Selective inhibitors of protozoan protein N-myristoyltransferases as starting points for tropical disease medicinal chemistry programs

Abstract

Inhibition of N-myristoyltransferase has been validated pre-clinically as a target for the treatment of fungal and trypanosome infections, using species-specific inhibitors. In order to identify inhibitors of protozoan NMTs, we chose to screen a diverse subset of the Pfizer corporate collection against Plasmodium falciparum and Leishmania donovani NMTs. Primary screening hits against either enzyme were tested for selectivity over both human NMT isoforms (Hs1 and Hs2) and for broad-spectrum anti-protozoan activity against the NMT from Trypanosoma brucei. Analysis of the screening results has shown that structure-activity relationships (SAR) for Leishmania NMT are divergent from all other NMTs tested, a finding not predicted by sequence similarity calculations, resulting in the identification of four novel series of Leishmania-selective NMT inhibitors. We found a strong overlap between the SARs for Plasmodium NMT and both human NMTs, suggesting that achieving an appropriate selectivity profile will be more challenging. However, we did discover two novel series with selectivity for Plasmodium NMT over the other NMT orthologues in this study, and an additional two structurally distinct series with selectivity over Leishmania NMT. We believe that release of results from this study into the public domain will accelerate the discovery of NMT inhibitors to treat malaria and leishmaniasis. Our screening initiative is another example of how a tripartite partnership involving pharmaceutical industries, academic institutions and governmental/non-governmental organisations such as Medical Research Council and Wellcome Trust can stimulate research for neglected diseases.

Figure 1. Structures of representative, previously reported NMT inhibitor series.

Three distinct series of C.albicans NMT inhibitors, exemplified by SC-58272, UK-370485 and compound 5 , and a series of inhibitors of both T.brucei and L.donovani NMT (e.g. DDD85646), have been reported in the chemical literature. Co-crystal structures with their respective targets have shown that each inhibitor binds in the same region of the binding site as the substrate peptides (Figure 2).

Figure 2. Overlay of structures of inhibitors (see Figure 1 ) based on alignment of binding site residues.

The crystal structural information for NMTs from fungal (C. albicans, Saccharomyces cerevisiae), human and protozoan (L. major and L donovani) NMTs was used as the basis for modelling work. The binding-site residues were aligned to create a view of the occupation of the NMT binding sites across species (based on 1iyl, 2wsa, 1iyk, and an unpublished structure for UK-370485 in C. albicans NMT). Images were created using the Pfizer molecule-modelling package MoViT and the inhibitors colored (Yellow: SC-58272: Orange: Compound 5 ; Green: UK-370485; Pink: DDD85646). Despite binding in the same region of their respective NMT, each compound occupies a different sub-region, which supported the case for high-throughput screening as a source of novel NMT inhibitor series.

Figure 3. Comparison of protein identity and similarity for the NMT isoforms in this study.

Figure showing the percentage sequence identity (Blue) and sequence similarity (Yellow) as defined by BLAST between the NMT proteins from L. donovani (LdonNMT, EMBL accession number FN555136), P. falciparum (PfalNMT, Swiss-Prot Q81LW6), T. brucei (TbNMT, EMBL FN554973) and the catalytic domains of Human NMT isoforms Hs1NMT1 (Swiss-Prot P30419) and Hs2NMT (Swiss-Prot O60551.

Figure 4. Bar chart showing numbers of active compounds in each chemical series.

Compounds with >40% activity against either Ldon [red], Pfal [yellow] or both NMT targets [blue], were sub-divided by chemical series. Most of the series were selective for one NMT orthologue. Molecules active against both targets were often highly active against one NMT and weakly active against the other primary screening target.

Figure 5. Structures of representative hit compounds from NMT screening.

Triage of the primary screening hits made use of local hit rate analysis; ligand efficiency against the primary target; selectivity in the primary screen; and broader NMT selectivity. Eight of the series were judged to contain lead compounds capable of further development towards new medicines to treat malaria or leishmaniasis. Table 1 summarises the primary screen activity of selected preferred examples from each series and a further series representative when analogue screening identified a superior compound. The broader NMT screening profile of each compound is detailed in Table 2.

Figure 6. Plot of dose-response activity against Pfal vs. Hs1 NMTs for primary screening hit and analogue screening set.

Broader selectivity screening was carried out using 5 NMT orthologues and a set of 2066 compounds. A plot showing the potency of each compound against Pfal and Hs1 showed that most compounds lacked selectivity. All compounds included in the broad selectivity screening were labelled with their chemical series. The series with potential for further follow-up are disclosed in this paper. The other series were discarded due to lack of enzyme activity or NMT-selectivity. Two series stood out from the general trend; representative structures for these series are shown in Figure 5. Two additional series provided potent, non-selective inhibitors, which may also be useful as starting points for medicinal chemistry programs.

Figure 7. Plot of dose-response activity against Ldon vs. Hs1 NMTs for primary screening hit and analogue screening set.

In contrast to Figure 6, a plot of potency against Ldon and Hs1 NMTs showed that activity did not correlate. Excellent selectivity was observed for several chemical series. The structures of representative members of four series are given in Figure 5.