The outer membrane component of the multidrug efflux pump from Pseudomonas aeruginosa may be a gated channel Eisaku Yoshihara, Hideaki Maseda and Kohjiro Saito Department of Molecular Life Science, School of Medicine Tokai University, Isehara, Japan OprM, the outer membrane component of the MexAB- OprM multidrug efflux pump of Pseudomonas aeruginosa, has been assumed to facilitate the export of antibiotics across the outer membrane of this organism. Here we purified to homogeneity the OprM protein, reconstituted it into lipo- some membranes, and tested its channel activity by using the liposome swelling assay. It was demonstrated that OprM is a channel-forming protein and exhibits the channel property that amino acids diffuse more efficiently than saccharides. However, antibiotics showed no significant diffusion through the OprM channel in the liposome membrane, suggesting that OprM functions as a gated channel. We reasoned that the protease treatment may cause the distur- bance of the gate structure of OprM. Hence, we treated OprM reconstituted in the membranes with a-chymotrypsin and examined its solute permeability. The results demon- strated that the protease treatment caused the opening of an OprM channel through which antibiotics were able to dif- fuse. To elucidate which cleavage is intimately related to the opening, we constructed mutant OprM proteins where the amino acid at the cleavage site was replaced with another amino acid. By examining the channel activity of these mutant proteins, it was shown that the proteolysis at tyrosine 185 and tyrosine 196 of OprM caused the channel opening. Furthermore, these residues were shown to face into the periplasmic space and interact with other component(s). We considered the possible opening mechanism of the OprM channel based on the structure of TolC, a homologue of OprM. Keywords: multidrug efflux pump; OprM; channel; gate; P. aeruginosa. Pseudomonas aeruginosa causes opportunistic infections in immunocompromised patients and exhibits natural and acquired resistance to structurally and functionally diverse antibiotics [1]. The broad specific antibiotic resistance of P. aeruginosa is mainly attributable to the synergy of a tight outer membrane barrier and expression of multidrug efflux pumps with a low specificity for antibiotics [2–7]. In P. aeruginosa, four operons encoding multidrug efflux pumps have been reported. Among these, the mexAB- oprM operon is expressed constitutively in the wild-type strain and overexpressed in the nalB-type mutant that exhibits an elevated level of resistance to fluoroquinolones, tetracycline, chloramphenicol, macrolides, trimethoprim, and b-lactams [8–13]. The MexCD-OprJ and MexEF- OprN pumps that are homologous to MexAB-OprM, are highly expressed in the nfxB-andnfxC-type mutants, respectively, and each contributes to multidrug resistance [11–15]. MexXY has been shown to act in conjunction with OprM and also confers multidrug resistance on P. aeruginosa [16–18]. These efflux pumps of P. aeruginosa consist of two inner membrane-associated components, MexB (MexD, MexF, and MexY) and MexA (MexC, MexE, and MexX), and an outer membrane component, OprM (OprJ, and OprN) [2–4,10,11,15,18]. MexB is a member of the resistance- nodulation-cell division (RND) family and has been presumed to act as a drug-proton antiporter. MexA is a periplasmic fusion protein and is assumed to link the inner and outer membranes. It has been predicted that OprM forms a channel allowing the antibiotics to pass through the outer membrane [2,3,19–21]. Although three components of the multidrug efflux pump are essential to confer multidrug resistance on P. aeruginosa [17,22,23], the biochemical and biophysical studies of the pump components are limited. Recently, Zgurskaya and Nikaido [24] reported a reconstituted system to investigate the function of AcrB, a member of the RND family and an inner membrane component of the multidrug efflux pump of E. coli. They reconstituted the purified AcrB into liposome membranes and measured the efflux activity. AcrB was shown to possess catalytic activity in extrusion of fluorescent phospholipids from the liposome membrane in the presence of DpH, and its activity was competitively inhibited by the substrates of the AcrAB-TolC pump. This reconstituted system does not directly measure the extrusion of its substrates, but appears to be applicable to other members of the RND family such as MexB. AcrA of E. coli is a member of the membrane fusion protein (MFP) family and is assumed to act as a linker between the inner membrane transporter and the outer membrane channel. Zgurskaya and Nikaido [25] demon- strated that AcrA exists as a highly asymmetric monomeric molecule in an elongated form that is sufficient to span the periplasmic space. Correspondence to E. Yoshihara, Department of Molecular Life Science, School of Medicine Tokai University, Isehara, 259-1193, Japan Fax: + 81 463 93 5437, Tel.: + 81 463 93 5436. E-mail: eisaku@is.icc.u-tokai.ac.jp Abbreviations: RND, resistance-nodulation-cell division; MFP, membrane fusion protein. Enzymes: a-chymotrypsin (EC 3.4.21.1) (Received 16 April 2002, revised 6 July 2002, accepted 24 July 2002) Eur. J. Biochem. 