Functional characterization of Pseudomonas fluorescens OprE and OprQ membrane proteins

BBRC Biochemical and Biophysical Research Communications 346 (2006) 1048–1052 www.elsevier.com/locate/ybbrc

Functional characterization of Pseudomonas fluorescens OprE and OprQ membrane proteins Thomas Jaouen a, Laurent Coquet b, Laure Marvin-Guy c, Nicole Orange Sylvie Chevalier a,*,1, Emmanuelle De´ b,1 a

a,*

,

Laboratoire de Microbiologie du Froid, UPRES 2123, Universite´ de Rouen, 55 rue Saint Germain, 27000 Evreux, France b UMR 6522 CNRS, IFRMP, Universite´ de Rouen, 76821 Mont Saint Aignan, France c Nestle´ Research Center, Nestec Ltd., Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland Received 31 May 2006 Available online 9 June 2006

Abstract Outer membrane (OM) proteins of the OprD family may enable bacteria of the genus Pseudomonas to adapt to various environments by modulating OM permeability. The OprE and OprQ porins from P. fluorescens strain MF0 were purified and identified by MALDITOF mass spectrometry and N-terminal and internal microsequencing. These proteins, when reconstituted in an artificial planar lipid bilayer, induced similar ion channels with low single-conductance values. Secondary structure prediction of both proteins showed similar folding patterns into a 16 transmembrane b-strands barrel but a highly variable amino-acid composition and length for their putative external loops implicated in porin function. Both proteins were overexpressed under poor oxygenation conditions, but not by using several amino acids as sole carbon source, indicating a different specificity for these proteins compared to the paradigm of this protein family, OprD. Ó 2006 Elsevier Inc. All rights reserved. Keywords: Pseudomonas; Porins; OprD; OprE; OprQ; Single channel conductance value; Expression

Bacteria from the genus Pseudomonas are present in all major natural environments, including water, soil, the plant rhizosphere, and the human body. Some of these environments are subjected to rapid variations, necessitating high levels of physiologic and genetic adaptability in bacteria living in such hostile habitats [1]. This adaptability involves many modifications, some of which affect outer membrane (OM) permeability. This permeability essentially depends on channel-forming proteins, which can be divided into three classes: general porins, specific porins, and highly substrate-specific gated channels translocating their ligands in an energy-dependent manner [2]. OM permeability is modified by changes to porins in two main ways: (1) modulation of porin channel size, as reported for OprF, the major non-specific porin of psychrotrophic *

1

Corresponding author. Fax: +33 02 32 29 15 66. E-mail address: [email protected] (N. Orange). These authors contributed equally to this study.

0006-291X/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2006.06.013

and mesophilic Pseudomonas species [3], and (2) differential expression of porins with narrow channels to decrease OM permeability [4], which seems to be particularly common for specific porins, such as OprD [5–7]. This porin OprD, which binds basic amino acids, is the paradigm for a 19-member family of OM proteins in P. aeruginosa [8]. These proteins, displaying growth condition-regulated production, seem to be specific for different classes of substrates, but may also act as non-specific porins for small substrates [7,9]. Member of this family, OprE, has been reported to be induced in anaerobic conditions in P. aeruginosa [10] and to form pores in vitro excluding molecules larger in size than di- or trisaccharides [11]. Little is known about OprQ, which may allow saccharide diffusion [12]. In the psychrotrophic P. fluorescens strain MF0, OprE, and OprQ were found to be overexpressed in an adaptive OprF-deficient mutant, presumably to compensate for the loss of this major non-specific porin [13]. This species, closely related to P. aeruginosa, colonizes many

