REVIEW. Extended-spectrum b-lactamases and the permeability barrier

Transcription

1 REVIEW Extended-spectrum b-lactamases and the permeability barrier L. Martínez-Martínez Service of Microbiology, University Hospital Marqués de Valdecilla, Santander, Spain ABSTRACT The outer membrane of Gram-negative bacteria represents a barrier for penetration of hydrophilic compounds. Loss of porins (water-filled protein channels) contributes to antimicrobial resistance, particularly when additional mechanisms of resistance are expressed. Many studies on the structure and regulation of porins in Escherichia coli K-12 are available, but there is little information concerning clinical isolates of this species. In Klebsiella pneumoniae, two major porins, OmpK35 and OmpK36, are produced, but many extended-spectrum b-lactamase (ESBL)-producing K. pneumoniae isolates do not express OmpK35. Loss of both OmpK35 and OmpK36 in ESBL-producing K. pneumoniae causes resistance to cefoxitin, increased resistance to expanded-spectrum cephalosporins, and decreased susceptibility to carbapenems, particularly ertapenem. Porin loss also decreases the susceptibility to other non-b-lactam compounds, such as fluoroquinolones, of ESBL-producing organisms. Keywords Enterobacteriaceae, extended-spectrum b-lactamases, outer membrane, permeability, porin, review Clin Microbiol Infect 2008; 14 (Suppl. 1): INTRODUCTION The cell envelope of Gram-negative bacteria is composed of two different membranes, separated by the periplasm space, where the peptidoglycan layer and b-lactamases are located [1]. The lipid composition of the inner membrane is similar to that of most biological membranes, including the cytoplasmic membrane of Gram-positive organisms, but the outer membrane is rather atypical, as it is an asymmetrical bilayer of lipopolysaccharide (outer leaflet) and phospholipids (inner leaflet). The outer membrane contains a relatively high proportion of saturated fatty acids and phosphatidylethanolamine, and its permeability to hydrophobic agents is reduced because of the interaction of the unsaturated fatty-acid chains of lipopolysaccharide molecules, stabilised by divalent cations, which leads to a more compact structure than if the membrane was made of only phospholipids. This reduced permeability prevents the penetration of multiple toxic compounds, helping the cell to survive in adverse environments [1]. Corresponding author and reprint requests: L. Martínez- Martínez, Service of Microbiology, University Hospital Marqués de Valdecilla, Santander, Spain lmartinez@humv.es Despite its relative impermeability, the outer membrane has to allow different hydrophilic compounds to enter (nutrients) or to leave (toxic waste) the cell, and outer membrane proteins (OMPs) are critical in this function [2]. The outer membrane contains lipoproteins (anchored to the phospholipid leaflet) and integral proteins. The latter are usually made of anti-parallel b-strands with a barrel-like structure functioning as transmembrane hydrophilic pores [3]. It is usual to distinguish, from a functional point of view, between non-specific and specific pores. It has been proposed that non-specific pores should be called porins, but in the literature this term is also often used for channels allowing penetration of specific substrates. Most studies on porins have been done in Enterobacteriaceae, particularly Escherichia coli (and especially the laboratory strain E. coli K-12), Klebsiella pneumoniae and Enterobacter spp. Information is also available for Pseudomonas aeruginosa and, to a lesser extent, Acinetobacter baumannii. Limited information has been reported for other organisms, e.g., Vibrio, Neisseria, Helicobacter or Campylobacter. Several reviews are available on the structure, biogenesis and regulation of these proteins [1,3 5]. There are two major non-specific porins, OmpF and OmpC, in E. coli K-12. Both are