269, 4738–4745 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03134.x OprM, OprJ and OprN are mutually homologous and also homologous to TolC, an outer membrane protein of E. coli. TolC is a versatile protein with multiple roles in extruding the antibacterial drugs or exporting toxins by forming complexes with different counterparts [26,27]. Bentz et al. [28] demonstrated that TolC forms a channel in planar lipid bilayer membranes and that the pore size of the TolC channel appeared to be insufficient for the exportation of drugs or toxins. Indeed, Koronakis et al. [29] recently revealed the crystal structure of TolC and showed that it forms a b-barrel channel in the outer membrane. We tried elucidating the function of the outer membrane component of the multidrug efflux pump of P. aeruginosa and showed that OprM forms a channel with some specificity for amino acids and peptides. Furthermore, we present data indicating that OprM functions as a gated channel allowing antibiotics to diffuse through the outer membrane of this organism. MATERIALS AND METHODS Purification of the outer membrane components of multidrug efflux pumps The oprM protein was isolated from P. aeruginosa TNP058 (nalB-type mutant). One litre of fully grown cells was diluted with 10 L of LB-medium (10 g of Bacto- tryptone,5gofyeastextractand10gofNaClperlitre, pH 7.3) and the flask was rotated at 200 r.p.m. at 37 °C for 4 h. The harvested cells were suspended in 50 mL of 20 m M sodium phosphate buffer, pH 7.6, and disrupted by passing through a French pressure cell three times at 1200 kg cm )2 . After removing unbroken cells by centri- fugation at 7000 g for 10 min, the supernatant was centrifuged at 100 000 g for 60 min at 15 °C. The pellet was suspended in 50 mL of 20 m M sodium phosphate buffer, pH 7.6, containing 0.8% of sodium lauroylsarco- sinate and incubated at 30 °C for 30 min to solubilize the inner membrane proteins. After centrifugation at 100 000 g for 60 min at 15 °C, the pellet was solubilized in 50 mL of 10 m M sodium phosphate buffer, pH 8.0, containing 2.5% of b-octylglucoside and centrifuged at 100 000 g for 60 min at 15 °C to remove the insoluble materials. The supernatant was subjected to DEAE-high performance liquid chromatography equilibrated in 10 m M sodium phosphate buffer, pH 8.0, containing 1% b-octylglucoside and the column was eluted with a 0–500 m M linear gradient of NaCl. Fractions containing OprM were pooled and subjected again to the DEAE- chromatography [30,31]. The purified protein was subjec- tedtoSDS/PAGEtoestimatethehomogeneityofthe purified protein. To confirm that the purified protein is OprM, we determined the N-terminal amino acid sequences of the products generated by protease treatment. The purified sample was mixed with a-chymotrypsin (10 lgÆmL )1 )in 10 m M sodium phosphate buffer, pH 8.0, containing 1% b-octylglucoside, and incubated at 37 °C for 5 h. Samples were subjected to SDS/PAGE (12% gel) and blotted electrically onto a poly(vinylidene difluoride) membrane. Protein bands were excised from the membrane and subjected to Edman degradation using a protein sequencer. Measurement of the permeability of OprM Proteoliposomes were prepared as described previously with minor modifications [30]. Briefly, an appropriate amount of OprM was mixed with 1 lmole of the lipids phosphatidyl- choline (PtdCho) and dicetylphosphate at the molar ratio of 97 : 3 in 250 lLof2%b-octylglucoside, and then dialysed against a large excess of deionized water at 4 °Covernight. The liposome suspension was transferred to a glass tube, dried under an N 2 gas stream, and retained in an evacuated desiccator for 30 min The lipid film was suspended in 133 lLof40mosMstachyosein1m M Mops buffer, pH 7.2 and then vortexed vigorously for 20 s. Diffusion rates of the test solutes were determined by measuring the time course of the D 450nm change [30,32]. The effect of the protease digestion of OprM on the channel activity was examined as follows. The liposome membranes with or without OprM were suspended in 20 mosM stachyose or 6% Dextran T-10 in 1 m M Mops buffer, pH 7.2, containing a-chymotrypsin (final concen- tration of 100 lgÆmL )1 ), and incubated at 37 °Cfor3h. Then the diffusion rates were measured by the liposome swelling assay. Construction of the mutant OprM proteins To introduce the mutation in the oprM gene, we used a site-directed mutagenesis system kit (Takara Shuzo). Oligo- nucleotide primers, Y24S (5¢-GTTCTGCCCGGAGGCC TGCC) and Y196S (5¢-GCCGACGTCGGAGCTGC GCT) were employed to substitute Tyr-24 or Tyr-196 to Ser, respectively. To construct the oprM gene coding the OprM with two mutations at tyrosine 185 and tyrosine 196, we used the pKF18 bearing the gene coding the mutant OprM (Y196C) and the oligonucleotides, Y185C (5¢-ACT CTTCTGGCAGGTGCCCAG) as the template and the primer, respectively. The nucleotide sequences of these mutant