T. Jaouen et al. / Biochemical and Biophysical Research Communications 346 (2006) 1048–1052

cold environments and was recently shown to be a potential opportunistic pathogen in humans, as it can adhere to nerve cells and trigger apoptosis [14,15]. OprE and OprQ have been shown to bind fibronectin, suggesting that they may be involved in virulence [16]. We therefore characterized these two proteins, investigating their structures and functions in P. fluorescens. Materials and methods Bacterial strains and growth conditions. Pseudomonas fluorescens MF0 was grown in nutrient broth (Difco), or on nutrient broth agar (1.5% w/v) plates, at 28 °C, under normal (200 mL culture in 2 L erlenmeyer) or poor (500 mL in 2 L) oxygenation. Cells were harvested in the late exponential growth phase by centrifugation at 8000g for 10 min at 4 °C. OM protein purification and characterization. OMs were isolated by the spheroplast procedure described in De´ et al. [17]. OprE and OprQ were purified according to Jaouen et al. [3]. The purity of the preparations was checked by SDS–PAGE and silver staining. Purified porins were concentrated by trichloroacetic acid/acetone precipitation and dissolved in 0.25 M Tris–HCl, 6 M guanidine. Samples were reduced by adding 10 mM dithiothreitol and alkylated with 50 mM iodoacetamide, and then were digested overnight at 37 °C with trypsin to a final protease/protein ratio of 1:5. The resulting peptides were pooled by lyophilization. The peptide mixture was dissolved in trifluoroacetic acid/acetonitrile/water (1:20:79, v/ v/v), separated by capillary reverse-phase HPLC (Applied Biosystems), and subjected to Edman degradation in an Applied Biosystems 492 automated protein sequencer. For proteins’ N-terminal sequences, porins were blotted onto Prosorb (Applied Biosystems) before Edman degradation. Mass spectra were recorded on an Autoflex (Bruker) MALDI timeof-flight mass spectrometer in the positive-ion mode. The acceleration voltage was set to 20 kV. Samples were prepared according to El Hamel et al. [18]. Reconstitution in planar lipid bilayers. Virtually solvent-free planar lipid bilayers were formed by the Montal and Muller technique [19], as described by Jaouen et al. [3]. Macroscopic current and selectivity experiments were carried out as described in De´ et al. [20]. A palmitoyloleoylphosphatidylcholine/dioleoylphosphoethanolamine solution (POPC/DOPE, 7/3, Lipid Products) at 0.5% (w/v) in hexane was used as lipid. General DNA procedures. Standard methods were used for general DNA procedures [21]. The porin genes were amplified with the following oligonucleotide primers: oprE, 5 0 (5 0 -CTGCAGGGACCAAATCAACT-3 0 ), 3 0 (5 0 -GAAGGCCTTGAGAGCTGTA-3 0 ); oprQ, 5 0 (5 0 -CGATTCACGAA ATTGACACG-3 0 ), 3 0 (5 0 -CAGAACACGTTGATCGGGTA-3 0 ), designed on the basis of a ClustalW alignment of the putative oprQ and oprE genes from the sequenced P. fluorescens PF5, SBW25, and PF0 genomes. Polymerase chain reaction (PCR) was carried out with 10 ng of purified chromosome and 1 U Taq DNA polymerase (Roche): after a hot start at 94 °C for 5 min, the DNA was subjected to 30 cycles of denaturation for 1 min at 94 °C, primer annealing for 1 min at 60 °C, and elongation for 2 min at 72 °C, followed by a final 10-min elongation step. Sequences of OprE and OprQ, deposited in the EMBL database under Accession Nos. AJ866545 and AJ866544, respectively, were analyzed with BlastP and ClustalW for identification and multiple protein sequence alignments (http://www.infobiogen.fr). Total RNAs were extracted using the RNA isolation kit (Roche Diagnostics) as described by the manufacturer. RNA integrity was checked by gel electrophoresis and RNA concentration was determined by OD260. RT-PCR experiments were achieved on 10 ng of total RNAs using the one-step RT-PCR procedure (Roche Diagnostics). Absence of contaminating DNA was checked by making a PCR directly on the RNA extracts. Primers used in RT-PCR experiments were designed using a ClustalW alignment of OprD, OprE, and OprQ, ensuring an amplification of the only searched gene. The following primers: oprQ5 0 , ACAAG AACGGCAAGCAAGAC; oprQ3 0 , CGGAACTGATCACCGTACT;

1049

oprE5 0 , GCCGGTGGTGACTACAAGAT; oprE3 0 , AGGCCGCTGTAC AGATTGTT; 16S5 0 , GGGGGTAGAATTCCAGGTGT; and 16S3 0 , CGGCAGTCTCCTTTGAGTTC were used in the same conditions as described above.

