl ␤-D-maltoside. For reconstitution of purified MsbA, liposomes were thawed slowly at 20°C and extruded by 11 passages through a 400-nm polycarbonate filter to form unilamellar liposomes. The unilamellar liposomes were diluted to 4 mg/ml in 50 mM potassium phosphate (pH 7.4) and destabilized by the stepwise addition of n-dodecyl ␤-D-maltoside to a maximum of 3.0 mM, as described for the functional reconstitution of LmrA (18). Purified protein was added to the detergent-destabilized liposomes at a protein:lipid ratio of 1:100 (w/w) and incubated at 20°C with mild stirring for 30 min. The mixture was added to methanol/ethanol/H 2 O washed Bio-Beads (Bio-Rad) at 80 mg of beads/ml and incubated twice for 2 h at 20°C changing Bio-Beads at each incubation. Subsequently, the mixture was transferred to fresh Bio-Beads and incubated for another 16 h at 4°C. Proteoliposomes were collected by centrifugation at 125,000 ϫ g for 30 min and resuspended in 50 mM potassium phosphate buffer (pH 7.4) to a protein concentration of 1 mg/ml. Proteoliposomes were used immediately for functional assays.
Growth Curves-Growth of E. coli WD2 cells in the absence or presence of ethidium bromide at concentrations up to 1 mM was monitored by measuring the A 660 at 30 and 42°C in accordance with published methods (19).
Ethidium Transport-E. coli WD2 cells were harvested in mid-exponential phase, washed 3 times in 50 mM potassium phosphate (pH 7.0) containing 5 mM MgSO 4 , and deenergized in the presence of 0.5 mM 2,4-dinitrophenol for 30 min at 30°C. Subsequently, cells were washed 3 times and resuspended in phosphate buffer to an A 660 of 0.5. The uptake of 2 M ethidium in cells was measured at 42°C in a 96-well plate using a Spectramax Gemini XS microplate reader (Molecular Devices). The excitation and emission wavelengths were 500 and 580 nm, respectively.
Hoechst 33342 Transport-Proteoliposomes containing purified and functionally reconstituted MsbA (5 g of membrane protein) were resuspended in 2 ml of 50 mM potassium phosphate (pH 7.4) supplemented with 5 mM ␤-mercaptoethanol, 5 mM MgSO 4 , 5 mM phosphocreatine, and 0.1 mg/ml creatine kinase. Hoechst 33342 (Molecular Probes, Leiden, The Netherlands) was then added to a final concentration of 0.5 M. After a steady state level of Hoechst 33342 fluorescence was obtained, 2 mM Mg-ATP or AMP-PNP was added. Measurements were performed in an LS 55B luminescence spectrometer (PerkinElmer Life Sciences) at excitation and emission wavelengths of 355 and 457 nm, respectively, and slit widths of 3 nm each.
ATPase Assay-The ATPase activity was based on a colorimetric ascorbic acid/ammonium molybdate assay to measure the liberation of P i from ATP, as described (19). Inside-out membrane vesicles were used at a final protein concentration of 1 mg/ml in 20 mM potassium HEPES (pH 7.4) containing 5 mM MgSO 4 and 5 mM Na-ATP. The solution was incubated on ice for 15 min with varying concentrations of drugs or Lipid A (Avanti, AL) as indicated in Figs. 5A and 6 and then incubated for 10 min at 30°C. The reaction was stopped by the addition of 0.2 mM ascorbic acid, and the concentration of P i was determined. ATPase activity measurements in the presence of 1 mM ortho-vanadate were obtained in parallel and subtracted from the readings.
Photoaffinity Labeling-Photoaffinity labeling of MsbA and LmrA with [ 3 H]azidopine (Amersham Biosciences) was performed with 90 g of total protein using inside-out membrane vesicles from E. coli DE3 expressing MsbA or L. lactis NZ9000 expressing LmrA in 20 mM potassium HEPES (pH 7.0). The final concentration of [ 3 H]azidopine (51 Ci/mmol) was 0.5 M in a total reaction volume of 50 l. After incubation of inside-out vesicles and [ 3 H]azidopine at 30°C for 10 min, the samples were irradiated on ice at 312 nm for 2.5 min in a Bio-Link BLX photo-reactor (BDH) with 5 UV lamps of 8 watts each. Subsequently, total membrane protein was analyzed by SDS-PAGE and autoradiography.