2 Martínez-Martínez ESBL and the permeability barrier 83 trimeric proteins. The functional pore of OmpF (with a diameter of about 12 Å and a preference for cations) is slightly larger than that of OmpC [6]. The porin b-strands (usually 16 per monomer) are tilted (30 60 ) in relation to the vertical axis of the pore, and transmembrane segments are connected with short turns (at the periplasmic space) or rather long loops (at the external cell surface). An important structural feature of OmpF is that loop 3 (33 residues long) folds into the pore and narrows it [1]. Two major porins, OmpK35 (a homologue of OmpF) [7,8] and OmpK36 (a homologue of OmpC) [9], have also been described in K. pneumoniae. The OmpK36 crystal structure has been resolved, indicating that it is rather similar to OmpF from E. coli [10]. In Salmonella enterica (with the notable exception of serovar Typhi), there is an additional porin, OmpD, with properties similar to those of OmpF and OmpC [11]. Enterobacter also contains genes analogous to ompf and ompc, and preliminary studies indicate that it also expresses OmpD [12]. Proteus spp., Providencia spp. and Morganella morganii seem to produce only a single major non-specific porin, but very few studies have yet been done on the porins of these three genera [13]. E. coli also express the OmpA protein, a homologue of the major porin of P. aeruginosa OprF [1]. OmpA-like proteins are also found in other enterobacteria; in K. pneumoniae, the corresponding protein is OmpK34. OmpA is monomeric and contains some helical structures. It allows the penetration of solutes, but it has a very low efficiency, about 100 times lower than the nonspecific trimeric porins [1]. Specific pores of enterobacteria, such as LamB (which allows entry of maltose derivatives, and acts as a receptor for phage lambda) and PhoE (specific for phosphate), have a general structure rather similar to that of the non-specific porins, but the actual residues lining the interior of the pore are sufficiently different to give a different functional specificity. In addition to the major porins, minor proteins with a porin structure have been described in several enterobacteria; these include NmpC and OmpN of E. coli, OmpK37 of K. pneumoniae and OmpS2 of Salmonella typhi [14]. They are quiescent porins, and expression is not detected in the outer membrane of strains grown under standard laboratory conditions, indicating that the corresponding genes are strongly downregulated. Multiple studies have concerned porins of P. aeruginosa, in particular the specific porin OprD, which is involved in basic amino-acid transport, and is important for carbapenem penetration [15]. More recently, several studies on porins of A. baumannii have been reported [16]. There is, however, very little information on the relevance of porin alterations as a mechanism complementing extended-spectrum b-lactamase (ESBL)-mediated resistance in these organisms. Moreover, ESBLs are much less prevalent than in enterobacteria, and for these reasons, non-fermenting organisms will not be covered further in this review. PORIN LOSS AND RESISTANCE TO ANTIMICROBIAL AGENTS: GENERAL ASPECTS The loss of non-specific porins has been related to increased resistance to hydrophilic antimicrobial agents, with studies in this area being focused on b-lactams and fluoroquinolones. However, the demonstration of the actual role of porins in antimicrobial resistance is complex and should not rely simply on the observation of (presumed) porin bands in SDS-PAGE gels. Porins are trimeric proteins that are highly stable to detergents and salts (due to their high b-sheet content) but are heat-modifiable; consequently, if an OMP in an unboiled sample runs with an apparent molecular mass of approximately three times that of the monomer in a boiled sample, it is probably a porin. The outer membrane profiles depend on the conditions (e.g., osmolarity, temperature) in which the organism was grown. It is generally accepted that OmpF and OmpK35 are expressed in low-osmolarity media, while their production is repressed in high-osmolarity media, where OmpC production is favoured [5,17,18]. The pattern of proteins also varies with the gel composition. Optimal resolution (separation) of the K. pneumoniae porins is obtained, depending on the strain, on resolving gels containing 11% acrylamide and either 0.21%, 0.35% or 0.55% bisacrylamide [18]. Adding urea (usually in the range of 4 8 M) to the resolving gel of SDS-PAGE systems helps in defining porin bands of E. coli

