, 2004), we cannot rule out that one or more erm genes of the Firmicutes might have been acquired from antibiotic-producing bacteria long ago. Subsequently, erm genes may have undergone nucleotide replacement while adapting their codon usage to a lower G+C content selleck screening library similar to their new hosts, which eventually resulted in changes in their amino acid sequences. Therefore, the early bifurcation of the main clade of Erm methylases could be an artifact
generated from present-day differences in amino acid sequences, which cannot be distinguished from evolutionary changes with currently available phylogenetic tree-constructing algorithms. Even though the tree topology supports the respective monophylies of the two protein families, the evolutionary relationship between Erm and KsgA/Dim1 remains to be unresolved because the tree cannot be precisely rooted because of the long-branch attraction problems and an insufficient signal for deep phylogeny
due to short sequences. The relatively longer branch lengths observed in the cluster of Erm methylases, compared with those in the cluster of corresponding bacterial KsgAs, reflect a more rapid evolution of the Erm sequences (Fig. 2). Such rate heterogeneity and weak phylogenetic signals frequently cause unavoidable problems (i.e., long-branch attraction) in reconstructing deep phylogenies and might have induced artifactual paralogies of the two protein families in our analyses. In addition, the fact that KsgA/Dim1 is one of the last common ancestors [i.e., a rare protein Protein Tyrosine Kinase inhibitor family conserved in all three domains of life (O’Farrell et al., 2006; Pulicherla et al., 2009)] suggests that ksgA may have been recruited in some bacteria under antibiotic
pressure and evolved into a new gene, erm, consistent with the irregular presence of erm in life, found in only certain pathogenic and soil bacteria. Indeed, it has been shown recently that certain antibiotic-resistance proteins share structural homologies with proteins having little or no relationship with antibiotic resistance, implying that those proteins might be immediate precursors or ancestors of antibiotic-resistance proteins DOK2 (Wright, 2007). In fact, the target nucleotide of Erm, an adenine in bacteria, is substituted by a guanine in eukaryotes and archaea (Bottger et al., 2001; Davidovich et al., 2008), indicating that there is no appropriate substrate for Erm proteins in eukaryotes and archaea. The phylogenetic tree also shows that horizontal erm gene transfer occurs not only within closely related genera and species but also between phylogenetically remote bacteria. The inclusion of the branches of the Erm methylases from Arcanobacterium, Bacteroides, Neisseria, and E. coli in the clade of the Firmicutes is evidence of horizontal gene transfer between phylogenetically distant bacteria (Fig. 4). The base composition also provides some important information on the acquisition of foreign DNA from different organisms.