Identification of arylamine N-acetyltransferase inhibitors as an approach towards novel anti-tuberculars

Abstract

New anti-tubercular drugs and drug targets are urgently needed to reduce the time for treatment and also to identify agents that will be effective against Mycobacterium tuberculosis persisting intracellularly. Mycobacteria have a unique cell wall. Deletion of the gene for arylamine N-acetyltransferase (NAT) decreases mycobacterial cell wall lipids, particularly the distinctive mycolates, and also increases antibiotic susceptibility and killing within macrophage of Mycobacterium bovis BCG. The nat gene and its associated gene cluster are almost identical in sequence in M. bovis BCG and M. tuberculosis. The gene cluster is essential for intracellular survival of mycobacteria. We have therefore used pure NAT protein for high-throughput screening to identify several classes of small molecules that inhibit NAT activity. Here, we characterize one class of such molecules—triazoles—in relation to its effects on the target enzyme and on both M. bovis BCG and M. tuberculosis. The most potent triazole mimics the effects of deletion of the nat gene on growth, lipid disruption and intracellular survival. We also present the structure-activity relationship between NAT inhibition and effects on mycobacterial growth, and use ligand-protein analysis to give further insight into the structure-activity relationships. We conclude that screening a chemical library with NAT protein yields compounds that have high potential as anti-tubercular agents and that the inhibitors will allow further exploration of the biochemical pathway in which NAT is involved.

Graphical abstract

Keywords

N-acetyltransferase / Mycobacterium tuberculosis / triazoles / screening

INTRODUCTION

There is an urgent need for new drugs for tuberculosis (TB), and new compounds have been described despite the fact that the targets are not clearly established (Al-Balas et al., 2009). Discovery of new targets is important to identify novel drugs (Cole and Azari, 2005; Makarov et al., 2009). It has been recognized that understanding important metabolic pathways is likely to be a useful route towards new therapies (www.tballiance.org) as a result of the success with an ATPase inhibitor (Diacon et al., 2009). Recent identification of menaquinone biosynthesis as a useful target pathway provides a growing portfolio of potential novel anti-tuberculars (Dhiman et al., 2009). The need for a wide range of candidate targets and molecular classes is insurance for the future (Maxmen and Barry, 2009; Lin et al., 2009). The biggest difficulty with anti-tubercular and indeed anti-microbiological agents is the development of resistance. The development of multi-therapies that work in a short time period is an important aim in order to understand more about the organism that causes tuberculosis and also to overcome the disease. Tuberculosis is a global problem compounded by the emergence of drug-resistant tuberculosis particularly in association with AIDS (Dye, 2006). Whilst tuberculosis is endemic in many developing countries, there are also urban areas in the United Kingdom where TB is now a common disease through the combination of high population density and immigration from countries with endemic TB (Story et al., 2006). Mycobacterium tuberculosis, the organism that causes TB, survives within macrophage in the host and is likely to be a contributory factor in the extended duration of therapy required for eradication of tuberculosis, as is the existence of persisting, but non-dividing organisms both in and remote from the human lung (Neyrolles et al., 2006).

Functional genomics has identified possible roles for several previously unknown gene products in M. tuberculosis (Rengarajan et al., 2005), and investigation of the role of candidate gene products has also resulted in novel functional information. Arylamine N-acetyltransferase (NAT) was initially identified as an interesting target in mycobacterial research because in humans one of the NAT isoenzymes (known as human NAT2) (Boukouvala and Fakis, 2005) had been demonstrated to be responsible for the acetylation and inactivation of the still widely used anti-tubercular drug isoniazid (INH) (Evans et al., 1960; Weber and Hein, 1985). The discovery of a nat homolog in M. tuberculosis led to the hypothesis that NAT had a role in drug resistance (Payton et al., 1999). The nat gene is transcribed and the NATenzyme is active in both M. bovis BCG and M. tuberculosis (Upton et al., 2001). To investigate drug resistance, a deletion of the nat gene in M. bovis BCG was characterized (Bhakta et al., 2004). The major target of isoniazid, a pro-drug which is activated inside mycobacteria, has been the subject of dispute, and it is likely that isoniazid has several targets (Raman et al., 2005). With isoniazid acetylation by NAT competing with activation of the pro-drug by the katG gene product (Nagy et al., 1997; Bhakta et al., 2004), NAT is likely to be a minor target of isoniazid because deleting the nat gene resulted in a small increase in the sensitivity of the Δnat strain to isoniazid. However, the effect of knocking out the gene was much more dramatic. The effects on the growth, morphology, lipid composition, sensitivity to antibiotics and most importantly the intracellular killing of the Δnat strain of M. bovis BCG within macrophage suggested that the NAT enzyme, per se, would be a useful target for anti-tubercular therapy (Bhakta et al., 2004). Deletion of the nat gene was the cause of the changes in the genetically modified strain, since complementation of the Δnat strain restored the original phenotype. The observation that the nat gene appears to be essential for survival of M. bovis BCG within macrophage (Bhakta et al., 2004) is consistent with the observation that the nat gene encodes for a protein that is involved in the synthesis of cell wall complex lipids, including mycolates (Bhakta et al., 2004). It has been shown that the nat gene is encoded as the last gene within an operon (Anderton et al., 2006). The flanking regions differ by only a single base (in the first gene of the operon) between M. bovis BCG and M. tuberculosis (Fig. 1). The first two genes of the operon, HsaA and HsaD, have been identified as essential genes for survival of M. tuberculosis within macrophage (Rengarajan et al., 2005). In addition, inhibitors of the HsaC enzyme encoded by the operon (Fig. 1) have a similar effect on the lipid profile as does deletion of the nat gene in M. bovis BCG, that is the mycolates and complex lipids are diminished but the phospholipids are not (Anderton et al., 2006). More recently, it has become clear that HsaC is essential for survival of M. tuberculosis in the host (Yam et al., 2009). It has been demonstrated that the operon is part of the KstR regulon that is essential for intracellular survival inside macrophage (Kendall et al., 2007). The genes encoding HsaA–D have a role in cholesterol catabolism (Van der Geize et al., 2008) and the utilization of propionyl CoA as a cofactor by the TBNAT enzyme links the NAT gene product functionally with the operon (Lack et al., 2009).

