Treatment of bacterial infections is becoming a substantial clinical problem due to the increasing number of resistant bacterial strains. Many failed experiments in the development of antibacterial agents over the years have forced the pharmaceutical industry to gradually began to phase out or at least limit the financial contribution for this purpose, which only worsened the situation in healthcare institutions. A large part of the development of antibacterial agents was therefore transferred to the academia and other research institutions. Bacterial cell wall is one of the most important targets in the development of antibacterial agents. It is essential for the survival of bacteria, because it maintains the osmotic pressure inside the cell, protects the organism from environmental factors, preserves the characteristic cell shape and acts as a semi-permeable membrane. The cell wall of the bacteria is largely constructed of peptidoglycan, consisting of repeated amino sugar units that are cross-linked to another peptidoglycan chain through a peptide bridge. The biosynthesis of peptidoglycan is performed both on the inside and outside of the cell membrane. Many enzymes are involved in this process and they represent potential targets for the development of antibacterial agents. Antibacterial agents that act on the peptidoglycan biosynthesis lead to cell lysis and death of bacteria. Their main advantage is a selective toxicity and specificity. The specificity of the mycobacterial cell wall is a thick layer of mycolyl-arabinogalactan-peptidoglycan complex, which is interspersed with free lipids and numerous proteins. The complexity of such cell wall thus represents the biggest obstacle and concurrently also a challenge in the development of new agents for the treatment of tuberculosis. In this doctoral thesis, we present the design of novel inhibitors of Mur enzymes, UppS and PPB2a, which participate in either the cytoplasmic or the final modification stages of the peptidoglycan biosynthesis. We also devoted our effords to develop inhibitors of the InhA enzyme, which involved in the synthesis of mycolic acids that represent a significant proportion of the cell wall of mycobacteria. We focused on the modern methods for drug design, such as structure-based design and screening of focused libraries. The development of computational methods has undergone a major upswing in the recent years and is currently the leading approach of new compounds design for many therapeutic targets. Within the doctoral thesis we used various well established computational methods and at the same time we also validated some recent programs developed by our partners from the National Institute of Chemistry in Ljubljana. The obtained compounds were subsequently biologically evaluated and used as a starting point for further optimization. The cytoplasmic Mur enzymes catalyze the formation of UDP-N-acetylmuramoyl-pentapeptide, the soluble precursor of the peptidoglycan chain. The inhibition of these enzymes thus prevents the formation of the cell wall of the bacterium and consequently causes its death. The ProBiS-CHARMMing web server was used to simulate the induced-fit effect of MurA from E. coli upon binding of the TAV ligand from the crystal structure of the MurA from E. cloacae. This led to the opening and enlarging of the active site of the enzyme, which was successfully used for docking of filtered `ZINC Drugs Now´ compound library. Three hit compounds were obtained, namely the dicarboxylic acid, quinazolinone and pyrrolopyridine, which in vitro inhibited MurA from E. coli in the micromolar concentrations (IC50 = 1 μM, 82 μM and 109 μM, respectively). A large series of quinazolinone type derivatives were synthesized by systematically modifying individual quinazolinone fragments and introducing different substituents. The vast majority of changes led to a decreased inhibitory potency. The removal of the methylene bridge at the site 2 between the two nitrogen of quinazolinone and the simultaneous introduction of the 5-nitrofuranyl group to this site, not only improved the inhibitory potency against MurA, but also gained antibacterial activity against E. coli and S. aureus. The 2-(5-nitro-furan-2-yl)-3H-quinazolin-4-one (IC50 = 47 μM, MIC = 1-8 μg/mL) and its 3-isopropyl derivative (IC50 = 87 μM, MIC = 2-8 μg mL) are considered a good starting point for further optimization of MurA inhibitors in the development of antibacterials. During the screening of GlaxoSmithKline focused library of kinase inhibitors on Mur ligases, 5 different structural classes of inhibitors were discovered and then re-synthesized. The inhibitory potency of styrene derivative MH-96 against MurC, MurD and MuF enzymes (IC50 = 39-104 μM) was confirmed using biochemical assays. In addition, the binding of this compound to MurD was confirmed using the STD-NMR. We postulate that this inhibitor could bind to the ATP binding site and act as an ATP competitive inhibitor; however additional experiments are needed to confirm this hypothesis. This compound thus represents a promising starting point for further optimization and development of antibacterial agents.
UppS synthase is involved in the formation of lipid II in the cytoplasm of the bacteria via catalyzing the formation of UPP. Although the enzyme is a validated antibacterial drug discovery target, the mechanism of action and inhibition of the enzyme is still fairly unexplored. Based on the currently most promising bisphosphonate inhibitors of UppS, more specifically BPH-629, an in silico pharmacophore model was designed and used to filter a library of several million compounds. Simultaneously with the library preparation, the protein construction was carried out by taking into account the importance of the two binding sites, among which site 1 coincide with the binding sites of the farnesyl pyrophosphate substrate. The biochemical evaluation further demonstrated the inhibitory potency of three benzoic acid derivatives in the micromolar range (IC50 = 24-45 μM). Carboxylic groups of the compounds mimic the pyrophosphate parts of the substrates, therefore a competitive mechanism of UppS inhibition was assumed. Despite the fact that the compounds do not posses antibacterial activity, they provide a good starting point for further optimization. Transpeptidase PBP2a is responsible for peptidoglycan cross-linking in the final stages of bacterial cell wall synthesis. Specific inhibition of PBP2a is crucial in the development of narrow-spectrum antibiotics, since the enzyme is typically expressed in methicillin resistant bacterial strain (MRSA) as a mechanism of defense against the effects of penicillins. Similarity search was performed based on previously mentioned query quinazolinone, which according to crystallographic data utilizes the binding to the allosteric site of PBP2a as mechanism of inhibition, similarly to a known drug ceftarolin. Further on, the pre-filtered compound library was docked in the allosteric site of the enzyme. Biochemical assays demonstrated the binding and inhibitory potencies of three hit compounds. Among these, quinazoline derivative Z729094432 inhibited PBP2a at 100 μM, while showing antibacterial activity against MRSA (MIC = 16 μg/mL), but not against S. aureus (MIC > 128 μg/mL), indicating a possible link between enzymatic inhibition and antibacterial activity. The discovery of PBP2a inhibitors with distinctive chemical structures supported by antibacterial activities, in our opinion, present a promising basis for further development of antibacterials. The enoyl acyl carrier protein reductase InhA is an important target in the development of drugs for the treatment of tuberculosis. The new inhibitors were designed using two methods. First, a similarity search based on the known nanomolar inhibitors, thiadiazole 8 and tetrahydropyrane 9, was performed using program LiSiCa. In the second case, the search for known ligands from the PDB bank was implemented using the ProBiS plugin, which collects ligands from active sites of all enzymes whose structure resembles the structure of InhA. 17 top-scoring compounds based on similarity and 8 ligands from the PDB database were purchased and biochemically evaluated. For the most potent compound, the ligand 09T, a mixed type of inhibition with a given Ki value of 4 ± 1 μM was determined. Although the compound is a less potent inhibitor than query compounds, it can serve as a good fragment for further optimization due to its relatively small molecular weight.
|
