Phylogenomic Classification and the Evolution of Clonal Complex 5 Methicillin-Resistant Staphylococcus aureus in the Western Hemisphere

Abstract

Clonal complex 5 methicillin-resistant Staphylococcus aureus (CC5-MRSA) includes multiple prevalent clones that cause hospital-associated infections in the Western Hemisphere. Here, we present a phylogenomic study of these MRSA to reveal their phylogeny, spatial and temporal population structure, and the evolution of selected traits. We studied 598 genome sequences, including 409 newly generated sequences, from 11 countries in Central, North, and South America, and references from Asia and Europe. An early-branching CC5-Basal clade is well-dispersed geographically, is methicillin-susceptible and MRSA predominantly of ST5-IV such as the USA800 clone, and includes separate subclades for avian and porcine strains. In the early 1970s and early 1960s, respectively, two clades appeared that subsequently underwent major expansions in the Western Hemisphere: a CC5-I clade in South America and a CC5-II clade largely in Central and North America. The CC5-I clade includes the ST5-I Chilean/Cordobes clone, and the ST228-I South German clone as an early offshoot, but is distinct from other ST5-I clones from Europe that nest within CC5-Basal. The CC5-II clade includes divergent strains of the ST5-II USA100 clone, various other clones, and most known vancomycin-resistant strains of S. aureus, but is distinct from ST5-II strain N315 from Japan that nests within CC5-Basal. The recombination rate of CC5 was much lower than has been reported for other S. aureus genetic backgrounds, which indicates that recurrence of vancomycin resistance in CC5 is not likely due to an enhanced promiscuity. An increased number of antibiotic resistances and decreased number of toxins with distance from the CC5 tree root were observed. Of note, the expansions of the CC5-I and CC5-II clades in the Western Hemisphere were preceded by convergent gains of resistance to fluoroquinolone, macrolide, and lincosamide antibiotics, and convergent losses of the staphylococcal enterotoxin p (sep) gene from the immune evasion gene cluster of phage ϕSa3. Unique losses of surface proteins were also noted for these two clades. In summary, our study has determined the relationships of different clades and clones of CC5 and has revealed genomic changes for increased antibiotic resistance and decreased virulence associated with the expansions of these MRSA in the Western Hemisphere.

Keywords: MRSA; convergent evolution; local adaptation; methicillin-resistant Staphylococcus aureus; phylogenomics.

FIGURE 1

Phylogeny of CC5. The ML phylogeny has recombination-corrected branch lengths and is outgroup-rooted. The major clades of CC5-Basal, CC5-I, and CC5-II are indicated with shading. The statistical support for these clades is described in the text. The positions of reference strains are indicated with stars. The four rings in the direction of inner-outer, show continent (ring 1), country (ring 2), SCCmec type (ring 3), and multilocus ST (ring 4) for each genome. STs with less than three genomes were combined into “Other STs” for display purposes. Scale bar indicates number of substitutions per site.

FIGURE 2

Convergent evolution within CC5. The ML phylogeny of Figure 1 is shown. The major clades, multilocus ST, and SCCmec type are shown for reference purposes and colored as in Figure 1. The CC5-I and CC5-II B expansion nodes are indicated with arrows. ML ancestral state reconstruction is shown along the phylogenetic backbone leading to the two expansion nodes; the four pie-charts at each of the backbone nodes indicate the likelihood of (in left-to-right order), fluoroquinolone, macrolide, and lincosamide resistance (red) or susceptibility (blue), and the presence (red) or absence (blue) of the sep toxin gene. Table adjacent to the tree indicates these traits in each genome. Scale bar indicates number of substitutions per site.