细菌多药外排系统及其研究方法的进展

【摘要】  位于细菌细胞膜上的多药外排泵可将多种结构无关药物排出，本文综述了细菌主要外排泵包括初级和次级转运系统的分子生物学特征及外排蛋白结构与药物识别和转运上的关系；同时了对外排蛋白进行结构与功能关系研究的主要的方法，包括定点诱变、螺旋空间排列和结晶等。

【关键词】  细菌 外排泵 多药耐药

    Introduction

    The problem of drug resistance is a main obstacle in the struggle between human being and infectious diseases. One of the most important bacterial resistant mechanisms is the multidrug efflux pumps situated on the membranes, which can be classified into two major types based on the energy source: primary transport system, such as ATP?binding cassette (ABC) transporter superfamily utilizes the energy produced by ATP hydrolysis to efflux diverse compounds; while secondary transport system pumps different substrates via the proton motive force (PMF) providing by transmembrane electrochemical proton gradient (ΔμH+) or the sodium motive force supplying by ion gradients (ΔμNa+). Despite the totally different evolutionary trait and energy sources, both can transport a broad spectrum of structurally unrelated substrates.    To date, due to the highly hydrophobic nature, the crystallization of the structure on most of the membrane transporters is yet to be cracked down, albeit recently, a few have been partially crystallized. Therefore, the most important analysis approach is still protein engineering methods involving site?directed mutagenesis and chemical modification. And currently there have developed some improvements on these methods.

    Bacterial multidrug efflux pump systems

    ATP?binding cassette (ABC) transporter

    To date, LmrA is the only well?characterised bacterial?origin ABC transporter, which is encoded by chromosomally located lmrA gene in Lactococcus lactis. In general, active ABC proteins demand minimal two transmembrane domains (TMDs) and two ABC units (nucleotide?binding domain, NBD). The former usually consists of six transmembrane segments (α?helices); the latter, a 200～250 amino acids, contains three conserved motifs Walker A, ABC signature and Walker B. Therefore, functional structural module is (TMD?NBD)2. Comparing with functional ABC transporter such as human multidrug resistance P?glycoprotein (P?gp), LmrA has only two domains, but homologous to each of the two halves of the P?gp, suggesting that it belongs to "half ABC transporter" and functions as a homodimer, which has been identified by various studies［1～3］.

    Although the wide?substrate transport mechanism of LmrA still remains unclear, two hypotheses have been proposed: hydrophobic vacuum cleaner model shows that toxic hydrophobic compounds are directly extruded from the inner leaflet of the membrane into the external water phase［4］; and the flippase model suggests that LmrA first recognizes substrates in the inner leaflet of the membrane then flip them to the outer leaflet from where they diffuse into the exterior. Possibly, substrate?binding resulting in conformational changes, which modulate the interaction between NBDs and ATP, triggers the transport. The binding and hydrolysis of ATP by one site lead to the movement of drug?binding site from inner membrane to the outside due to the prevention of ATP hydrolysis at the other site, so the drugs can be translocated between the high? and low?affinity binding sites, which forms a transport cycle［5］. So far, the confirmed substrates of LmrA include anticancer drugs (vinca alkaloids and anthracyclines), DNA intercalators, toxic peptides, fluorescent membrane probes and dyes as well as clinically important antibiotics such as aminoglycosides, lincosamides, macrolides, quinolones, streptogramins and tetracyclines［6］. Hence, the unusually broad antibiotic profile of LmrA and possible gene transfer of lmrA to other bacteria enable the non?pathogenic L.lactis potentially threatened［7］.

    Some other bacterial ABC transporters are also identified, such as the DrrAB doxorubicin/daunorubicin transporter of Streptomyces prucetius; the MsbA and BtuCD transporters of Escherichia coli, which efflux lipid A and vitamin B12, respectively. Albeit it seems that these pumps have little connection with bacterial resistance, the determination of the structure of MsbA to 4.5?［8］ and BtuCD to 3.2? deepens our understanding on multidrug transport mechanism.

    Secondary transport system

    On the basis of size and primary energy as well as secondary structure, the system is divided into four families: the major facilitator superfamily (MFS), the small multidrug (SMR) family, the resistance?nodulation?cell division (RND) family, and the multidrug and toxic compound extrusion (MATE) family.

