Deciphering the genetic basis for polyketide variation among mycobacteria producing mycolactones

Mycolactones are immunosuppressive and cytotoxic polyketides, comprising five naturally occurring structural variants (named A/B, C, D, E and F), produced by different species of very closely related mycobacteria including the human pathogen, Mycobacterium ulcerans. In M. ulcerans strain Agy99, mycolactone A/B is produced by three highly homologous type I polyketide megasynthases (PKS), whose genes (mlsA1: 51 kb, mlsA2: 7.2 kb and mlsB: 42 kb) are found on a 174 kb plasmid, known as pMUM001.

We report here comparative genomic analysis of pMUM001, the complete DNA sequence of a 190 kb megaplasmid (pMUM002) from Mycobacterium liflandii 128FXT and partial sequence of two additional pMUM replicons, combined with liquid chromatography-tandem mass spectrometric (LC-MS/MS) analysis. These data reveal how PKS module and domain differences affecting MlsB correlate with the production of mycolactones E and F. For mycolactone E these differences from MlsB in M. ulcerans Agy99 include replacement of the AT domain of the loading module (acetate to propionate) and the absence of an entire extension module. For mycolactone F there is also a reduction of one extension module but also a swap of ketoreductase domains that explains the characteristic stereochemistry of the two terminal side-chain hydroxyls, an arrangement unique to mycolactone F CONCLUSION: The mycolactone PKS locus on pMUM002 revealed the same large, three-gene structure and extraordinary pattern of near-identical PKS domain sequence repetition as observed in pMUM001 with greater than 98.5% nucleotide identity among domains of the same function. Intra- and inter-strain comparisons suggest that the extreme sequence homogeneity seen among the mls PKS genes is caused by frequent recombination-mediated domain replacement. This work has shed light on the evolution of mycolactone biosynthesis among an unusual group of mycobacteria and highlights the potential of the mls locus to become a toolbox for combinatorial PKS biochemistry.