Results and discussion Pseudomonas fluorescens MF0 is a psychrotrophic bacterium and potential opportunistic pathogen [15]. As OprE and OprQ have been reported to bind fibronectin [16], we characterized these two proteins, which are candidates for involvement in virulence. Purification and identification of OprE and OprQ Pseudomonas fluorescens MF0 was grown at 28 °C in conditions of poor oxygenation to induce overexpression of OprE [10]. Omps were isolated by the spheroplast procedure and analyzed by SDS–PAGE (Fig. 1-1). In these growth conditions, two major Omps are localized in the 45–46 kDa range (A, B in Fig. 1-1) which may include members of the large OprD family in P. aeruginosa, such as OprE and OprQ [8]. These two proteins were purified by preparative electrophoresis followed by electroelution in the non-ionic detergent Triton X-100 (Fig. 1-2), and their molecular masses (45,641 Da for A and 46884 Da for B) were determined by MALDI-TOF spectrometry (Fig. 1-3). The proteins were identified by Edman degradation sequencing of the N-terminus of the entire protein and of internal trypsin digestion peptides. In all cases, homogeneous sequencing results free of contaminating sequences were obtained, confirming the purity of our samples. BlastP searches gave high similarity scores with proteins of the large OprD family. The N-terminal sequence of the 45 kDa (A) protein (NDQDQSKGFIEDSH) was similar to those of OprQ and OprD from P. aeruginosa PAO1 (93% and 86%, respectively), P. fluorescens strain PF5 (79% and 93%), strain SBW25 (100% and 79%), and strain PF0 (86% and 71%). We then sequenced the N-terminal region of three internal peptides obtained by trypsin digestion (YGDQFPAVPV, LESGFTQGTV, and NAGGIGDGGN), to confirm the identity of this protein. These three peptides were 70–100%, 80–100%, and 50–100% similar to OprQ and only 40–60%, 60–90%, and 30–50% similar to OprD from P. aeruginosa PAO1, P. fluorescens strains PF5, SBW25, and PF0, respectively. The same procedure was applied to the 46 kDa protein. The N-terminal sequences of the entire 46 kDa protein (AGFVEDS KATLGLR) and of the three internal peptides (LPVIITNDGR, DLTLVGGQIEK, and VGVPGLTAPV) were clearly similar to those of the OprE protein (93%, 80%, 91%, and 80% similarity) of P. aeruginosa, P. fluorescens strains PF5 (100%, 80%, 82%, and 90%), SBW25 (100%, 100%, 91%, and 90%), and PF0 (100%, 100%, 91%, and 90%). Thus, the 45 kDa (A) protein was clearly identified as the homolog of OprQ, and the 46 kDa (B) protein as the homolog of OprE in P. fluorescens MF0.

1050

T. Jaouen et al. / Biochemical and Biophysical Research Communications 346 (2006) 1048–1052

(1)

(2)

MM

MM

A

B

94 94 67

67 43

→

B 43

A 30

→

(3) A

45641

B

46884

20

4

5

6 × 104 m/z

4

6 × 104 m/z

5

Fig. 1. (1) Outer membrane proteins profile of P. fluorescens MF0. SDS–PAGE (12%), followed by Coomassie blue staining. MM, molecular weight markers. The position of the OprQ (A) and OprE (B) proteins are indicated by arrows; (2) SDS–PAGE and (3) MALDI-TOF spectra of the purified OprQ (A) and OprE (B) proteins.

