Evolution of smooth tubercle Bacilli PE and PE_PGRS genes: evidence for a prominent role of recombination and imprint of positive selection

Abstract

Background: PE and PE_PGRS are two mycobateria-restricted multigene families encoding membrane associated and secreted proteins that have expanded mainly in the pathogenic species, notably the Mycobacterium tuberculosis complex (MTBC). Several lines of evidence attribute to PE and PE_PGRS genes critical roles in mycobacterial pathogenicity. To get more insight into the nature of these genes, we sought to address their evolutionary trajectories in the group of smooth tubercle bacilli (STB), the putative ancestor of the clonal MTBC.

Methodology/principal findings: By focussing on six polymorphic STB PE/PE_PGRS genes, we demonstrate significant incongruence among single gene genealogies and detect strong signals of recombination using various approaches. Coalescent-based estimation of population recombination and mutation rates (ρ and θ, respectively) indicates that the two mechanisms are of roughly equal importance in generating diversity (ρ/θ = 1.457), a finding in a marked contrast to house keeping genes (HKG) whose evolution is chiefly brought about by mutation (ρ/θ = 0.012). In comparison to HKG, we found 15 times higher mean rate of nonsynonymous substitutions, with strong evidence of positive selection acting on PE_PGRS62 (dN/dS = 1.42), a gene that has previously been shown to be essential for mycobacterial survival in macrophages and granulomas. Imprint of positive selection operating on specific amino acid residues or along branches of PE_PGRS62 phylogenetic tree was further demonstrated using maximum likelihood- and covarion-based approaches, respectively. Strikingly, PE_PGR62 proved highly conserved in present-day MTBC strains.

Conclusions/significance: Overall the data indicate that, in STB, PE/PE_PGRS genes have undergone a strong diversification process that is speeded up by recombination, with evidence of positive selection. The finding that positive selection involved an essential PE_PGRS gene whose sequence appears to be driven to fixation in present-day MTBC strains lends further support to the critical role of PE/PE_PGRS genes in the evolution of mycobacterial pathogenicity.

Figure 1. ML Phylogenetic trees showing the relationships of STB genotypes based on PE_PGRS and HKG polymorphism.

HKG sequences were imported from the study of Gutierrez et al. . Numbers on branches of the ML tree are bootstrap support values.

Figure 2. Split decomposition analysis of the six variable STB PE/PE_PGRS genes as well as their concatenated sequence.

The p-value of the PHI test is indicated for each splitgraph. The analysis was performed by taking into account sSNPs only.

Figure 3. Detection of recombination breakpoints within PE/PE_PGRS sequences using GARD.

The plots display potential recombination breakpoints within PE/PE_PGRS sequences. The probability of the breakpoints is evaluated by akaike information criterion (AIC) score and Kishino-Hasegawa topological incongruence test . Model supported breakpoints are indicated with an asterisk.

Figure 4. Detection of recombination breakpoints with GARD upon concatenation of PE/PE_PGRS sequences.

(A) GARD plot showing potential recombination breakpoints within the concatenated PE/PE_PGRS sequence (PE3-PE4-PE_PGRS26-PE_PGRS35-PE_PGRS51). (B) Position of potential recombination breakpoints (black vertical bars) identified on a gene-by-gene analysis (top) and within the concatenated sequence (red vertical bars) (bottom). The probability of the breakpoints is evaluated by akaike information criterion (AIC) score and Kishino-Hasegawa topological incongruence test . Model supported breakpoints are indicated with an asterisk.

Figure 5. Estimation of dN/dS ratios along branches of STB PE/PE_PGRS individual trees using the covarion-based approach described by Siltberg and Liberles .

dN and dS rates (top and bottom numbers, respectively) are shown on branches. Branches where positive selection was detected are drawn in red.