genes were determined using an ABI model 373A sequencer to confirm the introduction of the mutations. Proteolysis of OprM in the cells by a-chymotrypsin P. aeruginosa PAO1 cells were grown in 4 mL L broth at 37 °C for 4 h. Harvested cells were suspended in NaCl/P i and mixed with a-chymotrypsin (final concentration of 100 lgÆmL )1 ) in the absence or presence of 1 m M EDTA. The cells were incubated at 37 °Cfor60minandthen 0.5 m M p-toluenesulfonyl fluoride was added. The cells were disrupted by sonication and the membrane fraction was solubilized in the sample buffer. These samples were subjected to SDS/PAGE and the proteins were blotted on a poly(vinylidene difluoride) membrane. OprM and its fragments were visualized by immunostaining with poly- clonal antibodies directed against OprM [33]. Chemicals EggyolkPtdCho(typeV-E)anda-chymotrypsin were from Sigma; N-lauroylsarcosinate and dicetylphosphate were obtained from Nacalai Tesque Inc., Kyoto. Bacto-tryptone and yeast extracts were purchased from Difco. b-octylgluco- side was from Alexis Corp., Switzerland. All other chemicals used were of the highest purity available commercially. Ó FEBS 2002 Gating of the OprM channel of MexAB-OprM pump in P. aeruginosa (Eur. J. Biochem. 269) 4739 RESULTS Purification and identification of OprM OprM was isolated from P. aeruginosa nalB-type mutant overexpressing the MexAB-OprM pump and purified by ion-exchange HPLC in the presence of b-octylglucoside as described in Materials and methods. SDS/PAGE showed that the purified protein appeared to be homogeneous (Fig. 1, lane 2). However, in order to confirm that the purified protein is OprM, we determined the N-terminal amino acid sequences of fragment 1, 2, 3 and 4 that were produced by a-chymotrypsin treatment (Fig. 1, lane 3). The N-terminal amino acid sequences of fragment 1, 2 and 4 were determined to be GQNTGAAAVP, DVGVASALDL and GQNTGAAAVP, respectively, which matched completely the sequences from glycine 25 to proline 34, from aspartate 197 to leucine 206, and from glycine 25 to proline 34 of OprM [9]. The results demonstrate that the protein purified was OprM. On the other hand, the N-terminal sequence of fragment 3 could not be determined, suggesting that this fragment contains the N-terminus of the OprM protein since OprM has been demonstrated to be acylated [34]. Characterization of OprM The outer membrane components of the multidrug efflux pumps in P. aeruginosa havebeenassumedtoforma channel through which antibiotics pass to cross the outer membrane. Then, to demonstrate the channel activity and properties of OprM, we examined the permeability of various solutes through the OprM protein reconstituted into liposome membranes by using the liposome swelling assay [30]. First, we tested the glycine permeability, and found that Gly did not diffuse through the liposomes without any protein, but diffused through the liposome containing the OprM protein. The diffusion rates of Gly increased linearly depending on the amount of OprM reconstituted into the liposome membranes (Fig. 2 inset). These results clearly demonstrate that OprM forms the channel through which Gly can diffuse. Next we exam- ined the OprM channel properties by measuring the diffusion of various solutes such as amino acids, peptides and saccharides. As shown in Fig. 2, it was demonstrated that the OprM channel possesses a pore allowing the diffusion of these solutes so that amino acids and peptides (except the hydrophobic amino acids such as leucine and phenylalanine) diffused more efficiently than saccharides. This specific diffusion of the OprM channel may be unique among the outer membrane channels so far reported. Next, we tested the diffusion of antibiotics such as cephalexin (M r 374) and cephaloridine (M r 415). We chose these drugs as test solutes since they are zwitterionic b-lactams extruded by the MexAB-OprM pump (data not shown). The liposome swelling method used here cannot be applied to negatively or positively charged drugs. However, it was found that these antibiotics did not diffuse through the OprM channel, suggesting that OprM may function as a gated channel. Opening of the OprM channel Next we designed an experiment to demonstrate the gating function of the OprM channel. In general the gate domain of channel proteins seems to have a flexible region, and the flexible regions are prone to digestion by proteases. Accordingly, we expected that the protease treatment might affect the gating of the OprM channel. We chose a-chymotrypsin to digest OprM because this enzyme was shown to digest only at two sites in OprM (Fig. 1). Even when OprM exists in the liposome mem- branes, the same sites were digested by a-chymotrypsin (data not shown). Therefore we treated the OprM containing liposomes and measured the diffusion of various solutes including antibiotics through these lipo- somes. As shown in Fig. 3, it was clearly demonstrated that the protease treatment caused a noticeable increase in the diffusion rates of amino acids, peptides and saccha- rides, and surprisingly