3 84 Clinical Microbiology and Infection, Volume 14, Supplement 1, January 2008 (but causes blurred images when studying K. pneumoniae porins). Several functional assays can be used in defining a role for porin loss in resistance. In the case of b-lactams, permeability rates may be indirectly inferred spectrophotometrically from the ability of periplasmic b-lactamase to hydrolyse the assayed antimicrobial agent in comparison with hydrolysis by free, purified enzyme [6,19]. This indirect approach has the limitation that, in most cases, only a single, or just a few, compound(s) are tested, and results for other agents have simply been assumed to be related. In the liposome swelling assay, liposomes are reconstituted with purified porins, and the uptake of sugars with different molecular sizes is tested [8,14]. For a given porin, the penetration of sugars decreases as the size of the compound increases, allowing an indirect estimation of the functional diameter of the pore. Another, more direct, approach to prove the role of a lost OMP in antimicrobial resistance is to demonstrate a decrease in the MIC of antimicrobial agents against an organism when it reexpresses the lost OMP following introduction of an engineered plasmid encoding the corresponding gene [20]. ALTERED PERMEABILITY AND RESISTANCE IN ESBL-PRODUCING K. PNEUMONIAE Multiple studies indicate that porin loss from K. pneumoniae strains lacking other mechanisms of resistance does not cause clinically relevant resistance. MICs of b-lactams, fluoroquinolones, aminoglycosides, tetracycline and chloramphenicol against mutants deficient in antigen O, capsule or either OmpK35 or OmpK36 derived from an ampicillin-susceptible K. pneumoniae strain were within one dilution of those against the parental strain [21]. For the mutant simultaneously lacking the two porins, and expressing increased amounts of b-lactamase, there were no changes in the MICs of fluoroquinolones or aminoglycosides, and the MICs of ampicillin, cefoxitin, cefotaxime and ceftazidime increased four to >256 times [21]. ESBL-producing K. pneumoniae strains more often express only OmpK36 than do strains lacking these enzymes, which usually contain both OmpK35 and OmpK36 [18,22]. The absence of OmpK35 may be one of the factors contributing to antimicrobial resistance in ESBL-producing K. pneumoniae and may favour the selection of additional mechanisms of resistance. More generally, as OmpK35 is not normally expressed in high-osmolarity media, it is possible that its expression is repressed in vivo. This may be of therapeutic importance. Cefoxitin and other cephamycins are not good substrates for ESBLs and, in theory, might represent an option for treating infections caused by ESBL-producing organisms. However, several studies have demonstrated that cefoxitin is readily able to select resistant K. pneumoniae mutants that are deficient in the two major porins, both in vivo and in vitro [20]. There is no clear proof that this observation is valid for other cephamycins, e.g., cefotetan. In a study on 12 clinical isolates of K. pneumoniae producing ESBLs and lacking both OmpK35 and OmpK36, the MICs of cefoxitin were 32 mg L for all isolates, but those of cefotetan were 16 mg L for 11 isolates. This is in contrast to 19 K. pneumoniae isolates with plasmid-mediated AmpC-type b-lactamases (pampc), where MICs of both cefoxitin and cefotetan were 32 mg L for 19 and 15 isolates, respectively. As shown in Table 1, MICs of cefoxitin, but also of oxyimino-cephalosporins such as cefotaxime, are increased for porin-deficient mutants with ESBLs. Cloning and expression of a plasmidborne ompk36 gene in strains lacking both OmpK35 and OmpK36 (Table 2) demonstrated that porin loss was directly involved in resistance to cefoxitin and oxyimino-cephalosporins in both organisms [20,23]. In subsequent studies using the same approach, the role of OmpK35 was also evaluated [8]. Expression of this protein in an isolate lacking both OmpK35 and OmpK36 Table 1. MICs (mg L) of b-lactams for two pairs (LB3-LB4 and CSUB10S-CSUB10R) of extended-spectrum b-lactamase (ESBL)-producing Klebsiella pneumoniae strains of clinical origin, producing porins or not a Strain ESBL OmpK36 FOX CAZ CTX IMP MPM LB3 SHV > NT LB4 SHV-5 ) 128 > NT CSUB10S SHV > CSUB10R SHV-2 ) 128 > a All four organisms lack porin OmpK35. NT, not tested; FOX, cefoxitin; CAZ, ceftazidime; CTX, celotaxime; IMP, irnipenem; MPM, menopenem

