Characterization of Clioquinol and Analogs as Novel Inhibitors of Methionine Aminopeptidases from Mycobacterium tuberculosis

Summary

Mycobacterium tuberculosis, the causative agent of tuberculosis claims about five thousand lives daily world-wide, while one-third of the world is infected with dormant tuberculosis. The increased emergence of multi- and extensively drug-resistant strains of Mycobacterium tuberculosis (Mtb) has heightened the need for novel antimycobacterial agents. Here, we report the discovery of 7-bromo-5-chloroquinolin-8-ol (CLBQ14)-a congener of clioquinol (CQ) as a potent and selective inhibitor of two methionine aminopeptidases (MetAP) from Mycobacterium tuberculosis: MtMetAP1a and MtMetAP1c. MetAP is a metalloprotease that removes the N-terminal methionine during protein synthesis. N-terminal methionine excision (NME) is a universally conserved process required for the posttranslational modification of a significant part of the proteome. The essential role of MetAP in microbes makes it a promising target for the development of new therapeutics. Using a target-based approach in a high-throughput screen, we identified CLBQ14 as a novel MtMetAP inhibitor with higher specificity for both MtMetAP1s relative to their human counterparts. We also found that CLBQ14 is potent against replicating and aged non- growing Mtb at low micro molar concentrations. Furthermore, we observed that the antimycobacterial activity of this pharmacophore correlates well with in vitro enzymatic inhibitory activity. Together, these results revealed a new mode of action of clioquinol and its congeners and validated the therapeutic potential of this pharmacophore for TB chemotherapy.

1.1. Introduction

The lethal synergy of Mycobacterium tuberculosis (Mtb) with Human Immunodeficiency Virus (HIV) has heightened the need for the development of new antimycobacterials. Particularly, the therapeutic management of HIV patients co-infected with Mtb poses great challenges because of drug-drug interactions and toxicity.¹ Moreover, the rise of multi- and extensively drug-resistant strains of (Mtb) as well as dormant tuberculosis (TB) calls for novel anti-tuberculosis agents with new mechanisms of action.² In an attempt to identify novel inhibitors of TB and validate new targets, we have focused on methionine aminopeptidases (MetAP) that have been shown to be required for the viability of some organisms^3–5 including Mtb.¹⁸

MetAP is a metalloprotease that removes the N-terminal methionine from proteins and peptides.³ NME is an essential process in both prokaryotes and eukaryotes.^4,5 This universally conserved process is important for localization, stability and post-translational modifications of a significant number of proteins.⁴ In prokaryotes, protein synthesis is initiated by formyl-methionyl-tRNA. The formyl group is removed by peptide deformylase (PDF), a prerequisite step for the action of MetAP.⁴ In the past, drug discovery efforts have focused significantly on PDF. However, resistant mutants appear to limit the use of the PDF inhibitors, especially because the mutations in transformylase gene renders deformylation non-essential.⁶ Recent attention has been turned to MetAPs as antibacterial targets,^2,4,5 as it has been shown that deletion of MetAP from Escherichia coli, and Salmonella typhimurium is lethal.^7,8 MetAPs also play important roles in eukaryotes.^3–5 Inhibitors of MetAP enzymes have been developed as potential therapeutic agents for treating diseases such as cancer, rheumatoid arthritis, as well as malarial and fungal infections.^9–17

More recently, using a chemical genomics approach we successfully identified and characterized inhibitors of the two MtMetAPs from Mtb: MtMetAP1a and MtMetAP1c.¹⁸ Other groups have also characterized these mycobacterial enzymes and their inhibitors biochemically.^19–25 Together, these studies have revealed the potential usefulness of these enzymes as antimycobacterial targets.^18–25 Particularly, one study suggests that both MetAPs from Mtb might play different roles at different growth phases in the excision of N-terminal methionine from mycobacterial proteins.²³

There are two classes of MetAPs: MetAP1 and MetAP2.³ Eukaryotes possess both classes while prokaryotes have homologs of either MetAP1 (eubacteria) or MetAP2 (archeabacteria). M. tuberculosis possesses two MetAPs as mentioned above: MtMetAP1a and MtMetAP1c which share 33% sequence identity.^18,23,24 Both MtMetAPs have less than 48% similarity to human MetAP1 (hMetAP1) and less than 30% similarity to human MetAP2 (hMetAP2).²⁴ Moreover, the X-ray crystallography of MtMetAP1c revealed the existence of a highly conserved proline rich N-terminal extension which is absent in MtMetAP1a.²⁴ A recent study showed the indispensability of this N-terminal extension in Mtb.²⁵ We could take advantage of the differences in enzyme structure to design selective inhibitors. Hence, selectively targeting the mycobacterial MetAPs may be a viable strategy for the development of new and effective antimycobacterial agents.

