TB (tuberculosis) disease remains responsible for the death of over 1.5 million people each year. The alarming emergence of drug-resistant TB has sparked a critical need for new front-line TB drugs with a novel mode of action. In the present paper, we review recent genomic and biochemical evidence implicating Mycobacterium tuberculosis CYP (cytochrome P450) enzymes as exciting potential targets for new classes of anti-tuberculars. We also discuss HTS (high-throughput screening) and fragment-based drug-discovery campaigns that are being used to probe their potential druggability.
- cytochrome P450
- drug discovery
- fragment-based drug discovery
- high-throughput screening (HTS)
- Mycobacterium tuberculosis
TB (tuberculosis) and drug resistance
TB disease is caused primarily by pulmonary infection by the aerobic bacterium Mycobacterium tuberculosis . The infection is spread via the inhalation of aerosolized bacilli released from M. tuberculosis-infected individuals . Most people overcome the initial M. tuberculosis infection and develop asymptomatic latent TB, where the bacteria lie dormant/persistent within inactivated alveolar macrophages, hidden from the other immune defences of the body . Approximately 10% of individuals develop the active disease, where the M. tuberculosis bacilli multiply rapidly beyond the control of the host immune system . Without successful treatment, the bacteria disseminate and overrun the host, leading to death from systemic infection . The probability of developing active TB disease (initially or from reactivation of latent M. tuberculosis bacilli) is significantly increased when a person becomes immunocompromised, e.g. with concomitant HIV infection . Roughly one-third of the world's population is believed to be harbouring latent TB, and there are estimated to be 9 million new incidences of active TB each year, mostly in developing countries [2,3]. The WHO (World Health Organization) has declared the epidemic a global health emergency .
The front-line regimen of anti-bacterial drugs used to treat active TB has been in use for more than 40 years, and the last two decades have seen the increasing prevalence of multidrug-resistant (MDR-TB) and extensively drug-resistant (XDR-TB) strains of M. tuberculosis . MDR-TB is resistant to at least the two most efficacious oral first-line medications, isoniazid and rifampicin, whereas XDR-TB has evolved additional resistance to any fluoroquinolone drug and any of the second-line injectables (amikacin, kanamycin or capreomycin) . The successful treatment of drug-resistant TB requires replacing or adding to the front-line agents with second- and third-line bacteriocides . These drugs are more expensive, have increased side effects and/or lesser efficacy, leading to treatment durations of over 2 years and poor patient compliance . The WHO estimates that there are over 650000 MDR-TB cases emerging every year, and XDR-TB has become prevalent in more than 69 countries . New front-line anti-TB drugs, operating through a novel mechanism of action, are desperately needed to combat this growing burden.
CYPs (cytochrome P450s) and the M. tuberculosis CYPs
The success of drugs that inhibit human biosynthetic CYP enzymes, such as the aromatase inhibitors anastrozole and letrozole, has driven research towards understanding the roles of the unusually high number of CYPs (20) encoded in the 4.4 Mb M. tuberculosis H37Rv genome [6–9] (Figure 1A). The cohort represents a 200-fold greater CYP gene density than in humans (57 CYPs in a 3 Gb genome), and there are no CYPs in Escherichia coli, which has a similarly sized genome to M. tuberculosis . The typical function of CYP haem-thiolate proteins is to catalyse the oxidation of organic substrates via their haem prosthetic group (an iron–porphyrin complex) . This mono-oxygenation reaction involves the insertion of one oxygen atom from molecular dioxygen into the substrate, while the second oxygen undergoes reduction to water . The key stages of the catalytic cycle are: (i) substrate entry into the distal active site displaces the weak iron-co-ordinated water of the CYP resting state, (ii) the haem iron is reduced from ferric to ferrous (using an electron-transport partner system) and molecular dioxygen binds to the ferrous haem iron, (iii) a further single electron reduction of the oxy-complex and two protonation steps releases one oxygen as water and generates a highly reactive intermediate iron(IV)-oxo porphyrin π-radical cation ([FeIV=O]+•), termed Compound I, and finally (iv) the oxygen atom is incorporated into the substrate from Compound I via a radical rebound mechanism .