4 Martínez-Martínez ESBL and the permeability barrier 85 Table 2. MICs (mg L) of b-lactams for clinical isolates and laboratory mutants of Klebsiella pneumoniae strains expressing porins OmpK35 and or OmpK36 Strain Porin FOX CTX IMP MPM CSUB10S OmpK CSUB10R None a CSUB10R psha25 OmpK CSUB10R psha16 OmpK CSUB10S psha16 OmpK35, OmpK CSUB10R pq3 OmpK37 (low-level) CSUB10R pq7 OmpK37 (high-level) a None: absence of both OmpK35 and OmpK36. FOX, cefoxitin; CTX, celotaxime; IMP, irnipenem; MPM, menopenem caused a reduction of the MICs of most expanded-spectrum b-lactams.the greatest reductions were noted for cephamycins, oxyiminocephalosporins, zwitterionic cephalosporins and meropenem. A significant reduction was also obtained for imipenem. These reductions were four to eight times (cefepime, cefotetan, cefotaxime, and cefpirome), or 128 times (ceftazidime) higher than those observed when OmpK36 was expressed in the same organism. The level of resistance further increased when a high inoculum (10 7 CFU ml) was used, and this may be clinically relevant when the number of bacteria at the site of infection is high. Expression of OmpK35 in a K. pneumoniae strain already producing OmpK36 (resulting in an organism with both major porins) caused decreases in the MICs of b-lactams identical to, or one dilution step greater than, those for the derivative expressing only OmpK35 [8]. This suggests that, at least in K. pneumoniae, expression of just one of the two major porins may be sufficient for b-lactam penetration, and that resistance to cefoxitin and increased resistance to other compounds requires loss of both OmpK35 and OmpK36 (Table 2). Carbapenems are usually active against ESBLproducing strains, including those that lack major porins. This is related to the high stability of these molecules in the presence of ESBLs; in addition, imipenem is a zwitterion and is small, which facilitates diffusion through the outer membrane. Resistance (as defined by EUCAST; srga.org/eucastwt/mictab/miccarbapenems. html) to imipenem (MIC 8 mg L) and to meropenem (MIC 16 mg L) was described in a clinical isolate of K. pneumoniae producing SHV-2, lacking OmpK35, and producing reduced amounts of OmpK36 [24]. Using derivatives of a strain deficient in both OmpK35 and Ompk36 and producing different b-lactamases, resistance to imipenem and meropenem was observed when class A or D carbapenemases or some pampc types were produced [25,26]. For the construct expressing SHV-2, the MIC of imipenem was 16 mg Latan inoculum of 10 7 CFU ml [25]. Resistance to ertapenem was also obtained when carbapenemases (practically all tested), pampc, OXA-32 and some TEM-type ESBLs were expressed; the MICs of this compound ranged from 4 to 16 mg L when SHV-type, CTX-M-5 or -14 and other TEM types were expressed [26]. Ertapenem resistance has been observed in two isolates of K. pneumoniae from a patient treated with ertapenem [27]. The isolates lacked porins and produced an ESBL of the CTX-M-1 group. Two other clonally related isolates, producing the same ESBL and expressing porins, were susceptible to ertapenem. The porindeficient strains showed decreased activity of meropenem (MIC 4 8 mg L), but the activity of imipenem was not much affected (MIC mg L). In another independent study, carbapenem resistance was documented in two isolates of K. pneumoniae deficient in both OmpK35 and OmpK36 and producing CTX-M-1, cultured from patients treated with imipenem or meropenem [28]. The clinical implications of these data may be of great concern; while ertapenem may be more severely affected than other carbapenems, favouring the use of imipenem or meropenem may translate into additional selective pressure for multiresistant non-fermenters or into the expansion of carbapenemase-producing enterobacteria. Re-expression of the OmpK35 gene and, to a lesser extent, of the OmpK36 gene in K. pneumoniae deficient in both porins also caused a reduction in the MICs of ciprofloxacin, chloramphenicol, and tetracycline [8], although this reduction was minimal or absent for compounds to which the organism was already susceptible (clinafloxacin, amikacin, and gentamicin). This particular strain contained a Ser83Phe change in the A subunit of DNA gyrase and showed active efflux of

