L’émergence et la diffusion de la résistance aux antibiotiques au sein des bactéries à Gram négatif (BGN) sont aujourd’hui des enjeux de Santé Publique nationaux et internationaux. La multi-résistance aux antibiotiques concerne non seulement des espèces fréquemment responsables d’infections nosocomiales mais aussi des espèces hautement virulentes comme Yersinia pestis, agent de la peste et du bioterrorisme. Dans ce contexte, la mise au point de nouvelles molécules actives sur d’autres cibles bactériennes est primordiale. La métallo-enzyme LpxC catalyse la première étape irréversible de la biosynthèse du lipide A, constituant majeur de la membrane externe des bactéries à Gram négatif. Des inhibiteurs de LpxC sont ainsi développés depuis une vingtaine d’années mais leur spectre sur les BGN était initialement limité aux entérobactéries et leur activité partielle sur P. aeruginosa. Dans ce travail nous avons participé à l’optimisation de la structure chimique de ces molécules grâce à une approche dynamique des interactions enzymes/inhibiteurs utilisant la résonance magnétique nucléaire (RMN). Cette technique a permis l’élaboration d’un nouvel inhibiteur de LpxC, le LPC-058, caractérisé par une forte affinité pour l’enzyme (Ki = 3,5 ± 0,2 pM). Nous avons évalué in vitro l’activité antibiotique du LPC-058 et de trois autres composés (CHIR-090, LPC-011 et LPC-087) vis-à-vis de 369 souches cliniques responsables d’infections nosocomiales aux profils de résistance variés. Le LPC-058 présentait le plus large spectre d’activité en particulier sur A. baumannii et les valeurs de CMI les plus basses (CMI90 = 0,12 mg/L pour les entérobactéries et 0,5 mg/L pour P. aeruginosa). Il était bactéricide vis-à-vis de souches multi-résistantes et son action était synergique avec les C3G, l’imipénème, l’amikacine et la ciprofloxacine vis-à-vis de souches de K. pneumoniae, P. aeruginosa et A. baumannii productrice de carbapénémases, respectivement KPC-2, VIM-1 et OXA-23. Le LPC-058 présentait néanmoins une forte fixation protéique et, in vivo, son volume de distribution était limité au compartiment sanguin (Vd = 1,1 L/kg). Nous avons évalué son activité in vivo dans un modèle murin de peste bubonique car il s’agit de l’une des infections les plus virulentes pour l’homme. Nous avons obtenu une survie de 87 % après 5 jours de traitement à la posologie de 10 mg/kg q8h par voie veineuse. Le LPC-058 occasionnant des diarrhées chez le rongeur, nous avons évalué un de ses dérivés, le LPC-B, caractérisé par une moindre fixation protéique, un plus grand volume de distribution et l’absence d’effets secondaires chez la souris, même à fortes doses. Nous avons démontré qu’à la posologie de 200 mg/kg par voie veineuse, cet antibiotique était aussi efficace que la doxycycline (traitement de référence de la peste). L’ensemble de ces travaux souligne le rôle potentiel des inhibiteurs de LpxC dans la prise en charge des infections par des bactéries multi-résistantes ou hautement virulentes. / Antimicrobial resistance among Gram-negative bacteria (GNB) has become a national and international public health concern. Resistant strains are involved in nosocomial diseases and in highly virulent infections, such as plague caused by Yersinia pestis, a potential biological terrorism agent. In this context the development of new antimicrobial compounds efficient on new bacterial targets is critical. LpxC metallo-enzyme catalyzes the first commitment step of the lipid A biosynthesis, a major component of the Gram negative cell wall. LpxC inhibitors have been developed for twenty years but their activity was restricted to enterobacteria and weak against Pseudomonas aeruginosa. In this study, we have collaborated in the chemical optimization of the compounds thanks to a dynamic approach of enzyme/inhibitor interactions brought by nuclear magnetic resonance (NMR). This technology enabled the development of LPC-058, a new inhibitor, showing a high potency against LpxC (Ki = 3.5 ± 0.2 pM). We studied the in vitro efficacy of LPC-058 and three other compounds (CHIR-090, LPC-011 and LPC-087) against 369 clinical strains responsible for nosocomial infections with various antibiotic resistance profiles. In this part, LPC-058 displayed the broadest spectrum of efficacy, even on Acinetobacter baumannii with the lowest MIC values (MIC90 = 0.12 mg/L against enterobacteria and 0.5 mg/L against P. aeruginosa). It showed bactericidal activity against multi-resistant strains and synergistic activity in association with third generation cephalosporins, imipenem, amikacin and ciprofloxacin against carbapenemase producing Klebsiella pneumoniae, P. aeruginosa et A. baumannii strains (respectively KPC-2, VIM-1 and OXA-23). However, LPC-058 was constrained by strong protein interactions and a small volume of distribution (Vd = 1.1 L/kg). In vivo efficacy was studied in a murine model of bubonic plague. A 87% survival rate was obtained after five days of 10 mg/kg q8h intravenous administration. As LPC-058 treatment was associated to diarrheas in mice, we evaluated another derivate, LPC-B, characterized by a larger volume of distribution, minor protein fixation and less side effects, even for a high dose posology. We demonstrated a comparable efficacy between 200 mg/kg LPC-B treatment and doxycyclin administration (recommended in plague treatment). This work highlights the potential use of LpxC inhibitors in the management of infections caused by multi-resistant or highly virulent Gram-negative bacteria.

