Find in Library

<p>Abstract</p> <p>Background</p> <p><it>PPE38 </it>(Rv<it>2352c</it>) is a member of the large <it>PPE </it>gene family of <it>Mycobacterium tuberculosis </it>and related mycobacteria. The function of PPE proteins is unknown but evidence suggests that many are cell-surface associated and recognised by the host immune system. Previous studies targeting other <it>PPE </it>gene members suggest that some display high levels of polymorphism and it is thought that this might represent a means of providing antigenic variation. We have analysed the genetic variability of the <it>PPE38 </it>genomic region on a cohort of <it>M. tuberculosis </it>clinical isolates representing all of the major phylogenetic lineages, along with the ancestral <it>M. tuberculosis </it>complex (MTBC) member <it>M. canettii</it>, and supplemented this with analysis of publicly available whole genome sequences representing additional <it>M. tuberculosis </it>clinical isolates, other MTBC members and non tuberculous mycobacteria (NTM). Where possible we have extended this analysis to include the adjacent <it>plcABC </it>and <it>PPE39/40 </it>genomic regions.</p> <p>Results</p> <p>We show that the ancestral MTBC <it>PPE38 </it>region comprises 2 homologous <it>PPE </it>genes (<it>PPE38 </it>and <it>PPE71</it>), separated by 2 <it>esat-6 </it>(<it>esx</it>)-like genes and that this structure derives from an <it>esx/esx/PPE </it>duplication in the common ancestor of <it>M. tuberculosis </it>and <it>M. marinum</it>. We also demonstrate that this region of the genome is hypervariable due to frequent IS<it>6110 </it>integration, IS<it>6110</it>-associated recombination, and homologous recombination and gene conversion events between <it>PPE38 </it>and <it>PPE71</it>. These mutations result in combinations of gene deletion, gene truncation and gene disruption in the majority of clinical isolates. These mutations were generally found to be IS<it>6110 </it>strain lineage-specific, although examples of additional within-lineage and even within-cluster mutations were observed. Furthermore, we provide evidence that the published <it>M. tuberculosis </it>H37Rv whole genome sequence is inaccurate regarding this region.</p> <p>Conclusion</p> <p>Our results show that this antigen-encoding region of the <it>M. tuberculosis </it>genome is hypervariable. The observation that numerous different mutations have become fixed within specific lineages demonstrates that this genomic region is undergoing rapid molecular evolution and that further lineage-specific evolutionary expansion and diversification has occurred subsequent to the lineage-defining mutational events. We predict that functional loss of these genes could aid immune evasion. Finally, we also show that the <it>PPE38 </it>region of the published <it>M. tuberculosis </it>H37Rv whole genome sequence is not representative of the ATCC H37Rv reference strain.</p>