Portal de Periódicos da CAPES

Acidic Residues in the Lactococcal Multidrug Efflux Pump LmrP Play Critical Roles in Transport of Lipophilic Cationic Compounds

2002; Elsevier BV; Volume: 277; Issue: 29 Linguagem: Inglês

10.1074/jbc.m203141200

1083-351X

Piotr Mazurkiewicz, Wil N. Konings, Gerrit J. Poelarends,

Antimicrobial Resistance in Staphylococcus

The proton motive force-driven efflux pump LmrP confers multidrug resistance on Lactococcus lactis cells by extruding a wide variety of lipophilic cationic compounds from the inner leaflet of the cytoplasmic membrane to the exterior of the cell. LmrP contains one cysteine (Cys270), which was replaced by alanine. This cysteine-less variant was used in a cysteine scanning accessibility approach. All 19 acidic residues in LmrP were replaced one by one by cysteine and subsequently challenged with the large thiol reagent fluorescein maleimide. The labeling pattern strongly indicates that only three acidic residues (Asp142, Glu327, and Glu388) are membrane-embedded. The roles of these residues in drug recognition were evaluated based on transport experiments with two cationic substrates, ethidium and Hoechst 33342, after replacing each of these residues with cysteine, alanine, lysine, glutamate, or aspartate. The obtained results suggest that the negative charges at positions 142 and 327 are not critical for the transport function but are important for drug recognition by LmrP. Surprisingly, the residues Cys142 and Cys327become accessible for fluorescein maleimide upon binding of substrates, indicating a movement of these residues from a nonpolar to a polar environment. Substrate binding apparently results in a conformational change in this region of the protein and a reorientation of a lipid-embedded, hydrophobic substrate-binding site to an aqueous substrate translocation pathway. The proton motive force-driven efflux pump LmrP confers multidrug resistance on Lactococcus lactis cells by extruding a wide variety of lipophilic cationic compounds from the inner leaflet of the cytoplasmic membrane to the exterior of the cell. LmrP contains one cysteine (Cys270), which was replaced by alanine. This cysteine-less variant was used in a cysteine scanning accessibility approach. All 19 acidic residues in LmrP were replaced one by one by cysteine and subsequently challenged with the large thiol reagent fluorescein maleimide. The labeling pattern strongly indicates that only three acidic residues (Asp142, Glu327, and Glu388) are membrane-embedded. The roles of these residues in drug recognition were evaluated based on transport experiments with two cationic substrates, ethidium and Hoechst 33342, after replacing each of these residues with cysteine, alanine, lysine, glutamate, or aspartate. The obtained results suggest that the negative charges at positions 142 and 327 are not critical for the transport function but are important for drug recognition by LmrP. Surprisingly, the residues Cys142 and Cys327become accessible for fluorescein maleimide upon binding of substrates, indicating a movement of these residues from a nonpolar to a polar environment. Substrate binding apparently results in a conformational change in this region of the protein and a reorientation of a lipid-embedded, hydrophobic substrate-binding site to an aqueous substrate translocation pathway. multidrug resistance transmembrane segment The active extrusion of drugs out of cells is one of the major causes of failure of drug-based treatments of cancers and infectious diseases (1Travis J. Science. 1994; 264: 360-362Crossref PubMed Scopus (173) Google Scholar, 2Nikkado H. Curr. Opin. Microbiol. 1998; 1: 516-523Crossref PubMed Scopus (250) Google Scholar). These drug efflux processes are mediated by transport proteins, which can be specific for a given drug or class of drugs or can handle a wide variety of structurally unrelated compounds (3Roberts M.C. FEMS Microbiol. Rev. 1996; 19: 1-24Crossref PubMed Google Scholar, 4Paulsen I.T. Brown M.H. Skurray R.A. Microbiol. Rev. 1996; 60: 575-608Crossref PubMed Google Scholar, 5van Veen H.W. Konings W.N. Semin. Cancer Biol. 1997; 8: 183-191Crossref PubMed Scopus (93) Google Scholar). The latter group of transporters, the so-called multidrug resistance (MDR)1 transporters, can be divided in two major classes: (i) secondary MDR transporters that are driven by a proton or sodium motive force and (ii) ATP-binding cassette MDR transporters, which use the hydrolysis of ATP to fuel transport (6Putman M. van Veen H.W. Konings W.N. Mol. Biol. Rev. 2000; 64: 672-693Crossref PubMed Scopus (642) Google Scholar). Most bacterial MDR transporters known to date are secondary antiport systems that remove drugs from the cell in a coupled exchange with protons, as exemplified by the MDR transporter LmrP ofLactococcus lactis. LmrP is an 408-amino acid integral membrane protein with 12 putative transmembrane segments (TMSs) that belongs to the major facilitator superfamily (see Fig. 1) (7Bolhuis H. Poelarends G. van Veen H.W. Poolman B. Driessen A.J.M. J. Biol. Chem. 1995; 270: 26092-26098Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). The protein can be functionally overexpressed in L. lactis using the tightly regulated, nisin-controlled expression system (8Putman M. van Veen H.W. Poolman B. Konings W.N. Biochemistry. 