Development of biotin protein ligase inhibitors as new antibiotics to treat Staphylococcus aureus

Abstract

There is an urgent need to discover new antibiotics to combat the rise of antibiotic resistant bacteria, such as methicillin resistant Staphylococcus aureus (MRSA). Many of the antibiotics currently in clinical use are synthetic derivatives of chemical scaffolds identified over 50 years ago in the golden era of antibiotic drug discovery. These antibiotics are often subject to existing resistance mechanisms and, as such, represent a short term solution to the antibiotic resistance crisis. Therefore it is imperative that new classes of antibiotics are developed that exhibit new modes of action and that are not subject to existing resistance mechanisms. Most antibacterial discovery efforts are focussed on drug targets with no mammalian equivalent. These targets have been well explored and therefore new antibacterial targets need to be identified. One strategy to identify new antibiotics is to explore targets that have a closely related human homologue. However, it is important that such inhibitors exhibit extremely high selectivity for the bacterial target over the human equivalent. One example of such a target is the essential enzyme biotin protein ligase (BPL) which catalyses the attachment of the micronutrient biotin onto biotin-dependent enzymes. In bacteria biotin-dependent enzymes play important roles in fatty acid synthesis and the tricarboxylic acid cycle. Without protein biotinylation these enzymes are devoid of activity and unable to perform their essential metabolic functions. Hence, inhibitors of BPL with selectivity over the human homologue represent a potential new class of antibiotic to combat MRSA. Our group has previously reported the X-ray crystal structure of S. aureus BPL (SaBPL) that provides the essential information necessary for structure guided design of new inhibitors. Of particular importance are two adjacent binding sites for the ligands biotin and ATP which, when bound, conjugate to form the adenylated reaction intermediate, biotinyl-5ʹAMP. Whilst amino acid residues in the biotin-binding pocket are highly conserved, residues in the ATP binding pocket are more variable and can be exploited to create species selective inhibitors. Our laboratory has previously reported analogues of biotinyl-5ʹAMP as BPL inhibitors where the labile phosphoanhydride linker present in the native reaction intermediate has been replaced with a non-hydrolysable 1,4-disubstituted-1,2,3-triazole linker. The triazole linker can be readily synthesised by the Huisgen cycloaddition reaction that occurs between an acetylene and azide. This cycloaddition reaction can proceed in two ways. Firstly, copper or ruthenium catalysts can be used to produce the 1,4 or 1,5 regio-isomers respectively. Alternatively, in special cases, this reaction can be catalysed by an enzyme. This is known as in situ click chemistry. Our laboratory has identified a biotin triazole pharmacophore, containing the biotinyl moiety and a 1,4-disustituted triazole. Various groups that can probe available binding sites on SaBPL can be conjugated to the triazole through click chemistry. The most potent triazole inhibitor of SaBPL, BPL068, had an inhibition constant of 90 nM and, importantly exhibited >1000-fold selectivity over the human homologue (Soares da Costa et al, Journal of Biological Chemistry, vol. 287, p 17823-17832). Here, the biotin triazole was conjugated to a 2-benzoxalone moiety that was designed to bind in the ATP binding pocket. This compound inhibited growth of S. aureus and did not show any in vivo cytotoxicity against cultured mammalian cells. Although BPL068 exhibited antibacterial activity, the effect was not strong enough to determine a minimal inhibitory concentration (MIC), which is required for a pre-clinical candidate. The first aim of this project was to characterize new SaBPL inhibitors with the goal of improving the antibacterial activity of the parent compound. Here I have employed structure guided drug design and protein biochemistry techniques to design new SaBPL inhibitors with desirable properties for pre-clinical candidates. To facilitate the characterization of SaBPL inhibitors I developed a high-throughput enzyme assay to measure protein biotinylation, and a surface plasmon resonance assay to determine the kinetics of ligand binding (Chapter 4). With these techniques in hand I have characterised 40 rationally designed SaBPL inhibitors. Biotinol-5ʹAMP, a literature compound that has previously been developed as a research tool to characterize BPL function, was first characterized. Here the inhibition of BPLs from a panel of clinically important bacteria was measured using an in vitro protein biotinylation assay. The spectrum of whole cell antibacterial activity was also addressed, with S. aureus and Mycobacterium tuberculosis being most susceptible to this compound (Chapter 5). A series of triazole inhibitors of SaBPL, designed to probe the ribose binding pocket was also investigated. Here 25, 1,4-triazole based compounds with 1-benzyl substituents were synthesized and tested for inhibition of SaBPL. These compounds are smaller in molecular weight compared to the parent molecule, BPL068, allowing further optimization by extending into the ATP binding pocket. The most potent compound from this series had an inhibition constant of 280 nM and exhibited antibacterial activity against S. aureus. Furthermore, all compounds did not inhibit the human homologue or cultured mammalian cells (Chapter 6). A further series of compounds were next synthesized to optimise the triazole linker in BPL068. Firstly, compounds were synthesized in which the triazole linker has been replaced with alternative heterocycles with a view to improving its biological activity. A 1,2,4-oxadizole linker was found to inhibit SaBPL with an inhibition constant of 1.2 μM, with no inhibition of the human homologue (Chapter 7). A separate series of 1,4,5-trisubstituted-1,2,3 triazole analogues were also investigated (Chapter 8). Here, the hydrogen of the C5 atom in the triazole heterocycle was replaced with halogenated substituents to investigate whether halogenation of BPL068 could improve antibacterial activity. A 5-fluoro-1,2,3 triazole was found to inhibit SaBPL with an inhibition constant of 420 nM. Importantly, the fluorinated analogue exhibited an MIC of 8 μg/mL against a clinical isolate of S. aureus. This compound is the first example of a triazole based BPL inhibitor in which an MIC could be determined. Following the identification of BPL antibacterials, the mechanism of action needed to be addressed. Therefore the second aim of my project was to develop probes that could facilitate mechanism of action and uptake studies. Here a fragment based approach was employed using in situ click chemistry. In situ click chemistry relies on the ability of the target enzyme to select out and synthesize its own inhibitors from a series of small molecule building block precursors. This technique exploits the Huisgen cycloaddition reaction that proceeds between an acetylene and an azide to produce the 1,2,3-triazole heterocycle. Here, I have demonstrated that the in situ click chemistry approach could be adopted to identify inhibitors of a panel of 4 BPLs from clinically important bacteria. Next, azide-functionalized analogues of 2 fluorophores were synthesized and tested for chemical ligation to biotin acetylene using BPL as a catalyst. The ‘clicked’ compounds were confirmed for inhibition of SaBPL and entry into S. aureus. This newly developed probe will be used in solution based and con-focal microscopy studies to probe the mechanism of entry and action in S. aureus (Chapter 9). In summary, I have characterized a new series of SaBPL inhibitors that have improved antibacterial activity and still maintain the selectivity required for a pre-clinical candidate. Using in situ click chemistry, I also developed a new inhibitor that will be used to probe the mechanism of entry and action of BPL inhibitors in S. aureus. The work demonstrated in this thesis will be used to help optimize BPL inhibitors, ultimately leading to the development of a pre-clinical candidate.