Functional analysis We characterized the ionophore properties of the OprE and OprQ proteins from P. fluorescens by reconstituting the purified proteins in POPC/DOPE planar lipid bilayers. After addition of the proteins to the cis compartment of the measurement cell and the application of voltage, the current fluctuated, indicating channel formation and reflecting the open and closed states of the OprE and OprQ channels. No substrates were observed in these transitions (see the associated amplitude histograms in Fig. 2). Several experiments allowed to determine the major single-channel conductance value of (30 ± 4) pS for OprQ (Fig. 2A, n = 15) and of (25 ± 4) pS for OprE (Fig. 2B, n = 15), possibly corresponding to the incorporation of a single monomer.

These results are consistent with reported conductance values for the paradigm protein of the same family, the OprD monomer of P. aeruginosa, which presents a single-channel conductance of 20 pS in 1 M KCl [22,23] or 30 pS in 1 M NaCl [24]. During reconstitution experiments, higher conductance states of 50, 75, and 150 pS could be detected, suggesting insertion into the membrane of oligomeric forms of the proteins. Ion selectivity was investigated by setting up an NaCl gradient across the bilayer. We determined the zero current potential and then calculated permeability ratios (PNa/PCl) by applying the Goldman–Hodkin–Katz equation [25]. The ratios obtained for both proteins were 2.3 ± 0.3, indicating poor cationic selectivity. A similar selectivity was calculated for P. aeruginosa OprD [22], suggesting that Counts

A

3000

c o

2000 1000

4 pA

0 0

500 ms

B

20

40

60

G (pS) Counts 3000

c o

2000 1000

4 pA

0

500 ms

0

20

40

60

G (pS)

Fig. 2. Ionophore behavior of the P. fluorescens MF0 proteins reincorporated into POPC/DOPE planar lipid bilayers. Recordings and their associated amplitude histograms present the major single-channel conductance of OprQ (A) and OprE (B) at an applied voltage of 100 mV in 1 M NaCl, 10 mM Hepes, pH 7.4. ‘‘c’’ and ‘‘o’’ for closed state and open state.

T. Jaouen et al. / Biochemical and Biophysical Research Communications 346 (2006) 1048–1052

these proteins may have similar net charges at the entrance to the pore or at the pore eyelet [26,27]. As the general ionophore properties of OprE and OprQ were similar to those of the OprD of P. aeruginosa PAO1, we investigated the effect of OprD-specific substrates (arginine and imipenem) on the OprE and OprQ channels. These molecules are known to induce closure of the OprD channel at micromolar concentrations [22,28]. In single and macroscopic current experiments, these molecules had no effect on the OprE and OprQ channels, and could therefore not be considered as the specific substrates for these two porins. Gene sequencing and deduced amino-acid sequence analysis We assessed the structural similarity of these OM proteins to other porins, by amplifying the oprE and oprQ genes of P. fluorescens MF0 by PCR, sequencing them and carrying out computer-based sequence analysis. A topological model has been proposed for the P. aeruginosa OprD porin and successfully tested by means of PCR-mediated site-specific deletions [9,22,23]. According to this model, the porin subunit consists of 16 transmembrane b-strands with long loops (L1–L8) on the external side of the membrane and short turns facing the periplasmic space. Based on the high similarities in the size and sequences of members of the OprD family (46–58% similarity in P. aeruginosa), Hancock and Brinkman [8] suggested that this model could be applied to all these paralogous proteins, including OprE and OprQ. We have therefore made an alignment of the sequences of OprE, OprQ (from P. fluorescens and P. aeruginosa), and OprD (P. aeruginosa) which confirms that the studied proteins have structures consistent with the proposed 16 b-strand-model (data not shown). As expected, the most variable parts of the sequences are the external loops L2, L3, L5, L7, and L8, which have been shown to be directly involved in P. aeruginosa OprD function. A ClustalW alignment of the loops of the P. fluoresPo r i n