Five of the 20 M. tuberculosis CYPs have been structurally and functionally characterized to date [8,9] (Figure 1). Their roles have been difficult to assign because of a distinct lack of sequence similarity with other CYPs of known function, and owing to extensive divergence within the M. tuberculosis CYP cohort itself . The first and only M. tuberculosis CYP to be functionally assigned on the basis of amino acid sequence similarity was CYP51B1, which shows 30–40% identity with eukaryotic CYP51s . These enzymes catalyse 14α-demethylation during the eukaryotic biosynthesis of sterols . Expression and isolation of M. tuberculosis CYP51B1 by Bellamine et al.  led to confirmation of sterol demethylase activity for the substrates lanosterol and 24,25-dihydrolanosterol, and the plant sterol obtusifoliol  (Figure 1B). This was the first discovery of a bacterial CYP51 and the first evidence of a P450 species conserved throughout phylogeny from the prokaryotic era . Although oxidative demethylation is a general feature of steroid synthesis in eukaryotes, the relevance of this activity to M. tuberculosis remains unclear, as several components for a complete sterol biosynthetic pathway are absent from M. tuberculosis .
The subset of M. tuberculosis CYPs comprising CYP124, CYP125 and CYP142 are proposed to co-operate in a pathway responsible for metabolism of host-derived steroids by the bacterium [14–18]. All three enzymes are able to oxidize the aliphatic tail of cholesterol/cholest-4-en-3-one at C-27 to the alcohol, aldehyde and then carboxylic acid  (Figure 1B). CYP125 generates oxidized sterols of the (25S) configuration, whereas the opposite (25R) stereochemistry is obtained with CYP124 and CYP142 . The catalysis is thought to be the initial event in the degradation of sterol side chains as metabolic fuel for the bacteria [14,15]. M. tuberculosis is capable of growth in vitro using cholesterol as the sole source of carbon, and CYP125-knockout (ΔCYP125) strains under these conditions fail to grow and instead accumulate cholest-4-en-3-one [14,15,19]. Genetic overexpression/complementation of the ΔCYP125 strains with CYP124 or CYP142 partially or fully rescues their growth defect on cholesterol . Interestingly, CYP124 is also capable of hydroxylating fatty acids and other long-alkyl-chain lipids  (Figure 1B). M. tuberculosis is known to produce a wide array of lipids for components of its complex fatty/waxy cell wall and for modulating the host immune response during pathogenesis . CYP124 oxygenates regiospecifically at the aliphatic terminal ω-position and has a preference for long-chain methyl-branched lipids . This dual functionality of CYP124 may explain its low catalytic efficiency as a sterol 27-oxidase compared with CYP125 and CYP142 .
The most recent M. tuberculosis CYP to be functionally characterized is CYP121, which Belin et al.  show catalyses an unusual intramolecular C–C bond-forming reaction between the two tyrosine residues of the cYY (cyclodityrosine) substrate  (Figure 1B). cYY is derived from two molecules of tyrosine-bound tRNA by action of the cyclodipeptide synthetase encoded by Rv2275, the gene adjacent to CYP121 (Rv2276) [22,23]. Although the physiological role of cYY and its product of transformation by CYP121 (termed mycocyclosin) remains to be determined, cyclic dipeptides are known to have a wide range of biological activities in prokaryotes, including downstream enzyme inhibition and regulation of gene expression [21,24]. Phylogenetic analysis reveals that CYP121 is exclusive to M. tuberculosis, and this is presumably to accomplish the unique C–C bond-forming catalysis that is not required by other bacteria . To date, the only other M. tuberculosis CYP for which there are compelling data on its functionality is CYP128, which is likely to be involved in a process leading to the sulfation of MK-9(H2) (dihydromenaquinone-9), the major quinol electron carrier of M. tuberculosis .
Discriminating between bacterial and host CYPs
Although bacterial and eukaryotic CYPs both share a common haem-utilizing mechanism, there are distinct structural differences between them and their electron-transfer machinery [10,26]. Mammalian CYPs have well-known essential functions in the phase-I metabolism of xenobiotics, as well as in the biosynthesis of a variety of endogenous compounds, particularly steroid hormones . They are typically membrane-bound and incorporated into the cellular endoplasmic reticulum or inner mitochondrial membrane, receiving electrons derived from NAD(P)H via membrane-associated flavin (FAD/FMN-containing) reductases . In contrast, bacterial CYPs are largely soluble and cytosolic, and are generally reduced by a two-component system comprising an FAD-containing ferredoxin reductase and an iron–sulfur cluster containing ferredoxin . Most bacterial CYPs are orphan enzymes, but those of known functionality are found to catalyse diverse reactions ranging from the breakdown of carbon sources for microbial growth to the synthesis of bioactive secondary metabolites, such as anti-fungals and anti-parasitics . These structural differences (i.e. active-site alterations associated with divergent substrates, and surface differences to accommodate distinct redox partners) illustrate the potential for good inhibitor selectivity between the two CYP classes. It is also worth noting that between catabolic and biosynthetic CYPs, it is the latter that are usually the focus for therapeutic intervention owing to their high substrate specificity . Hence the best M. tuberculosis CYPs as anti-TB targets are likely to be those catalysing biosynthetic reactions with no comparable reaction in humans.
M. tuberculosis CYPs present attractive anti-TB targets at the genetic level
The CYP121, CYP125 and CYP128 isoforms are arguably the most promising anti-TB drug targets within the M. tuberculosis CYP complement, on the basis of genetic analysis (Figure 1A). Recent studies show that their genes are essential for M. tuberculosis growth in vitro, ex vivo cultured in macrophages and/or in vivo during infection in mice [27–31]. CYP121 appears to be exclusive to M. tuberculosis, and construction of an M. tuberculosis CYP121-knockout mutant is only achievable when a complementing vector carrying CYP121 is also present . The unusual catalytic mechanism and substrate selectivity of CYP121 suggests a unique active site, which may facilitate the development of specific inhibitors. Although CYP125 is non-essential in vitro, transposon-insertion mutagenesis indicates that it is pivotal for M. tuberculosis virulence during macrophage infection in mouse models . Furthermore, microarray transcriptome analysis of M. tuberculosis cultured in vitro and ex vivo shows that CYP125 expression is clearly induced on intraphagosomal residence . It is speculated that the exploitation of host cholesterol by CYP125 enables M. tuberculosis persistence in the cholesterol-rich macrophage [14,15,19]. Hence a drug inhibiting CYP125 could be an important candidate for eliminating non-replicating latent M. tuberculosis. Since CYP124 and CYP142 possess the same sterol 27-oxidase activity and complement the growth defect of a ΔCYP125 M. tuberculosis strain on cholesterol, they too may present additional secondary targets for latent M. tuberculosis  (Figure 1A). CYP128 is also of particular interest because genome-wide transposon insertion distinguishes it as essential for M. tuberculosis survival in vitro . However, attempts to generate soluble CYP128 for biochemical characterization have proved unsuccessful to date . CYP128 is proposed to function as a menaquinone hydroxylase on the basis of its gene location in an operon with the sulfotransferase stf3, which produces a derivative of MK-9(H2) sulfated on its polyisoprene tail (S881) . Prior oxidation of MK-9(H2) by CYP128 at the same position is thought to enable the sulfation .
Azole anti-fungal CYP inhibitors show cross-activity as anti-tuberculars
The therapeutic potential of M. tuberculosis CYPs is further evident from the potent anti-mycobacterial activity of azole anti-fungals. These compounds are well-known CYP51 inhibitors used clinically for the treatment of skin or systemic fungal infections [10,32–34]. Many of the large flexible azole anti-fungals possess broad overlapping CYP-inhibition profiles, and bind to the characterized M. tuberculosis CYPs in vitro with high affinity [dissociation equilibrium constants (Kd) as low as double-digit nanomolar] [16,17,20,27,33–36]. They inhibit CYP catalysis by reversibly co-ordinating the ferric haem iron through their azole nitrogen in the sixth distal ligand position (type-II binding, example in Figures 2A and 2D) [10,33]. This replaces the weakly distal co-ordinated water of the CYP resting state and competitively inhibits substrate access to the active site . Several azole anti-fungals also exhibit potent antimycobacterial activity [37–42]. Treating M. tuberculosis with these compounds, particularly econazole, has proved effective in eliminating both active and latent M. tuberculosis in vitro, ex vivo and from infected mice [37–42]. Hence, assuming that their anti-tubercular efficacy is due directly to CYP inhibition, the data strengthen the evidence for essential functions of M. tuberculosis CYPs. Most azole anti-fungals remain unsuitable as scaffolds for front-line oral anti-TB drug candidates, owing to their poor bioavailability and unwanted non-specific affinity for host CYPs, which result in severe systemic toxicology and induced drug–drug interactions [32–34]. Those administered clinically for systemic fungal infections (e.g. fluconazole and voriconazole) also typically bind weakly or non-detectably to the M. tuberculosis CYPs [16,17,20,27,32,35,36]. Furthermore, M. tuberculosis mutants that are resistant to these agents have been isolated and show up-regulation of a transmembrane transporter protein believed to act as an azole efflux pump .
M. tuberculosis CYP inhibitor development by HTS (high-throughput screening) and fragment-based approaches
The development of M. tuberculosis CYP inhibitors has progressed beyond azole anti-fungals in the area of compound screening. Podust et al. [44,45] conducted HTS programmes against CYP51B1 and CYP130, in order to identify new ligand chemotypes and to explore the topology of their active sites. The screens monitored for ligand-induced shifts in the CYP haem Soret absorbance region (~350–450 nm). A red shift in the Soret absorbance is indicative of type-II binding via ligand–haem co-ordination, and a type-I blue shift indicates substrate-like binding near the haem iron . Some 20000 compounds were screened against both enzymes, resulting in the identification of many high-affinity type-II hits and a few weaker type-I hits for CYP51B1. The most interesting top non-azole type-II hits were a series of N-pyridinyl-benzeneacetamides for CYP51B1 (example in Figures 2B and 2D), and a diverse group of primary arylamines for CYP130. Members of both series had binding affinities comparable with those of reference azole anti-fungals for each isoform (single-digit micromolar Kd values). The best CYP51B1 hits were found to be selective over in vitro binding to M. tuberculosis CYP125 and CYP130, and some CYP130 hits were selective compared with CYP51B1, CYP125 and CYP142. X-ray crystal structures show that the hits' type-II binding mode is by co-ordination to the haem iron through their pyridine and amine nitrogen lone pairs. The N-pyridinyl-benzeneacetamides maintain interactions with CYP51B1 residues Tyr76 and His259, which are invariable in the CYP51 family, and the benzene moiety contacts a strictly species-specific hydrophobic cavity. A second-generation commercial N-(4-pyridyl)-formamide analogue of the CYP51B1 hits, compound LP10, was also identified to bind preferentially to the more pharmacologically attractive M. tuberculosis CYP125 isoform [Kd(CYP125)=1.7 μM, Kd(CYP51B1)=18 μM] . LP10 exhibits an unusual reverse type-I binding mode with CYP125, where its pyridine is sterically hindered from co-ordinating the haem iron and instead interacts with a three water hydrogen-bond network around the iron . These new HTS inhibitory scaffolds/templates all remain to be optimized structurally and synthetically to the CYP51B1, CYP125 and CYP130 active sites.
