Deciphering electron transfer and cytochrome P450 activity in Mycobacterium marinum

Abstract

Cytochrome P450s are haem-monooxygenase enzymes, responsible for the catalytic hydroxylation of a large variety of organic molecules. The bacterium Mycobacterium marinum, has a larger genome than its close relatives, the causative agents of human tuberculosis (Mycobacterium tuberculosis) and Buruli ulcer (Mycobacterium ulcerans), which have undergone substantial reductive evolution. The genome of M. marinum contains an unusually large number of P450 genes (47). Twelve ferredoxin genes are associated with the CYPome and eleven of these are uncharacterised ferredoxins of the 3/4Fe-4S type. In their iron-sulfur cluster binding motif (CXX?XXC(X)nCP), these ferredoxins (Fdx1 – Fdx11) have non-standard residues at the ? position of the sequence. Instead of the cysteine residue expected of a [4Fe-4S] ferredoxin, or the alanine/glycine residue expected in a [3Fe-4S] ferredoxin, they contain histidine, asparagine, tyrosine, serine, threonine and phenylalanine residues. In the course of this work, they have been purified aerobically and anaerobically. When isolated anaerobically, three of these ferredoxins were determined, by non-denaturing ESI-MS and EPR to contain 3Fe-4S clusters. The reduction potentials for the three varied from +150 mV to -360 mV, which are highly anomalous for [3Fe-4S] ferredoxins. Similar ferredoxins were found to accompany P450s in the biosynthetic gene clusters of other bacteria, especially in Actinomycete species. These ferredoxins were demonstrated to support the activity of a number of the M. marinum P450s, some of which were from previously uncharacterised families. CYP147G1, in combination with the electron transfer partners Fdx3 and FdR1 was demonstrated to act as a ω-1 fatty acid hydroxylase. CYP147G1 selectivity favoured the ω carbon when branched methyl substrates were used. The same ferredoxin reductase, FdR1, was also shown to support the activity of CYP278A1 (with Fdx2), and CYP150A5 (with Fdx8), both of which were shown to regioselectively hydroxylate β-ionone. CYP150A5 binds terpenes and polycyclic substrates. An additional CYP150 enzyme, CYP150A6, was crystallised and structurally resolved to 1.6 Å in the substrate-free form. CYP268A2, when reconstituted with a non-native electron transfer chain, hydroxylated the branched fatty acetate derivatives, pseudoionone and geranyl acetate, at the terminal position. The structure of CYP268A2 with trans-pseudoionone bound in the active site was solved by X-ray crystallography to a resolution of 2.0 Å and from this the selectivity of the enzyme was rationalised. Several M. marinum P450s that have close counterparts in M. tuberculosis were selected for comparison, in order to investigate whether the substrate and inhibitor binding affinities were preserved between species. The P450s investigated were analogues of the steroid metabolising P450s in M. tuberculosis. CYP125A6 and CYP125A7 have a single counterpart in M. tuberculosis (CYP125A1). The sequence identity and cholesterol binding affinity of CYP125A7 indicates it more closely resembles CYP125A1. However, CYP125A7 interacts differently to CYP125A1 with a range of inhibitors. CYP142A3 bound sterols with similar affinities as the M. tuberculosis CYP142A1. CYP124A1 from M. marinum was structurally characterised by X-ray crystallography, and showed a very closely preserved active site when compared to the M. tuberculosis analogue. These results suggest that individual P450 enzymes have maintained similar substrate specificities and roles between Mycobacterium species. However, for effective inhibitor design cross-species differences should be noted.