Abstract:
Mycobacterium tuberculosis is the causative agent of tuberculosis (TB), an ancient disease that causes approximately 2 million deaths worldwide annually. This research focuses on the biosynthetic pathway for the siderophore mycobactin in M. tuberculosis. The genes that encode mycobactin biosynthetic enzymes are found in two gene clusters, mbt and mbt-2, and are essential for the growth of the organism, both in vitro and in macrophages. This indicates that mycobactin biosynthetic enzymes are excellent drug targets. In this research, a number of the enzymes involved in mycobactin biosynthesis, MbtG (Rv2378c), MbtL (Rv1344), MbtM (Rv1345), MbtN (Rv1346), have been selected for structural studies. Expression trials showed that MbtM was only slightly soluble when expressed in E. coli whereas MbtL was soluble but could not be purified to homogeneity. The structure of MbtK (Rv1347c) had already been determined prior to this study, and was selected for fragment screening studies. It was not possible to obtain additive-free MbtK crystals, however, and for this reason the enzyme was deemed not suitable for fragment screening by X-ray crystallography. The crystal structure of MbtN, which is annotated as an acyl-CoA dehydrogenase (ACAD), was solved by molecular replacement using synchrotron data and refined at 2.3 Å resolution. The overall fold of MbtN proved to conform to the fold characteristic of other ACADs, comprising an N-terminal !–helical domain, a middle "–sheet domain, and a C-terminal !–helical domain, with an FAD cofactor bound to each monomer. The flavin ring of the bound FAD was found to be substantially bent, consistent with reduction of the cofactor. A notable feature of the structure is the presence of a mystery ligand connected to the flavin N(5) atom, interpreted as a polyethylene glycol fragment derived from the crystallization solution. This covalent adduct could represent a trapped complex formed during “turnover” or could result from a reaction caused by photoreduction of the flavin in the intense synchrotron X-ray beam. Kinetic studies and deductions from the three-dimensional structure suggested that MbtN has a preference for substrates with longer acyl chain lengths, which is in accordance with the findings that mycobactins have long fatty acyl substituents. A basic patch that is important for electrostatic interaction with acyl carrier protein (ACP) has been identified. An interesting feature is the different orientation of the putative acyl-binding pocket of MbtN. In silico docking suggests that it is possible that the MbtN pocket preferentially binds substrate with a cis-like conformation. Preliminary NMR data have shown inconclusive results, and the stereochemical course of the MbtN reaction remains to be investigated. Expression of MbtG gave only very small amounts of soluble protein, and crystallization attempts were unsuccessful. Attention was then turned to the homologous enzyme, MbsG from M. smegmatis (75 % sequence identity). MbsG was found to be much more soluble and gave significantly higher yield than MbtG. Size exclusion chromatography coupled with multi-angle laser light scattering (SEC-MALLS) analysis showed that neither protein existed in a distinct oligomer state, instead forming higher order states (> 10-mers), and native-PAGE analyses showed the presence of two populations consisting of multiple species. Despite the heterogeneity in solution, MbsG was crystallized in two forms: native and methylated. Different cryoprotectants were tested, and it is not known whether the cause of poor diffraction was poor crystals, bad cryoprotection or flash cooling. Based on circular dichroism (CD) analysis, MbsG contains a high percentage of !–helices and a small percentage of "–strands, which is quite different from the secondary structure contents of already structurally characterized orthologues PvdA and SidA. In summary, this study provides valuable information that can be used for future investigation into the functions of these proteins.