Figure 1.
Phylogenetic reconstruction based on 16S rRNA sequences to map the taxonomic distribution of bioenergetic pathways.
272 prokaryotic species are shown, whose full genome sequence is available, and which represent the full diversity of bacteria and archaea, colour-coded based on their bioenergetic mode. Bootstrap values for highly supported nodes have been replaced by symbols, as indicated. The full species names, as well as details and accession numbers for all sequences used are given in Table S1. The tree shown was produced by RaxML, and its topology broadly agrees with the one produced by PhyML (the analysis based on MrBayes did not converge after 5 million generations when all sequences were included; however, when the bacteria and the archaea were examined separately, the MrBayes analysis also agreed with the RaxML and PhyML results).
Table 1.
The distribution of bioenergetic modes across taxonomic lineages suggests rampant horizontal gene transfer, or multiple independent origins.
Figure 2.
Phylogenetic reconstruction of ATPF0A.
The tree shown is the best Bayesian topology, based on 215 sequences and 232 amino acid positions (length after trimming; median sequence length before trimming: 254). Numerical values at the nodes of the tree (x/y/z) indicate statistical support by MrBayes, PhyML and RAxML (posterior probability, bootstrap and bootstrap, respectively). Values for highly supported nodes have been replaced by symbols, as indicated. Species names are colour-coded based on their bioenergetic mode as in Figure 1. Full details and accession numbers for all protein sequences used are given in Table S1. The tree is rooted at the N-ATPase clade, previously reported to be the result of horizontal gene transfer in a variety of species, all of which also contain a canonical ATPF0F1 (apart from the two Methanosarcina species shown, which also have a canonical ATPV). The tree confidently separates the major bacterial taxonomic lineages, but with limited support for their branching order: strong support is provided for a subgroup containing the verrucomicrobia and chloroflexi, while another subgroup containing the alpha-proteobacteria, actinobacteria, chlorobi, bacteroidetes and planctomycetes also has reasonable support. This group also includes the spirochaete Leptospira interrogans and the gemmatimonadete Gemmatimonas aurantiaca, as well as Candidatus Nitrospira defluvii which groups with the alpha-proteobacteria. Reasonable support is also provided for the grouping of dictyoglomi and cyanobacteria, and for a subgroup containing the fusobacteria, firmicutes, tenericutes, thermotogae, and beta-gamma-proteobacteria. Two species-specific duplications (in Saccharopolyspora erythraea and Pelobacter carbinolicus) are highlighted with a red “>”. Two further duplications are highlighted with a red “-” after the species name; in Photobacterium profundum the duplication either occurred before the split from other closely-related species or represents HGT from other gamma-proteobacteria; the duplication in Desulfococcus oleovorans possibly represents HGT from thermotogae (also see Figures S1 and S2).
Figure 3.
Phylogenetic reconstruction of ATPF1A.
The tree shown is the best Bayesian topology, based on 215 sequences and 502 amino acid positions (length after trimming; median sequence length before trimming: 508). Numerical values at the nodes of the tree (x/y/z) indicate statistical support by MrBayes, PhyML and RAxML (posterior probability, bootstrap and bootstrap, respectively). Values for highly supported nodes have been replaced by symbols, as indicated. Species names are colour-coded based on their bioenergetic mode as in Figure 1. Full details and accession numbers for all protein sequences used are given in Table S1. The tree is rooted at the N-ATPase clade, previously reported to be the result of horizontal gene transfer in a variety of species, all of which also contain a canonical ATPF0F1 (apart from the two Methanosarcina species shown, which also have a canonical ATPV). The tree confidently separates the major bacterial taxonomic lineages, but with limited support for their branching order: reasonable support is only provided for one subgroup containing the chlorobi, and the bacteroidetes. The spirochaete Leptospira interrogans groups with the planctomycetes. Two species-specific duplications (in Photobacterium profundum and Pelobacter carbinolicus) are highlighted with a red “>”. Two further duplications within the tenericutes are highlighted with a red “-” after the species name; this duplication likely happened before the split between Mycoplasma agalactiae and Ureaplasma parvum.
Figure 4.
ATPF0F1 gene locus organization per lineage.
The ATPF0F1 gene locus organization was checked for all species in the IMG database [47], and is summarized per lineage. The gene order shown follows the order in which the genes are transcribed in each genome (upstream to downstream). Semicolons indicate that the separated gene groups are on non-adjacent genetic locations (and can be very far upstream or downstream; e.g. separated by only 4 intervening ORFs in Geobacter sp. FRC-32, and by up to 5026 intervening ORFs, or 6 Mb, in Nostoc sp. PCC 7120; see Table S1). When the locus is split, the genes are shown in the order they are usually found in when the locus is intact. ATPF0B (K02109) is often duplicated, so one copy is called 0B, and the other 0B′, based on the gene order. ATPI (K02116) is also often duplicated, and is designated “I” “sI” and “R” based on the presence of distinct pfam domains, as discussed in the text. Question marks indicate that the ATPI subunit is sometimes not clearly assigned to the orthology group. “X” denotes hypothetical intervening ORFs. Notable variations within some lineages are shown. *Especially for lineages represented by relatively few species, please see TableS1 for variations between the species examined within each lineage.
Table 2.
Gene duplications of ATP synthase subunits in the species analyzed.