1999; 38: 1002-1008Crossref PubMed Scopus (58) Google Scholar). The number of compounds recognized and transported by LmrP is remarkably vast and includes many structurally diverse lipophilic cations such as ethidium, Hoechst 33342, and tetraphenylphosphonium (6Putman M. van Veen H.W. Konings W.N. Mol. Biol. Rev. 2000; 64: 672-693Crossref PubMed Scopus (642) Google Scholar, 7Bolhuis H. Poelarends G. van Veen H.W. Poolman B. Driessen A.J.M. J. Biol. Chem. 1995; 270: 26092-26098Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar, 9Bolhuis H. Molenaar D. Poelarends G. van Veen H.W. Poolman B. Driessen A.J.M. Konings W.N. J. Bacteriol. 1994; 176: 6957-6964Crossref PubMed Google Scholar, 10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). A common property of most known substrates of LmrP is their ability to readily intercalate into the lipid bilayer because of their high lipophilicity. This led to the hypothesis that LmrP recognizes its substrates within the membrane and not from the cytoplasm. Transport experiments using the model substrate 1-(4-trimethylammoniumphenyl)-6-phenyl-1, 3,5-hexatriene p-toluenesulfonate indeed revealed a relation between the transport rate and the amount of 1-(4-trimethylammoniumphenyl)-6-phenyl-1, 3,5-hexatrienep-toluenesulfonate associated with the inner leaflet of the lipid bilayer, providing convincing evidence for drug efflux from the membrane to the external medium (10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Transport experiments in intact cells and in inside-out membrane vesicles further demonstrated that extrusion of cationic drugs by LmrP is driven by both the membrane potential (Δψ) and the transmembrane proton gradient (ΔpH), indicating that LmrP mediates electrogenic nH+/drug (n ≥ 2) antiport (7Bolhuis H. Poelarends G. van Veen H.W. Poolman B. Driessen A.J.M. J. Biol. Chem. 1995; 270: 26092-26098Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar, 10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Interestingly, LmrP-mediated Hoechst 33342 transport is competitively inhibited by quinine and verapamil, noncompetitively by nicardipin and vinblastin, and uncompetitively by tetraphenylphosphonium (11Putman M. Koole L.A. van Veen H.W. Konings W.N. Biochemistry. 1999; 38: 13900-13905Crossref PubMed Scopus (88) Google Scholar). These findings are indicative for the presence of multiple drug interaction sites in the multidrug transporter LmrP. However, the molecular basis of multidrug recognition by LmrP is still unclear. To improve our understanding of the molecular mechanism of multidrug transport by LmrP, it is important to identify residues that are critical for multidrug recognition. The excretion by LmrP of lipophilic cationic compounds predicts the involvement of negatively charged amino acid residues in transmembrane segments of LmrP in substrate binding. In this study, a scanning cysteine accessibility approach (12Dunten R.L. Sahin-Toth M. Kaback H.R. Biochemistry. 1993; 32: 12644-12650Crossref PubMed Scopus (60) Google Scholar, 13Kaback H.R. Sahin-Toth M. Weinglass A.B. Nat. Rev. Cell Biol. 2001; 2: 610-620Crossref PubMed Scopus (251) Google Scholar) was used to demonstrate that three residues (Asp142, Glu327, and Glu388) of 19 acidic residues in LmrP are located within the membrane domain. The roles of these membrane-embedded acidic residues in multidrug recognition were analyzed. Residues Asp142, Glu327, and Glu388 were systematically mutagenized, and the transport activities of the mutant proteins were measured as the excretion of the cationic drugs ethidium and Hoechst 33342. The results show that the negative charge at position 388 is important neither for active transport nor for multidrug recognition. In contrast, the negative charges at positions 142 and 327 are critical for the multidrug transport character of LmrP. These residues are located in a part of the transporter that becomes solvent-accessible upon interaction with substrates. L. lactis NZ9000 (ΔlmrA), which lacks the gene encoding the ATP-binding cassette-type MDR transporter LmrA (a kind gift from O. Gajic and J. Kok, Department of Genetics, University of Groningen), was used in combination with the nisin-controlled expression system (14de Ruyter P.G.G.A. Kuipers O.P. de Vos W.M. Appl. Environ. Microbiol. 1996; 62: 3662-3667Crossref PubMed Google Scholar, 15Kuipers O.P. de Ruyter P.G.G.A. Kleerebezem M. de Vos W.M. J. Bacteriol. 1998; 64: 15-21Google Scholar) for overexpression of LmrP and the mutant proteins. L. lactis cells were grown at 30 °C in M17 medium (Difco) supplemented with 0.5% (w/v) glucose and 5 μg/ml chloramphenicol when appropriate. Expression of LmrP variants from pNZ8048-derived plasmids was induced by adding ∼10 ng of nisin A/ml at anA 660 of about 0.6, and the cells were harvested 90 min after induction. General procedures for cloning and DNA manipulation were performed essentially as described by Sambrook et al. (16Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). The PCR overlap extension method (17Higuchi R. Krummel B. Saiki R.K. Nucleic Acids Res. 1988; 16: 7351-7367Crossref PubMed Scopus (2102) Google Scholar) was used to introduce mutations in the lmrP gene on the pHLP5 expression plasmid (8Putman M. van Veen H.W. Poolman B. Konings W.N. Biochemistry. 