allowed cephalexin and cephalo- ridine to permeate through the OprM channel. These results suggest that the protease treatment rendered the channel open probably by affecting the gate structure of OprM. One may argue that the activation of the OprM channel by the protease treatment is caused by the gross destruction of the OprM protein. However, this possibility seems very unlikely since the protease treatment caused no significant change in the fluorescence spectrum of OprM, indicating that no gross structural alteration of the OprM protein occurred by the protease treatment (data not shown). Fig. 1. SDS/PAGE of purified OprM. OprM was purified as described in Materials and methods and was subjected to SDS polyacrylamide gel electrophoresis after heating at 100 °C for 5 min. Lane 1, molecular mass markers: myoglobin (17 kDa), carbonic anhydrase (30 kDa), aldolase (42 kDa), albumin (66 kDa), a-galactosidase (116 kDa) and myosin(200 kDa);Lane2,purifiedOprM;Lane3,OprMtreatedwith a-chymotrypsin. Fragments are indicated by F1, F2, F3 and F4 according to the molecular masses. 4740 E. Yoshihara et al. (Eur. J. Biochem. 269) Ó FEBS 2002 The region involved in the opening of the OprM channel Next we investigated how the protease treatment caused the opening of the OprM channel. As OprM was digested at tyrosine 24 and tyrosine 196 by a-chymotrypsin, we examined which proteolysis is responsible for the channel opening. To answer this issue, we constructed mutant OprM proteins, that are resistant to the a-chymotrypsin digestion, by replacing the amino acid at position 24 or 196 with another one. The mutant OprM (Y24S) and OprM (Y196S) genes were constructed by site-directed mutagenesis as described in Materials and methods, and these mutant proteins were purified by the method described above. First, we checked whether these mutant proteins were resistant to a-chymotrypsin. Upon digestion of the mutant OprM (Y24S) by a-chymotrypsin, it was shown that fragment 2 and 3 alone were generated (Fig. 4B), the result indicating that there is no cleavage at the amino acid 24 as expected. On the contrary, the a-chymotrypsin digestion of OprM (Y196S) generated all fragments and caused the open state of channel through which antibiotics permeate (data not shown), suggesting that another site(s) near tyrosine 196 may be digested by a-chymotrypsin. Since tyrosine 185 resides near tyrosine 196, we considered the possibility that tyrosine 185 may be digested by a-chymotrypsin. Then, tyrosine 185 of OprM (Y196S) was substituted with cysteine and the OprM (Y185C/Y196S) protein was purified des- cribed above. The digestion of OprM (Y185C/Y196S) by a-chymotrypsin produced only fragment 1 (Fig. 4A), indi- cating that the digestion occurred at tyrosine 24 but not at the positions 185 and 196 of OprM (Y185C/Y196S) as expected. To examine the effect of these mutations on the OprM channel activity, we measured the diffusion rates of various solutes through these mutant proteins. As depicted in Fig. 5(A), OprM (Y185C/Y196S) exhibited permeability Fig. 2. Permeability of OprM reconstituted into the liposome membrane. Proteoliposomes were prepared from 1 lmol of lipids (PtdCho and dicetylphosphate in a 97 : 3 molar ratio) and 5 lg of the purified OprM, and the diffusion rates of amino acids, peptides and saccharides were examined by the liposome swelling assay as described previously [30]. Solutes used here were 1, glycine; 2, alanine; 3, serine; 4, proline; 5, threonine; 6, glycylglycine; 7, leucine; 8, methionine; 9; phenylalanine; 10, glycylglycylglycine; 11, ribose; 12; arabinose; 13, galactose; 14; glucose and 15, a-methylglucoside. Inset: effect of the amount of OprM on the diffusion rates of glycine. The proteoliposomes were prepared from the appropriate amount of OprM and the diffusion rates of glycine through these liposomes were determined. Fig. 3. Permeability of the OprM channel treated with a-chymotrypsin. Proteoliposomes were prepared according to the procedure as des- cribed in the legend of Fig. 2 except that the liposome was suspended in Dextran T-10 solution containing a-chymotrypsin (100 lgÆmL )1 ). The proteoliposomes were incubated at 37 °C for 2 h and then subjected to the liposome swelling assay. The data for the treated and untreated OprM with protease are represented by closed and open circles, respectively. Solutes used here were 1, glycine; 2, proline; 3, threonine; 4, glycylglycine; 5, arabinose; 6, ribose; 7, glucose; 8, glycylglycylgly- cine; 9, a-methylglucoside; 10, cephalexin; and 11, cephaloridine. Ó FEBS 2002 Gating of the OprM channel of MexAB-OprM pump in P. aeruginosa (Eur. J. Biochem. 