5 86 Clinical Microbiology and Infection, Volume 14, Supplement 1, January 2008 fluoroquinolones. As both mechanisms affected ciprofloxacin more efficiently than clinafloxacin, the difference in MIC between the two compounds decreases, again suggesting that loss of porins has a greater effect when another mechanism of resistance is also present. Porin-deficient ESBLproducing K. pneumoniae strains more often express active efflux of quinolones than those expressing porins, and active efflux is more common among ESBL-positive strains than among ESBL-negative strains [22]. It may be possible that efflux works better in porin-deficient strains. Organisms producing both ESBL and the Qnr proteins involved in quinolone resistance have increasingly been reported in recent years [29]. These studies on K. pneumoniae were extended to a collection of 50 clinical isolates producing at least one porin and 15 isolates lacking both OmpK35 and OmpK36. It was shown again that loss of porins is critical for cefoxitin resistance, and that resistance to other expanded-spectrum b-lactam antibiotics was increased in porin-deficient organisms. In terms of MIC 50 and MIC 90, meropenem was more active than imipenem against both porin-expressing and porin-deficient strains, but MIC ranges of meropenem for the latter were broader, indicating that, for a subset of strains deficient in porins, the activity of meropenem is more affected than that of imipenem [30]. The mechanisms controlling porin loss in K. pneumoniae are still poorly understood. This organism contains a homologue of mara (see below for the regulatory effect of this gene in E. coli) [31]. Independently, it has been reported that direct interruption of the ompk36 gene by insertion sequences and, less frequently, point mutations or deletions explains the loss of the OmpK36 porin in both laboratory and clinical strains. Similar mechanisms may control loss of OmpK35 [32]. The minor OmpK37 porin allows the penetration of carbapenems (but not other b-lactams) into the cell. Its predicted secondary structure has an insertion of one bulky residue (Tyr118) in loop 3, presumably narrowing the pore [14]. When it is overexpressed in isolates deficient in both OmpK35 and OmpK35, MICs of both imipenem and meropenem decrease, but as the amount of this protein in the outer-membrane is negligible when it is expressed from its natural promoter, it is difficult to have a precise idea about the importance of its down-regulation in carbapenem resistance [14]. The importance of other specific pores in the resistance of K. pneumoniae to antimicrobial agents has been little explored. It is not clear whether LamB expression (observed in Mueller Hinton media) directly facilitates the penetration of hydrophilic compounds, and whether its loss may be involved in increased resistance. Comparison of clonally related carbapenem-susceptible and carbapenem-resistant clinical isolates of K. pneumoniae lacking both OmpK35 and OmpK36 showed down-regulation of ompk37, phob and phoe in the latter [33]. Expression of the native phoe gene under the control of the heterologous lacz promoter reversed carbapenem resistance, whereas this was not the case with the native promoter [33]. It is possible that expression of PhoE is responsible for carbapenem susceptibility in the porin-deficient background (because of compensation for the loss of major non-specific porins), and that down-regulation of phoe was involved in the carbapenem resistance. ALTERED PERMEABILITY AND RESISTANCE IN E. COLI PRODUCING ESBL Multiple studies on the structure, function and regulation of porins in E. coli K-12 have been reported. OmpF and OmpC are the major porins of this organism. They are under the control of several systems, including the two-component system ompr envz [5] and the sox and mar operons. The latter interferes with the expression of ompf by the RNA antisense micf and, additionally, activates the acrab genes involved in active efflux of antimicrobial agents [34]. Despite these data, and in contrast to what is known about ESBL-producing K. pneumoniae, there is scarce information concerning the true number and nature of porins expressed by clinical isolates of E. coli, and on the relationship of porin expression, if any, to antimicrobial resistance, particularly for ESBL-producing clinical isolates. Preliminary reports indicated that loss of both OmpF and OmpC causes resistance to b-lactams when they are substrates for an efficient b-lactamase [35,36]. Loss of OmpF alone in mutants expressing TEM-3 or -9 caused moderate changes (two- to four-fold increases) in MICs, as compared with the equivalent mutant expressing OmpF.