Herein, we report the discovery of clioquinol (CQ) and its analogues as novel inhibitors of MtMetAP enzymes which accounts for their inhibition of Mtb growth. Clioquinol’s anti-TB activity has been previously reported^{42, 43}; however the mechanism of action was unknown. We unexpectedly identified 7-bromo-5-chloroquinolin-8-ol (CLBQ14), the bromine analogue of CQ in a high-throughput screen for inhibitors of MtMetAP1c. CQ is a derivative of 8-hydroxyquinoline that was used to treat infections.^26,27 However, it was withdrawn from the clinic because of untoward effects of subacute myelo-optic neuropathy (SMON) experienced in Japan in the 1950’s.^28–30 More recently, CQ and its derivatives are being studied as therapeutics for neurodegenerative diseases, cancer, and infectious diseases.^31–40 As mentioned above, this pharmacophore has been identified by other groups in drug screens and activity assays for drug sensitive, drug resistant and dormant M. tuberculosis.^41–43 However, its mode of action has remained elusive. In this study, we assessed the effect of CLBQ14, CQ and two additional analogues on the activity of both MtMetAPs. We also determined the specificity of the CLBQ14 and CQ for the mycobacterial enzymes in comparison to the human isoforms of MetAP. Furthermore, we determined the activity of the inhibitors in Mtb. The results suggest that CLBQ14 is a new potent and selective inhibitor of the MetAPs from Mtb.

2. Materials and Methods

2.1. Materials

The M. tuberculosis culture medium, Middlebrook 7H9, was purchased from Becton Dickinson (Sparks, MD). The compounds were provided by ASDI (Newark, DE). We prepared all stock solutions in Dimethylsulfoxide (DMSO).

2.2. High-Throughput Screening

In a high-throughput screening assay as described previously,¹⁸ we identified 7-bromo-5-chloroquinolin-8-ol. Briefly, we screened about 175,000 compounds against MtMetAP1c at concentrations of 30 μM in 384-well plates, using the chromogenic substrate Met-Pro-pNA.⁴⁵ The compounds were dissolved in DMSO. The initial screen was conducted using a titertek instrument (Titertek Instruments, Inc., Alabama, USA) with liquid handling capabilities coupled to a spectrophotometer. The total reaction volume was 50 μL and each reaction contained 40 mM Hepes buffer (pH 7.5), 100 mM NaCl, 100 μg/mL BSA, 0.1 U/mL ProAP, 1.5 mM CoCl₂, 600 μM substrate (Met-Pro-pNA), and 252 nM MtMetAP1c. The enzyme was pre-incubated with compounds for 20 min at room temperature followed by addition of 600 μM substrate. The reaction was incubated at room temperature for 30 min and monitored at 405 nm on a spectrophotometer. The Compounds that showed greater than 30–40% inhibition were chosen as “hits”.

2.3. Determination of IC₅₀ of 7-bromo-5-chloroquinolin-8-ol (CLBQ14) and its analogues.

We obtained CQ and two additional analogues and determined the concentration needed for 50% enzyme inhibition in 96-well plates at final concentrations ranging from 100 μM −300 nM. The total reaction volume as described¹⁸ was 50 μL and each reaction contained each MetAP1 respectively and 40 mM Hepes buffer (p.H 7.5), 100 mM NaCl, 100 μg/mL BSA, 0.1 U/mL BcProAP, 1.5 μM CoCl₂, 600 μM substrate (Met-Pro-pNA). The enzyme was pre-incubated with compounds for 20 min at room temperature followed by addition of substrate. The reaction was incubated at room temperature for 30 min and monitored at 405 nm on a spectrophotometer. The background hydrolysis was subtracted and the data was fitted to a four-parameter logistic (variable slope) equation using GraphPad prism software.

2.4. Determination of Minimum Inhibitory Concentration in Replicating M. tuberculosis

The MetAP inhibitors were serially diluted in DMSO and added to 7H9 broth and OADC (without Tween 80) to give final concentrations of 50, to 0.05 μg/mL. A culture of M. tuberculosis H37Rv was grown till O.D. = 1.0, and diluted to 1/100. Then each tube containing a compound was inoculated with 0.1 mL of culture to give a total assay volume of 5 mL. The controls were DMSO, Isoniazid (a positive control) and a blank (drug free media). The 15-ml conical assay tubes containing mycobacteria were incubated at 37°C and 5% CO₂. M .tuberculosis growth was monitored for two weeks.