RESULTS
Complementation Experiments-E. coli WD2 expresses a genome-encoded temperature-sensitive MsbA protein containing an alanine to threonine substitution at position 270 in transmembrane segment 5 (14). This MsbA variant is rapidly inactivated at 42°C. As a result, at the non-permissive temperature the biogenesis of the outer membrane is impaired in the mutant organism due to a deficiency in the MsbA-mediated transport of phospholipids and Lipid A to the outer membrane. In contrast, E. coli WD2 exhibits normal growth at 30°C. To test if LmrA expression could rescue the mutant cells at 42°C, E. coli WD2 was transformed with the empty vector pGK13 or plasmid pGKLmrA, a derivative of pGK13 that contains the lmrA gene downstream of its endogenous lactococcal promotor to enable the expression of LmrA in E. coli (4). As a positive control for complementation E. coli WD2 was also transformed with pACYC (vector) or plasmid pZZ34, a derivative of pACYC carrying wild type msbA (10,14). At 30°C, the growth rates of E. coli WD2 cells expressing MsbA or LmrA were similar to those of control cells (Fig. 1). Surprisingly, at 42°C the growth rates of MsbA-expressing cells and LmrA-expressing cells were significantly higher than those of non-expressing cells. These results suggested that LmrA could functionally replace the temperature-sensitive A270T mutant MsbA protein in E. coli WD2 at the non-permissive temperature.
Resistance to Ethidium Bromide-LmrA is known to confer ethidium resistance on E. coli by mediating the active extrusion of ethidium from the cells (4). In view of the functional complementation of E. coli msbA by lactococcal lmrA, it was interesting to determine whether MsbA expression conferred ethidium resistance on E. coli WD2 cells. The growth rate of E. coli WD2 cells harboring pZZ34 (msbA) or pACYC184 (vector) was measured in liquid cultures containing increasing concentrations of ethidium bromide. At 30°C, no significant differences were observed between the concentrations of ethidium necessary to reduce the growth rates of MsbAexpressing cells and non-expressing control cells by 50% (IC 50 ) (Fig. 2). However, at 42°C MsbA-expressing cells showed a significantly higher IC 50 value compared with control cells. Hence, the expression of MsbA in E. coli conferred ethidium resistance on the cells under these conditions.
To test whether ethidium efflux from the cell was the underlying mechanism of drug resistance in E. coli WD2 expressing MsbA at 42°C, fluorimetric ethidium transport assays were performed (Fig. 3). Suspensions of deenergized E. coli WD2 cells containing pZZ34 (msbA) or pACYC184 (vector) were first allowed to accumulate ethidium. When metabolic energy was then generated in the cells through the addition of glucose, ethidium extrusion was observed in cells expressing MsbA but not in control cells (Fig. 3A). Upon the addition of glucose, a significant ethidium efflux was also observed for E. coli WD2 cells expressing LmrA compared with control cells (Fig 3B).
MsbA Interacts with Multiple Drugs-To further investigate drug-protein interactions in MsbA, the photoaffinity compound [ 3 H]azidopine was used. In previous work, this 1,4-dihydropyridine derivative was shown to be photo-incorporated into LmrA 2 and the human multidrug resistance P-glycoprotein MDR1 (20). Inside-out membrane vesicles were prepared from E. coli DE3 cells harboring pET28b (vector) or plasmid pWTD1, which allowed the expression of His 6 -tagged MsbA under control of an isopropyl-1-thio-␤-D-galactopyranoside-inducible T7 promoter (17). After photo-cross-linking of the membrane vesicles in the presence of [ 3 H]azidopine, a ϳ65-kDa band was detected on the autoradiogram in E. coli membrane vesicles containing MsbA and in lactococcal membrane vesicles containing LmrA but not in E. coli control membrane vesicles (Fig. 4).
