L'utilisation d'antibiotiques pour lutter contre les infections bactériennes a favorisé l'émergence de souches résistantes. Comprendre les mécanismes de résistance est crucial pour lutter contre ces pathogènes. Cette thèse propose une étude structurale d'une pompe à efflux multidrogues de Pseudomonas aeruginosa qui se compose d'un transporteur MexB, d'une protéine canal OprM et d'une protéine adaptateur MexA. Les partenaires du complexe tripartite stabilisés en nanodisques ont permis la formation du complexe in vitro. L'optimisation des conditions de production du complexe a permis de cribler les différents paramètres régissant son assemblage. L'étude structurale par cryo-ME révèle un complexe de 30 nm de long en conformation de repos. L'étude de la stabilisation des protéines membranaires par nanodisques a conduit au développement d'un système minimal, débarrassé des lipides. Ce système minimal a révélé la nécessité d'une phase lipidique autour de MexB pour l'assemblage du complexe. / Antibiotics use against bacterial infections has led to the emergence of resistance. Understanding the mechanisms underlying resistance to antibiotics is critical to fight against these pathogens. This thesis presents a structural study of a multidrug efflux pump in Pseudomonas aeruginosa, composed of a transporter MexB, an exit duct OprM and an adaptor protein MexA. The proteins reconstituted in nanodiscs allowed tripartite complex formation in vitro. Optimization of yield led to the identification of key parameters governing complex assembly. Structural cryo-EM study revealed a 30 nm long complex in a resting state. The study of membrane protein stabilization by nanodisks led to the development of a minimal system devoid of lipids. This system showed a lipid phase around MexB is required for complex formation.

