Homologous recombination among cp32 plasmids.

aClustalX2 0.3 [214] was used to create a neighbor-joining tree for proteins encoded in the genes the four cp32 “variable regions”. The variable regions were identified by inspection of matrix comparison plots (Figure S6A), and their positions are shown in Figure S6B. Robust branches (amino acid “sequence types”) were identified for the proteins encoded by the genes in each region, and these different protein types were given arbitrary numbers. The trees and definitions of sequence types are shown in figures S6C–F. Where a region 3 has two genes, the left (transcriptionally upstream) gene is in column “a” and the right gene is in column “b”. A dash “–” indicates no gene is present; PF, paralogous family. The genes in the rightmost region 4 “other” position are as follows: 1, mlp type 3 gene; 2, homolog of b31_m39; 3, homologs of 297_w43 (very weak similarity to b31_m39) and 297_w44 (PFam55); 4, homolog of 297_m41; 5, unique 297_o29. The different B31_S27, Rev and BapA proteins are all so similar that subtypes were not evident.