1051

cens OprQ (OprE) and P. aeruginosa OprQ (OprE), and OprD is shown in Fig. 3A and B. Mutagenesis experiments suggested that loops L5, L7, and L8 may be involved in channel constriction [22,23]. But, as the ionophore properties (selectivity and conductance values) of the two studied porins are similar to those of OprD, one must assume that the sequence variations observed here in these particular loops may not modify the pore constriction. Concerning loops L2 and L3, which are required for the binding of basic amino acids and imipenem in the OprD [22,23], sequence variations appear to have a more drastic effect, since the specific substrates of P. aeruginosa PAO1 OprD (i.e., arginine and imipenem) could not use the OprE and OprQ channels of P. fluorescens, as demonstrated by the functional studies. RT-PCR experiments We investigated then the expression of OprE and OprQ by semi-quantitative RT-PCR experiments on total RNA extracted from cells cultured in minimal medium. The use, as the sole carbon source, of a 20 mM solution of an amino acid (arginine, glutamate, alanine, and histidine) known to induce expression of the P. aeruginosa PAO1 OprD [5] had no effect on the level of oprE and oprQ transcripts (data not shown). However, as shown in Fig. 4, both genes were clearly overproduced in poorly oxygenated cultures, whereas the 16S rDNA gene, used as control, was

1

2

oprQ oprE 16S rDNA

Fig. 4. RT-PCR of oprQ, oprE, and 16S genes. Cells were grown under normal (1) or poor oxygenation (2) and harvested in the late log growth phase.

Presumed function

Loops L2

L3

OprD Pa

KLDGTSDKTGTGNLPVMNDGK-------------PRDDYSRAGGAVKVRISKT

PQTA TGFQ LQS SEFEG LDLE AG H FTEGK EPTTVKSRGELYA-- --- TYAGET

OprQ Pa OprQ Pf

RLDGGKGRSGAAGIDFFKQGD----------SGSAADDLSKGGAAVKFRISNT RLDGGKGRAGAGGIDFFKQGNGTTNPDGSNNPGSAPHDLAKGGAAVKFRVSNT

PESYS GTL ITSKE IE GLELN AG RFTAESRKSA EGR DS------ -- ----- GG SETYT GTSI VSK EIAG LLLDA G HFTKEAR KSMESTY S-- ----- --- ---GR

OprE Pa OprE Pf

RLDGGG-RAGKSG-----LDRQPGTVFPLESNGEPVHDFASLGLTAKAKVSNT RLDSGGGTNGATKPSGANPGSYGGTIFPSEHNGKAVNDYSSLGLTAKAKISQT

PVTFEGGQVTSTDLKDFTLVAGQLEHSKGRNSTDNRSLSIAGANGSSASSRD PQTFQ GGQI TTNDI KDL TLVG GQIEK AKG RNSSNNE NLSLS GAN SRG ATWRD

L5

L7

Specificity

L8

OprD Pa OprQ Pa OprQ Pf OprE Pa OprE Pf

RTNDEGK--------------------AKAGDISNT

P GL TFM VRYI N GK DI DGT KMSDN NVGYKN YG YGEDG KH HE

HRAN--ADQGEG---DQNEF

RTKLDSDFADQN----------------FNGNRDNK KTKLDRSFALENQ--------------KDADARDNK

PGLTYRVAYVRGDNIKTAETS-------------NGKERE PGLTYNVAYVRGTNIDDGTNRG------------DGTERE

LRVSNDARSYND---DGNEI LRVSNNADQYNV---GGNEL

DSSSDGKNGSRSGRADGYVSSGYYGSGVTKGEVDNR NSSDDGANGNES---AYFTTGNYQGFRAGKGKVDNN

PGLTFNTIYLSGDKI KTA R------- -------G DQ SEWE PGLTAGVVYLKGSDIDTVANTASRTQF------NGQSEWE

FRSGLPAAGSSNNQRDQDEN WRTDIAGAR------DQDEN

Constriction

Fig. 3. ClustalW alignment of the external loops sequences of OprQ and OprE from P. fluorescens MF0 (Pf) and P. aeruginosa PAO1 (Pa) with the model protein OprD from P. aeruginosa. The predicted functions of these loops (specificity or constriction) are indicated, according to the OprD topological model. Identical residues are black shaded and similar ones are light grey shaded.