The 65-kDa bands were also detected on an immunoblot probed by using an anti-His 6 -tag antibody (data not shown). These data provide evidence for a direct interaction between azido-2 L. Balakrishnan and H. W. van Veen, unpublished data. The specificity of His 6 -tagged MsbA for drugs was also assessed through measurements of the MsbA-associated ATPase activity in inside-out membrane vesicles. In contrast to control membrane vesicles, MsbA-containing membrane vesicles displayed a significant amount of vanadate-sensitive ATPase activity that was stimulated up to 3-fold in the presence of daunomycin (Fig. 5A), a chemotherapeutic drug transported by LmrA and human P-glycoprotein MDR1. The concentration of daunomycin required for half-maximal stimulation (SC 50 ) of the MsbA-associated ATPase activity was about 25 M. Vinblastine and Hoechst 33342, which are substrates for LmrA and Pglycoprotein MDR1, stimulated the MsbA ATPase activity 4and 6-fold, respectively, with SC 50 values of 28 and 49 M, respectively. Interestingly, these drugs stimulated the LmrA ATPase activity to a similar extent as observed for the MsbA ATPase activity but at significantly lower SC 50 values of 5 M for daunomycin, 6 M for vinblastine, and 1.7 M for Hoechst 33342.
In addition to the observation that the MsbA ATPase activity is stimulated by Hoechst 33342, evidence for the interaction between MsbA and Hoechst 33342 was obtained in Hoechst 33342 transport measurements in proteoliposomes prepared from lactococcal lipids containing Ni 2ϩ -nitrilotriacetic acid affinity-purified and functionally reconstituted His 6 -tagged MsbA. The protein was greater than 97% pure as judged by silver-stained SDS-PAGE (data not shown). Hoechst 33342 is only fluorescent when it is present in the phospholipid environment of the membrane and essentially non-fluorescent in the aqueous phase (18,21,22). The addition of Hoechst 33342 to the proteoliposomes resulted in a rapid increase in fluorescence up to a steady state level due to the partitioning of the dye in the membrane. The subsequent addition of Mg-ATP resulted in a rapid quenching of the Hoechst 33342 fluorescence in proteoliposomes containing MsbA (Fig. 5B). The quenching of Hoechst 33342 was not observed in MsbA-containing proteoliposomes in the presence of the non-hydrolyzable ATP analogue AMP-PNP or in empty liposomes in the presence of Mg-ATP or AMP-PNP. These observations point to the MsbA-dependent transport of Hoechst 33342 from the phospholipid bilayer into the aqueous lumen of the membrane vesicles, similar to previous observations for LmrA and Pglycoprotein MDR1 (5,21,22). Taken together these data demonstrate the interaction between MsbA and cytotoxic drugs.
LmrA Interacts with Lipid A-In view of the genetic evidence that LmrA can substitute for MsbA in E. coli WD2 and, hence, appears able to transport Lipid A in E. coli, it was of interest to study the interaction between Lipid A and LmrA through measurements of the LmrA-associated ATPase activity in inside-out membrane vesicles prepared from L. lactis. As shown in Fig. 6, Lipid A stimulated the vanadate-sensitive ATPase activity in LmrA-containing inside-out membrane vesicles 2.1fold with a SC 50 of 3 M. No effect of Lipid A was observed on the vanadate-sensitive ATPase activity in control inside-out membrane vesicles without LmrA. Similar to the experiments on LmrA, Lipid A also stimulated the MsbA-associated ATPase activity in E. coli DE3 membrane vesicles 2-fold with an SC 50 of about 7 M.

DISCUSSION
An increasing number of transporters are appreciated for their ability to transport a variety of lipids in addition to their ability to transport multiple drugs. ABC transporters such as the human multidrug resistance P-glycoprotein MDR1 and multidrug resistance-associated protein MRP1 (ABCC1) are A, effect of daunomycin on the vanadate-sensitive ATPase activity in inside-out membrane vesicles prepared from E. coli DE3 cells containing plasmid pET28b (vector) (f) or pWTD1 (His 6 msbA) (Ⅺ). The data points represent the mean Ϯ S.E. of three independent experiments using different batches of membrane vesicles. B, Hoechst 33342 transport in control liposomes and proteoliposomes containing purified and functionally reconstituted His 6 MsbA. Proteoliposomes were diluted to a protein concentration of 2.5 g/ml in 50 mM potassium phosphate (pH 7.4) containing 5 mM MgSO 4 and an ATP-regenerating system. Upon the addition of 0.5 M Hoechst 33342, the increase in the fluorescence of the dye was followed in time until a steady state was reached. Active transport of Hoechst 33342 was then initiated by the addition of 2 mM Mg-ATP at the arrow. Traces obtained for proteoliposomes in the presence of the non-hydrolyzable ATP analogue AMP-PNP were identical to that displayed for control liposomes in the presence of Mg-ATP. A.U., arbitrary units. able to transport a broad range of short-chain lipid analogues including 2-6-(7-nitro-2,1,3-benzoxadiazol-4-yl) aminocaproyl (C 6 -NBD) phosphatidylcholine, C 6 -NBD glucosylceramide, and C 6 -NBD sphingomyelin (23)(24)(25)(26). Evidence has also been obtained that P-glycoprotein MDR1 mediates the translocation of the native glycosphingolipid precursor glucosylceramide (26). Although human P-glycoprotein MDR3 (ABCB4) is known to transport phosphatidylcholine across the canalicular membrane of the hepatocyte, the protein has also been shown to mediate the transport of multiple P-glycoprotein MDR1 substrates, such as vinblastine, paclitaxel, and digoxin (27). In recent work we have demonstrated that the breast cancer resistance protein (BCRP, also termed ABCG2) is able to transport sterols and steroids in addition to its ability to interact with multiple cationic substrates for LmrA and P-glycoprotein MDR1 (19).
In this work, we have compared the substrate specificities of LmrA and MsbA, which are two bacterial homologues of the human multidrug resistance P-glycoprotein MDR1. LmrA has previously been characterized as a multidrug transporter in Gram-positive L. lactis (4,5), whereas MsbA is known as a transporter for Lipid A in Gram-negative E. coli (10,14). Our observations that (i) LmrA was able to substitute for the temperature-sensitive A270T mutant MsbA protein in E. coli WD2 at non-permissive temperatures and (ii) the LmrA ATPase activity was significantly stimulated by Lipid A suggested that LmrA can transport Lipid A.
Evidence was also obtained for the interaction between MsbA and substrates for LmrA and P-glycoprotein MDR1. First, MsbA could be photoaffinity-labeled with azidopine. Second, the MsbA ATPase activity was significantly stimulated by Hoechst 33342, daunomycin, and vinblastine. Third, MsbA mediated the transport of Hoechst 33342 in proteoliposomes containing purified and functionally reconstituted protein. Finally, the expression of MsbA in E. coli WD2 at 42°C was associated with an enhanced ethidium resistance and ethidium efflux. However, this latter observation may also be explained by the MsbA-associated transport of lipopolysaccharides and phospholipids to the outer membrane. These lipids may increase the integrity and stability of this compartment at elevated temperatures (20) and act as molecular chaperones for the assembly of endogenous drug transporters such as AcrAB-TolC (13,28,29).
Interestingly, we were unable to detect MsbA-mediated Hoechst 33342 transport in inside-out membrane vesicles prepared from E. coli in standard assays, including 2 M Hoechst 33342, developed for the detection of the LmrA-associated transport activity in lactococcal membrane vesicles. Moreover, it was observed that (i) the SC 50 value for the stimulation of the LmrA ATPase activity by Hoechst 33342 in lactococcal membrane vesicles was almost 30-fold lower than that for the MsbA ATPase activity in E. coli membrane vesicles, (ii) the SC 50 values for the stimulation of the LmrA ATPase activity by vinblastine or daunomycin were almost 4.7-and 5-fold lower, respectively, than the corresponding SC 50 values for stimulation of the MsbA ATPase activity, and (iii) purified and functionally reconstituted MsbA mediated the ATP-dependent transport of Hoechst 33342 in proteoliposomes prepared from lactococcal lipids rather than E. coli lipids. Taken together, these observations imply that MsbA exhibits a reduced affinity for drugs compared with LmrA, which may reflect a true difference in the K m for drug transport between these transporters and/or may relate to a competition between drugs and endogenous lipid (such as Lipid A) for transport by MsbA in E. coli membrane vesicles.
The MsbA activity in E. coli has been reported to be associated with the transport of Lipid A and glycerophospholipids from the inner leaflet of the cytoplasmic membrane (the location of biosynthesis) to the outer membrane (10,13,14). Although the molecular mechanism is not known, MsbA may function as a lipid flippase, catalyzing the transbilayer movement of lipids in the cytoplasmic membrane. Alternatively, MsbA may be involved in the transperiplasmic movement of lipopolysaccharides and phospholipids from the outer leaflet of the cytoplasmic membrane to the outer membrane. In the latter model, MsbA may operate in a similar fashion as Lol-CDE, an E. coli ABC transporter that mediates the transport of lipoproteins from the outer leaflet of the cytoplasmic membrane to the outer membrane in conjunction with the periplasmic chaperone LolA (30). Previous work on LmrA and P-glycoprotein demonstrated that these proteins mediate the extrusion of lipophilic compounds from the inner leaflet of the plasma membrane rather than the outer leaflet (21,22,31). The functional substitution of MsbA by LmrA in E. coli WD2 at non-permissive temperatures implies that this mode of transport is also relevant for MsbA and, hence, that MsbA catalyzes the transbilayer movement of Lipid A in the cytoplasmic membrane.
However, pumping from the inner leaflet of the cytoplasmic membrane does not explain how MsbA would mediate the translocation of its lipid substrates to the outer membrane. In this context, the hemolysin A secretion system HlyBD-TolC in E. coli (32,33) represents an interesting example of how membrane proteins, including the MsbA homologue HlyB, and a periplasmic accessory protein act together in an export reaction across the cell envelope. Although MsbA and LmrA are homologous proteins, the subtle differences between their primary structures make precise compulsory protein-protein interactions between the MsbA or LmrA and additional components of the Lipid A translocation machinery in E. coli less likely. Instead, the transport of Lipid A in E. coli could involve lipid transfer proteins, as observed in eukaryotic cells (34,35), which would shuttle Lipid A from the outer leaflet of the inner membrane to the inner leaflet of the outer membrane by diffusion or in a transporter-dependent fashion (e.g. analogous to LolCDE/ LolA (30)). The transport of Lipid A from the cytoplasmic membrane to the outer membrane in E. coli could also involve endogenous multidrug transporters, such as AcrAB-TolC (28,29), which can transport lipophilic substrates from the outer leaflet of the cytoplasmic membrane. Alternatively, MsbA-mediated Lipid A transport from the cytoplasmic membrane to the outer membrane may occur at contact sites between these membranes, known as Bayer junctions (36).
Our insight into the detailed mechanisms by which phospholipids are transported across and between membranes in mi- FIG. 6. Lipid A stimulates the LmrA-associated ATPase activity. The vanadate-sensitive ATPase activity was measured in insideout membrane vesicles prepared from L. lactis harboring plasmid pNZ8048 (vector) (f) or pNZHLmrA (His 6 lmrA) (Ⅺ) in the presence of increasing concentrations of Lipid A. croorganisms and mammalian cells is limited (37). This aspect of the biogenesis of biological membranes has lagged far behind the tremendous advances in other areas, such as understanding solute transport and membrane protein traffic and sorting, made over the last two decades. We conclude that LmrA and MsbA have overlapping substrate specificities, providing another example of the close link that exists between lipid transporters and multidrug transporters. Our conclusion implies the presence of structural elements in MsbA supporting drug-protein interactions and may also point to a physiological role of LmrA in lipid transport in L. lactis.