To obtain more information on the role of the NAT enzyme as a target for anti-tubercular therapy, a high-throughput screen has been established using recombinant NAT protein (Brooke et al., 2003a, b; Russell et al., 2009). The pure recombinant NAT enzyme from M. tuberculosis is relatively insoluble with maximum yields approaching approximately 2 mg per Liter of culture (Upton et al., 2001; Sikora et al., 2008; Lack et al., 2009; Fullam et al., 2009). In order to optimize the screening procedure, we have therefore used NAT proteins that have similar substrate specificities but are produced more abundantly as recombinant proteins at around 20 mg per liter of culture and that store well when pure (Westwood et al., 2005; Fullam et al., 2007,2009). NAT isoenzymes have been characterized from a range of organisms and crystal structures have been obtained for the enzyme from Salmonella typhimurium (Sinclair et al., 2000), Pseudomonas aeruginosa (Westwood et al., 2005), the mycobacterial species Mycobacterium smegmatis (Sandy et al., 2002) and more recently from Mycobacterium marinum (Fullam et al., 2007; PDB code 2vfb), which is 74% identical to the enzyme from M. tuberculosis (Fullam et al., 2009). A preliminary manual screen of a limited range of compounds using the NAT enzyme from M. smegmatis had already proved successful in identifying NAT inhibitors (Brooke et al., 2003b). We report here that specific NAT inhibitors mimic the effects of deleting the nat gene in M. bovis BCG and that these compounds are also effective in inhibiting the growth of M. tuberculosis.

RESULTS

Screen design

The substrate specificity profiles of NAT enzymes from a range of bacterial NATs are extremely similar to each other but are distinct from the NAT isoenzymes from humans and other mammals (Westwood et al., 2005; Westwood et al., 2006; Kawamura et al., 2005, 2008). In order to identify compounds that would be unlikely to be metabolized by the human NAT isoenzymes, NAT inhibitors that would be effective against the prokaryotic NAT enzymes but were not inhibitors of the eukaryotic NAT enzymes were sought. NAT enzymes were chosen where sufficient stable recombinant enzyme (tens of milligrams of each protein) was available. The eukaryotic enzymes were pure hamster NAT2 (shNAT2) and its human homolog human NAT1 (hNAT1) (Kawamura et al., 2005; Wang et al., 2005). The prokaryotic NAT enzymes were from M. smegmatis (Sandy et al., 2002) (MSNAT), S.typhimurium (Sinclair et al., 2000) (STNAT) and P. aeruginosa (Westwood et al., 2005) (PANAT). The NAT inhibitor screening method is as described previously (Brooke et al., 2003a, b; Russell et al., 2009). The assay method relies on a colorimetric reaction linked to the NAT-catalysed hydrolysis of acetyl Coenzyme A (acetylCoA), which occurs at a greatly enhanced rate in the presence of a substrate for the acetylation reaction. Conditions for a linear initial rate of acetylCoA hydrolysis were determined for each NAT enzyme included in the screen. Recombinant NAT from M. marinum (MMNAT) was also used for quantitative analyses of compounds as specific prokaryotic NAT inhibitors following the screening process, and a crystal structure for the M. marinum enzyme is also now available (Fullam et al., 2007).

Inhibition of the hydrolysis of acetylCoA was determined using a library of 5000 drug-like compounds (Russell et al., 2009). The compounds in the library had been compiled to represent a structurally diverse set of chemically tractable entities without reactive groups such as aldehydes, Schiff bases or aminals, or known cytotoxic agents (such as polynitro aromatics). The structures are also amenable to both resynthesis and rapid diversification (Russell et al., 2009). Each compound was screened in triplicate against each enzyme with two substrates. Hits in the screen were compounds that inhibited, with both substrates, at least two of the prokaryotic NATenzymes more than 50% at 30 µM and that did not inhibit either of the eukaryotic NATenzymes more than 20% at the same concentration. This was also coupled with Z-score analysis to compensate for any systematic errors related to plate position (Brooke et al., 2003a). The strategy used was designed to maximize the hit rate and to minimize both false-positive and false-negative results.

Using this screening strategy, a number of structurally distinct small molecules were identified as hits. The sultam inhibitors, which were identified in the previous one-enzyme screen (Brooke et al., 2003b), did not appear as hits in this more discriminating five-enzyme screen. In order to confirm the inhibitory activity, the hits were re-screened using more stringent criteria against the same panel of five NATenzymes. Only compounds that satisfied the criteria of showing over 80% inhibition of at least two prokaryotic enzymes and showing less than 20% inhibition of the eukaryotic enzymes at 30 μM in the re-screen were carried forward (Fig. 2). At the completion of the second round of screening, of the six compounds that were identified in distinct chemical classes, two compounds were weak NAT inhibitors upon retesting with the M. marinum NAT enzyme such that less than 80% inhibition was found at 30 µM. Neither of these compounds was therefore pursued further.

