Genomic repeats, genome plasticity and the dynamics of Mycoplasma evolution

Abstract

Mycoplasmas evolved by a drastic reduction in genome size, but their genomes contain numerous repeated sequences with important roles in their evolution. We have established a bioinformatic strategy to detect the major recombination hot-spots in the genomes of Mycoplasma pneumoniae, Mycoplasma genitalium, Ureaplasma urealyticum and Mycoplasma pulmonis. This allowed the identification of large numbers of potentially variable regions, as well as a comparison of the relative recombination potentials of different genomic regions. Different trends are perceptible among mycoplasmas, probably due to different functional and structural constraints. The largest potential for illegitimate recombination in M.pulmonis is found at the vsa locus and its comparison in two different strains reveals numerous changes since divergence. On the other hand, the main M.pneumoniae and M.genitalium adhesins rely on large distant repeats and, hence, homologous recombination for variation. However, the relation between the existence of repeats and antigenic variation is not necessarily straightforward, since repeats of P1 adhesin were found to be anti-correlated with epitopes recognized by patient antibodies. These different strategies have important consequences for the structures of genomes, since large distant repeats correlate well with the major chromosomal rearrangements. Probably to avoid such events, mycoplasmas strongly avoid inverse repeats, in comparison to co-oriented repeats.

Figure 1

Distribution of repeats in the four Mycoplasma genomes. Repeats were classed into five classes, SSRs, close repeats and three classes for large repeats (inverse, direct and multiple). The x-axis represents the position of the first occurrence of the repeat in the chromosome and the y-axis represents the position of the second occurrence. Naturally, the two copies overlap for SSR and almost do so for close repeats. For multiple repeats all possible pairs among the group are shown.

Figure 2

Comparison of the vsa locus of the UAB CTIP and KD735-15 strains of M.pulmonis. Boxes indicate regions of SSR common to both sequences and diagonal lines regions of high similarity between the two loci. The legends to boxes are of the form A × (B, C), where A indicates the length of the repeated motifs and B and C represent the number of consecutive motifs identified in the UAB CTIP strain and in the KD735-15 strain. The expression site is indicated by an arrow. The gene nomenclature is taken from the original papers describing these loci (7,31). CHP, conserved hypothetical protein.

Figure 3

Analysis of the potential for homologous recombination between the P1 and MgpB proteins and their pseudogenes scattered on the chromosome. The y-axis represents the number of repeats larger than 25 bp found elsewhere in the genome for sliding windows of 60 bp (10 bp steps) along the genes coding for P1 (A) and MgpB (C). HAb (B) indicates the positions of the binding sites of human IgG serum antibodies of patients suffering from M.pneumoniae disease (35). Similarity between mgpB and P1 at the DNA level (D) was computed in sliding windows of 30 bp (10 bp steps) of the mgpB gene (see Materials and Methods for details).

Figure 4

Synteny between the chromosomes of M.pneumoniae and M.genitalium (inner square) and distribution of large direct repeats capable of engaging in homologous recombination (histograms). Dots represent the position of orthologous genes in the respective genomes.