    (1) Major facilitator superfamily

    So far, more than twenty MFS families have been recognised according to the degrees of sequence simi?larity. The clinically important multidrug transporters are in the family 2 and 3, which possess 14 and 12 TMSs (transmembrane segments), respectively. They are also referred to as the DHA14 and DHA12 families due to the ability of catalysing drug: H+ antiport. Multiple sequence alignment suggested that N?terminal region is related to proton translocation producing energy for transport since the high similarity among them of the proteins; while C?terminal halves may contribute to substrate recognition and binding because of the less similarity among different proteins. Additionally, it is noticeable that the majority of the bacterial drug exporters belong to MFS.

    Many highly conserved motifs have been found in the two families, which are probably essential to direct transport. Their location, conserved amino acids and postulated function are listed in Tab.1［9］.

    i) DHA 14 family of MFS

    The multidrug transporters, including QacA, QacB, EmrB, Bmr3, LfrA and VceB, from Staphylococcus aureus, Escherichia coli, Bacillus subtilis, Mycobacterium smegmatis, and Vibrio cholerae, have been identified belonging to DHA14. Here we briefly describe some features of QacA, QacB and EmrB.

    QacA confers resistance to various antiseptics and disinfectants in Staphylococcus aureus. It possesses 514 amino acids encoded by plasmid?carrying (such as pSK1) qacA determinant. QacB, only mediating resistance to monovalent cations, is encoded by qacB also on plasmids (such as pSK 23 and pSK 156). Sequence analysis disclosed a difference of only seven base pairs between qacA and qacB, resulting in the difference of start codon and three non?conservative amino acids at positions 291, 323 and 380. Site?directed mutagenesis showed that the aspartate at the position 323 is essential for resistance to divalent compounds of QacA, because of the substitution by alanine (the corresponding residue in QacB) resulting in the lost of resistance to divalent cations. Sequence analysis and the emergence of the qacA S.aureus clinical isolates have suggested that qacA evolved from qacB by gaining a high affinity?binding site for divalent cations. A 188?amino acid, QacR, has been identified to regulate the expression of qacA and qacB as a trans?acting repressor. QacR, encoded by qacR which is located upstream of qacA and qacB genes, belongs to the TetR family of regulatory proteins. The QacA substrates have been observed in vitro to induce the expression of qacA due to the ability to prevent QacR binding to the operator sequence and result to de?repression of the qacA promoter. The region between the qacR and qacA contains two invert repeats (IR). IR1 overlaps the proposed qacA promoter and consists of a palindrome with a 6?bp gene flanked by two 15 bp arms. IR1 has been demonstrated to be the operator site for QacR but the function of IR2 remains unknown［10］. Our studies on distribution of the gene qacA/B in 126 clinical isolates of S.aureus showed that the genes were found in 12/126 S.aureus isolates (9.5%), which means that the prevalence of antiseptic resistance genes has yet to be a serious problem in our hospital. Meanwhile, functional analysis on the amino acid residues in and flanking cytoplasmic end of TMS13 and TMS5 including E407, S408, M409, Y410, D411, L412, L146, A147, V148, W149 and S150 showed that E407, Y410, D411 L146, W149 and S150 are likely to be functionally important residues probably involved in the substrate binding and translocation of QacA (data unpublished).

    EmrB functions as an inner membrane efflux transporter belonging to the DHA?14 family of MFS, encodedby emrB. It has been suggested that EmrB functions cooperatively with EmrA and EmrR, analogously operating as a tripartitie system in gram?negative bacteria. Studies showed that they are encoded by three cotranscribed genes of the emr locus, which situated at 57.5 min on the E.coli chromosome and confers resistance to hydrophobic uncouplers, such as carbonyl cyanide m?chlorophenylhydrazone (CCCP) and tetrachlorosalicylanilide, to organomercurials, and to some hydrophobic antibiotics, such as nalidixic acid and thiolactomycin［11，12］.

    ii) DHA 12 family of MFS

    There are a number of multidrug transporters in bacteria belonging to DHA 12 family of MFS. Among them, NorA from Staphylococcus aureus is a typical one, which is predicted to have 388 amino acids and a molecular mass of 42,385 Da. The NorA protein is coded by the norA gene, which is located on the SmaI D fragment of the S.aureus chromosome. Overexpression of norA in E.coli as a cloned gene and in S.aureus produces resistance to a broad range of substrates, including fluoroquinolones, similar in breadth to that encoded by the bmr gene, a homolog of norA in Bacillus subtilis［13］. The bmr gene encodes another kind of DHA 12 transporter, named Bmr. Other members of DHA 12 family include PmrA from Streptococcus pneumoniae, EmrD from Escherichia coli, LmrP from Lactococcus lactis, et al［14］.