2.5. Activity of inhibitors on aged non-growing M. tuberculosis

We determined the activity of 7-bromo-5-chloroquinolin-8-ol in aged non-growing M. tuberculosis at concentrations ranging from 0.5 to 100 μM for three weeks using a persister model as described.⁴⁴ Briefly, a 2 month old M. tuberculosis H37Ra culture grown in 7H9 medium (Difco) with 10% albumin-dextrose-catalase (ADC) and 0.05% Tween 80 was resuspended in acid 7H9 medium (pH5.5) without ADC. The bacterial cell suspension was used as inocula for assaying the activity of the compounds for persister bacilli. The inhibitor was diluted from the stock solution (10 mM in DMSO) to 10 μM (final) followed by incubation with the bacilli in 200 μl in acid pH5.5 7H9 medium without ADC in 96-well plates for 3 days without shaking under 1% oxygen in a hypoxic chamber. The assay was done in duplicate. Rifampin (5μg/ml) was used as a positive control. After 3 day drug exposure, the viability of the bacilli was determined by adding 20 μl of 1 mg/ml XTT (2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide) and incubated at 37°C up to 7 days when the plates were read at OD485 nm.

3. Results

3.1. Effects of 7-bromo-5-chloroquinolin-8-ol (CLBQ14) and its analogues on MetAP activity

In our high-throughput screening efforts to identify inhibitors of MetAPs from M. tuberculosis, we discovered CLBQ14, a derivative of clioquinol, as a hit. Clioquinol is a compound in the hydroxyquinoline class that was used to treat some infections in the past^26,27 and is now under evaluation for the treatment of Alzheimer’s disease, cancer and other disorders.^31–40 We determined the effects of CLBQ14, CQ and two additional analogs on MtMetAPs. We found that all four compounds inhibited both MtMetAPs with IC₅₀ values in the low micro molar concentration, with CLBQ14 being the most potent and CQ being the least potent amongst the analogues tested (Table 1). In addition, all four inhibitors displayed little or insignificant specificity for either MtMetAP1a or MtMetAP1c. These results revealed a new mode of action for CQ. More importantly, this is the first report indicating that MetAP is a target of clioquinol and its analogues. All three analogues, CLBQ14–1, CLBQ14–2 and CLBQ14 are more potent than CQ, with IC₅₀ values ranging from 3.76 μM to 5.44 μM (Table 1).

Table 1.

Effects of 7-bromo-5-chloroquinolin-8-ol (CLBQ14) and its analogues on MetAP enzyme activity.

			IC50 (μM)
Inhibitor ID	Chemical Structure	MtMetAP1a	MtMetAP1c	HsMetAP1	HsMetAP2
CLBQ14–1		4.30	3.76	ND	ND
CLBQ14–2		4.87	4.92	ND	ND
CLBQ14		5.44	4.92	105.29	112.20
Clioquinol (CQ)		9.25	11.16	84.70	80.40

N.D., not determined

3.2. Selectivity of CLBQ14 and CQ for M. tuberculosis MetAPs over HsMetAPs

Next, we determined the selectivity of CLBQ14 and CQ for MtMetAPs over the human MetAPs (HsMetAP1 and HsMetAP2). Both CLBQ14 and CQ were specific for the mycobacterial enzymes compared to the human MetAPs (Table 1). Of the two, CLBQ14 exhibited greater than 19 folds selectivity for both MtMetAPs over human MetAPs (Table 1). CQ showed less than 10 folds selectivity for the mycobacterial enzymes relative to their human counterparts (Table 1). The selectivity of both CLBQ14 and CQ for MtMetAPs over their human counterparts raised the possibility to selectively target the mycobacterial enzymes with a sufficiently wide therapeutic window with new analogues.

3.3. Effects of MtMetAP1 Inhibitors on M. tuberculosis growth

We determined the Minimum Inhibitory Concentrations (M.I.C.) of CLBQ14 and CQ in M. tuberculosis (Table 2). In the replicating M. tuberculosis culture, CLBQ was more potent than CQ with a two-fold selectivity (Table 2). The relative potency of the two compounds in Mtb is in agreement with their relative activity against the two MtMetAP enzymes in vitro. Moreover, we observed an increase in potency in the aged non-growing M. tuberculosis for CLBQ14 with M.I.C. values of 0.65 – 1.62 μg/mL (Table 2). These results suggest that CLBQ14 and CQ might be promising leads for TB therapeutics. These observations also provide support for earlier reports of the effects of the 8-hydroxyquinolines in M. tuberculosis.^41–43

