Menaquinone is biosynthesized from chorismate by the action of at least eight enzymes (Figure 1).⁵ The first studies on menaquinone biosynthesis focused on Escherichia coli, Mycobacterium phlei and Bacillus subtilis, and the pathway is best understood in E. coli, where the first six enzymes are present in an operon. These and other genetic experiments delineated many of the components of the pathway and also demonstrated the essential role menaquinone plays in bacterial viability.^5b,6

Our initial efforts to target this pathway have focused on MenE,⁷ an acyl-CoA synthetase (ligase) that is essential in M. tuberculosis.^6b MenE converts o-succinyl-1-benzoate (OSB) to OSB-CoA via a two-step process involving initial ATP-dependent adenylation of OSB to form a reactive OSB-AMP intermediate, followed by thioesterification with CoA to form OSB-CoA.

Acyl-CoA synthetases⁸ belong to a superfamily of structurally and mechanistically related adenylate-forming enzymes that also includes non-ribosomal peptide synthetase (NRPS) adenylation domains⁹ and firefly luciferase.¹⁰ Analogous adenylation reactions are also catalyzed by structurally unrelated aminoacyl-tRNA synthetases.¹¹ We and others have used 5′-O-(N-acylsulfamoyl)adenosines (acyl-AMS) and related compounds to inhibit such adenylate-forming enzymes by mimicking the cognate, tightly-bound acyl-AMP intermediates.^10,12,13,14 These molecules were inspired by a class of sulfamoyladenosine natural products that includes nucleocidin and ascamycin.¹⁵ To avoid potential liabilities of the aromatic carboxylate moiety with respect to cell permeability or chemical instability via spirodilactone formation (observed for OSB-CoA), we posited that it might be replaced with a neutral methyl ester, since this carboxylate is not directly involved in the reaction mechanism.¹⁶ Thus, we envisioned that MeOSB-AMS (1) or its sulfamide analog MeOSB-AMSN (2) might be effective inhibitors of MenE and menaquinone biosynthesis (Figure 2).

We also considered that the corresponding vinyl sulfonamide MeOSB-AVSN (3) might inhibit MenE through covalent binding to the incoming CoA thiol nucleophile during the second half-reaction (Figure 3), forming a mimic of the tetrahedral intermediate. Michael acceptors have been used extensively to inhibit cysteine proteases,¹⁷ and also to target protein thiol nucleophiles in polyketide and non-ribosomal peptide synthetases.¹⁸ Based on studies of Roush and coworkers on the inherent reactivities of various sulfonyl-based Michael acceptors,¹⁹ we selected the vinyl sulfonamide moiety to provide the requisite balance of reactivity and selectivity to bind CoA in the MenE active site without reacting promiscuously with other nucleophiles.

Synthesis of these inhibitors began with the preparation of MeOSB (11, Figure 4). OSB was first synthesized by Roser in 1884 from phthalic anhydride and succinic acid.²⁰ MeOSB has also been synthesized by selective monohydrolysis of the corresponding CDI-derived bis(acylimidazole), followed by methanolysis.¹⁶ To provide more efficient and flexible access to OSB and analogs thereof, we developed a new synthesis from the known vinyl bromide 7, prepared by alkylation of t-butyl acetate with 2,3-dibromopropene (Figure 4).²¹ Suzuki cross-coupling with aryl boronate 8 provided styrene 9. Ozonolysis of the vinyl group afforded the orthogonally protected OSB diester 10. Acid deprotection of the t-butyl ester then yielded the desired aromatic monoester MeOSB (11). This modular approach should provide access to a wide range of OSB analogs. Indeed, the exo-methylene intermediate 9 provided immediate access to the corresponding OSB analog 12, which we envisioned would allow us to remove the potentially enolizable ketone functionality in OSB-AMP analogs 4–6 (Figure 2) and to assess its importance in binding.