    (2) The small multidrug resistance (SMR) family

    SMR family possesses the smallest multidrug transporters, which are typically about 110 amino acid residues and have been identified a model of a four TMS. The small size makes it suitable for mutagenesis studies. The well?characterized SMR transporter is SMR protein from Staphylococcus, mediating resistance to many organic cations. Some residues such as Glu?24, Pro?31, Cys?42, and Glu?80 appear to affect the substrate specificity［15］.

    Another important one is the EmrE protein from E.coli, which has been characterized, purified and reconstituted in a functional form. Glu?14, the only membrane?embedded charged residue, was shown to be part of a binding site for both protons and substrates. Most recently, one comformational state of this protein in complex with a substrate, tetraphenylphosphonium, was crystallized at 3,7?angstrom resolution. Two EmrE polypeptides form a homodimeric transporter that binds substrate at the dimerization interface. The two subunits have opposite orientations in the membrane and adopt slightly different folds, forming an asymmetric antiparallel dimer［16］.

    (3) The resistance?nodulation?cell division (RND) family

    RND family transporters are most commonly found in Gram?negative bacteria, and typically operate as part of a tripartite system that includes a periplasmic membrane fusion protein (MFP) and an outer membrane factor, bypassing the outer membrane barrier. The secondary structure was proposed to consist of 12 TMS. Sequence alignment has previously identified three highly conserved motifs shared by RND proteins, which may be structurally or functionally essential. A typical member of RND family is AcrB of E.coli, which functions together with AcrA (MFP) and TolC (outer membrane factor). More recently, the AcrAB?TolC bound with a substrate has been successfully crystallized［17～19］.

    Another clinically important pathogen, P.aeruginosa, is highly resistant to a variety of antimicrobials. So far, it has been recognized that its seven RND efflux systems combined with intrinsic low membrane permeability contribute to the property. These pumps are named MexAB?OprM［20］, MexCD?OprJ［21］, MexEF?OprN［22］, MexXY?OprM［23］, MexJK?OprM［24］, MexGHI?OpmD［25］ and MexVW?OprM［26］, respectively.

    (4) The multidrug and toxic compound extrusion (MATE) family

    In 1998, Brown et al. found a new family of putative secondary transporters unique to plants and microbes termed the multidrug and toxic compound extrusion (MATE) family. Hydropathy analysis suggests that proteins of the MATE family have a common topology consisting of 12 transmembrane (TM) domains. On the basis of phylogenetic analysis the MATE family is divided into three groups. Members of the two more related groups are found in prokaryotes. Expression of the Vibrio parahaemolyticus norM gene in Escherichia coli increases resistance to the antibiotic norfloxacin as well as to ciprofloxacin, and to the structurally unrelated antimicrobial agents ethidium, streptomycin, and kanamycin. Expression of ydhE, which encodes an E.coli ortholog of NorM, from a multicopy plasmid, confers on E.coli resistance to a group of antimicrobial agents that are similar to but not overlapping with those to which NorM confers resistance［27,28］.

    Analysis approach on membrane transporters

    Here we mainly focus on the analysis approaches of secondary transporters, which account for the majority of the important pumps in bacteria. To a greater degree, these protocols are also suitable for primary transporters.

    Site?directed cysteine mutagenesis

    The central question is the mechanism of multidrug recognition and transport. How can they interact with scores of molecules having no structural similarity? Apparently, the determination of crystal structure bound with substrates will tell us the answer. Unfortunately, it is hard to get the crystallized high?resolution structure. Alternatively, site?directed cysteine mutagenesis is utilized to mutate the targeted amino acid residues, which, coupled with maleimide compounds, can determine the boundary of the membrane and the functionally essential residues.

    The once most popular method to determine the membrane?spanning topology is PhoA (reporter?enzyme) fusions［29］. But the native structure is not really representable. Tamura et al. developed a new protocol based on the site?directed mutagenesis, by introducing cysteine into a cysteine?free derivative of a membrane protein, in which two kinds maleimide agents, membrane?permeable N?ethyl maleimide (NEM) and membrane?impermeable 4?acetamido?4′?meleimidylstilbene?2, 2′?disulfonic acid (AMS), were utilised. Thus, the inside/outside localization of introduced cysteine residues can be decided by competitive binding of these two compounds in intact E.coli cells expressing Cys mutants. That is to say, the maleimide reactivity of the transporter will be a measure of the reactivity of the newly introduces cys residue. This will be determined by measuring incorporation of labelled NEM into the targeted protein, using immunoprecipitation with exporter antibodies. The reactivity of a cysteine sulphyhydral group embedded in the hydrophobic interior should be lower than that of one exposed to the aqueous phase. In addition, residues locates in the periplasm should not be NEM?reactive after treatment with AMS［30］. Moreover, if a complete cysteine?scanning mutagenesis combined with the application of NEM, could be performed, the exact membrane border might be determined or the water?channel?facing helices and the membrane?embedded helices could be distinguished［31］.