For the compounds that were investigated further, effects on growth of M. tuberculosis H37Rv and M. bovis BCG on agar have been determined (Anderton et al., 2006). Table 1 shows data for the remaining four compounds: compound 1 inhibits PANAT at less than 10 μM but is less effective against the M. marinum enzyme. It is not a good inhibitor of growth of M. tuberculosis and therefore this compound was not pursued further. Of the three remaining compounds, pyrazole 2 had an IC₅₀ of 18.5 ± 1.0 µM against M. marinum NAT and had an MIC of less than 10 µg/mL against both M. bovis BCG and M. tuberculosis (Table 1). In addition, compound 3 was observed to inhibit MMNAT with an IC₅₀ of 6.0 ± 1.0 µM and to inhibit the growth of M. bovis BCG and M. tuberculosis with a minimum inhibitory concentration (MIC) of less than 5 µg/mL. Compound 3 had previously tested positive as an anti-tubercular agent in an extensive survey of 240 compounds, but no mechanistic information was provided (Jeney et al., 1956). Compound 4, a 1,2,4 triazole-3-thione derivative, was shown on resynthesis to have an IC₅₀ of 6.4 μM with PANAT and around 10 μM with MMNAT and to inhibit growth of M. tuberculosis with an MIC of less than 10 μg/ml. We have concentrated on compounds 3 and 4. Since compound 3 has a chiral centre, the present study describes the analysis of compound 4 and its analogues on NAT and M. tuberculosis. Structures incorporating the 1,2,4-triazole-3-thione motif and related examples (Macaev et al., 2005; Foroumadi et al., 2006) have been reported recently to have anti-tubercular activity although no mechanism of action has been proposed.

To summarize, a series of compounds that inhibit bacterial NAT were shown to suppress growth of M. tuberculosis at 10 µg/mL or lower concentration, and we report here on a detailed analysis of the 1,2,4-triazole-3-thione derivative (compound 4).

Compound 4 and its structural analogues

All of the compounds shown in Table 1 were resynthesised to confirm their identity, but we focus here on the resynthesis of compound 4 and a series of analogues. Compound 4 was synthesised as follows: p-tolyl chloride was condensed with thiosemicarbazide in pyridine (Kane et al., 1994; Singh, 1996). Cyclisation of the condensation product was effected by subsequent treatment with 1 M aqueous sodium bicarbonate to produce the precursor 5-(4'-methylphenyl)-1,2,4-triazole-3-thione (5) in 62% yield and 97% purity over two steps. Alkylation of 5 with methyl 3-bromopropionate produced the target methyl ester (4) in 52% isolated yield (Scheme 1).

Methyl ester 4 was subjected to saponification with lithium hydroxide. The free acid formed (6) was found to be a weaker inhibitor of NAT activity than 4 (IC₅₀ value of 20.1 ± 0.9 µM), and showed no inhibition of the growth of M. bovis BCG in vitro at any of the concentrations tested (up to 100 µg/mL).

The substituents at both the 3-position and the 5-position of the 1,2,4-triazole ring of 4 were varied to allow preliminary investigations into the structure-activity relationships. Consequently, a range of 1,2,4-triazole-3-thiones (7–12) was synthesised from their corresponding acid chlorides. With the exception of compound 7, the purified yields varied between 44% and 98% (Table 2). The purity of all synthesised compounds was > 95%, as assessed by HPLC (described in Materials and Methods). The 1,2,4-triazole-3-thiones were then treated with a range of electrophiles to generate an array of compounds with a range of substituents on the 3-position of the 1,2,4-triazole ring. An excess of the electrophile was used in order to drive the reaction to completion. Unreacted electrophile was subsequently removed by adding a nucleophilic scavenger resin, in order to avoid the need for further purification steps (Ley, 2000). Therefore, treatment of the crude reaction mixtures with polymer-supported methylthiourea (for activated electrophiles) or polymer-supported thiophenolate (for non-activated electrophiles) furnished the S-substituted 1,2,4-triazoles 13–18 in 18%–45% isolated yield and > 95% purity (Table 2).

Compound 4 exhibited competitive inhibition of NAT with respect to the substrate p-anisidine (ANS) at inhibitor concentrations of up to 10 µM (Fig. 3A) and the inhibition constant, K_ic, was found to be 6.0±0.3 µM (Fig. 3B). The synthetic precursor to compound 4, 1,2,4-triazole-3-thione, 5, showed very similar inhibitory properties (Table 2) with competitive inhibition at concentrations up to 10 µM, with a K_ic value of 8.6 ± 0.5 µM.

An aryl substituent at the C5 position on the 1,2,4-triazole ring was found to be essential for inhibition of NAT. For example, 5-tert-butyl-1,2,4-triazole-3-thione (7) showed no inhibition of NAT activity at any concentration tested (up to 50 µM, Table 2). Of the 5-aryl-1,2,4-triazoles synthesised, the 4’-methoxyphenyl (8) and 4’-chlorophenyl (9) derivatives were considerably less active as NAT inhibitors than the parent and precursor 4’-methylphenyl compounds, 4 and 5 (Table 2). However, the 3’-chlorophenyl-1,2,4-triazole-3-thione (10) showed an increase in activity as a NAT inhibitor relative to the 4’-methylphenyl triazole, with an IC₅₀ value of 244 ± 30 nM, a 60-fold increase in potency. However, the compounds that possessed a nucleophilic sulfur (where substituent R² is H, 5, 7–12, Table 2) while not observed to react with DTNB or Coenzyme A, were not pursued further due to their potential non-selective reactivity with intracellular electrophiles.

In exploring the preliminary structure-activity relationships for NAT ligands (Brooke et al., 2003a, b), we observed that the incorporation of a large hydrophobic substituent markedly improved the ligand potency. To probe this effect further, a series of 1,2,4-triazoles with long chain aliphatic or planar aromatic substituents on sulfur were evaluated. Clear structure-activity relationships could be ascertained among these compounds. The length of the chain of the S-substituent was estimated following energy minimization with Chem3D Pro (Cambridgesoft). The assumption was made that the carbon chain exists in an extended conformation where multiple conformations are accessible. The NAT inhibition potency increased with extended chain length up to approximately 10 Å, with the S-4-phenylbenzyl (16) and S-octyl (17) derivatives being the most potent in the series (Fig. 4A). The S-decyl derivative 18 has an approximate chain length of 13 Å and was a less potent inhibitor of NAT than the shorter-chain S-alkyl derivatives. The IC₅₀ against NAT of the S-decyl derivative (18) was 13.1 ± 1.1 µM, compared with an IC₅₀ of 3.8 ± 0.4 µM for the S-octyl-1,2,4-triazole (17).