<p>The incessant increase of antibiotic resistance among Gram-negative pathogens is a serious threat to public health worldwide. A lack of new antimicrobial agents, particularly those against multidrug-resistant Gram-negative bacteria further aggravates the situation, highlighting an urgent need for development of effective antibiotics to treat multidrug-resistant Gram-negative infections. Past efforts to improve existing classes of antimicrobial agents against drug-resistant Gram-negative bacteria have suffered from established (intrinsic or acquired) resistance mechanisms. Consequently, the essential LpxC enzyme in the lipid A biosynthesis, which has never been exploited by existing antibiotics, has emerged as a promising antibiotic target for developing novel therapeutics against multidrug-resistant Gram-negative pathogens. </p><p>In Chapter I, I survey the medically significant Gram-negative pathogens, the molecular basis of different resistance mechanisms and highlight the benefits of novel antibiotics targeting LpxC. In Chapter II, I discuss a structure-based strategy to optimize lead compounds for LpxC inhibition, revealing diacetylene-based compounds that potently inhibit a wide range of LpxC enzymes. The elastic diacetylene scaffold of the inhibitors overcomes the resistance mechanism caused by sequence and conformational heterogeneity in the LpxC substrate-binding passage that is largely defined by Insert II of LpxC. In Chapter III, I describe the structural basis of inhibitor specificity of first-generation LpxC inhibitors, including L-161,240 and BB-78485 and show that bulky moieties of early inhibitors create potential clashes with the &#61538;a-&#61538;b loop of Insert I of non-susceptible LpxC species such as P. aeruginosa LpxC, while these moieties are tolerated by E. coli LpxC containing long and flexible Insert I regions. These studies reveal large, inherent conformational variation of distinct LpxC enzymes, providing a molecular explanation for the limited efficacy of existing compounds and a rationale to exploit more flexible scaffolds for further optimization of LpxC-targeting antibiotics to treat a wide range of Gram-negative infections. </p><p>In Chapters IV and V, a fragment-based screening and structure-guided ligand optimization approach is presented, which has resulted in the discovery of a difluoro biphenyl diacetylene hydroxamate compound LPC-058 with superior activity in antibacterial spectrum and potency over all existing LpxC inhibitors. In Chapter VI, I describe our efforts to improve the cellular efficacy of LPC-058 by reducing its interaction with plasma proteins, such as human serum albumin (HSA). The binding mode of LPC-058 was captured in the crystal structure of HSA/LPC-058 complex. The acquired structural information facilitated the development of the dimethyl amine substituted compound LPC-088 that displays significantly improved cellular potency in presence of HSA.</p> / Dissertation

Burkholderia cepacia complex (Bcc) species are opportunistic pathogens in patients with cystic fibrosis (CF), which are able to cause lethal infections. The Bcc are inherently resistant to most classes of antibiotics, which makes successful treatment problematic. Lipid A (also known as endotoxin), the hydrophobic anchor of lipopolysaccaride (LPS), is the bio-active component of LPS. One of several unique characteristics of the lipid A of the Bcc, is the permanent attachment of 4-amino-4-deoxy-L-arabinose (L-Ara4N) to the lipid A molecule. Also, the genes involved in L-Ara4N biosynthesis are necessary for viability in B. cenocepacia. Here we present research on lipid A biosynthesis, modi cation, and transport in the Bcc and highlight promising antimicrobial targets. The synthetic antibiotic CHIR-090 is an inhibitor of LpxC, an enzyme involved in the lipid A biosynthetic pathway. I investigated the activity of CHIR-090 against the Bcc and found that sensitivity to this antibiotic was both species- and strain-specific. CHIR-090 displayed MICs between 0.1 and 12.5 μg/ml against a panel of B. multivorans, the most prevalent Burkholderia species in CF. The species- and strain-specific sensitivity towards CHIR-090 was further explored and a strong correlation was found between the presence of a unique open reading frame, named LpxC2, in resistant species. To address the problem of multiple drug-resistance of the Bcc, we investigated the activity of the pyridoxal 50-phosphate (PLP)-dependent enzyme inhibitor cycloserine (CS) against the Bcc. CS is used as a second line of defense against M. tuberculosis. The activity of the D-enantiomer of CS (DCS) against the Bcc was tested and displayed MICs between 2 and 128 μg/ml and acted bactericidal towards the Bcc. Additionally, DCS inhibition of recombinant ArnB from B. cenocepacia J2315, a PLP-dependent enzyme necessary for viability in the Bcc, was studied. ArnB was inhibited reversibly by DCS. ArnB was further explored as a promising drug-target in the Bcc, but only CS has been identified as an inhibitor so far. In this thesis it was attempted to find the reason why is L-Ara4N modification of lipid A necessary for viability in B. cenocepacia. Therefore, two proteins were characterised, which are involved in lipid A transport: LptA, the periplasmic lipid A binding protein, and LptB, the cytoplasmic ATP-ase. LptA was found to be able to bind both modified and unmodified lipid A in vitro and therefore is not L-Ara4N specific. Furthermore, LptA could bind deep-rough-, rough-, and smooth- LPS, similar to that described for Escherichia coli LptA. The kinetic parameters of LptB were determined in vitro (kcat = 5.71 min-1 and KM = 0.88 mM), and were comparable to E. coli LptB. The ATP-ase activity of LptB was not influenced by the presence of any forms of LPS (modified or non-modified). Therefore, we concluded that both B. cenocepacia J2315 LptA and LptB are not L-Ara4N specific.