269) 4741 almost the same as that of the wild-type OprM (open circles), showing that the mutation appears not to affect the channel activity. On the contrary, OprM (Y24S) was shown to have slightly less permeability towards amino acids and peptides compared with the wild-type OprM (Fig. 5B, open circles), which diminished the specificity of the OprM channel for amino acids and peptides. As it was shown that these mutations produced no, or a less significant, effect on the channel activity of OprM, we then investigated the effect of protease digestion on the channel activity of these mutant proteins. When OprM (Y185C/Y196S) was digested by a-chymotrypsin and subjected to the liposome swelling assay, it was shown that the protease treatment caused a slight effect on the saccharide diffusion but no significant effect on the diffusion of amino acids and peptides (Fig. 5A). Furthermore, the treatment caused no effect on the permeability of antibiotics through the OprM channel. These results indicate that the cleavage at tyrosine 24 is unrelated to the channel opening. Next, we examined the effect of protease digestion on the channel activity of OprM (Y24S). As show in Fig. 5B, it was clearly demonstrated that the protease treatment enhanced noticeably the permeability of amino acids, peptides and saccharides. More importantly, the protease treatment allowed the diffusion of cephalexin and cephaloridine through the OprM channel. All these data indicate that the digestion at tyrosine 185 and tyrosine 196 but not at tyrosine 24 is responsible for opening of the OprM channel. Tyrosine-185 and tyrosine 196 face into the periplasm As it was shown that the tyrosine 185 and tyrosine 196 were directly involved in the opening of the OprM channel, the next issue to address was whether tyrosine 185 and tyrosine 196 faced into the periplasmic space or the extracellular medium. To clarify this, we examined the accessibility of a-chymotrypsin to tyrosine 185 and 196 of OprM existing in the cell. Hence, P. aeruginosa cells were suspended in NaCl/ P i and digested by a-chymotrypsin as described in the legend of Fig. 6. The digestion of OprM was assessed by an immuno Western blot using a polyclonal anti-OprM antibody. As shown in Fig. 6, the protease treatment caused no digestion of the OprM protein. Next, we suspended the cells in the presence of EDTA to permeabilize the outer membrane, and then added a-chymotrypsin to the Fig. 4. SDS/PAGE of the mutant OprM (Y185C/Y196S) or OprM (Y24S) treated with a-chymotrypsin. (A) The mutant OprM (Y185C/ Y196S) was treated with a-chymotrypsin (100 lgÆmL )1 )at37°Cfor 2 h and subjected to SDS/PAGE: Lane 1, mol markers; lane 2, OprM (Y185C/Y196S) without the protease treatment; lane 3, OprM (Y185C/Y196S) treated with a-chymotrypsin. (B) Lane 1, OprM (Y24S) without protease teatment; lane 2, OprM (Y24S) treated with a-chymotrypsin (100 lgÆmL )1 )at37°Cfor2h. Fig. 5. Permeability of the mutant OprM (Y185C/Y196S) channel (A) and OprM (Y24S) (B) with or without the protease treatment. The mutant OprM (Y185C/Y196S) or OprM (Y24S) channel was reconstituted in liposome membranes with or without a-chymotrypsin in Dextran T-10, incubated at 37 °C for 2 h and then subjected to the liposome swelling assay. Open and closed circles represent the protease-untreated and treated data, respectively. Solutes used here were 1, glycine; 2, alanine; 3, serine; 4, threonine; 5, glycylglycine; 6, methionine; 7, phenylalanine; 8, ribose; 9, glucose; 10, a-methylglucoside; 11, cephalexin; and 12, cephaloridine. 4742 E. Yoshihara et al. (Eur. J. Biochem. 269) Ó FEBS 2002 cell suspension. It was demonstrated that only fragment 1 was produced (Fig. 6, lane 3), suggesting that a-chymo- trypsin can access tyrosine 24 under these conditions. It seems unlikely that the EDTA treatment may alter the accessibility of the surface region of OprM since if so, both sites of OprM should be digested. These results suggest that tyrosine 24 faces into the periplasmic space, and that tyrosine 185 and 195 may be protected from the access of a-chymotrypsin. As it was considered that this protected state is produced by the interaction between OprM and other pump component(s), we reconstituted the purified OprM protein into liposomes followed by a-chymotrypsin digestion. When the digestion products were examined by SDS/PAGE (Fig. 7), it was found that all fragments were produced under these conditions. Since fragment 4 is produced when a-chymotrypsin digested at both sites of OprM, these results indicate that tyrosine 185 and 195, and tyrosine 24 face into the same side of the membrane, that is, tyrosine 185 and 195 of OprM reside in the periplasmic space. DISCUSSION In order to investigate the assumed channel activity of the outer membrane component of the MexAB-OprM pump, we purified the OprM protein, reconstituted it into liposome membranes and examined the permeability through these proteoliposomes by using the liposome swelling assay [30]. When glycine was used as a test solution, it was clearly demonstrated that glycine could not diffuse through the liposome without OprM but diffused through those containing OprM. Its diffusion rates increased linearly to the amount of OprM reconstituted into the liposome membranes (Fig. 2). These data indicate unequivocally that OprM is a channel-forming protein, and are consistent with the data presented by Wong et al. [35]. They reported that OprM possesses channel