Synthesis and in vitro antimycobacterial activity of B-ring modified diaryl ether InhA inhibitors

Tuberculosis (TB) is responsible for more than 1.6 million deaths per annum with 8.8 million new cases being reported each year. These numbers make TB one of the leading infectious causes of death, eclipsed only by AIDS. In addition, according to the World Health Organization, the number of multi-drug-resistant and extensively drug-resistant TB cases is growing with almost a half million new cases being reported each year. Therefore, there is an urgent need to develop novel TB chemotherapeutic agents.1 The current front-line treatment strategy utilizes isoniazid (INH), a pro-drug which inhibits the synthesis of mycolic acids that are essential components required for the integrity of the bacterial cell wall.2 INH inhibits InhA, the FabI enoyl reductase (ENR) in the fatty acid synthesis (FAS-II) pathway. However, before INH can inhibit InhA, it must be activated by KatG, a catalase-peroxidase enzyme. The activated form of INH then reacts with NAD+ to form the INH-NAD adduct (Scheme 1).3–7 A significant number of the strains resistant to INH arise from mutations in KatG.8–11 Therefore, the development of an InhA inhibitor which can by-pass this initial activation step should have activity against INH-resistant strains of Mycobacterium tuberculosis (MTB). Scheme 1 Enoyl reductase inhibitors. The diphenyl ether triclosan (Scheme 1) is a potent inhibitor of ENR’s from many organisms including Escherichia coli and Plasmodium falciparium.12–20 However this compound only inhibits InhA with a Ki value of 0.2 µM.21 Using structure-based drug design we developed a series of alkyl diphenyl ethers that are potent inhibitors of InhA, with Ki values as low as 1 nM and MIC90 values of 1–2 µg/mL against M. tuberculosis H37Rv.22 Importantly the alkyl diphenyl ethers display similar MIC values against INH-resistant strains of MTB.22 However, despite their promising in vitro activity, these compounds have relatively low solubility and have ClogP values greater than 5, which is likely one reason why they have limited in vivo efficacy.23 Based on the observed relationship between lipophilicity and in vivo efficacy, especially as it pertains to antibacterial compounds,24, 25 we synthesized a series of analogues that incorporated functionalities designed to increase the polarity of the parent diphenyl ether InhA inhibitors. The effect on compound polarity was estimated by calculating the logP value (ClogP) for each compound synthesized. In this study we describe two classes of molecules in which alterations have been made to the diphenyl ether ‘B’ ring. In one series of compounds we have replaced the B ring with isosteric heterocycles that incorporate nitrogen atoms within the ring, thereby causing little steric perturbation to the overall structure of the molecule (Scheme 2). The second series of compounds have nitro, amino, amide and piperazino functionalities incorporated at the ortho, meta, or para positions of the B ring (Scheme 3 and Scheme 4). This second series of compounds was synthesized not only to improve solubility but also to systematically identify positions on the B ring which could be substituted without diminishing biological activity. Scheme 2 Reagents and conditions: (a) K2CO3, DMAc, Δ, Y; (b) (CuOTf)2·PhH, Cs2CO3, EtOAc, toluene, 120°C, Y; (c) Hexyl ZnCl2, Pd(P(t-Bu)3)2, THF/NMP, 130°C; (d) BBr3, DCM, 0°C to rt. Scheme 3 Reagents and Conditions: (a) K2CO3, DMAc, Δ; (b) Hexyl ZnCl2, Pd(P(t-Bu)3)2, THF/NMP, 130°C; (c) Zn, HCl, EtOH, rt; (d) acyl chloride, NEt3, DCM; (e) BBr3, DCM, 0°C to rt. Scheme 4 Reagents and Conditions: (a) K2CO3, DMAc, Δ; (b) N-methyl piperazine, NaBH(OAc)3, DCE; (c) Hexyl ZnCl2, Pd(P(t-Bu)3)2, THF/NMP, 130°C; (d) BBr3, DCM, 0°C to rt. The synthesis of the heterocyclic diaryl ether compounds was initiated either by nucleophilic aromatic substitution or by Buchwald-Hartwig cross-coupling of the appropriate nitrogen heterocycle with 4-bromo or chloro-2-methoxy phenol producing 1a–f (Scheme 2).26, 27 This was followed by palladium catalyzed Negishi coupling of the diaryl ethers with