Table 2:

Activity of MtMetAP1 Inhibitors -CLBQ14 and CQ on M. tuberculosis

M.I.C. (μg/mL)
Inhibitor ID	M. tuberculosis	Aged non-growing M. tuberculosis
CLBQ14	5.0	0.65 – 1.62
Clioquinol (CQ)	10.0	ND

N.D., not determined

4. Discussion

The availability of pure, active and well-characterized MtMetAP1a and MtMetAP1c enabled us to identify novel inhibitors of MetAP from M. tuberculosis using a high-throughput screening approach. We screened a structurally diverse small molecule library of over 175,000 compounds and identified CLBQ14, a structural relative of CQ as a potent and selective inhibitor of MtMetAP1a and MtMetAP1c. We have been able to characterize the effect of CLBQ14 and three of its analogues on both bacterial and human MetAPs. CLBQ14 displayed higher specificity for the mycobacterial MetAPs than their human counterparts (Table 1). More importantly, we found that these class of compounds were active as inhibitors of M. tuberculosis (Table 2). CLBQ14 showed potent activity in aged non-growing M. tuberculosis (Table 2), validating the potential of this pharmacophore as a therapeutic agent for TB chemotherapy.

Despite the reported adverse effects of clioquinol^28–30, this pharmacophore has attracted much attention recently in studies aimed at discovering novel therapeutics for neurodegenerative diseases, cancer, and other disorders.^31–40 More recently, this class of compounds were also identified in drug screens for drug sensitive, drug resistant and dormant M. tuberculosis by other reports.^41–43 However, their exact mode of action was unknown. Our results show that the clioquinol pharmacophore has activity against a novel target in mycobacterial: MtMetAPs. Moreover, we observed that the bromine analogue of clioquinol-CLBQ14 is slightly more potent and highly specific for the two mycobacterial MetAPs relative to the two human MetAPs (Table 1). Particularly, this selectivity suggests that CLBQ14 might not have the same adverse effects experienced with CQ in the past^28–30, potentially alleviating some of the concerns with CQ.

We recently reported the identification of the 2,3-dichloro-1,4-naphthoquinone pharmacophore as potent inhibitors of the two MetAPs from Mtb.¹⁸ Like the 2,3-dichloro-1,4-naphthoquinones, the 8-hydroxyquinolines have been used to treat bacterial infections in the past. Similarly, we observed correlations from our in vitro MetAP assay with the Mtb activity. However, unlike the 2,3-dichloro-1,4-naphthoquiones, the hydroxyquinolines show significantly high specificity for the mycobacterial enzymes compared to the human MetAP isoforms. These results suggest that the hydroxyquinolines could be further optimized for MtMetAP1 inhibition and anti-tuberculosis effects. Moreover, this report suggests that CLBQ14 shows more promise with selective toxicity.

In comparison to the 2,3-dichloro-1,4-naphthoquinones, the CLBQ14 structural class is more active against both replicating and aged non-growing Mtb. The observation that our in vitro enzymatic studies translate and correlate proportionally to the activity in replicating Mtb (Table 2) suggest that MtMetAPs are likely to be relevant targets of both CQ and CLBQ in vivo. Furthermore, it is important to note that the M.I.C. values from replicating M. tuberculosis correlated well with that of dormant tuberculosis for the 2,3-dichloro-1,4-naphthoquinones which we have validated genetically.¹⁸ Similarly, the increase in potency of CLBQ14 shows promise for use of this novel mode of action for therapeutic treatment of dormant TB (Table 2). Our identification of a new mode of action for CQ and discovery of CLBQ as a potent and selective MtMetAP1 inhibitor makes this pharmacophore and target a promising combination for further optimization.

Although there has been recent reports on the identification of MetAP inhibitors from Mtb,^18,19,21 this is the first report to show a potent and selective inhibitor with potent antimycobacterial activity in Mtb as well as dormant Mtb. Moreover, these inhibitors could be used as tools in the future to understand the physiologic role of MetAP in NME, an essential process in all organisms. Further SAR and crystal structure studies will aid the rational design and synthesis of more potent inhibitors. Because MetAP is a novel target and its activity is a requirement for N-terminal processing of some proteins, its inhibitors have the potential to treat dormant TB, and limit the development of MDR-TB and XDR-TB. In addition, potent and selective MtMetAP inhibitors have the prospects of shortening the duration of TB therapy if evaluated in conjunction with current treatment options as well as reduce the drug to drug interactions presently encountered by TB-HIV co-infected patients.