activity with an average single- channel conductance of about 80 pS in 1 M KCl. Next, in order to reveal the characteristics of the OprM protein, we measured the diffusion rates of various solutes such as amino acids, peptides and saccharides through the OprM channel. Consequently it was demonstrated that all the amino acids, peptides and saccharides so far tested diffused through the OprM channel, but the permeability of the amino acids and peptides appeared to be superior to that of the saccharides (Fig. 2). To our knowledge, this kind of specificity seems to be unique among the outer membrane channels so far reported. Interestingly this specificity of the OprM channel disappeared by the substitution of tyrosine 24 to serine (Fig. 5B). Furthermore, OprJ and OprN, being the outer membrane components of MexCD-OprJ and MexEF-OprN pumps, were shown to exhibit similar specificity (unpublished data) 1 . Therefore, this kind of specificity may contribute to the function of the efflux pump. Since it was demonstrated that OprM is a channel protein but antibiotics could not diffuse through the purified OprM protein, we considered the possibility that OprM may be a gated channel. To date, many gated channels have been reported and the calcium-activated potassium channel from rat adrenal chromaffin cells is one of them. Solaro et al. [36] have shown that the activation of this channel induced by Fig. 6. Proteolysis of OprM in the cells with or without EDTA treat- ment. The experiments were carried out according to the method described in Materials and methods. Lane 1 represents the cells with- out the a-chymotrypsin treatment. Lanes 2 and 3 represent the cells treated with a-chymotrypsin in the absence or presence of 1 m M EDTA, respectively. Fig. 7. Proteolysis of OprM in the liposome membranes. The liposomes with OprM were prepared as described in materials and methods except that the liposomes were dialyzed against 20 m M sodium phos- phate buffer, pH 8.0. The liposome suspension was mixed with a-chymotrypsin (final concentration of 100 lgÆmL )1 ) and incubated at 37 °C for 2 h. Then the mixture was added to the sample buffer and subjected to SDS/PAGE. Proteins were stained by Coomassie blue dye. Lane 1, molecular mass markers; lane 2, proteoliposmes without the protease treatment; lane 3, proteoliposomes with the protease treatment. Ó FEBS 2002 Gating of the OprM channel of MexAB-OprM pump in P. aeruginosa (Eur. J. Biochem. 269) 4743 applying a voltage was followed by an inactivation process, and the tryrpsin treatment of this channel diminished the inactivation process. These results indicate that the inacti- vation process is carried out by plunging the inactivation domain into the open channel; the inactivation domain lost this gating function by the protease treatment. Accordingly, we thought that protease treatment might be applicable to show the gating function of the OprM channel. When OprM reconstituted into the liposome membranes was digested by a-chymotrypsin and the permeability of the treated OprM examined, it was demonstrated that this treatment enabled antibiotics to permeate through the OprM channel, probably by inducing the opening of it (Fig. 3). Next, in order to investigate how the OprM channel is opened by the protease treatment, we determined which digestion is responsible for the channel opening. Conse- quently it was demonstrated that the digestion at tyrosine 185 and tyrosine 196 but not at tyrosine 24 is responsible for the channel opening. The N-terminal inactivation domain of the potassium channel has been demonstrated to participate in the inactivation process [37,38]. However, this gating mechanism is unlikely to hold for the OprM channel because the digestion at tyrosine 24 located near the N-terminus of OprM is unrelated to the channel opening. Recently Li and Poole [21] reported the mutational analysis of OprM of P. aeruginosa. They showed that two of several deletions which were stably expressed, spanning residues G199 to A209 and A278 to N286 were unable to restore the antibiotic resistance of the cell, indicating that these regions are essential to the proper function of OprM. Since tyrosine 196 is very near from glycine 199, it is conceivable that the region from tyrosine 185 to alanine 209 might take an important role in the OprM function, e.g. the channel gating. OprM is a homologue of TolC of E. coli whose crystal structure has been solved recently by Koronakis et al.