6 Martínez-Martínez ESBL and the permeability barrier 87 Remarkably, the MICs of both imipenem and meropenem were slightly higher (still within the range of clinical susceptibility) for the strain lacking OmpF [36]. Electrophoretic analysis of porins from 50 E. coli isolates producing ESBLs grown in lowand high-osmolarity media indicated that only four (8%) of them presented a profile similar to that of E. coli K-12. Four other major profiles were noted among this collection of organisms, characterised by different models of expression of OMP with mobilities similar to those of OmpF and OmpC of the K12 strain. There were no clear correlations between porin profiles and resistance to cefoxitin or to expanded-spectrum cephalosporins. ENTEROBACTER AND OTHER ENTEROBACTERIACEAE There are several reports on the association of porin loss and increased resistance to antimicrobial agents in other species, including S. enterica, Enterobacter spp., and Serratia spp. [19,37,38]. Although these descriptions do not specifically refer to ESBL-producing organisms, it seems likely that the basic observations outlined above for K. pneumoniae and E. coli are also valid for these related bacteria. Nevertheless, some peculiarities are expected to be discovered in these genera and, in fact, some original observations have already been made. Omp35 and Omp36 have been characterised in detail in Enterobacter. Omp35 (like OmpK35 of K. pneumoniae) contains several negatively charged amino-acids in the L3 region of the pore, and these are of critical relevance for higher permeability, in comparison with Omp36 [12]. Reduced permeability for antimicrobial agents has been documented in an Enterobacter aerogenes porin with a structural change in loop 3 [39]. It is possible that similar changes occur in K. pneumoniae strains with reduced susceptibility to cefoxitin but still expressing porins in their outer membrane (unpublished data). Studies in Enterobacter (and other enterobacteria) have shown that, independently of porin changes caused by environmental alterations (e.g., in ph, osmolarity, O 2 concentration), the function of porins may be altered by polyamines because of the interaction between their amino groups and amino-acids lining the internal surface of the channel. This has been demonstrated for OmpF of Enterobacter cloacae, where spermine causes decreased penetration of cefepime and norfloxacin [40]. OmpX has been shown to be related to decreased expressions of major porins in Enterobacter [41]. FUTURE PROSPECTS Porin loss may increase the level of resistance to b-lactams in ESBL-producing organisms, especially if there is simultaneous expression of other b-lactamases. This may result in very complex phenotypes that are difficult to recognise in the clinical laboratory, sometimes masking the typical ESBL phenotype. New testing approaches are required and should eventually be incorporated into the expert systems of automatic susceptibility testing devices. Available information suggests that porin loss may contribute to resistance in K. pneumoniae differently than in other Enterobacteriaceae, including E. coli. Thus, it may be unwise to extrapolate from observations in one organism and, similarly, data obtained with laboratory strains do not always apply to clinical isolates. REFERENCES 1. Nikaido H. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 2003; 67: Nikaido H. Proteins forming large channels from bacterial and mitochondrial outer membranes: porins and phage lambda receptor proteins. Methods Enzymol 1983; 97: Jeanteur D, Lakey JH, Pattus F. The porin superfamily: diversity and common features. In: Ghuysen J-M, Hakenbeck R, eds, Bacterial cell wall. Amsterdam: Elsevier, 1994; Schulz GE. The structure of bacterial outer membrane proteins. Biochim Biophys Acta 2002; 1565: Pratt LA, Hsing W, Gibson KE, Silhavy TJ. From acids to osmz: multiple factors influence synthesis of the OmpF and OmpC porins in Escherichia coli. Mol Microbiol 1996; 20: Nikaido H, Rosenberg EY, Foulds J. Porin chanels in Escherichia coli: studies with b-lactams in intact cells. J Bacteriol 1983; 153: Hernández-Allés S, Albertí S, Rubires X, Merino S, Tomás JM, Benedí VJ. Isolation of FC3-11, a bacteriophage specific for the Klebsiella pneumoniae porin OmpK36, and its use for the isolation of porin deficient mutants. Can J Microbiol 1996; 41: Doménech-Sánchez A, Martínez-Martínez L, Hernández- Allés Set al. Role of Klebsiella pneumoniae OmpK35 porin in antimicrobial resistance. Antimicrob Agents Chemother 2003; 47:

7 88 Clinical Microbiology and Infection, Volume 14, Supplement 1, January Albertí S, Rodríguez-Quiñones F, Schirmer T et al. A porin from Klebsiella pneumoniae: sequence homology, threedimensional structure, and complement binding. Infect Immun 1995; 63: Dutzler R, Rummel G, Albertí S et al. Crystal structure and functional characterization of OmpK36, the osmoporin of Klebsiella pneumoniae. Structure 1999; 7: Santiviago CA, Toro CS, Bucarey SA, Mora GC. A chromosomal region surrounding the ompd porin gene marks a genetic difference between Salmonella typhi and the majority of Salmonella serovars. Microbiology 2001; 147: Bornet C, Saint N, Fetnaci L et al. Omp35, a new Enterobacter aerogenes porin involved in selective susceptibility to cephalosporins. Antimicrob Agents Chemother 2004; 48: Mitsuyama J, Hiruma R, Yamaguchi A, Sawai T. Identification of porins in outer membrane of Proteus, Morganella, and Providencia spp. and their role in outer membrane permeation of beta-lactams. Antimicrob Agents Chemother 1987; 3: Doménech-Sánchez A, Hernández-Allés S, Martínez- Martínez L, Benedí VJ, Albertí S. Identification and characterization of a new porin gene of Klebsiella pneumoniae: its role in b-lactam antibiotic resistance. J Bacteriol 1999; 181: Trías J, Nikaido H. Outer membrane protein D2 catalyzes facilitated diffusion of carbapenems and penems through the outer membrane of Pseudomonas aeruginosa. Antimicrob Agents Chemother 1990; 34: Vila J, Marti S, Sanchez-Cespedes J. Porins, efflux pumps and multidrug resistance in Acinetobacter baumannii. J Antimicrob Chemother 2007; 59: Medeiros AA, O Brien TF, Rosenberg EY, Nikaido H. Loss of OmpC porin in a strain of Salmonella typhimurium causes increased resistance to chephalosporins during therapy. J Infect Dis 1987; 156: Hernández-Allés S, Albertí S, Alvarez D et al. Porin expression in clinical isolates of Klebsiella pneumoniae. Microbiology 1999; 145: Cornaglia G, Guan L, Fontana R, Satta G. Diffusion of meropenem and imipenem through the outer membrane of Escherichia coli K-12 and correlation with their antibacterial activities. Antimicrob Agents Chemother 1992; 36: Martínez-Martínez L, Hernández-Allés S, Albertí S, Tomás JM, Benedí VJ, Jacoby GA. In vivo selection of porin-deficient mutants of Klebsiella pneumoniae with increased resistance to cefoxitin and expanded-spectrum cephalosporins. Antimicrob Agents Chemother 1996; 40: Hernández-Allés S, Conejo MC, Pascual A, Tomás JM, Benedí VJ, Martínez-Martínez L. Relationship between outer membrane alterations and susceptibility to antimicrobial agents in isogenic strains of Klebsiella pneumoniae. J Antimicrob Chemother 2000; 46: Martínez-Martínez L, Pascual A, Conejo MC et al. Energydependent accumulation of norfloxacin and porin expression in clinical isolates of Klebsiella pneumoniae and relationship to extended-spectrum b-lactamase production. Antimicrob Agents Chemother 2002; 46: Ardanuy C, Liñares J, Domínguez MA, Hernández-Allés S, Benedí VJ, Martínez-Martínez L. Outer membrane profiles of clonally related Klebsiella pneumoniae isolates from clinical samples and activities of