Introdução: Isolados clínicos de Pseudomonas aeruginosa multirresistentes estão associados a elevadas taxas de mortalidade. A resistência ao imipenem é uma urgência global, uma vez que é considerado o tratamento de escolha para infecções associadas a bactérias Gram negativas multirresistentes. Assim, elucidar os mecanismos de resistência é de vital importância para realizar um controle epidemiológico efetivo da disseminação deste tipo de isolado. Objetivos: Caracterizar os principais mecanismos de resistência ao imipenem em 76 isolados clínicos brasileiros de Pseudomonas aeruginosa, recuperados em 2004/2007, de 4 centros hospitalares do Estado de São Paulo. Material e métodos: Foram investigados: i) o perfil de resistência com determinação da CIM do imipenem; ii) a detecção de metalo-betalactamases (MBL) através de métodos fenotípicos e genotípicos; iii) a sensibilidade e especificidade do método de dupla difusão do disco na detecção de MBL; iv) a presença de genes codificadores de metilases 16S RNAr e sua associação com fenótipos aminoglicosídeo resistentes; v) alterações da permeabilidade por perda da porina OprD; vi) a presença ou ausência do gene oprD por PCR; vii) triagem fenotípica para expressão de bombas de efluxo através da determinação da CIM de quinolonas, cefalosporinas e carbapenêmicos na presença/ausência de inibidores específicos, realizando uma análise comparativa com o método de disco combinado; viii) os genes associados às bombas de efluxo mexA e mexE, através de PCR; ix) caracterizar a expressão das bombas de efluxo MexAB-OprM e MexEF-OprN, x) a clonalidade dos isolados por tipagem genotípica, através de ERIC-PCR, avaliando a relação genética (dendrograma) e sua associação com o predomínio de um determinado mecanismo de resistência. Resultados: Dentre os isolados de P. aeruginosa resistentes ao imipenem estudados (n=76, CIM50 e CIM90 = 32 µg/mL e > 512 µg/mL, respectivamente) 82% apresentaram um fenótipo de multirresistência. O principal mecanismo de resistência ao imipenem foi a produção de MBL, detectada em 74% dos isolados, e destes, 62% carregavam o gene blaSPM-1 e 12% carregavam o gene blaVIM-like. O método de dupla difusão do disco identificou a produção de MBL em 61% dos isolados. A combinação CAZ/MAA apresentou maior sensibilidade na detecção de MBL associada à SPM-1 (89%), mostrando uma especificidade de 86%. A presença do gene rmtD foi confirmada em 66% das amostras resistentes aos aminoglicosídeos, sendo que a presença concomitante do gene rmtD e do gene blaSPM-1 foi confirmada em 61% dos isolados. A deleção da porina OprD foi observada em 71% dos isolados. Dentre os isolados MBL positivos, 66% apresentaram ausência desta porina e, dentre as amostras MBL negativas, 85% não apresentaram OprD. Assim, para a resistência ao imipenem foi confirmada a contribuição de dois mecanismos, mediados pela presença de MBL e ausência de porina OprD. Em 13% (10/76) isolados, a deleção da porina OprD esteve associada à presença de seqüências de inserção (SI) em uma região anterior ao gene oprD. Por outro lado, a ausência de amplificação da região 736/1394 do gene oprD, em 11% (9/76) dos isolados, sugeriu a presença de polimorfismos. O gene mexA esteve presente em 92% dos isolados, enquanto que o gene mexE esteve presente em 82% dos isolados. A triagem de bombas de efluxo por disco combinado e análise da CIM na presença de reserpina, CCCP e PAβN, utilizando levofloxacina, meropenem, aztreonam, imipenem ou levofloxacina, não teve correlação com a superexpressão dos sistemas MexAB-OprM e MexEF-OprN. Ambos os métodos careceram de especificidade e sensibilidade quando comparados ao PCR em tempo real. A superexpressão dos sistemas mexA e mexE foi confirmada em 35% (7/20) isolados MBL negativos, enquanto que 11% (6/56) isolados MBL positivos apresentaram superexpressão do gene mexA ou mexE, sendo que 7% (4/56) isolados MBL positivos superexpressaram ambos os genes. A superexpressão dos sistemas MexAB-OprM e MexEFOprN como único mecanismo de resistência ao meropenem e imipenem foi confirmada em 10% (2/20) dos isolados MBL negativos. Nos 76 isolados, a tipagem genotípica por ERIC-PCR, identificou a presença de 24 clusters (considerando 90% de similaridade na análise do dendrograma). Conclusão: A convergência de múltiplos mecanismos de resistência em P. aeruginosa parece ser um evento favorável para a seleção de clones endêmicos multirresistentes disseminados na região Sudeste do Brasil. / Introduction: Clinical isolates of Pseudomonas aeruginosa are associated with high mortality rates. Resistance to imipenem is a global concern, since it is a drug of choice for the treatment of infections produced by multidrug-resistant Gram-negative bacteria. Thus, research on resistance mechanisms is crucial to carry out an effective program for infection control and epidemiology of imipenem-resistant strains. Objective: to characterize the major mechanisms of imipenem resistance in 76 clinical isolates of P. aeruginosa recovered from clinical samples collected, from 2004 to 2007, in four hospitals in the State of São Paulo, Brazil. Material and methods: Isolates were screened for: i) resistance profile to antibacterial agents, determining the MIC of imipenem; ii) the detection of metallo-beta-lactamase (MBL) by phenotypic and genotypic methods, iii) MBL detection by using a double-disk diffusion test (D-test), determining the sensitivity and specificity of the assay; iv) the presence of genes encoding 16S rRNA methylases and their association with aminoglycoside-resistant phenotypes, v) changes in the bacterial permeability due to porin (OprD) loss; vi) the presence or absence of the oprD gene by using PCR; vii) phenotypic expression of efflux pumps by determining the MIC of quinolones, cephalosporins and carbapenems in the presence/absence of specific inhibitors, performing a comparative analysis with a combined-disk method, viii) genes encoding efflux pumps proteins (mexC and mexX) by PCR; ix) MexAB-OprM and MexEF efflux pumps expression; x) clonal relatedness, by ERIC-PCR genotyping, regarding the predominance of major resistance genotypes. Results: Among imipenem-resistant P. aeruginosa strains (n=76, MIC50 e MIC90 = 32 µg/mL e > 512 µg/mL, respectively) 82% showed a multidrug-resistant phenotype. The main mechanism of imipenem resistance was the MBL production detected in 74% strains, of which 62% harbored the blaSPM-1 gene, and 12% harbored the blaVIM-like gene. The D-test identified MBL production in 61% strains. In this regard, CAZ/MAA was the most sensitive combination for MBL detection associated to SPM-1 enzyme (89%), exhibiting 86% specificity. The presence of the rmtD 16S rRNA methylase gene was confirmed in 66% aminoglycoside-resistant strains. Moreover, presence of both rmtD and blaSPM-1 genes was identified in 61% strains. Loss of OprD porins was observed in 71% strains. In this regard, 66% MBL positive strains and 85% MBL negative strains showed OprD loss. Thus, MBL production and OprD loss contributed to imipenem resistance in P. aeruginosa. Most likely, in 13% (10/76) strains the porin loss was associated to insertion sequences (SI) inserted upstream of the oprD gene. On the other hand, in 11% (9/76) strains the absence of a PCR product targeting the 736/1394 region of the oprD gene, suggested the presence of polymorphisms. The mexA gene was identified in 92% strains, whereas the mexE gene was identified in 82% strains. Results obtained from efflux pump screening by using a combined-disk assay and MIC determination in the presence of reserpine, CCCP e PABN (using levofloxacin, meropenem, aztreonam or imipenem) was not correlated with results obtained from MexAB-OprM and MexEF-OprN overexpression analysis by RT-PCR. In this regard, both combined-disk and MIC assay showed lack of specificity and sensitivity in comparison to RT-PCR. Overexpression of mexA and mexE genes was confirmed in 35% (7/20) MBL-negative and 11% (6/56) MBL-positive strains, respectively, being 7% (4/56) MBL-positive strains overexpressed both genes. The overexpression of MexAB-OprM and MexEF-OprN efflux pumps, as only mechanism of resistance to meropenem and imipenem was observed in 10% (2/20) MBL-negative strains. ERICPCR typing revealed the presence of 24 clusters among 76 imipenem-resistant P. aeruginosa strains (≥ 90% similarity). Conclusion: The convergence of multiple mechanisms of resistance in Pseudomonas aeruginosa seems to be a favorable event for the selection of multiresistant clones endemic in the southeastern region of Brazil.