[29]. TolC is assembled into a trimer and forms a unique structure called a Ôchannel-tunnelÕ. TolC traverses the outer membrane by forming a b-barrel, whereas a-helical bundles form a long tunnel with a tapered proximal end. They propose that this channel may transport its substrates by opening the proximal end of the tunnel. We compared the amino acid sequences between TolC and OprM to consider the gating of the OprM channel. Based on the alignment of OprM and TolC presented by Li and Poole [21], it was found that tyrosine 185 and tyrosine 196 of OprM correspond to isoleucine 133 and phenylalanine 144 of TolC, respectively. The crystal structure of TolC shows that these residues are located at the C-terminal side of the helix 3 of TolC and this portion participates in forming the tapered proximal end of TolC. From these results we propose the possible opening mechanism that tyrosine 185 and tyrosine 196 are located at the tapered end of OprM and the digestion at these residues dilates this portion. However, what triggers the opening of the OprM channel during the MexAB- OprM pump transporting drugs into the extracellular environment? The digestion profile of OprM in the cellls or the liposome membranes by a-chymtrypsin showed that tyrosine 185 and 196 face into the periplasm and probably interact with other component(s). Furthermore, MexA is thought to link the outer channel with an inner membrane component. Accordingly, it is considered that the interaction of MexA with this portion of OprM may trigger the opening of this channel. However, we cannot exclude the possibility of MexB as a partner of OprM because this protein possesses two large loops extruding into the periplasm and so can contact with OprM [39]. On the contrary, Zhao et al. [40] reported that a TonB homologue in P. aeruginosa influenced multidrug resist- ance, suggesting interaction between the Ton B homologue and OprM channel. Therefore, the Ton B homologue may contribute to the triggering of the channel opening. To the best of our knowledge, this is the first biochemical study demonstrating that the outer membrane component of the multidrug efflux pumps works as a gated channel. Our study reported here may help to understand the molecular function of the multidrug efflux pumps. How- ever, many issues still remain to be clarified, including how protein-to-protein interaction triggers the opening of the OprM channel. ACKNOWLEDGEMENTS This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture, the Ministry of Health, Labour and Welfare for Antibiotic Resistance Research Project, the Japan Association for the Promotion of Science, and Tokai University School of Medicine, Project Research. REFERENCES 1. Brown, M.R.W. (1975) Resistance of Pseudomonas Aeruginosa. John Wiley and Sons, London. 2. Ma, D., Cook, D.N., Hearst, J.E. & Nikaido, H. (1994) Efflux pumps and drug resistance in gram-negative bacteria. Trends Microbiol. 2, 489–493. 3. Nikaido, H. (1994) Prevention of drug access to bacterial targets: permeability barriers and active efflux. Science 264, 382–388. 4. Nikaido, H. (1996) Multidrug efflux pumps of gram-negative bacteria. J.Bacteriol. 178, 5853–5859. 5. Poole, K. (2000) Efflux-mediated resistance to fluoroquinolones in gram-negative bacteria. Antimicrob. Agents Chemother. 44, 2233– 2241. 6. Hancock, R.E. (1997) The bacterial outer membrane as a drug barrier. Trends Microbiol. 5, 37–42. 7. Zgurskaya, H.I. & Nikaido, H. (2000) Multidrug resistance mechanisms: drug efflux across two membranes. Mol. Microbiol. 37, 219–225. 8. Gotoh, N., Tsujimoto, H., Poole, K., Yamagishi, J. & Nishino, T. (1995) The outer membrane protein OprM of Pseudomonas aer- uginosa is encoded by oprK of the mexA-mexB-oprK multidrug resistance operon. Antimicrob. Agents Chemother. 39, 2567–2569. 9. Poole, K., Krebes, K., McNally, C. & Neshat, S. (1993) Multiple antibiotic resistance in Pseudomonas aeruginosa: evidence for involvement of an efflux operon. J.Bacteriol. 175, 7363–7372. 10. Poole, K., Heinrichs, D.E. & Neshat, S. (1993) Cloning and sequence analysis of an EnvCD homologue in Pseudomonas aeruginosa: regulation by iron and possible involvement in the secretion of the siderophore pyoverdine. Mol. Microbiol. 10,529– 544. 11. Kohler, T., Kok, M., Michea-Hamzehpour, M., Plesiat, P., Gotoh, N., Nishino, T., Curty, L.K. & Pechere, J.C. (1996) Multidrug efflux in intrinsic resistance to trimethoprim and sul- famethoxazole in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 40, 2288–2290. 4744 E. Yoshihara et al. (Eur. J. Biochem. 269) Ó FEBS 2002 12. Li, X.Z., Nikaido, H. & Poole, K. (1995) Role of mexA-mexB- oprM in antibiotic efflux in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 39, 1948–1953. 13. Li, X.Z., Zhang, L., Srikumar, R. & Poole, K. (1998) b-Lactamase inhibitors are substrates for the multidrug efflux pumps of Pseu- domonas aeruginosa. Antimicrob. Agents Chemother. 42, 399–403. 14. Kohler, T., Michea-Hamzehpour, M., Henze, U., Gotoh, N., Curty, L.K. & Pechere, J.C. (1997) Characterization of MexE- MexF-OprN, a positively regulated multidrug efflux system of Pseudomonas aeruginosa. Mol. Microbiol. 23, 345–354. 15. Poole, K., Gotoh, N., Tsujimoto, H., Zhao, Q., Wada, A., Yamasaki, T., Neshat, S., Yamagishi, J., Li, X.Z. & Nishino, T. (1996) Overexpression of the mexC-mexD-oprJ efflux operon in nfxB-type multidrug-resistant strains of Pseudomonas aeruginosa. Mol. Microbiol. 21, 713–724. 16. Aires, J.R., Kohler, T., Nikaido, H. & Plesiat, P. (1999) Involve- ment of an active efflux system in the natural resistance of Pseu- domonas aeruginosa to aminoglycosides. Antimicrob. Agents Chemother. 