hexyl zinc chloride to give 2a–f.28 Boron tribromide cleavage of the methyl ether was used subsequently to generate the respective phenols, 3a–f.29 Structural characterization of all compounds was performed using 1 H NMR and ESI/MS. The synthesis of the nitro, amino and amide-substituted compounds was performed using the series of reactions shown in Scheme 3. Nucleophilic aromatic substitution reactions with fluoronitrobenzenes were first used to generate compounds 4a–c.27 This was followed by Negishi coupling giving 5a–c followed by boron tribromide cleavage to give compounds 13a–c or zinc-mediated reduction giving anilines 6a–c.28–30 Cleavage of the methyl ether gave 14a–c while acylation of the anilines with acyl chlorides afforded compounds 7, 8 and 9a–c.29, 31 Boron tribromide cleavage then gave the final compounds 10, 11 and 12a–c.29 The piperazine derivatives were synthesized in a similar fashion starting with nucleophilic aromatic substitution with the 2- or 4-fluorobenzaldehyde to give 13a and b (Scheme 4).27 Subsequently, reductive amination with methyl piperazine and sodium triacetoxyborohydride produced 14a and b,32 whereas Negishi coupling followed by boron tribromide cleavage gave the final compounds 16a and b.28, 29 The in vitro activities of the ultimate products were evaluated using enzyme inhibition and whole cell antibacterial assays as described previously (Table 1–Table 3).22, 33, 34 In general, addition of a bulky substituent at either the ortho, meta or para position of the B ring of 19 or incorporation of most aromatic nitrogen heterocycles resulted in a significant reduction in both enzyme inhibition and antibacterial activity (Table 1 and Table 3). In contrast, introduction of either amino or nitro substituents at the ortho and para positions had only a minimal effect on activity (Table 2). The two most active compounds, 3c and 14a, have MIC90 values of 3.13 µg/mL, similar to that of 19, and have ClogP values of 4.97 and 5.24, respectively, compared to 6.47 for the parent compound (Table 1 and Table 2). In addition it is also worth noting that the pyrazine derivative 3e, has a ClogP value that is more than an order of magnitude lower than 19, but still only shows a 3-fold increase in MIC90 compared to the parent (Table 1). In general the MIC values correlated with the IC50 values for enzyme inhibition. Thus ortho and para amino substituents (14a,c) were well tolerated in addition to the meta nitro substituent (13b). In these three cases the IC50 values obtained using 100 nM InhA approached 50% of the enzyme concentration, indicating that these compounds are tight-binding enzyme inhibitors. Additional IC50 values were determined using 10 and 50 nM InhA in the enzyme assays. Subsequent linear regression analyses of the IC50 values as a function of enzyme concentration yielded estimates for Kiapp of 21 ± 3 nM (13b), 16 ± 12 nM (14a) and 40 ± 3 nM (14c). Thus, introduction of a meta nitro (13b) or an ortho amino (14a) group into the B ring of the parent compound 19 has only a minor effect on the affinity of the inhibitor for the enzyme. Compound 14a is of particular interest because this derivative has a MIC90 value that is close to the value determined for 19. These data provide important information on the structural flexibility of the inhibitor binding-site that will be useful in directing the design of additional compounds. Table 1 Inhibition and solubility data for B-ring heterocycles. Table 2 Inhibition and solubility data for nitro and aniline compounds. Table 3 Inhibition and solubility data for amide and piperazine compounds. In conclusion, a series of hexyl diaryl ethers were synthesized in which the B ring of compound 19 has been substituted with a variety of groups, or replaced with nitrogen-containing aromatic heterocycles. Several of these new compounds possess MIC90 and Kiapp values similar to that of 19 while having significantly improved ClogP values. Studies are currently underway to determine whether the modifications that we have introduced have resulted in an increase in compound bioavailability and an improvement in their in vivo antibacterial activity.