The corresponding vinyl sulfonyl chlorides 20 and 21 were also prepared by a similar route (Figure 5), featuring selective Horner–Wadsworth–Emmons coupling of ketoaldehyde 15 with sulfonyl phosphonate 17²² to afford the vinyl sulfonate 18. The exo-methylene analog 19 was similarly prepared from 16. The esters were purified and converted to vinyl sulfonyl chlorides 20 and 21, which were used without further purification.

With these OSB analogs in hand, MeOSB-AMS (1) and its exo-methylene analog 4 were synthesized by analogy to our established procedures,^14h via N-acylation of a protected 5′-O-sulfamoyladenosine derivative with 11 and 12, respectively, followed by deprotection.²³ Sulfamide analogs 2 and 5 were synthesized similarly from a protected 5′-N-sulfamoylaminodeoxyadenosine.²³ The vinyl sulfonamide analogs 3 and 6 were prepared by acylation of a protected 5′-aminodeoxy-adenosine with 20 and 21, respectively.²³

To test these compounds for inhibition of MenE, we used a coupled assay with MenE and MenB, the DHNA-CoA synthetase that follows MenE in the bio-synthetic pathway.^4,23 E. coli MenE and M. tuberculosis MenB were separately cloned and expressed with N-terminal His₆-tags in E. coli (BL21) cells, then purified to homogeneity using affinity chromatography. Reactions were initiated by adding MenE (final concentration 20 nM) to a solution containing MenB (7.2 μM), ATP (240 μM), CoA (240 μM), OSB (240 μM) and inhibitor (0–200 μM). Formation of DHNA-CoA was monitored at 392 nm and IC₅₀ values were determined.

We were gratified to find that both the sulfamate MeOSB-AMS (1) and sulfamide MeOSB-AMSN (2) were effective inhibitors of MenE (Table 1). Moreover, the vinyl sulfonamide analog MeOSB-AVSN (3) proved to be the most potent inhibitor, with an IC₅₀ of 5.7 ± 0.7 μM; kinetic analysis indicated that this compound is a slow-binding inhibitor, suggesting a conformational change during binding. In contrast, none of the corresponding exo-methylene analogs (4–6) inhibited the enzyme at up to 200 μM concentration. No inhibition was observed when assays were performed using a limiting concentration of MenB (100 nM) in the presence of excess MenE (5 μM), indicating that the compounds do not inhibit MenB directly. In a preliminary experiment, 1–6 (up to 300 μM) did not inhibit M. smegmatis growth, suggesting that additional pharmacological issues may need to be addressed. Further investigations of cellular activity are ongoing.

It is interesting to note that the vinyl sulfonamide analog MeOSB-AVSN (3) is the most potent inhibitor of MenE. In contrast to the sulfamate and sulfamide analogs 1 and 2, this compound lacks the carbonyl and adjacent heteroatom of the acyl phosphate group in OSB-AMP, which may be involved in hydrogen bonding interactions, based on the cocrystal structure of a related fatty acyl-CoA synthetase with myristoyl-AMP.^8d These results also contrast with the relative potencies of related inhibitors of the NRPS salicylate adenylation enzyme MbtA.^18b This may be due to a variety of factors, including possible structural differences between these enzymes,²⁴ different binding requirements for the inhibitors or resulting covalent adducts, and/or the different thiol nucleophiles involved: CoA in the case of MenE and a protein (MbtB) phosphopantetheine group in the case of MbtA. Our results also suggest that the OSB ketone group is required for inhibition, as shown by the complete lack of activity in exo-methylene analogs 4–6.

In conclusion, we have designed, synthesized, and evaluated a series of mechanism-based inhibitors of the OSB-CoA synthetase MenE, which is used in bacterial menaquinone biosynthesis. This work expands the scope of sulfonyladenosine-based inhibitors to the acyl-CoA synthetase class of the adenylate-forming enzyme superfamily and sets the stage for future assessment of these inhibitors and additional analogs in cellular and animal models of infection to evaluate the potential of targeting MenE in antibacterial drug discovery.