43, 2624–2628. 17. Masuda, N., Sakagawa, E., Ohya, S., Gotoh, N., Tsujimoto, H. & Nishino, T. (2000) Contribution of the MexX-MexY-oprM efflux system to intrinsic resistance in Pseudomonas aeruginosa. Anti- microb. Agents Chemother. 44, 2242–2246. 18. Mine, T., Morita, Y., Kataoka, A., Mizushima, T. & Tsuchiya, T. (1999) Expression in Escherichia coli of a new multidrug efflux pump, MexXY, from Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 43, 415–417. 19. Paulsen, I.T., Park, J.H., Choi, P.S. & Saier, M.H. Jr (1997) A family of gram-negative bacterial outer membrane factors that function in the export of proteins, carbohydrates, drugs and heavy metals from gram-negative bacteria. FEMS Microbiol. Lett. 156, 1–8. 20. Wong, K.K., Brinkman, F.S., Benz, R.S. & Hancock, R.E. (2001) Evaluation of a structural model of Pseudomonas aeruginosa outer membrane protein OprM, an efflux component involved in intrinsic antibiotic resistance. J. Bacteriol. 183, 367–374. 21. Li, X.Z. & Poole, K. (2001) Mutational analysis of the OprM outer membrane component of the MexA-MexB-OprM multi- drug efflux system of Pseudomonas aeruginosa. J.Bacteriol. 183, 12–27. 22. Yoneyama, H., Ocaktan, A., Tsuda, M. & Nakae, T. (1997) The role of mex-gene products in antibiotic extrusion in Pseudomonas aeruginosa. Biochem. Biophys. Res. Commun. 233, 611–618. 23. Srikumar, R., Li, X.Z. & Poole, K. (1997) Inner membrane efflux components are responsible for b-lactam specificity of multidrug efflux pumps in Pseudomonas aeruginosa. J. Bacteriol. 179, 7875– 7881. 24. Zgurskaya, H.I. & Nikaido, H. (1999) Bypassing the periplasm: reconstitution of the AcrAB multidrug efflux pump of Escherichia coli. Proc.NatlAcad.Sci.USA96, 7190–7195. 25. Zgurskaya, H.I. & Nikaido, H. (1999) AcrA is a highly asym- metric protein capable of spanning the periplasm. J. Mol. Biol. 285, 409–420. 26. Wandersman, C. & Delepelaire, P. (1990) TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc. Natl Acad. Sci. USA 87, 4776–4780. 27. Thanabalu, T., Koronakis, E., Hughes, C. & Koronakis, V. (1998) Substrate-induced assembly of a contiguous channel for protein export from E.coli: reversible bridging of an inner-membrane translocase to an outer membrane exit pore. EMBO J. 17, 6487– 6496. 28. Benz, R., Maier, E. & Gentschev, I. (1993) TolC of Escherichia coli functions as an outer membrane channel. Zentralbl. Bakteriol. 278, 187–196. 29. Koronakis, V., Sharff, A., Koronakis, E., Luisi, B. & Hughes, C. (2000) Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export. Nature 405, 914– 919. 30. Yoshihara, E. & Nakae, T. (1989) Identification of porins in the outer membrane of Pseudomonas aeruginosa that form small dif- fusion pores. J. Biol. Chem. 264, 6297–6301. 31. Yoshihara, E., Yoneyama, H. & Nakae, T. (1991) In vitro assembly of the functional porin trimer from dissociated mono- mers in Pseudomonas aeruginosa. J. Biol Chem. 266, 952–957. 32. Yoshihara, E., Gotoh, N. & Nakae, T. (1988) In vitro demon- stration by the rate assay of the presence of small pore in the outer membrane of Pseudomonas aeruginosa. Biochem. Biophys. Res. Commun. 156, 470–476. 33. Yoneyama,H.,Ocaktan,A.,Gotoh,N.,Nishino,T.&Nakae,T. (1998) Subunit swapping in the Mex-extrusion pumps in Pseudo- monas aeruginosa. Biochem. Biophys. Res. Commun. 244, 898–902. 34. Nakajima, A., Sugimoto, Y., Yoneyama, H. & Nakae, T. (2000) Localization of the outer membrane subunit OprM of resistance- nodulation-cell division family multicomponent efflux pump in Pseudomonas aeruginosa. J. Biol. Chem. 275, 30064–30068. 35. Wong, K.K. & Hancock, R.E. (2000) Insertion mutagenesis and membrane topology model of the Pseudomonas aeruginosa outer membrane protein OprM. J. Bacteriol. 182, 2402–2410. 36. Solaro, C.R. & Lingle, C.J. (1992) Trypsin-sensitive, rapid inactivation of a calcium-activated potassium channel. Science 257, 1694–1698. 37. Hoshi, T., Zagotta, W.N. & Aldrich, R.W. (1990) Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science 250, 533–538. 38. Zagotta, W.N., Hoshi, T. & Aldrich, R.W. (1990) Restoration of inactivation in mutants of Shaker potassium channels by a peptide derived from ShB. Science 250, 568–571. 39. Guan, L., Ehrmann, M., Yoneyama, H. & Nakae, T. (1999) Membrane topology of the xenobiotic-exporting subunit, MexB, of the MexA,B-OprM extrusion pump in Pseudomonas aeruginosa. J. Biol. Chem. 274, 10517–10522. 40. Zhao, Q., Li, X.Z., Mistry, A., Srikumar, R., Zhang, L., Lomovskaya, O. & Poole, K. (1998) Influence of the TonB energy- coupling protein on efflux-mediated multidrug resistance in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 42, 2225–2231. Ó FEBS 2002 Gating of the OprM channel of MexAB-OprM pump in P. aeruginosa (Eur. J. Biochem. 269) 4745 . The outer membrane component of the multidrug efflux pump from Pseudomonas aeruginosa may be a gated channel Eisaku Yoshihara, Hideaki Maseda and Kohjiro. have been reported and the calcium-activated potassium channel from rat adrenal chromaffin cells is one of them. Solaro et al. [36] have shown that the activation