Pyridoxal 5’-phosphate (PLP) is an essential enzyme cofactor required for the correct functioning of PLP-dependent enzymes (PLP-DE), which catalyse a plethora of vital cellular processes. PLP-DE are ubiquitously present in all living organisms, including pathogens such as Mycobacterium tuberculosis (Mtb). Understanding their role in pathogen biology could significantly advance our understanding of the pathogen and facilitate the discovery of novel therapeutic agents.
In this doctoral thesis, the development of an activity-based profiling approach for the identification of PLP-DE in Mtb is described. After synthesising a PLP analogue chemical probe, it was used in chemical proteomics experiments where selective enrichment of PLP-DE was achieved and 51 % of predicted Mtb PLP-DE could be identified. Two of the identified proteomics hits, Rv1769 and Rv2148c, were completely unannotated proteins that have not been computationally predicted to bind PLP. In a series of follow-up experiments, Rv2148c was expressed and purified, and some of its biophysical parameters as well as its crystal structure were determined, which confirmed that Rv2148c does indeed bind a PLP molecule and that it is a fold type III PLP-DE. Several attempts to determine the function of Rv2148c were made, but this remains a work in progress.
In addition, a secondary project focusing on the mode of action of D-cycloserine (DCS), an Mtb antibiotic, was undertaken. DCS was thought to irreversibly inactivate Mtb alanine racemase (MtAlr), a fold type III PLP-DE, by forming a covalent adduct with PLP. However, after realising that MtAlr is not fully inhibited in vivo, a systematic study to reveal the mechanism of inhibition was undertaken. It was found that the PLP-DCS adduct can undergo hydrolysis, which is followed by chemical rearrangement and release of the product into the cellular milieu, leading to the reactivation of MtAlr. DCS inhibition was therefore found to be a reversible process.