1052

T. Jaouen et al. / Biochemical and Biophysical Research Communications 346 (2006) 1048–1052

constitutively expressed whatever the conditions used. Such a result has also been described for P. aeruginosa OprE [10], but it is the first time that an OprQ homolog is also involved in growth under oxygenation stress. Conclusion We have characterized, for the first time, the OprE and OprQ proteins belonging to the OprD family in a psychrotrophic P. fluorescens strain. Structurally, they present a similar 16 b-barrel-fold predicted for all members of this family. Despite the high level of sequence variation in the long external loops, which are thought to determine pore characteristics, the ionophore properties of these proteins are similar to the OprD ones. But, these loops interact directly with the external medium, and sequence variations probably reflect the specific adaptation of a Pseudomonas strain to a particular environment, conserved in the genome under selection pressure. The OprD-family proteins are probably functional for a wide range of nutrients, as illustrated by the great variability in L2 and L3, even among members of the same species. By regulating the production of these proteins as a function of growth conditions and the nature of carbon or nitrogen sources, Pseudomonas species may increase its chances of adapting to various environmental conditions. This regulation may reflect the metabolic versatility and opportunism of this bacterial genus. References [1] A.J. Spiers, A. Buckling, P.B. Rainey, The causes of Pseudomonas diversity, Microbiology 146 (2000) 2345–2350. [2] H. Nikaido, Molecular basis of bacterial outer membrane permeability revisited, Microbiol. Mol. Biol. Rev. 67 (2003) 593–656. [3] T. Jaouen, E. De´, S. Chevalier, N. Orange, Pore size dependence on growth temperature is a common characteristic of the major outer membrane protein OprF in psychrotrophic and mesophilic Pseudomonas, Appl. Environ. Microbiol. 70 (2004) 6665–6669. [4] E. Yoshihara, T. Nakae, Identification of porins in the outer membrane of Pseudomonas aeruginosa that form small diffusion pores, J. Biol. Chem. 264 (1989) 6297–6301. [5] M.M. Ochs, C.D. Lu, R.E. Hancock, A.T. Abdelal, Amino acidmediated induction of the basic amino acid-specific outer membrane porin OprD from Pseudomonas aeruginosa, J. Bacteriol. 181 (1999) 5426–5432. [6] M.M. Ochs, M.P. McCusker, M. Bains, R.E. Hancock, Negative regulation of the Pseudomonas aeruginosa outer membrane porin OprD selective for imipenem and basic amino acids, Antimicrob. Agents Chemother. 43 (1999) 1085–1090. [7] S. Tamber, M.M. Ochs, R.E. Hancock, Role of the novel OprD family of porins in nutrient uptake in Pseudomonas aeruginosa, J. Bacteriol. 188 (2006) 45–54. [8] R.E. Hancock, F.S. Brinkman, Function of Pseudomonas porins in uptake and efflux, Annu. Rev. Microbiol. 56 (2002) 17–38. [9] H. Huang, D. Jeanteur, F. Pattus, R.E. Hancock, Membrane topology and site-specific mutagenesis of Pseudomonas aeruginosa porin OprD, Mol. Microbiol. 16 (1995) 931–941.