The simulated annealing of compound 4 and its structural analogs to NAT from M. tuberculosis was then evaluated by docking with the program GOLD (Verdonk et al., 2003). The NATs from M. marinum and M. tuberculosis share 74% identity and 83% similarity at the amino acid level (Fullam et al., 2007), and the root mean square deviation between the 1080 backbone peptide atoms of the crystal structure of M. marinum NAT and the equivalent atoms in the model of M. tuberculosis NAT is 0.48 Å. The predicted binding orientation of 4 into the NAT from M. tuberculosis is shown in Fig. 3. The triazole ring was predicted to form π-stacking interactions with Phe¹³⁰, and to form H-bonds with the backbone carbonyls of Phe¹³⁰ and Gly¹³¹, residues in the proposed 'P-loop' (Sinclair et al., 2000). The carbonyl of the ester moiety of 4 is in position to H-bond with Gln¹³³. These anchor points result in the orientation of the 4’-methylphenyl moiety towards the active site Cys⁷⁰ residue, in a hydrophobic pocket bounded by Phe³⁸, Phe²⁰⁴, Val⁹⁵ and Thr¹⁰⁹. The correlation observed between chain length and NAT inhibitory potency (Fig. 4) was closely matched by the docking scores for the same series of compounds (Fig. 5). An increase in the calculated binding affinity was observed in the series: S-methyl (13), S-propyl (14), S-benzyl (15), S-4-phenylbenzyl (16) and S-octyl (17) 1,2,4-triazoles. When the S-substituent was a decyl chain (18), the docking score was lower than for the S-octyl derivative. Fig. 5 shows the relative binding orientations of this series of 1,2,4-triazoles, with compound 4 included for comparison. All except compound 18 were predicted to bind in essentially identical orientations. The major predicted binding interactions between these compounds and the M. tuberculosis NAT protein were very similar to those described above for compound 4: π-stacking of the 1,2,4-triazole ring with Phe¹³⁰ and H-bonding of one nitrogen of the 1,2,4-triazole ring with the carbonyl oxygens of the 'P-loop' residues, Phe¹³⁰ and Gly¹³¹ (Sinclair et al., 2000). The 4’-methylphenyl group in all molecules (except the S-decyl derivative 18) was predicted to point towards the active site cysteine residue, as described for 4, above.

Inhibition of growth of M. bovis BCG and M. tuberculosis

The series of 1,2,4-triazole-3-thiones was tested for inhibitory effects on the growth of M. bovis BCG and M. tuberculosis H37Rv. There was excellent correlation between the inhibitory effects on the growth of both organisms (Fig. 4B), and for all compounds there appeared to be greater potency in inhibition of M. tuberculosis growth. The growth inhibitory effects were specific, since there was no inhibition of growth of E. coli cultures at up to 100 μg/mL (Fig. 4C). Compound 4 was a modestly potent inhibitor of both NAT activity and the growth of mycobacteria. The relationship between inhibition of NAT activity and inhibition of growth of M. bovis BCG by compound 4 and its structural analogs was investigated. Carboxylic acid 6 was an inhibitor of the NAT enzyme in vitro. However, 6 was found not to be an inhibitor of mycobacterial growth. This may be due to poor cell wall permeability of the carboxylate salt of 6, which is expected to be the predominant form of 6 under the assay conditions. In contrast, there was good correlation between inhibitory potency against NAT and mycobacterial growth inhibition for the remaining S-substituted analogues of compound 4 (Fig. 4). Inhibitory activities on both the NAT enzyme and on growth of M. bovis BCG were affected similarly by the length of the S-substituent, with maximum inhibition corresponding to an extended chain length of around 9 to 10 Å.

NAT inhibitors mimic the effects of Δnat strain

Compound 4, which is a competitive inhibitor of NAT at less than 10 µM, was a good inhibitor of the growth of M. bovis BCG on agar. Treatment of M. bovis BCG with compound 4 resulted in a major effect on the lipid composition (Fig. 6C–E). The effect was very similar to that observed when the nat gene was deleted (Bhakta et al., 2004): there was a diminution in the amount of the mycolic acids and complex lipids, although the phospholipid composition appeared to be relatively unchanged (Fig. 6). In addition, it was observed that compound 4 acted synergistically with gentamycin, mimicking the effect of the genetic deletion of nat in M. bovis BCG. Gentamycin alone is not inhibitory against the growth of wild-type M. bovis BCG at 10 µg/mL (Fig. 6B), but does affect the growth of the Δnat strain of M. bovis BCG at 10 µg/mL (Bhakta et al., 2004).

In order to determine whether NAT inhibitors would also be active on the intracellular growth of M. bovis BCG, the effects were tested initially on macrophages alone. Compound 4 was not toxic to macrophages (Fig. 6F). However, when macrophage were infected with M. bovis BCG that had been exposed to compound 4, the bacilli were killed within the macrophage, as has been demonstrated with the Δnat strain (Fig. 6G) (Bhakta et al., 2004). The same effects were observed at the highest concentration when compound 4 was added after the macrophages had been infected (data not shown).

DISCUSSION

There is an urgent need for novel therapies for tuberculosis, and here we report a new potential drug target in M. tuberculosis: the arylamine N-acetyltransferase enzyme. Following from previous work on genetic deletion of the nat gene in M. bovis BCG (Bhakta et al., 2004), we now report on the identification and anti-mycobacterial activity of novel small-molecule inhibitors of NATs. From high-throughput screening, we have identified four compounds from chemically distinct structural classes that are active both as inhibitors of purified prokaryotic NAT enzymes and as inhibitors of the growth of M. tuberculosis.

We have synthesised an array of analogs of one of these high-throughput screening hits, the 1,2,4-triazole 4. Structure-activity relationships were identified for a series of S-substituted 1,2,4-triazole-3-thiones where the S-substituent was an aliphatic or planar aromatic group. These structure-activity relationships show that the maximal NAT inhibitory activity occurs when the S-substituent extended chain length is around 9–10 Å (Fig. 4). A similar pattern of interaction between a series of substrates (4-alkoxyanilines) and NAT from M. smegmatis has also been observed (Brooke et al., 2003b). These NAT substrates/ligands were shown to have anti-tubercular activity although by an unexplained mechanism (Nodzu et al., 1954). For the 1,2,4-triazole series described (Table 2, Fig. 4), the increase in alkyl chain length allows for more enthalpically favorable hydrophobic contacts within the active site, at the expense of an increasing entropic penalty for binding to the protein. The in silico docking experiments successfully predict the observed structure-activity relationships for the 1,2,4-triazole series (Fig. 5), and provide a rationale for the observed trends in decreased NAT inhibitory potency when the alkyl chain length is ten carbon units. Binding energy is most favorable and NAT inhibitory potency is maximal when the chain length is equivalent to eight carbons. Furthermore, the structure-activity relationships observed for inhibition of mycobacterial growth mirrored those for NAT inhibition. The availability of NAT inhibitors also allows for studies to investigate the role of NAT in mycobacterial cell wall synthesis and we are pursuing both the class that we describe here and the other classes of inhibitor that we have identified.

Recently, efficient methods for generating sufficient amounts of the NAT enzyme from M. tuberculosis have been identified (Sikora and Blanchard, 2008; Fullam et al., 2009) and the enzyme shows a very similar substrate specificity profile to the NAT enzyme from M. marinum (Fullam et al., 2009). The studies presented here where NAT has been used in a high-throughput screen have identified compounds that have anti-tubercular activity and the triazoles affect mycobacteria within macrophage. These studies also demonstrate an inhibitory effect on growth on agar, whilst the nat deleted mutant strain of M. bovis BCG does show slower growth. Although the nat gene does not appear to be essential for growth on agar, it does appear to be essential for survival of M. bovis BCG inside macrophage. Whilst these studies do not eliminate the possibility that the inhibitors may also act against an additional target in mycobacteria, they suggest that this approach has identified a novel series of compounds to pursue as anti-tubercular agents. The recent finding that the other genes of the operon in which NAT is encoded are essential for survival in vivo (Yam et al., 2009) and the observation of the role of the gene cluster in cholesterol metabolism (van der Giese et al., 2007; Lack et al., 2009) controlled by the KstR regulon (Kendall and Stoker, 2007) further underline the importance of this group of genes as targets for exploring anti-tubercular agents.

There is a move to identify compounds that affect more than one target and, in the fight against TB, potentially useful targets should not be ignored, especially when the target is important for intracellular survival of mycobacteria within macrophage.

MATERIALS AND METHODS

Protein production and enzymic assays

The NAT enzymes from M. smegmatis (Payton et al., 1999), S. typhimurium (Sinclair et al., 2000), P. aeruginosa (Westwood et al., 2005), M. marinum (Fullam et al., 2007), hamster NAT2 (Kawamura, 2005) and human NAT1 (Wang, 2005) were produced as recombinant proteins and purified as previously described. Activity assays were performed as previously described (Kawamura, 2005; Brooke, 2003a, b).

Computational docking of inhibitors with NAT

The program Modeller 8v2 (Sali and Blundell, 1993) was used to construct a model of the TBNAT protein from the known crystal structures of NATs from M. smegmatis (PDB code 1GX3) (Sandy et al., 2002), S. typhimurium (PDB code 1E2T) (Sinclair et al., 2000), M. marinum (PDB code 2VFB), P. aeruginosa (PDB code 1W4T) (Westwood et al., 2005) and Mesorhizobium loti NAT1 (PDB code 2BSZ) (Holton et al., 2005). A comparison of the model with M. marinum NAT (PDB code 2VFB) is shown in Supplementary Fig. 1. The protein file was prepared in pdb format by using the AddH tool within the Chimera molecular modeling package (Pettersen et al., 2004), and docking was performed with GOLD v3.0.1 (Verdonk et al., 2003). Ligand structures were drawn in ChemDraw (Cambridgesoft) and then converted to 3-dimensional structures in Chem3D. The molecules were energy-minimized with the MM2 force field and saved in sdf format for docking. The active site of the protein was defined by the Sγ atom of Cys⁷⁰ with an automatic cavity-detection radius of 20 Å. The docking solutions were ranked with the GoldScore algorithm. Lowest-energy docking solutions were merged with the protein structure and saved in the pdb format, prior to visualization with Aesop as previously described (Sinclair et al., 2000).

Mycobacterial growth inhibition in vitro

Mycobacteria (M. bovis BCG and M. tuberculosis H37Rv) were grown as spot cultures in 6-well plates on solid medium (Middlebrook 7H10 medium supplemented with 10% (v/v) oleic acid-albumin-dextrose-catalase (OADC)) as previously described (Anderton et al., 2006), with test compounds at the concentrations indicated in the text. Test compounds were added to the melted, partially cooled 7H10-OADC agar medium as solutions in DMSO, and the final concentration of DMSO in each well was 0.1% (v/v). The minimum inhibitory concentration (MIC) is defined as the concentration of inhibitor at which no growth of mycobacteria was detected after a period of 2–3 weeks.

Effect of NAT inhibitors on the infection of macrophages by M. bovis BCG

M. bovis BCG was grown for approximately 4 days in 7H9 medium (containing 10% (v/v) albumin-dextrose-catalase (ADC), 0.2% (v/v) glycerol and 0.05% (v/v) Tween-80) until the optical density at 600 nm was 1.0. The mycobacterial cells were washed and resuspended in 7H9-ADC medium (as above) supplemented with compound 4 (at 20 µg/mL) or isoniazid (INH) (0.01 µg/mL) in DMSO at a final concentration of 0.1% (v/v). A control was included where DMSO alone was used. After incubation with the test compound for 2 h, the mycobacterial cells were washed and resuspended in RPMI (containing 0.05% Tween-80) for infection of RAW 264.7 mouse macrophages. The RAW 264.7 cells were grown routinely in RPMI medium (supplemented with 10% (v/v) fetal bovine serum (FBS)) without antibiotics for 48 h at 37°C in a humidified, 5% CO₂ atmosphere prior to the infection. RAW 264.7 cells (5 × 10⁴ per well) were seeded onto 24-well tissue-culture plates and incubated (at least 30 minutes) with the test compound (at the concentrations shown in Fig. 5G) or DMSO alone, prior to infection with M. bovis BCG (5 × 10⁵ mycobacterial cells per well, prepared as described above). After 2 h, the adherent RAW 264.7 cells were washed twice with RPMI, and then resuspended in RPMI containing test compound or DMSO alone (at the concentrations shown in Fig. 5G). The infected RAW 264.7 cells were cultured for 72 h in RPMI-FBS medium containing either the test compounds or vehicle alone and then washed twice with RPMI alone and lysed by the addition of sterile water. The surviving M. bovis BCG cells in the lysed RAW 264.7 cell suspension were cultured on 7H10-OADC agar medium for 3 weeks at 37°C. The number of colony forming units (CFUs) in each case was determined and used for comparison between tests and controls.

Chemical methods: synthesis and characterization data

See supplementary information.

References

[1] Al-Balas, Q., Anthony, N.G., Al-Jaidi, B., Alnimr, A., Abbott, G., Brown, A.K., Taylor, R.C., Besra, G.S., McHugh, T.D., Gillespie, S.H., et al. (2009). Identification of 2-aminothiazole-4-carboxylate derivatives active against Mycobacterium tuberculosis H37Rv and the beta-ketoacyl-ACP synthase mtFabH. PLoS ONE 4, e5617.

[2] Anderton, M.C., Bhakta, S., Besra, G.S., Jeavons, P., Eltis, L.D., and Sim, E. (2006). Characterization of the putative operon containing arylamine N-acetyltransferase (nat) in Mycobacterium bovis BCG. Mol Microbiol 59, 181–192.

[3] Besra, G.S. (1998). Preparation of cell-wall fractions from mycobacteria. Methods Mol Biol 101, 91–107.

[4] Bhakta, S., Besra, G.S., Upton, A.M., Parish, T., Sholto-Douglas-Vernon, C., Gibson, K.J., Knutton, S., Gordon, S., DaSilva, R.P., Anderton, M.C., et al. (2004). Arylamine N-Acetyltransferase Is Required for Synthesis of Mycolic Acids and Complex Lipids in Mycobacterium bovis BCG and Represents a Novel Drug Target. J Exp Med 199, 1191–1199.

[5] Brooke, E.W., Davies, S.G., Mulvaney, A.W., Pompeo, F., Sim, E., and Vickers, R.J. (2003a). An approach to identifying novel substrates of bacterial arylamine N-acetyltransferases. Bioorg Med Chem 11, 1227–1234.

[6] Brooke, E.W., Davies, S.G., Mulvaney, A.W., Okada, M., Pompeo, F., Sim, E., Vickers, R.J., and Westwood, I.M. (2003b). Synthesis and in vitro evaluation of novel small molecule inhibitors of bacterial arylamine N-acetyltransferases (NATs). Bioorg Med Chem Lett 13, 2527–2530.

[7] Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S.V., Eiglmeier, K., Gas, S., Barry, C.E., et al. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544.

[8] Cole, S.T., and Alzari, P.M. (2005). Microbiology. TB–a new target, a new drug. Science 307, 214–215.

[9] Dhiman, R.K., Mahapatra, S., Slayden, R.A., Boyne, M.E., Lenaerts, A., Hinshaw, J.C., Angala, S.K., Chatterjee, D., Biswas, K., Narayanasamy, P., et al. (2009). Menaquinone synthesis is critical for maintaining mycobacterial viability during exponential growth and recovery from non-replicating persistence. Mol Microbiol 72, 85–97.

[10] Diacon, A.H., Pym, A., Grobusch, M., Patientia, R., Rustomjee, R., Page-Shipp, L., Pistorius, C., Krause, R., Bogoshi, M., Churchyard, G., et al. (2009). The diarylquinoline TMC207 for multidrug-resistant tuberculosis. N Engl J Med 360, 2397–2405.

[11] Dye, C. (2006). Global epidemiology of tuberculosis. Lancet 367, 938–940.

[12] Evans, D.A., Manley, K.A., and McKusick, V.A. (1960). Genetic control of isoniazid metabolism in man. Br Med J 5197, 485–491.

[13] Foroumadi, A., Kargar, Z., Sakhteman, A., Sharifzadeh, Z., Feyzmohammadi, R., Kazemi, M., and Shafiee, A. (2006). Synthesis and antimycobacterial activity of some alkyl [5-(nitroaryl)-1,3,4-thiadiazol-2-ylthio]propionates. Bioorg Med Chem Lett 16, 1164–1167.

[14] Fullam, E., Westwood, I.M., Anderton, M.C., Lowe, E.D., Sim, E., and Noble, M.E.M. (2007). Divergence of cofactor recognition across evolution: Coenzyme A binding in a prokaryotic arylamine N-acetyltransferase. J Mol Biol 375, 178–191.

[15] Fullam, E., Kawamura, A., Wilkinson, H., Abuhammad, A., Westwood, I., and Sim, E. (2009). Comparison of the Arylamine N-acetyltransferase from Mycobacterium marinum and Mycobacterium tuberculosis. Protein J 28, 281–293.

[16] Holton, S.J., Dairou, J., Sandy, J., Rodrigues-Lima, F., Dupret, J.M., Noble, M.E., and Sim, E. (2005). Structure of Mesorhizobium loti arylamine N-acetyltransferase 1. Acta Crystallogr C 61, 14–16.

[17] Jacobson, R., Ramsay, J., Aller, H. and Thirugnanam, M. (1987). 1-Dimethylcarbamoyl-3-substituted-5-substituted-1H-1,2,4-triazoles. European Patent Office, EP0213718.

[18] Jeney, E., and Zsolnai, T. (1956). Studies in search of new tuberculostatic drugs. I. Hydrazine derivatives, carbolic acid, phenols, quaternary ammonium compounds and their intermediaries. Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene 1 Abt Medizinisch-hygienische Bakteriologie, Virusforschung und Parasitologie 167, 55–64.

[19] Kane, J.M., Staeger, M.A., Dalton, C.R., Miller, F.P., Dudley, M.W., Ogden, A.M., Kehne, J.H., Ketteler, H.J., McCloskey, T.C., Senyah, Y., et al. (1994). 5-Aryl-3-(alkylthio)-4H-1,2,4-triazoles as selective antagonists of strychnine-induced convulsions and potential antispastic agents. J Med Chem 37, 125–132.

[20] Kawamura, A., Graham, J., Mushtaq, A., Tsiftsoglou, S.A., Vath, G.M., Hanna, P.E., Wagner, C.R., and Sim, E. (2005). Eukaryotic arylamine N-acetyltransferase. Investigation of substrate specificity by high-throughput screening. Biochem Pharmacol 69, 347–359.

[21] Kawamura, A., Westwood, I., Wakefield, L., Long, H., Zhang, N., Walters, K., Redfield, C., and Sim, E. (2008). Mouse N-acetyltransferase type 2, the homologue of human N-acetyltransferase type 1. Biochem Pharmacol 75, 1550–1560

[22] Kendall, S.L., Withers, M., Soffair, C.N., Moreland, N.J., Gurcha, S., Sidders, B., Frita, R., Ten Bokum, A., Besra, G.S., Lott, J.S., et al. (2007). A highly conserved transcriptional repressor controls a large regulon involved in lipid degradation in Mycobacterium smegmatis and Mycobacterium tuberculosis. Mol Microbiol 65, 684–699.

[23] Lack, N., Lowe, E.D., Liu, J., Eltis, L.D., Noble, M.E.M., Sim, E., and Westwood, I.M. (2008). Structure of HsaD, a steroid-degrading hydrolase, from Mycobacterium tuberculosis. Acta Crystallogr Sect F Struct Biol Cryst Commun 64, 2–7.

[24] Lack, N., Kawamura, A., Fullam, E., Laurieri, N., Beard, S., Russell, A. J., Evangelopoulos, D., Westwood, I., and Sim, E. (2009a). Temperature stability of proteins essential for the intracellular survival of Mycobacterium tuberculosis. Biochemical J 418, 369–378.

[25] Lack, N.A., Yam, K.C., Lowe, E.D., Horsman, G.P., Owen, R.L., Sim, E., and Eltis, L.D. (2009b). Characterization of a C-C hydrolase from Mycobacterium tuberculosis involved in cholesterol metabolism. J Biol Chem. In press.

[26] Ley, S.V., Baxendale, I.R., Bream, R.N., Jackson, P.S., Leach, A.G., Longbottom, D.A., Nesi, M., Scott, J.S., Storer, I., and Taylor S.J. (2000). Multi-step organis synthesis using solid-supported reagents and scavengers: a new paradigm in chemical library generation. J Chem Soc, Perkin Trans 1 23, 3815–4195.

[27] Lin, G., Li, D., de Carvalho, L.P., Deng, H., Tao, H., Vogt, G., Wu, K., Schneider, J., Chidawanyika, T., Warren, J.D., et al. (2009). Inhibitors selective for mycobacterial versus human proteasomes. Nature 461, 621–626.

[28] Macaev, F., Rusu, G., Pogrebnoi, S., Gudima, A., Stingaci, E., Vlad, L., Shvets, N., Kandemirli, F., Dimoglo, A., and Reynolds, R. (2005). Synthesis of novel 5-aryl-2-thio-1,3,4-oxadiazoles and the study of their structure-anti-mycobacterial activities. Bioorg Med Chem 13, 4842–4850.

[29] Makarov, V., Manina, G., Mikusova, K., Mollmann, U., Ryabova, O., Saint-Joanis, B., Dhar, N., Pasca, M.R., Buroni, S., Lucarelli, A.P., et al. (2009). Benzothiazinones kill Mycobacterium tuberculosis by blocking arabinan synthesis. Science 324, 801–804.

[30] Malbec, F., Milcent, R. and Barbier, G. (1984). Derivatives of 2,4-dihydro-1,2,4-triazole-3-thione and 2-amino-1,3,4-thiadiazole from thiosemicarbazones of esters. J Heterocycl Chem 21, 1689–1698.

[31] Maxmen, A., and Clifton E.B. (2009). TB's strategic opponent. J Exp Med 206, 494–495.

[32] Moskowitz, H., Mignot, A. and Miocque, M. (1980). On the synthesis of thiazolotriazolylacetic acids. J Heterocycl Chem 17, 1321–1323

[33] Nagy, J.M., Cass, A.E., and Brown, K.A. (1997). Purification and characterization of recombinant catalase-peroxidase, which confers isoniazid sensitivity in Mycobacterium tuberculosis. J Biol Chem 272, 31265–31271.

[34] Neyrolles, O., Hernandez-Pando, R., Pietri-Rouxel, F., Fornes, P., Tailleux, L., Barrios Payan, J.A., Pivert, E., Bordat, Y., Aguilar, D., Prevost, M.C., et al. (2006). Is adipose tissue a place for Mycobacterium tuberculosis persistence? PLoS ONE 1, e43.

[35] Nodzu, R., Watanabe, H., Kuwata, S., Nagaishi, C., and Teramatsu, T. (1954). Chemotherapy of tuberculosis. IV. Syntheses of m- and p-aminophenol alkyl ethers and their bacteriostatic actions on Mycobacterium tuberculosis. Yakugaku Zasshi 74, 872–875.

[36] Pangborn, A.B., Giardello, M.A., Grubbs, R.H., Rosen, R.K., and Timmers, F.J. (1996). Safe and convenient procedure for solvent purification. Organometallics 15, 1518–1520.

[37] Payton, M., Auty, R., Delgoda, R., Everett, M., and Sim, E. (1999). Cloning and characterization of arylamine N-acetyltransferase genes from Mycobacterium smegmatis and Mycobacterium tuberculosis: increased expression results in isoniazid resistance. J Bacteriol 181, 1343–1347.

[38] Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Meng, E.C., and Ferrin, T.E. (2004). UCSF Chimera–a visualization system for exploratory research and analysis. J Comput Chem 25, 1605–1612.

[39] Raman, K., Rajagopalan, P., and Chandra, N. (2005). Flux balance analysis of mycolic acid pathway: targets for anti-tubercular drugs. PLoS Comput Chem 1, e46.

[40] Rengarajan, J., Bloom, B.R., and Rubin, E.J. (2005). Genome-wide requirements for Mycobacterium tuberculosis adaptation and survival in macrophages. Proc Natl Acad Sci U S A 102, 8327–8332.

[41] Russell, A.J., Westwood, I.M., Crawford, M.H., Robinson, J., Kawamura, A., Redfield, C., Laurieri, N., Lowe, E.D., Davies, S. G., and Sim, E. (2009). Selective small molecule inhibitors of the potential breast cancer marker, human arylamine N-acetyltransferase 1, and its murine homologue, mouse arylamine N-acetyltransferase 2. Bioorg Med Chem 17, 905–918.

[42] Sali, A., and Blundell, T.L. (1993). Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234, 779–815.

[43] Sandy, J., Mushtaq, A., Kawamura, A., Sinclair, J., Sim, E., and Noble, M. (2002). The structure of arylamine N-acetyltransferase from Mycobacterium smegmatis–an enzyme which inactivates the anti-tubercular drug, isoniazid. J Mol Biol 318, 1071–1083.

[44] Sikora, A.L., Frankel, B.A., and Blanchard, J.S. (2008). Kinetic and chemical mechanism of arylamine N-acetyltransferase from Mycobacterium tuberculosis. Biochemistry 47, 10781–10789.

[45] Sim, E., Sandy, J., Evangelopoulos, D., Fullam, E., Bhakta, S., Westwood, I., Krylova, A., Lack, N., and Noble, M. (2008) Arylamine N-acetyltransferases in mycobacteria. Current Drug Metabolism 9, 510–519.

[46] Sinclair, J.C., Sandy, J., Delgoda, R., Sim, E., and Noble, M.E. (2000). Structure of arylamine N-acetyltransferase reveals a catalytic triad. Nat Struct Biol 7, 560–564.

[47] Singh, R.Fiakpui, C., Galpin, J., Stewart, J., Singh, M.P., and Micetich, R.G. (1996). Synthesis and structure-activity relationship of C-3 substitued triazolylthiomethyl cephems. Eur J Med Chem 31, 301–309.

[48] Stewart, G.R., Newton, S.M., Wilkinson, K.A., Humphreys, I.R., Murphy, H.N., Robertson, B.D., Wilkinson, R.J., and Young, D.B. (2005). The stress-responsive chaperone alpha-crystallin 2 is required for pathogenesis of Mycobacterium tuberculosis. Mol Microbiol 55, 1127–1137.

[49] Story, A., van Hest, R., and Hayward, A. (2006). Tuberculosis and social exclusion. BMJ 333, 57–58.

[50] Upton, A.M., Mushtaq, A., Victor, T.C., Sampson, S.L., Sandy, J., Smith, D.M., van Helden, P.V., and Sim, E. (2001). Arylamine N-acetyltransferase of Mycobacterium tuberculosis is a polymorphic enzyme and a site of isoniazid metabolism. Mol Microbiol 42, 309–317.

[51] Van der Geize, R., Yam, K., Heuser, T., Wilbrink, M.H., Hara, H., Anderton, M.C., Sim, E., Dijkhuizen, L., Davies, J.E., Mohn, W.W., et al. (2007). A gene cluster encoding cholesterol catabolism in a soil actinomycete provides insight into Mycobacterium tuberculosis survival in macrophages. Proc Natl Acad Sci U S A 104, 1947–1952.

[52] Verdonk, M.L., Cole, J.C., Hartshorn, M.J., Murray, C.W., and Taylor, R.D. (2003). Improved protein-ligand docking using GOLD. Proteins 52, 609–623.

[53] Wallace, A.C., Laskowski, R.A., and Thornton, J.M. (1995). LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein Eng 8, 127–134.

[54] Wang, W., Zhang, C., Marimuthu, A., Krupka, H.I., Tabrizizad, M., Shelloe, R., Mehra, U., Eng, K., Nguyen, H., Settachatgul, C., et al. (2005). The crystal structures of human steroidogenic factor-1 and liver receptor homologue-1. Proc Natl Acad Sci U S A 102, 7505–7510.

[55] Weber, W.W., and Hein, D.W. (1985). N-acetylation pharmacogenetics. Pharmacol Rev 37, 25–79.

[56] Westwood, I.M., Holton, S.J., Rodrigues-Lima, F., Dupret, J.M., Bhakta, S., Noble, M.E., and Sim, E. (2005). Expression, purification, characterization and structure of Pseudomonas aeruginosa arylamine N-acetyltransferase. Biochem J 385, 605–612.

[57] Westwood, I.M., Kawamura, A., Fullam, E., Russell, A.J., Davies, S. G., and Sim, E. (2006). Structure and mechanism of arylamine N-acetyltransferases. Curr Top Med Chem 6, 1641–1654.

[58] Yam, K.C., D'Angelo, I., Kalscheuer, R., Zhu, H., Wang, J.X., Snieckus, V., Ly, L.H., Converse, P.J., Jacobs, W.R., Strynadka, N., et al. (2009). Studies of a ring-cleaving dioxygenase illuminate the role of cholesterol metabolism in the pathogenesis of Mycobacterium tuberculosis. PLoS Pathog 5, e1000344.

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg 2010