<p>Drug transport across the intestinal epithelium is roughly predicted from permeability values obtained from Caco-2 cell monolayers. This thesis examines the important role of <i>pH</i> and extracellular additives for increasing the reliability and predictivity of the <i>in vitro</i> screening system, Caco-2.</p><p>It was shown that the passive transport of ionizable compounds may be biased by a false efflux or uptake component, when applying a physiological <i>pH</i>-gradient across the membrane. <i>pH</i> also affected the amount of compound available at the transporter-binding site. Therefore, <i>pH</i> dependence should be considered in studies of such compounds and of drug-drug interactions involving efflux transporters. It was also shown that proton-dependent apical uptake or basolateral efflux should be studied both with and without a <i>pH</i> gradient over the whole monolayers. </p><p>The two extracellular additives, bovine serum albumin (BSA) and the solubilizing agent, Cremophor^® EL, also influenced Caco-2 permeabilities. BSA applied to the receiver side increases, and to the donor side decreases drug permeation according to the drug’s protein binding capacity. Thus, the absorptive transport for both passive and active compounds is favoured, giving a physiologically sound improvement of the Caco-2 cell model. Inclusion of BSA increased both the predictivity and quality of permeability studies, particularly of highly lipophilic, BCS class II compounds. Passive and active transport processes could also be distinguished after accounting for unbound concentrations. The overall effect of Cremophor^® EL on the permeability to a drug was compound-specific and probably dependent on micellar incorporation. Cremophor^® EL can therefore not be recommended. </p><p>Neither <i>pH</i> nor BSA affect the functionality of transporters such as P-glycoprotein. However, efflux ratios of ionizable or protein bound drugs are altered in the presence of a <i>pH</i>-gradient or BSA, indicating that an experimental system without protein or <i>pH</i> gradient can over- or underestimate active and passive efflux in drug transport.</p>