cephalosporins and carbapenems. Antimicrob Agents Chemother 1998; 42: Crowley B, Benedí VJ, Doménech-Sánchez A. Expression of SHV-2 beta-lactamase and of reduced amounts of OmpK36 porin in Klebsiella pneumoniae results in increased resistance to cephalosporins and carbapenems. Antimicrob Agents Chemother 2002; 46: Martínez-Martínez L, Pascual A, Hernández-Allés S et al. Roles of b-lactamases and porins in activities of carbapenems and cephalosporins against Klebsiella pneumoniae. Antimicrob Agents Chemother 1999; 43: Jacoby GA, Mills DM, Chow N. Role of beta-lactamases and porins in resistance to ertapenem and other beta-lactams in Klebsiella pneumoniae. Antimicrob Agents Chemother 2004; 48: Elliott E, Brink AJ, van Greune J et al. In vivo development of ertapenem resistance in a patient with pneumonia caused by Klebsiella pneumoniae with an extended-spectrum beta-lactamase. Clin Infect Dis 2006; 42: e95 e Mena A, Plasencia V, Garcia L et al. Characterization of a large outbreak by CTX-M-1-producing Klebsiella pneumoniae and mechanisms leading to in vivo carbapenem resistance development. J Clin Microbiol 2006; 44: Robicsek A, Jacoby GA, Hooper DC. The worldwide emergence of plasmid-mediated quinolone resistance. Lancet Infect Dis 2006; 6: Doménech-Sánchez A, Pascual A, Suárez AI, Álvarez D, Benedí VJ, Martínez-Martínez L. Activity of nine antimicrobial agents against clinical isolates of Klebsiella pneumoniae producing extended-spectrum beta-lactamases and deficient or not in porins. J Antimicrob Chemother 2000; 46: George AM, Hall RM, Stokes HW. Multidrug resistance in Klebsiella pneumoniae: a novel gene, rama, confers a multidrug resistance phenotype in Escherichia coli. Microbiology 1995; 141: Hernández-Allés S, Benedií VJ, Martínez-Martínez L et al. Development of resistance during antimicrobial therapy caused by insertion sequence interruption of porin genes. Antimicrob Agents Chemother 1999; 43: Kaczmarek FM, Dib-Hajj F, Shang W, Gootz TD. Highlevel carbapenem resistance in a Klebsiella pneumoniae clinical isolate is due to the combination of bla(act-1) b-lactamase production, porin OmpK35 36 insertional inactivation, and down-regulation of the phosphate transport porin Phoe. Antimicrob Agents Chemother 2006; 50: Cohen SP, McMurry LM, Levy SB. Cross-resistance to fluoroquinolones in multiple-antibiotic-resistant (Mar) Escherichia coli selected by tetracycline or chloramphenicol: decreased drug accumulation associated with membrane changes in addition to OmpF reduction. Antimicrob Agents Chemother 1989; 33: Hiraoka M, Okamoto R, Inoue M, Mitsuhashi S. Effects of beta-lactamases and omp mutation on susceptibility to beta-lactam antibiotics in Escherichia coli. Antimicrob Agents Chemother 1989; 33: Jacoby GA, Carreras I. Activities of b-lactam antibiotics against Escherichia coli strains producing extended-

8 Martínez-Martínez ESBL and the permeability barrier 89 spectrum b-lactamases. Antimicrob Agents Chemother 1990; 34: Hopkins JM, Towner KJ. Enhanced resistance to cefotaxime and imipenem associated with outer membrane protein alterations in Enterobacter aerogenes. J Antimicrob Chemother 1990; 25: Sánchez L, Ruiz N, Leranoz S, Viñas M, Puig M. The role of outer membrane in Serratia marcescens intrinsic resistance to antibiotics. Microbiology 1997; 13: Dé E, Baslé A, Jaquinod M et al. A new mechanism of antibiotic resistance in Enterobacteriaceae induced by a structural modification of the major porin. Mol Microbiol 2001; 41: Chevalier J, Malléa M, Pagès JM. Comparative aspects of the diffusion of norfloxacin, cefepime and spermine thourgh the F porin channel of Enterobacter cloacae. Biochem J 2000; 348: Stoorvogel J, van Bussel MJ, Tommassen J, van der Klundert JA. Biological characterization of an Enterobacter cloacae outer membrane protein (OmpX). J Bacteriol 1991; 173: