Mechanistic and chemoenzymatic studies of the C-C bond forming enzymes in menaquinone biosynthesis and searching for the missing thiosesterase in the pathway

Menaquinone (MK) or vitamin K2, is a lipid-soluble essential molecule which plays important roles in electron transport, oxidative phosphorylation, active transport, and endospore formation in bacteria. In the majority of Gram-positive bacteria, including Mycobacterium tuberculosis (M. tuberculosis), MK is the sole quinone in the electron transport chain, and the pathway leading to the biosynthesis of MK is absent in human. Thus, the bacterial enzymes catalyzing the synthesis of MK are potential targets for the development of novel antibacterial drugs. To better understand MK biosynthesis for use as a target for the development of new antibiotics, two key enzymes of the pathway, MenD and MenB, are chosen for investigations of their catalytic mechanism in my thesis research. In addition, MenD, a thiamine dependent enzyme catalyzing C-C bond formation, is explored for its potential in chemoenzymatic synthesis. Moreover, experiments are designed to search for the missing 1,4-Dihydroxy-2-naphthoyl-CoA (DHNA-CoA) thioesterase in the biosynthetic pathway. Firstly, our interest centers on 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexadiene-1-carboxylate (SEPHCHC) synthase, MenD, the enzyme that catalyzes the first committed step of the MK biosynthetic route. We identified eight strictly conserved amino acid residues (Arg-32, Arg-106, Lys-299, Arg-409, Arg-428, Ile-489, Phe-490, and Leu-493) in the active site of Bacillus subtilis MenD by modeling of substrate binding, sequence, and structure comparisons. Detailed steady-state kinetic parameters for wild-type MenD and eight point mutants were determined. Kinetic characterization of these mutants illustrate these residues are essential in substrate recognition or stabilizing the cofactor position. Five basic residues (Arg-32, Arg-106, Arg-409, Arg-428, and Lys-299) are postulated to interact with carboxylate and hydroxyl groups to align substrates for catalysis in combination with a cluster of non-polar residues (Ile-489, Phe-490, and Leu-493) on one side of the active site. This data supports the proposed two-stage mechanism, and allow for a detailed description of the structure–reactivity relationship that governs MenD function. Secondly, we found MenD is a unique member of the ThDP dependent enzyme family. It catalyzed a 1,4-Michael type addition or Stetter like reaction in physiological conditions. The ThDP dependent enzyme are well known for catalyzing “umpolung” benzoin condensations, and wildly used in chemoenzymatic synthesis of molecules of therapeutic value. In an attempt to develop this enzyme into a biocatalyst for similar Stetter reaction, we unexpectedly found that it is able to partially reduce nitro-compounds and to catalyze C-N ligation with the resulting nitroso-compound. In addition, various donor or acceptor substrates can be applied to MenD catalyzed chemoenzymatic reactions, indicating its possible new utility in asymmetric synthesis. DHNA-CoA synthase, MenB, converts o-succinylbenzoyl-CoA (OSB-CoA) to DHNA-CoA, the sixth step in the biosynthesis of MK. We tried to validate the catalytically essential aspartate (Asp-185) of M. tuberculosis MenB (MtbMenB) is the catalytic base responsible for abstraction of a proton from the succinyl R-carbon of the OSB-CoA substrate to initiate the MenB catalyzed reaction. Through structure alignment of MtbMenB and Salmonella Typhimurium MenB (SaMenB), we found that an exogenous bicarbonate takes a position similar to that of the MtbMenB Asp-185 side chain carboxylate and plays an essential catalytic role equivalent to that of the basic group. In an attempt to collect direct evidence demonstrating the real role of Asp-185 for this proposed role, we carried out H/D exchange experiments on the α-protons of a series of substrate analogs. However, there are no obvious H/D exchanges for the OSB-CoA analogs in the presence of MtbMenB. These unsuccessful verification test results may due to the fact that the substrate analogs lacking aromatic carboxylate are not good mimics of the substrate, and they do not seem to be bound properly at the enzyme active site. However, this may be an indirect evidence to illustrate that the aromatic carboxyl group is essential to position the α-H in the proper position in the enzyme active site. Next, we found that the DHNA-CoA product and its analogues bind and inhibit the synthase from Escherichia coli (E. coli) with significant ultraviolet-visible spectral changes, which are similar to the changes induced by deprotonation of the free inhibitors in a basic solution. Dissection of the structure-affinity relationships of the inhibitors identifies the hydroxyl groups at positions 1 (C1-OH) and 4 (C4-OH) of DHNA-CoA as the dominant and minor sites, respectively, for the enzyme-ligand interaction that polarizes or deprotonates the bound ligands to cause the observed spectral changes. In the meantime, the active site residues involved in catalysis or ligand binding are defined through spectroscopic studies with active site mutants: C4-OH of the enzyme-bound DHNA-CoA interacts with residues (Arg-91, Tyr-97, Tyr-258 and Ser-161) through a hydrogen bonding network. The Asp-163 side chain is most likely hydrogen-bonded to C1-OH of DHNA-CoA to provide the dominant polarizing effect. Moreover, mutation of Asp-163 completely eliminates the enzyme activity; strongly supporting the possibility that this residue may provides a strong stabilizing hydrogen bond to the tetrahedral oxyanion, the second high-energy intermediate in the intracellular Claisen condensation reaction. Last part of my work is search for the missing thioesterase in MK biosynthesis. We reconstituted the activity of recombinant YbgC, which is proposed to be the missing DHNA-CoA thioesterase in E. coli MK biosynthesis, and found no DHNA-CoA hydrolase activity for YbgC. To further explore this issue, we tested other hotdog-fold proteins as the possible DHNA-CoA synthase in MK biosynthesis using gene knockout mutants. Our preliminary results point to the possibility that the other hotdog-fold protein named YdiI rather than YbgC is most likely the missing DHNA-CoA thiosesterase and the hydrolysis of DHNA-CoA may be accomplished by nonspecific thiosesterases from other pathways