    After incorporated into cysteine, the functional important amino acid residues can also be suggested by this strategy via functional analyses. Generally, two assays are employed. One is the minimum inhibitory concentration (MIC) test, and the other is the assay of the capability of expelling representative substrates, named transport assay, by which a series of efflux curves were obtained and initial velocity was calculated by averaging the linear part of each curve. And these data with corresponding concentration were used to estimate the kinetic parameters of transport, Km and Vmax. Km shows the binding affinity of the protein for the substrate, whereas Vmax represents the maximum velocity of the transport. The range of substrate concentrations was chosen to be equally spread above and below the Km and at least ten different substrate concentrations were used in one experiment.

    Helix packing

    Using different molecular and biochemical or biophysical techniques, helix?packing have been used to determine the helix arrangement in the membrane plane. A helix?packing model has been proposed for the lactose permease LacZ from E.coli after exhaustive studies. And a highly probable model for 12?TMS MFS transporters has been proposed based on the structure of OxlT and the systematical studies of TetA(B) and LacZ.

    To do cross?linking, two approaches, both holding their potential advantages and disadvantages, have been tried. The first approach is intramolecular disulfide linkage of double Cys mutants, using full?length protein with two cysteine residues introduced in different TMS. The advantage of this approach is the protein is normally expected to retain its native conformation, hence it will reflect the real structure. But this approach may suffer the difficulties in detcting the potential mobility change due to cross?linking between the two cysteine residues. The second approach is intermolecular cross?linking, in which the protein is expressed as two non?overlapped fragments. Cross?linking of the two?cysteine residues introduced in this two fragments respectively will result in full?length of the transporter. The results will be very clear. But the potential problem lays in the fact that the two fragments may not be able to assemble into a structure resemble the native conformation. To carry out intermolecular cross?linking, a split construct requires to be made. To make this mutant, a two?step PCR approach was used to introduce the stop codons, the ribosome?binding site and restriction enzyme cutting sites, which will give rise to the formation of the non?overlapped?fragment split and facilitate the future fragment shuffling. The formation of a disulfide linkage is detected by the band shift on SDS polyacrylamide gel electrophoresis after oxidation with Cu2+/o?phenanthroline.

    To make sure that the two fragments are able to assemble into a 3D structure similar to native structure, MIC analysis and transport assay are usually carried out. Strain containing this construct should be resistant to different efflux substrates to a certain degree. Significant transport activities should also be retained, indicating that the two fragments are able to assembled into a structure similar to the native. Then the construct is able to employ to make double cysteine mutants.

    The next question is where to start. For a transporter, there are billions of helix?arrange possibilities if arrange randomly. It would be a big challenge even to choose a start point. Based on site?directed mutagenesis or complete cysteine?scanning mutagenesis, the roles and positions of individual residues have been revealed. Thus the TMS possibly containing substrate?binding region or lining the central water?filled channel will be the candidate.

    Since the function of the transporter is pumping out all kinds of structure?unrelated compounds, it is more authentic using cross?linking pattern when different substrate bound to. The binding may reveal the probable movement between different TMS and conformational change. Studies have suggested that proton translocation and substrate transport are separated and only conformationally linked［32～34］.

    Crystallization

    It goes without saying that structural studies, such as X?ray crystallography and solid?state NMR, of the transporter will provide direct insight of its drug?recognition and transport mechanisms［35］.

    To perform such detailed studies requires large amounts of purified protein, which must be structurally integral and functionally active. First is the over?expression of the protein, solubilizing the protein from the bacterial membrane with detergents and purifying it. To confirm the structural integrity and activity of this purified protein, the isolated purified protein should be able to reconstitute back into synthetic membranes (liposomes), followed by the measure of its transport activity upon reconstitution. These processes are an essential prerequisite for subsequent structural studies.