The fragment-based approach to small-molecule drug discovery has been developed over the last decade to complement HTS [47–49]. This powerful methodology involves building larger more potent lead drug candidates from weaker binding low-molecular-mass fragment molecules (typically <250 Da) [48,49]. The main advantages of fragment-based drug discovery are that (i) a vast amount of chemical space can be explored by screening a relatively small library (typically 1000–2000 fragments), and (ii) fragment hits must make optimal interactions with the target in order for them to bind with sufficient affinity for detection, and this in turn gives them desirable high ligand efficiency (binding energy divided by molecular complexity) [48,49]. We recently initiated a fragment-based drug-discovery campaign targeting several of the M. tuberculosis CYPs, and noted clear isoform selectivity in the fragment screen (S.A. Hudson K.J. McLean, S. Surade, Y.-Q. Yang, D. Leys, A. Ciulli, A.W. Munro and C. Abell, unpublished work). The detailed findings are yet to be reported, but we are excited to show preliminary data of an X-ray crystal structure of an M. tuberculosis CYP complexed with a highly ligand efficient type-II fragment hit, MB189 (Figures 2C and 2D).
Conclusions and future directions
HTS and fragment-based approaches provide solid foundations for the rational design and synthesis of more potent and selective M. tuberculosis CYP inhibitors. Future work should be prioritized in the first instance to those isoforms that present the most attractive anti-TB targets, i.e. CYP121, CYP125 and CYP128. Isoform-specific inhibitors can also be employed as tools for chemical biology. For example, comparisons between the gene induction/repression profiles of M. tuberculosis cultures grown with and without specific inhibitors could shed light on how CYPs relate to M. tuberculosis infection, growth and persistence. It would be particularly interesting to observe the correlation between responses of CYP124, CYP125 and CYP142 at the transcriptional level in wild-type M. tuberculosis strains. Moreover, biophysical characterization of ligand/fragment–CYP binding modes may provide a clearer understanding of the CYP active-site properties and enzymatic mechanism. We have recently designed a novel library of fragments called Biofragments that are derived from known bacterial CYP substrates, and hope that screening this library against orphan M. tuberculosis CYPs will provide clues as to their likely functional role at the substrate level (S.A Hudson, E.H. Mashalidis and C. Abell, unpublished work). In all, we look forward to the many new and exciting discoveries to be made through exploring the M. tuberculosis CYPome, and particularly with respect to developing CYP isoform-specific inhibitors as novel anti-TB drug candidates.
S.A.H. was supported by a Sir Mark Oliphant Cambridge Australia Scholarship awarded by the Cambridge Commonwealth Trust and Cambridge Overseas Trust, University of Cambridge, and Christ's College, University of Cambridge.
We give many thanks to Dr Alessio Ciulli (Department of Chemistry, University of Cambridge), Professor David Leys (Faculty of Life Sciences, University of Manchester) and Professor Sir Tom Blundell (Department of Biochemistry, University of Cambridge) for their helpful discussions.
Frontiers in Biological Catalysis: Biochemical Society Annual Symposium No. 79 held at Robinson College, Cambridge, U.K., 10–12 January 2012. Organized and Edited by David Leys (Manchester, U.K.), Andrew Munro (Manchester, U.K.), Emma Raven (Leicester, U.K.) and Martin Warren (University of Kent, U.K.).
Abbreviations: CYP, cytochrome P450; cYY, cyclodityrosine; HTS, high-throughput screening; MDR-TB, multidrug-resistant tuberculosis; MK-9(H2), dihydromenaquinone-9; TB, tuberculosis; WHO, World Health Organization; XDR-TB, extensively drug-resistant TB
- © The Authors Journal compilation © 2012 Biochemical Society