1999; 38: 1002-1008Crossref PubMed Scopus (58) Google Scholar), which encodes LmrP with an C-terminal His6 tag. All PCR-amplified DNA fragments were sequenced to verify that only the intended changes were introduced. DNA sequencing was performed at the BioMedical Technology Centre (University of Groningen, Groningen, The Netherlands). Membrane vesicles with an inside-out orientation were prepared from L. lactis NZ9000 (ΔlmrA) cells expressing LmrP variants by the French press procedure as described by Putman et al. (8Putman M. van Veen H.W. Poolman B. Konings W.N. Biochemistry. 1999; 38: 1002-1008Crossref PubMed Scopus (58) Google Scholar). The vesicles were frozen in liquid nitrogen and stored at −80 °C at a protein concentration of 15–35 mg/ml in 50 mm potassium phosphate, pH 7.0, containing 10% (w/v) glycerol. Expression levels of His6-tagged proteins were determined by running membrane vesicles (20 μg of total protein) on SDS-polyacrylamide gels, followed by transfer to polyvinylidene difluoride membranes and detection with monoclonal antibodies directed against the His6 tag (Dianova, Hamburg, Germany). The antibodies were visualized by using a Western light chemiluminescence detection kit (Tropix, Bedford, MA) and a Lumi-imager F1 (Roche Molecular Biochemicals). For ethidium transport experiments, whole cells expressing LmrP variants were washed with 50 mmpotassium phosphate, pH 7.0, containing 5 mmMgSO4 and resuspended to an A 660 of 0.5 in the same buffer. The cells (2 ml) were energized by the addition of 25 mm of glucose. After 2 min, 10 μmethidium was added to the cell suspension, and the accumulation of ethidium was measured indirectly by following the fluorescence of the ethidium-polynucleotide complex in the cells. Fluorescence was monitored with a PerkinElmer LS 50B fluorometer using excitation and emission wavelengths of 500 and 580 nm, respectively. Transport of Hoechst 33342 (Molecular Probes Inc.) in inside-out membrane vesicles was measured as described by Putman et al.(8Putman M. van Veen H.W. Poolman B. Konings W.N. Biochemistry. 1999; 38: 1002-1008Crossref PubMed Scopus (58) Google Scholar). Inside-out membrane vesicles (0.5 mg of protein/ml) were resuspended in 50 mm potassium Hepes, pH 7.0, containing 2 mm MgSO4, 8.5 mm NaCl, 0.1 mg/ml creatine kinase, and 5 mm phosphocreatine. After incubation for 30 s at 30 °C, 1 μm of Hoechst 33342 was added. Upon the addition of 0.5 mm Mg-ATP, LmrP was activated through the generation of a proton motive force by F0F1 H+-ATPase. The amount of membrane-associated Hoechst 33342 was measured fluorimetrically (PerkinElmer LS 50B) using excitation and emission wavelengths of 355 and 457 nm, respectively, and slit widths of 5 nm each. Membrane vesicles (2 mg of protein in 300 μl of 50 mm potassium phosphate, pH 7.0) were incubated in the presence of 50 μm fluorescein maleimide for 5 min at 30 °C. The reaction was stopped with 5 mmdithiothreitol. The vesicles were solubilized with 1% dodecyl-β-d-maltoside for 30 min on ice. Subsequently, debris was removed by centrifugation. To purify His6-tagged LmrP variants, supernatant containing solubilized proteins was mixed with 20 μl of Ni2+-nitrilotriacetic acid-agarose, which was equilibrated with buffer A (50 mm potassium phosphate, 20 mm imidazole, pH 8.0, 100 mm NaCl, and 0.05% dodecyl-β-maltoside) and gently shaken for 1 h at 4 °C. Resin was settled down by pulse centrifugation in a table centrifuge, supernatant was removed by aspiration, and resin was washed with 1 ml of buffer A. The protein was eluted with 60 μl of buffer B (50 mm potassium phosphate, 250 mm imidazole, pH 7.0, 100 mm NaCl, and 0.05% dodecyl-β-maltoside). The protein concentration in the eluted fraction was estimated with the RC-DC protein assay kit from Bio-Rad. Similar amounts of protein (2 μg) were subjected to SDS-PAGE. Proteins labeled with fluorescein maleimide were detected by UV excitation and were visualized with a Lumi-imager. To test the effect of substrates on labeling of mutant proteins by fluorescein maleimide, the membrane vesicles were incubated with different concentrations of ethidium and Hoechst 33342 for 5 min at 30 °C prior to the addition of the fluorophore. LmrP contains one native cysteine residue, Cys270, that is located in putative transmembrane segment VIII (Fig. 1). A cysteine-less mutant of LmrP was constructed by replacing the native cysteine with alanine (C270A) using site-directed mutagenesis. The mutant could be expressed in L. lactis and was present in the cytoplasmic membrane as judged from Western blots using anti-His6 tag antibodies (data not shown). In previous studies, the positively charged compounds ethidium and Hoechst 33342 proved to be useful drugs for studying the transport activity of LmrP in whole cells and inside-out membrane vesicles, respectively (8Putman M. van Veen H.W. Poolman B. Konings W.N. Biochemistry. 