    The progress in the study of high?resolution crystallography is quite encouraging during the recent years. Crystal structures of the three components of a drug efflux pump have now been solved: the outer membrane TolC exit duct in the year 2000［36］, the inner membrane AcrB antiporter in 2002［37］ and the periplasmic adaptor MexA in 2004, all belonging to the RND family［38］. Another one is EmrE, belonging to SMR family. However, the situation is still challenging in many other transporters, including multidrug exporters.

    Conclusion

    So far, the physiological role of the bacterial multidrug transporters is yet to be recognized. But the significance lies on the clinical relevance and the apparently paradoxical ability of high affinity mutidrug recognition. The deep understanding of the structure and function will certainly facilitate the new design of future antibiotics. And indeed the efficacy of efflux pump inhibitors is promising. Hence, the improvement and innovation of the study approach are imperative.

【】
  ［1］ Veen van H W, Venema K, H. Bolhuis H, et al. Multidrug resistance mediated by a bacterial homolog of the human drug transporter MDR1 ［J］. Proc Natl Acad Sci USA,1996,93(20):10668

［2］ Higgins C F. ABC transporters: from microorganisms to man ［J］. Annu Rev Cell Biol,1992,8:67

［3］ Poelarends G J, Vigano C, Ruysschaert J M, et al. Bacterial multidrug resistance mediated by ABC?transporters ［A］. In: ABC proteins: From bacteria to man ［M］. Ed. by Holland B, Kuchler K, Higgens C, et al. London: Academic Press,2002:243

［4］ Raviv Y, Pollard H B, Bruggeman E P, et al. Photosensitized labeling of a functional multidrug transporter in living drug resistant tumor cells ［J］. J Biol Chem,1990,265(7):3975

［5］ van Veen H W, Higgins C F, Konings W N. Multidrug transport by ATP binding cassette transporters: a proposed two?cylinder engine mechanism ［J］. Res Microbiol,2001,152(3～4):365

［6］ Margolles A, Putman M, van Veen H W, et al. The purified and functionally reconstituted multidrug transporter LmrA of Lactococcus lactis mediatesthe transbilayer movement of specific fluorescent phospholipids ［J］. Biochemistry,1999,38(49):16298

［7］ van Veen H W, Putman M, Margolles A, et al. Structure?function analysis of multidrug transporters in Lactococcus lactis ［J］. Biochim Biophys Acta,1999,1461(2):201

［8］ Chang G, Roth C B. Structure of MsbA from E.coli: a homolog of the multidrug resistance ATP binding cassette (ABC) transporters ［J］. Science,2001,293(5536):1793

［9］ Paulsen I T, Brown M H, Skurray R A. Proton?dependent multidrug efflux systems ［J］. Microbiol Rev,1996,60(4):575

［10］ Brown M H, Skurray R A. Staphylococcal multidrug efflux protein QacA ［J］. J Mol Microbiol Biotechmol,2001,3(2):163

［11］ Furukawa H, Tsay J T, Jackowski S, et al. Thiolactomycin resistance in Escherichia coli is associated with the multidrug resistance efflux pump encoded by emrAB ［J］. J Bacteriol,1993,175(12):3723

［12］ Borges?Walmsley M I, Beauchamp J, Kelly S M, et al. Identification of oligomerization and drug?binding domains of the membrane fusion protein EmrA ［J］. J Biol Chem,2003,278(15):12903

［13］ Kaatz G W, Seo S M. Effect of substrate exposure and other growth condition manipulations on norA expression ［J］. J Antimicrob Chemother,2004,54(2):364

［14］ Klyachko K A, Schuldiner S, Neyfakh A A. Mutations affecting substrate specificity of the Bacillus subtilis multidrug transporter Bmr ［J］. J Bacteriol,1997,179(7):2189

［15］ Grinius L, Dreguniene G, Goldberg E B, et al. A staphylococcal multidrug resistance gene product is a member of a new protein family ［J］. Plasmid,1992,27(2):119

［16］ Pornillos O, Chen Y J, Chen A P, et al. X?ray structure of the EmrE multidrug transporter in complex with a substrate ［J］. Science,2005,310(5756):1950

［17］ Murakami S, Nakashima R, Yamashita E, et al. Crystal structure of bacterial multidrug efflux transporter AcrB ［J］. Nature,2002,419(6907):587

［18］ Yu E W, Aires J R, Nikaido H. AcrB multidrug pump of Escherichia coli: composite substrate?binding cavity of exceptional flexibility generates its extremely wide substrate specificity ［J］. J Bacteriol,2003,185(19):5657