<p>One of the aims of this thesis was to investigate the involvement of efflux proteins, such as the P-glycoprotein (Pgp), in the drug transport in different regions of the rat and the human intestine. The intestinal extrusion of intracellularly formed CYP3A4 metabolites, including whether this extrusion might be mediated by Pgp, was also studied. The model drugs used were local anaesthetics (LA), which have been evaluated for inflammatory bowel disease, such as ropivacaine, lidocaine and bupivacaine. The intestinal permeability to LAs was found to be high throughout all intestinal regions of the rat and human intestine. Results from the Ussing chamber model indicated only minor efflux involvement as the drug permeability was higher in the serosa to mucosa transport direction than in the opposite direction. However, the involvement of efflux in the absorption of LAs could not be verified using in situ single-pass perfusion of rat jejunum. The extrusion of the ropivacaine metabolite, 2´,6´-pipecoloxylidide (PPX), was polarized to the mucosal reservoir of the Ussing chamber for both rat and human intestinal samples, and was probably not caused by any Pgp involvement. The expression levels of CYP3A4 and efflux transporters were consistent with the enzymes’ activity in human intestine. PPX formation was mediated by CYP3A4 in human intestine, and cyp2c and cyp2d in rat intestine. Species differences were observed, as PPX was formed in rat colon, but not human colon. In conclusion, the permeability of ropivacaine, lidocaine and bupivacaine was not subjected to efflux transport of significance for their intestinal uptake. The transport of ropivacaine metabolites to the mucosal compartment was probably not mediated by Pgp. The Ussing chamber model showed consistent results with those from intestinal microsomes as far as intestinal metabolism is concerned, making it a suitable model for investigations of the interplay of efflux and metabolism. </p>

One of the aims of this thesis was to investigate the involvement of efflux proteins, such as the P-glycoprotein (Pgp), in the drug transport in different regions of the rat and the human intestine. The intestinal extrusion of intracellularly formed CYP3A4 metabolites, including whether this extrusion might be mediated by Pgp, was also studied. The model drugs used were local anaesthetics (LA), which have been evaluated for inflammatory bowel disease, such as ropivacaine, lidocaine and bupivacaine. The intestinal permeability to LAs was found to be high throughout all intestinal regions of the rat and human intestine. Results from the Ussing chamber model indicated only minor efflux involvement as the drug permeability was higher in the serosa to mucosa transport direction than in the opposite direction. However, the involvement of efflux in the absorption of LAs could not be verified using in situ single-pass perfusion of rat jejunum. The extrusion of the ropivacaine metabolite, 2´,6´-pipecoloxylidide (PPX), was polarized to the mucosal reservoir of the Ussing chamber for both rat and human intestinal samples, and was probably not caused by any Pgp involvement. The expression levels of CYP3A4 and efflux transporters were consistent with the enzymes’ activity in human intestine. PPX formation was mediated by CYP3A4 in human intestine, and cyp2c and cyp2d in rat intestine. Species differences were observed, as PPX was formed in rat colon, but not human colon. In conclusion, the permeability of ropivacaine, lidocaine and bupivacaine was not subjected to efflux transport of significance for their intestinal uptake. The transport of ropivacaine metabolites to the mucosal compartment was probably not mediated by Pgp. The Ussing chamber model showed consistent results with those from intestinal microsomes as far as intestinal metabolism is concerned, making it a suitable model for investigations of the interplay of efflux and metabolism.

Potassium (K+) is an essential nutrient and the most abundant cation in plant cells. Plants possess two transport systems for K+ acquisition: a high-affinity system (HATS), operating at external K+ concentrations ([K+]ext) below 1 mM, and showing reduced transport activity in the presence of ammonium (NH4+); and, a low-affinity system (LATS), operating at [K+]ext above 1 mM, that is not affected by NH4+. K+ transport and compartmentation were investigated in barley (Hordeum vulgare L.) and rice (Oryza sativa L.) using the non-invasive technique of compartmental analysis by tracer efflux (CATE), to simultaneously determine unidirectional membrane fluxes, ion concentrations, and exchange characteristics in subcellular compartments. These studies revealed striking differences in unidirectional K+ fluxes between HATS and LATS. It was found that flux measurements, using traditional direct influx (DI) protocols, accurately represented HATS influx, but underestimated LATS influx by as much as seven-fold. In both barley and rice, LATS K+ fluxes were found to undergo rapid, futile cycling, with the ratio of efflux:influx 3 to 5 times greater, and the cytosolic exchange rate 2 to 3 times faster than under HATS. Based upon plasma-membrane electrical potential measurements, efflux was found to be active under LATS conditions. LATS-mediated conditions for K+ were found to provide relief from NH4+ toxicity in barley by immediately reducing NH4+ influx by more than 50%, and significantly reducing NH4+ futile cycling. Employing the K+ channel inhibitors cesium, lanthanum, and tetraethylammonium, NH4+ was shown to have both K+-sensitive and –insensitive influx pathways at high [NH4+]ext. Based on current models of flux energetics, the combined uptake of K+ and NH4+ was found to utilize 60% of root oxygen consumption. Barley and rice both showed signs of NH4+ toxicity at low [K+]ext, but rice recovered at much lower [K+]ext, suggesting a crucial role of K+ in the NH4+-tolerance of rice. These experiments address fundamental aspects of K+ fluxes, and help provide a physiological framework for future studies of K+ transport and mineral nutrition.