1999; 38: 1002-1008Crossref PubMed Scopus (58) Google Scholar). Ethidium transport in whole cells can be measured indirectly by monitoring the fluorescence of the intracellular ethidium-polynucleotide complex. Ethidium accumulation by LmrA-deficient, glucose-energized, nonexpressing control cells and cells expressing wild-type LmrP demonstrated that wild-type LmrP can reduce the accumulation of ethidium in the cell by excreting this cationic compound (Fig. 2 A). Also the C270A variant of LmrP is capable of limiting the accumulation of ethidium to lower levels than the control cells, indicating that the C270A mutant retained significant transport activity (Fig.2 A). Hoechst 33342 is highly fluorescent when it is present in the lipid environment of membranes but is essentially nonfluorescent in the aqueous phase. The transport of Hoechst 33342 from the membrane into the aqueous phase can therefore be followed in membrane vesicles by monitoring the Hoechst 33342 fluorescence over time. Upon the addition of ATP, which results in the formation of a proton motive force through proton pumping by the F1F0H+-ATPase, Hoechst 33342 was transported out of the membrane of inside-out membrane vesicles of L. lactis by LmrP as indicated by the decrease in Hoechst 33342 fluorescence (Fig.2 B). The Hoechst 33342 transport measurements further demonstrated that the C270A mutant retained significant transport activity (Fig. 2 B). In conclusion, the C270A mutant (referred to as “Cys-less LmrP”) is still capable of transporting multiple drugs and therefore provides a good starting point for generating mutant proteins with single cysteine replacements. Previous studies revealed that LmrP recognizes its lipophilic, cationic substrates within the membrane and not from the aqueous phase (10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Thus, it seems likely that cationic drug-LmrP interaction involves electrostatic interactions with membrane-embedded, negatively charged residues. The distribution of the 19 acidic residues within the predicted topology model of LmrP is shown in Fig. 1. This topology model still awaits experimental confirmation, and therefore it is not known precisely which acidic residues are membrane-embedded. A cysteine scanning accessibility approach was used to identify the acidic residues located in the membrane domain of LmrP. For this, the codon of each acidic residue was mutated to that of cysteine. Inside-out membrane vesicles derived from cells expressing the single cysteine mutants were analyzed by SDS-PAGE and Western blotting using anti-His6 tag antibodies (data not shown). Of 19 mutants, 17 were expressed to approximately the same level as wild-type LmrP, suggesting that no major structural perturbation occurred upon substitution of any of these residues. Only one mutant, E166C, was not expressed, whereas mutant D340C was found at a low level in the membrane of L. lactis (data not shown). Hoechst 33342 transport measurements demonstrated that the single cysteine mutants can be divided into three groups based on their level of transport activity: (i) mutants that had essentially Cys-less LmrP transport activity: E133C, D198C, and E388C; (ii) mutants that had moderately reduced Hoechst 33342 transport activity: E3C, D8C, E199C, E202C, E255C, D286C, and E345C; and (iii) mutants that had strongly reduced Hoechst 33342 transport activity: D68C, D128C, D142C, E189C, D214C, D235C, and E327C (Table I).Table ITransport activities of cysteine mutants of LmrP for Hoechst 33342 in inside-out membrane vesiclesMutantInitial rate %E3C38D8C31D68C5D128C8E133C93D142C6E189C3D198C79E199C40E202C38D214C14D235C2E255C24D286C36E327C14E345C32E388C89For clarity of presentation activities are presented as initial rates of transport relative to the Cys-less LmrP (100%). The initial rates were determined over the first 17 s after addition of 0.5 mm Mg2+-ATP. Open table in a new tab For clarity of presentation activities are presented as initial rates of transport relative to the Cys-less LmrP (100%). The initial rates were determined over the first 17 s after addition of 0.5 mm Mg2+-ATP. Also, the accumulation of ethidium was measured in cells expressing single cysteine mutants. Essentially Cys-less LmrP transport activity was found for mutants E3C, D8C, E133C, E189C, D198C, E199C, E202C, D214C, E255C, D286C, E345C, and E388C (data not shown). The D68C, D128C, D235C, and E327C mutants had all lost the ability to transport ethidium from the cell (Fig. 3). These four mutants have not only lost the transport activity but even seem to facilitate the influx of ethidium. Only one mutant, D142C, was more active then the Cys-less transporter. Evidence has been provided that cysteines in putative TMSs are inaccessible to the large, membrane-permeable thiol reagent fluorescein maleimide (18Meuller J. Rydström J. J. Biol. Chem. 1999; 274: 19072-19080Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar, 19Slotboom D.J. Konings W.N. Lolkema J.S. J. Biol. Chem. 2001; 276: 10775-10781Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar). In contrast, cysteine residues that are modified by fluorescein maleimide must be exposed to the aqueous environment. Thus, fluorescein maleimide can be used to identify membrane-embedded residues in a membrane protein. Therefore, each single cysteine mutant was challenged with fluorescein maleimide in inside-out membrane vesicles. Subsequently, the vesicles were solubilized, and His6-tagged LmrP mutants were purified and subjected to SDS-PAGE. Fluorescence of cysteine mutants labeled with fluorescein maleimide was detected by UV excitation. The cysteine-less mutant C270A, used as a negative control, was not modified by fluorescein maleimide (Fig.4). The single cysteine mutants E3C, D8C, D68C, D128C, E133C, E189C, D198C, E199C, E202C, D214C, D235C, E255C, D286C, and E345C were all labeled by fluorescein maleimide. The accessibility of these residues for fluorescein maleimide indicates a location in the protein structure that is well exposed to the aqueous environment and is consistent with a position in a loop structure. In contrast, the single cysteine mutants D142C, E327C, and E388C could not be labeled by fluorescein maleimide. Wild-type LmrP, harboring one native cysteine residue (Cys270), was also not labeled by fluorescein maleimide. These observations indicate that these four cysteine residues are either buried in the protein structure or exposed to the lipid environment. Thus, of the 19 negatively charged residues in LmrP, three acidic residues are located within the membrane domain of the transporter. The fluorescein maleimide labeling pattern of the single cysteine mutants is consistent with the previously proposed model for the membrane topology of LmrP (Fig. 1) (7Bolhuis H. Poelarends G. van Veen H.W. Poolman B. Driessen A.J.M. J. Biol. Chem. 1995; 270: 26092-26098Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). The acidic residues at positions 142 (TMS V), 327 (TMS X), and 388 (TMS XII) thus appear to be located in the membrane domain of LmrP. To investigate in more detail whether the negative charges are important for drug transport activity of LmrP, each of these residues was replaced by other residues with (i) a conserved negative charge (mutants D142E, E327D, and E388D), (ii) a positive charge (mutants D142K, E327K, and E388K), or (iii) by a neutral residue (mutants D142A, E327A, and E388A). Inside-out membrane vesicles were prepared from cells expressing these mutants, and the amounts of mutant protein were determined by Western blotting using anti-His6 tag antibodies. All of the mutant proteins were expressed to levels similar to that of wild-type LmrP (data not shown). The functional properties of these mutants were studied by transport assays using the cationic substrates ethidium and Hoechst 33342. The acidic residues located in the loop regions of LmrP were not studied further, because these residues are probably not directly involved in drug binding. Replacing Asp142, Glu327, and Glu388 with the basic residue lysine almost completely abolished transport activity for both cationic substrates (Figs.5 and6). However, when the negative charge at these positions was preserved (mutants D142E, E327D, and E388D), transport of ethidium and Hoechst 33342 was not affected. Interestingly, mutants E388A (Figs. 5 and 6) and E388C (Table I and data not shown) retained Hoechst 33342 and ethidium transport activities comparable with those of the wild type. These results demonstrate that a negative charge at position 388 is not essential for transport nor for multidrug recognition. In contrast, the replacement of Asp142 and Glu327 with alanine or cysteine resulted in almost complete loss of Hoechst 33342 transport activity (Table I and Fig. 6), indicating that a negative charge at these positions is essential for transport of this divalent cationic drug. Surprisingly, mutants D142A and E327A retained significant transport activity for the monovalent cation ethidium (Fig. 5), whereas mutant D142C displayed an even higher ethidium transport activity then the Cys-less transporter (Fig. 3). Mutant E327C had lost the ability to transport ethidium from the cell (Fig. 3). The negative charges at positions 142 and 327 therefore appear to be essential for active transport of divalent negatively charged Hoechst 33342 but not for the monovalent ethidium. Thus, these residues dictate the multidrug transport character of LmrP.Figure 6Hoechst 33342 transport measurements with inside-out membrane vesicles prepared from cells expressing LmrP variants harboring mutations at positions of membrane-embedded negatively charged residues. Hoechst 33342 (1 μm) was added, and its binding to the membrane vesicles (0.5 mg of total protein/ml) was followed in time until a steady state was reached. Hoechst 33342 transport was initiated by the addition of 0.5 mm Mg-ATP. WT, wild type.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Because mutant D142C displayed a high ethidium transport activity, whereas mutant E327C lost the ability to transport ethidium, we tested whether the double cysteine mutant D142C/E327C could transport ethidium. This mutant was constructed and could be expressed in L. lactis as judged from Western blots using anti-His6 tag antibodies (Fig.7 A). Cells expressing the double cysteine mutant displayed a steady state ethidium accumulation level similar to that of nonexpressing control cells (Fig.7 B). As expected, this mutant protein also did not exhibit significant Hoechst 33342 transport activity (Fig. 7 C). Taken together, these results provide further support for the finding that the negative charges at positions 142 and 327 in LmrP play critical roles in transport of lipophilic cationic compounds. Our results indicated that the cysteines introduced at positions 142 and 327 are not accessible to fluorescein maleimide and thus are likely to be located in the membrane-embedded domain of LmrP. Interestingly, the solvent accessibility of these cysteines was found to be increased upon binding of substrate by LmrP (Fig. 8). For mutant D142C, the efficiency of labeling by fluorescein maleimide was strongly increased upon preincubation with Hoechst 33342. Surprisingly, preincubation with ethidium did not result in an increased accessibility to fluorescein maleimide. The opposite behavior was found for mutant E327C (Fig. 8). Preincubation with Hoechst 33342 did not result in an increased labeling efficiency, but exposure to ethidium resulted in an increased accessibility of the cysteine at position 327 to fluorescein maleimide. An increase in fluorescein maleimide labeling efficiency upon incubation with substrates was not observed for the wild-type protein and mutant E388C, both of which contain a membrane-embedded cysteine, nor for mutant D198C, which contains a solvent-accessible cysteine (Fig. 8). To our surprise, the double cysteine mutant, which is transport inactive, displayed an increased fluorescein maleimide labeling efficiency in the presence of both ethidium (1 mm) and Hoechst 33342 (0.2 mm) (Fig. 8), indicating that this mutant protein can still interact with substrates. However, because high substrate concentrations were used in these experiments, we repeated the labeling experiments in the presence of low concentrations of drugs. As shown in Fig. 9, efficient labeling of D142C already occurs at 10 μm of Hoechst 33342, whereas 200 μm of Hoechst 33342 is necessary to obtain efficient labeling of mutant D142C/E327C. This suggests that mutant D142C has a higher affinity for Hoechst 33342 than the double cysteine mutant. Labeling of mutants E327C and D142C/E327C takes place only at ethidium concentrations higher than 50 μm, which suggests that both mutants have low affinity for this substrate. Transport of cationic substrates Hoechst 33342 and ethidium was studied in inside-out membrane vesicles and cells of L. lactis expressing wild type and mutants of LmrP. Mutants of LmrP were obtained by systematic replacement of the acidic residues in a cysteine-less mutant C270A by cysteine. In these experiments the residues Asp68, Asp128, Asp235, and Glu327 (summary in Table II) were identified to be critical for transport function of LmrP. Cysteine scanning accessibility experiments revealed that three acidic residues (Fig. 4) Asp142, Glu327, and Glu388are located in the membrane-embedded domain of LmrP. These predictions are consistent with the presented membrane topology of LmrP (Fig.1).Table IISummary of transport activities of cysteine mutants of LmrP for ethidium and Hoechst 33342MutantTransport activityEthidiumHoechst 33342Wild type++++++C270A+++E3C++D8C++D68C−i−D128C−i−E133C+++D142C++−E189C+−D198C+++E199C++E202C++D214C+−D235C−i−E255C++D286C++E327C−i−E345C++E388C+++D142C/E327C−−+++, wild-type LmrP transport activity; −, lack of transport activity, as observed for LmrP nonexpressing cells and membrane vesicles; i, mutants the expression of which results in increased influx of ethidium into cells under energized conditions. Open table in a new tab +++, wild-type LmrP transport activity; −, lack of transport activity, as observed for LmrP nonexpressing cells and membrane vesicles; i, mutants the expression of which results in increased influx of ethidium into cells under energized conditions. Because the lactococcal multidrug transporter LmrP recognizes its lipophilic cationic substrates within the membrane (10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar) rather than from the aqueous phase, the residues critical for multidrug recognition must be located in the membrane-embedded domain of the protein. Therefore, we have focused on the drug recognition properties of LmrP by analyzing the consequences of substitutions of the negatively charged residues at positions 142, 327, and 388 (TableIII). Membrane-embedded acidic residues in multidrug transporters (4Paulsen I.T. Brown M.H. Skurray R.A. Microbiol. Rev. 1996; 60: 575-608Crossref PubMed Google Scholar, 20Edgar R. Bibi E. EMBO J. 1999; 18: 822-832Crossref PubMed Scopus (125) Google Scholar, 21Yerushalmi H. Schuldiner S. J. Biol. Chem. 2000; 275: 5264-5269Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar) and buried negatively charged residues in multidrug-binding proteins (22Zheleznova E.E. Markham P.N. Neyfakh A.A. Brennan R.G. Cell. 1999; 96: 353-362Abstract Full Text Full Text PDF PubMed Scopus (168) Google Scholar, 23Schumacher M.A. Miller M.C. Grkovic S. Brown M.H. Skurray R.A. Brennan R.G. Science. 2001; 294: 2158-2163Crossref PubMed Scopus (327) Google Scholar) have been found to be important for transport and/or binding of lipophilic cations. Introduction of neutral or negatively charged residues at position 388 in LmrP had no influence on the transport of the lipophilic cations ethidium and Hoechst 33342. Only when a positively charged residue was introduced, was the transport activity abolished. In contrast, mutations at positions 142 and 327 had drastic effects on drug transport. The most important outcome of our experiments is that the negative charges at positions 142 and 327 are essential only for the transport of the divalent cation Hoechst 33342. In principle, this could reflect the possibility that these residues play a critical role in energy utilization and coupling or that their replacement may have caused detrimental structural modifications. These possibilities seem unlikely, however, because mutants 142A/C and 327A transport the monovalent cation ethidium quite efficiently. Hence, the mutations at positions 142 and 327 alter the drug specificity pattern by exerting a differential effect on the transport of monovalent and divalent cations.Table IIITransport activities of LmrP mutants of membrane embedded negatively charged residuesMutantTransport activityEthidiumHoechst 33342Wild type++++++D142E++++++D142A+−D142K−−E327D++++++E327A++−E327K+−E388D++++++E388A++++++E388K−−+++, WT LmrP transport activity; −, lack of transport activity, as observed for LmrP nonexpressing cells and membrane vesicles. Open table in a new tab +++, WT LmrP transport activity; −, lack of transport activity, as observed for LmrP nonexpressing cells and membrane vesicles. The studies of Putman et al. (11Putman M. Koole L.A. van Veen H.W. Konings W.N. Biochemistry. 1999; 38: 13900-13905Crossref PubMed Scopus (88) Google Scholar) indicated the presence of multiple substrate-binding sites in LmrP. Because of overlapping fluorescence spectra, it was unfortunately not possible to determine whether ethidium inhibits Hoechst 33342 transport competitively, noncompetitively, or uncompetitively. The possibility that Hoechst 33342 binds to another binding site than ethidium has to be considered. The results of labeling experiments of cysteine mutants with fluorescein maleimide in the presence of substrates indeed suggest such a possibility. The cysteine at position 142 was labeled by fluorescein maleimide in the presence of Hoechst 33342 but not in the presence of ethidium. The high transport activity of mutant D142C for ethidium demonstrates that this mutant must interact with this monovalent cation. The opposite phenomenon was observed for mutant E327C, which was labeled in the presence of ethidium but not in the presence of Hoechst 33342. Surprisingly, increased labeling by fluorescein maleimide showed that the double cysteine mutant D142C/E327C interacts both with ethidium and Hoechst 33342. These observations, together with the substrate concentration dependence of labeling, demonstrate that residues Asp142 and Glu327 are not essential for binding of these cationic drugs but increase the affinity of LmrP for substrates. Probably other residues in the binding pocket,i.e. those containing aliphatic side chains or aromatic rings, supply hydrophobic interactions that are sufficient to allow substrate binding with low affinity. The lack of protection of cysteine residues by substrates against labeling by fluorescein maleimide may suggest that these residues are in a location distant to the bound substrate. On the other hand, the possibility that each of the two substrates protects only one of the two cysteines (ethidium, D142C; Hoechst 33342, E327C) against labeling by fluorescein maleimide cannot be excluded. Paulsen et al. (4Paulsen I.T. Brown M.H. Skurray R.A. Microbiol. Rev. 1996; 60: 575-608Crossref PubMed Google Scholar) described a very similar charge-related phenomenon with the multidrug resistance proteins QacA and QacB fromStaphylococcus aureus. These closely related multidrug efflux pumps both confer resistance to various organic cations. However, QacB mediates significantly lower levels of resistance to divalent cations than QacA. The authors observed that the presence of an acidic residue at position 323 within putative TMS X of the QacA protein plays a critical role in conveying high level resistance to divalent organic cations. Additionally, it was demonstrated that QacB mutants containing an acidic residue at position 322 were also able to confer resistance to divalent cations. Hence, it was proposed that the region of the QacA/B protein containing the essential acidic residue plays a role in recognition or binding of divalent cations. The importance of negatively charged residues for recognition of cationic drugs was clearly demonstrated in a recent structural study of the S. aureus multidrug-binding protein QacR (23Schumacher M.A. Miller M.C. Grkovic S. Brown M.H. Skurray R.A. Brennan R.G. Science. 2001; 294: 2158-2163Crossref PubMed Scopus (327) Google Scholar). This regulatory protein represses transcription of the qacA gene and is induced by structurally diverse cationic lipophilic drugs. Schumacher et al. (23Schumacher M.A. Miller M.C. Grkovic S. Brown M.H. Skurray R.A. Brennan R.G. Science. 2001; 294: 2158-2163Crossref PubMed Scopus (327) Google Scholar) reported the crystal structures of six QacR-drug complexes. The combined structures revealed that QacR contains several separate but linked binding sites within one extended drug-binding pocket, which contains four negatively charged residues (e.g. glutamates) that neutralize the positive charges on the structurally distinct drug molecules. The results of our studies suggest that the acidic residues at positions 142 and 327 in LmrP are involved in a drug-binding pocket, which may consist of multiple drug interaction sites to account for the multidrug transport character of LmrP. The observation that introduced cysteines at positions 142 and 327 could not be labeled by fluorescein maleimide suggests a hydrophobic nature for this pocket. An important outcome of the fluorescein maleimide labeling studies is the observation that upon interaction with substrates, the membrane-embedded residues Cys142 and Cys327 move from a solvent-inaccessible to a solvent-exposed environment. This movement may be related to a translocation of these residues from a hydrophobic substrate-binding site to a hydrophilic substrate translocation pathway, which is in agreement with the for LmrP proposed “hydrophobic vacuum cleaner” model of drug transport (10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Whereas the presence of one negative charge in the multidrug-binding pocket may be sufficient for the recognition of the monovalent cation ethidium by LmrP, the recognition of the divalent cation Hoechst 33342 clearly depends on the presence of both negative charges. Possibly, each of these negative charges interacts directly with one of the positively charged moieties of the divalent cation. Interestingly QacA seems to contain several acidic residues within its putative membrane-embedded domain (4Paulsen I.T. Brown M.H. Skurray R.A. Microbiol. Rev. 1996; 60: 575-608Crossref PubMed Google Scholar). It would be interesting to investigate whether more than one negatively charged residue plays a role in the recognition of divalent cations. Ethidium accumulation in glucose-energized L. lactis cells expressing the LmrP mutants D68C, D128C, D235C, and E327C was faster then in control cells (Fig. 3), suggesting that these mutant proteins facilitate ethidium influx into the cell. Because all of these mutants appear to be (almost completely) inactive with regard to extrusion of ethidium and Hoechst 33342, it is possible that these residues play important roles in active transport and coupling. Cysteine scanning accessibility experiments revealed that residues Asp68, Asp128, and Asp235 are located in loop regions (Fig. 1), which makes a direct interaction of these residues with substrates very unlikely. Another possible explanation could be that the conformation of LmrP is affected by these mutations, resulting in a change of gating mechanism, allowing the influx of ethidium through the translocation pathway. Interestingly, residue Asp68 is part of a conserved motif, GXXXDRXGRK, located in the cytoplasmic loop between TMSs II and III of the major facilitator superfamily transporters (24Paulsen I.T. Skurray R.A. Gene (Amst.). 1993; 124: 1-11Crossref PubMed Scopus (105) Google Scholar). On the basis of mutagenesis of the TetA(B) tetracycline transporter, it has been proposed that this motif may be involved in the reversible conformational change required for opening and/or closing of the substrate translocation pathway (25Yamagutchi A. Ono N. Akasaka T. Noumi T. Sawai T. J. Biol. Chem. 1990; 265: 15525-15530Abstract Full Text PDF PubMed Google Scholar). Furthermore, Someya et al. (26Someya Y. Kimura-Someya T. Yamaguchi A. J. Biol. Chem. 2000; 275: 210-214Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar) indicated that Arg70 in TetA(B), which is also located in this conserved motif, requires the negative charge of Asp120, corresponding to Asp128 in LmrP, for proper positioning of transmembrane segments in the membrane. Taken together, these observations suggest that residues Asp68, Asp128, Asp235, and possibly Glu327(also implicated in substrate binding) might be part of the structural framework responsible for gating the translocation pathway of LmrP. Further mutagenesis studies are required to investigate this possibility. In summary, the results of this study showed that a number of acidic residues are important for LmrP-mediated transport of lipophilic cationic compounds. The residues Asp68, Asp128, and Asp235 are located in putative loop regions of the protein and are probably not involved in direct interactions with drugs. Further studies must be done to determine the exact role of these residues in the function of LmrP. The other critical residues Asp142 and Glu327 are located at the intracellular side of the TMSs V and X, respectively, consistent with the fact that LmrP extrudes drugs from the inner leaflet of the lipid bilayer (10Bolhuis H. van Veen H.W. Brands J.R. Putman M. Poolman B. Driessen A.J.M. Konings W.N. J. Biol. Chem. 1996; 271: 24123-24128Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Based on the kinetics of transport-competition experiments, Putman et al. (11Putman M. Koole L.A. van Veen H.W. Konings W.N. Biochemistry. 1999; 38: 13900-13905Crossref PubMed Scopus (88) Google Scholar) proposed the presence of at least two drug interaction sites in LmrP, which may represent distinct drug-binding sites on the protein or may represent drug-binding regions within a common hydrophobic binding pocket. We favor the hypothesis that the acidic residues Asp142 and Glu327 are part of a common hydrophobic multidrug-binding pocket, in which both charges are essential for the recognition site of divalent cations such as Hoechst 33342.