［19］ Yu E W, McDermott G, Zgurskaya H I, et al. Structural basis of multiple drug?binding capacity of the AcrB multidrug efflux pump ［J］. Science,2003,300(5621):976

［20］ Poole K, Krebes K, McNally C, et al. Multiple antibiotic resistance in Pseudomonas aeruginosa: evidence for involvement of an efflux operon ［J］. J Bacteriol,1993,175(22):7363

［21］ Poole K, Gotoh N, Tsujimoto H, et al. Overexpression of the mexC?mexD?oprJ effux operon in nfxB?type multidrug resistant strains ［J］. Mol Microbiol,1996,21(4):713

［22］ Kohler T, Michea?Hamzehpour T, Henze U, et al. Characterization of MexE?MexF?OprN, a positively regulated multidrug efflux system of Pseudomonas aeruginosa ［J］. Mol Microbiol,1997,23(2):345

［23］ Mine T, Morita Y, Kataoka A, et al. Expression in Escherichia coli of a new multidrug efflux pump, MexXY, from Pseudomonas aeruginosa ［J］. Antimicrob Agents Chemother,1999,43(2):415

［24］ Chuanchuen R, Narasaki C T, Schweizer H P. The MexJK efflux pump of Pseudomonas aeruginosa requires OprM for antibiotic efflux but not for efflux of triclosan ［J］. J Bacteriol,2002,184(18):5036

［25］ Aendekerk S, Ghysels B, Cornelis P, et al. Characterization of a new efflux pump, MexGHI?OpmD, from Pseudomonas aeruginosa that confers resistance to vanadium ［J］. Microbiology,2002,148(Pt8):2371

［26］ Li Y, Mima T, Komori Y, et al. A new member of the tripartite multidrug efflux pumps, MexVW?OprM, in Pseudomonas aeruginosa ［J］. J Antimicrob Chemother,2003,52(4):572

［27］ Brown M H, Paulsen I T, Skurray R A. The multidrug efflux protein NorM is a prototype of a new family of transporters ［J］. Mol Microbiol,1998,31(1):393

［28］ Morita Y, Kodama K, Shiota S, et al. NorM, a putative multidrug efflux protein, of Vibrio parahaemolyticus and its homolog in Escherichia coli ［J］. Antimicrob Agents Chemother,1998,42(7):1778

［29］ Boyd D, Beckwigh M. Determinants of membrane protein topology ［J］. Proc Natl Acad Sci USA,1987,84(23):8525

［30］ Kimuara T, Lhnuna M, Sawai T, et al. Membrane topology of the transposon 10?encoded metal?tetracycline/H+ antiporter as studied by site?directed chemical labeling ［J］. J Biol Chem,1997,272(1):580

［31］ Tamura N, Konishi S, Iwaki S, et al. A complete cysteine?scanning mutagenesis and site?directed chemical modification of the TN10?encoded metal?tetracycline/H+ antiporter ［J］. J Biol Chem,2002,276(23):20330

［32］ Kubo Y, Konishi S, Dawabe T, et al. Proximity of periplasmic loops in the metal?tetracycline/H+ antiporter of Escherichia coli observed on site?derected chemical cross?linking ［J］. J Biol Chem,2000,275(8):5270

［33］ van Montfort B A, Schuurman?Wolters G K, Wind J, et al. Mapping of the dimer interface of the Escherichia coli mannitol permease by cysteine cross?linking ［J］. J Biol Chem,2002,277(17):14717

［34］ Ermolova N V, Smirnova I N, Kasho V N, et al. Interhelical packing modulates conformational flexibility in the lactose permease of Escherichia coli ［J］. Biochemistry,2005,44(21):7669

［35］ Lorch M, Lehner I, Siarheyeva A, et al. NMR and fluorescence spectroscopy approaches to secondary and primary active multidrug efflux pumps ［J］. Biochem Soc Trans,2005,33(Pt 4):873

［36］ Koronakis V, Sharff A, Koronakis E, et al. Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export ［J］. Nature,2000,405(6789):914

［37］ Murakami S, Nakashima R, Yamashita E, et al. Crystal structure of bacterial multidrug efflux transporter AcrB ［J］. Nature,2002,419(6709):587

［38］ Akama H, Matsuura T, Kashiwagi S, et al. Crystal structure of the membrane fusion protein, MexA, of the multidrug transporter in Pseudomonas aeruginosa ［J］. J Biol Chem,2004,279(25):25939