Potassium (K+) is an essential nutrient and the most abundant cation in plant cells. Plants possess two transport systems for K+ acquisition: a high-affinity system (HATS), operating at external K+ concentrations ([K+]ext) below 1 mM, and showing reduced transport activity in the presence of ammonium (NH4+); and, a low-affinity system (LATS), operating at [K+]ext above 1 mM, that is not affected by NH4+. K+ transport and compartmentation were investigated in barley (Hordeum vulgare L.) and rice (Oryza sativa L.) using the non-invasive technique of compartmental analysis by tracer efflux (CATE), to simultaneously determine unidirectional membrane fluxes, ion concentrations, and exchange characteristics in subcellular compartments. These studies revealed striking differences in unidirectional K+ fluxes between HATS and LATS. It was found that flux measurements, using traditional direct influx (DI) protocols, accurately represented HATS influx, but underestimated LATS influx by as much as seven-fold. In both barley and rice, LATS K+ fluxes were found to undergo rapid, futile cycling, with the ratio of efflux:influx 3 to 5 times greater, and the cytosolic exchange rate 2 to 3 times faster than under HATS. Based upon plasma-membrane electrical potential measurements, efflux was found to be active under LATS conditions. LATS-mediated conditions for K+ were found to provide relief from NH4+ toxicity in barley by immediately reducing NH4+ influx by more than 50%, and significantly reducing NH4+ futile cycling. Employing the K+ channel inhibitors cesium, lanthanum, and tetraethylammonium, NH4+ was shown to have both K+-sensitive and –insensitive influx pathways at high [NH4+]ext. Based on current models of flux energetics, the combined uptake of K+ and NH4+ was found to utilize 60% of root oxygen consumption. Barley and rice both showed signs of NH4+ toxicity at low [K+]ext, but rice recovered at much lower [K+]ext, suggesting a crucial role of K+ in the NH4+-tolerance of rice. These experiments address fundamental aspects of K+ fluxes, and help provide a physiological framework for future studies of K+ transport and mineral nutrition.

No description available.

Drug transport across the intestinal epithelium is roughly predicted from permeability values obtained from Caco-2 cell monolayers. This thesis examines the important role of pH and extracellular additives for increasing the reliability and predictivity of the in vitro screening system, Caco-2. It was shown that the passive transport of ionizable compounds may be biased by a false efflux or uptake component, when applying a physiological pH-gradient across the membrane. pH also affected the amount of compound available at the transporter-binding site. Therefore, pH dependence should be considered in studies of such compounds and of drug-drug interactions involving efflux transporters. It was also shown that proton-dependent apical uptake or basolateral efflux should be studied both with and without a pH gradient over the whole monolayers. The two extracellular additives, bovine serum albumin (BSA) and the solubilizing agent, Cremophor® EL, also influenced Caco-2 permeabilities. BSA applied to the receiver side increases, and to the donor side decreases drug permeation according to the drug’s protein binding capacity. Thus, the absorptive transport for both passive and active compounds is favoured, giving a physiologically sound improvement of the Caco-2 cell model. Inclusion of BSA increased both the predictivity and quality of permeability studies, particularly of highly lipophilic, BCS class II compounds. Passive and active transport processes could also be distinguished after accounting for unbound concentrations. The overall effect of Cremophor® EL on the permeability to a drug was compound-specific and probably dependent on micellar incorporation. Cremophor® EL can therefore not be recommended. Neither pH nor BSA affect the functionality of transporters such as P-glycoprotein. However, efflux ratios of ionizable or protein bound drugs are altered in the presence of a pH-gradient or BSA, indicating that an experimental system without protein or pH gradient can over- or underestimate active and passive efflux in drug transport.

Diss. (sammanfattning) Umeå : Umeå universitet, 2010. / Härtill 5 uppsatser.