A seven-gene, multilocus, genus-wide approach to the phylogeny of mycobacteria using supertrees.

This is the first study that estimates mycobacterial phylogeny using the maximum-likelihood method (PhyML-aLRT) on a seven-gene concatenate (hsp65, rpoB, 16S rRNA, smpB, sodA, tmRNA and tuf) and the super distance matrix (SDM) supertree method. Two sets of sequences were studied: a complete seven gene sequence set (set R, type strains of 87 species) and an incomplete set (set W, 132 species) with some missing data. Congruencies were computed by using the consense program (phylip package). The evolution rate of each gene was determined, as was the evolution rate of each strain for a given gene. Maximum-likelihood trees resulting from concatenation of the R and W sets resulted in a similar phylogeny, usually showing an early separation between slow-growing (SG) and rapidly growing (RG) mycobacteria. The SDM tree for the W set resulted in a different phylogeny. The separation of SG and RG was still evident, but it was located later in the nodes. The SG were therefore positioned as a subgroup of RG. Maximum-likelihood phylogenetic reconstruction was less affected by increasing the number of strains (with incomplete data), but did seem to cushion the variability of the evolution rate (ER), whereas the SDM method seemed to be more accurate and took into account both the differing ER values and the incomplete data. With regard to ER, it was observed that the 16S rRNA gene was the gene that displayed the slowest evolution, whereas smpB was the most rapidly evolving gene. Surprisingly, these two genes alone accurately separated the SG from the RG on the basis of their ER values. This study focused on the differences in ER between genes and in some cases linked the ER to the phenotypic classification of the mycobacteria.

Identification of Mycobacterium using the EF-Tu encoding (tuf) gene and the tmRNA encoding (ssrA) gene.

The partial nucleotide sequences encoding the elongation factor Tu (tuf gene) (652 bp) and transfer-mRNA (tmRNA or ssrA gene) (340 bp) were determined to assess the suitability of these two genes as phylogenetic markers for the classification of mycobacteria, and thus as alternative target molecules for identifying mycobacteria. A total of 125 reference strains of the genus Mycobacterium and 74 clinical isolates were amplified by PCR and sequenced. Phylogenies of the two genes constructed by the neighbour-joining method were created and compared to a concatenated tree of 16S rDNA, hsp65, sodA and rpoB genes. The phylogenetic trees revealed the overall natural relationships among Mycobacterium species. The tmRNA phylogeny was similar to that of 16S rDNA, with low resolving power. The tuf gene provided better resolution of each mycobacterial species, with a phylogeny close to that of hsp65. However, none of these methods differentiated between the members of the Mycobacterium tuberculosis complex or the subspecies of the Mycobacterium avium complex. The correct identification of clinical isolates confirms the interest of these genes, especially tuf. It is suggested from these findings that tmRNA might be useful as another housekeeping gene in a polyphyletic approach to Mycobacterium species, but not as a first-line marker of species. tuf gene analysis suggests that this gene could be used effectively for phylogenetic analysis and to identify mycobacteria.

Mycobacterium bovis infection, Lyon, France.

In a 5-year retrospective study, we used spoligotyping and mycobacterial interspersed repetitive units to type 13 strains of Mycobacterium bovis isolated from human sources. Despite the relatively high incidence of human tuberculosis caused by M. bovis (2%), these tools showed no clonal evolution and no relationships between the isolates.