[10] Y. Yamano, T. Nishikawa, Y. Komatsu, Cloning and nucleotide sequence of anaerobically induced porin protein E1 (OprE) of Pseudomonas aeruginosa PAO1, Mol. Microbiol. 8 (1993) 993–1004. [11] M. Obara, T. Nakae, Porin of Pseudomonas aeruginosa forms low conductance ion channel in planar lipid bilayers, Biochem. Biophys. Res. Commun. 186 (1992) 645–651. [12] K. Okamoto, N. Gotoh, H. Tsujimoto, H. Yamada, E. Yoshihara, T. Nakae, T. Nishino, Molecular cloning and characterization of the oprQ gene coding for outer membrane protein OprQ of Pseudomonas aeruginosa, Microbiol. Immunol. 43 (1999) 297–301. [13] S. Chevalier, J.F. Burini, M.A. Freulet-Marriere, C. Regeard, G. Schoofs, J. Guespin-Michel, R. De Mot, N. Orange, Characterization of an OprF-deficient mutant suggests that OprF is an essential protein for Pseudomonas fluorescens strain MF0, Res. Microbiol. 151 (2000) 619–627. [14] L. Picot, S. Mezghani-Abdelmoula, A. Merieau, P. Leroux, L. Cazin, N. Orange, M.G.J. Feuilloley, Pseudomonas fluorescens as a potential pathogen: adherence to nerve cells, Microbes Infect. 3 (2001) 985–990. [15] M.G.J. Feuilloley, S. Mezghani-Abdelmoula, L. Picot, O. Lesouhaitier, A. Me´rieau, J. Gue´rillon, N. Boujedaini, L. Cazin, N. Orange, Involvement of Pseudomonas and related species in central nervous system infections, Recent Res. Dev. Microbiol. 7 (2001) 55–71. [16] J. Rebiere-Huet, J. Guerillon, A.L. Pimenta, P. Di Martino, N. Orange, C. Hulen, Porins of Pseudomonas fluorescens MFO as fibronectin-binding proteins, FEMS Microbiol. Lett. 215 (2002) 121–126. [17] E. De´, R. De Mot, N. Orange, N. Saint, G. Molle, Channel-forming properties and structural homology of major outer membrane proteins from Pseudomonas fluorescens MFO and OE 28.3, FEMS Microbiol. Lett. 127 (1995) 267–272. [18] C. El Hamel, S. Chevalier, E. De, N. Orange, G. Molle, Isolation and characterisation of the major outer membrane protein of Erwinia carotovora, Biochim. Biophys. Acta 1515 (2001) 12–22. [19] M. Montal, P. Mueller, Formation of bimolecular membranes from monolayers and study of their properties, Proc. Natl. Acad. Sci. USA 69 (1972) 3561–3566. [20] E. De´, M. Jullien, G. Labesse, J.M. Pages, G. Molle, J.M. Bolla, MOMP (major outer membrane protein) of Campylobacter jejuni: a versatile pore-forming protein, FEBS Lett. 469 (2000) 93–97. [21] J.E.F. Sambrook, T. Maniatis, Molecular Cloning: A Laboratory Manual, second ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1997. [22] H. Huang, R.E. Hancock, The role of specific surface loop regions in determining the function of the imipenem-specific pore protein OprD of Pseudomonas aeruginosa, J. Bacteriol. 178 (1996) 3085–3090. [23] M.M. Ochs, M. Bains, R.E. Hancock, Role of putative loops 2 and 3 in imipenem passage through the specific porin OprD of Pseudomonas aeruginosa, Antimicrob. Agents Chemother. 44 (2000) 1983–1985. [24] J. Ishii, T. Nakae, Lipopolysaccharide promoted opening of the porin channel, FEBS Lett. 320 (1993) 251–255. [25] B. Hille, Selective permeability: independance, in ‘‘Ionic channels of excitable membranes’’, Sinauer Associates Inc., Sunderland, MA, 1984, pp. 226–248. [26] A. Karshikoff, V. Spassov, S.W. Cowan, R. Ladenstein, T. Schirmer, Electrostatic properties of two porin channels from Escherichia coli, J. Mol. Biol. 240 (1994) 372–384. [27] E. De´, A. Basle, M. Jaquinod, N. Saint, M. Mallea, G. Molle, J.M. Pages, A new mechanism of antibiotic resistance in Enterobacteriaceae induced by a structural modification of the major porin, Mol. Microbiol. 41 (2001) 189–198. [28] J. Ishii, T. Nakae, Specific interaction of the protein-D2 porin of Pseudomonas aeruginosa with antibiotics, FEMS Microbiol. Lett. 136 (1996) 85–90.