The SAR11 group of alpha-proteobacteria is not related to the origin of mitochondria

Abstract

Although free living, members of the successful SAR11 group of marine alpha-proteobacteria contain a very small and A+T rich genome, two features that are typical of mitochondria and related obligate intracellular parasites such as the Rickettsiales. Previous phylogenetic analyses have suggested that Candidatus Pelagibacter ubique, the first cultured member of this group, is related to the Rickettsiales+mitochondria clade whereas others disagree with this conclusion. In order to determine the evolutionary position of the SAR11 group and its relationship to the origin of mitochondria, we have performed phylogenetic analyses on the concatenation of 24 proteins from 5 mitochondria and 71 proteobacteria. Our results support that SAR11 group is not the sistergroup of the Rickettsiales+mitochondria clade and confirm that the position of this group in the alpha-proteobacterial tree is strongly affected by tree reconstruction artefacts due to compositional bias. As a consequence, genome reduction and bias toward a high A+T content may have evolved independently in the SAR11 species, which points to a different direction in the quest for the closest relatives to mitochondria and Rickettsiales. In addition, our analyses raise doubts about the monophyly of the newly proposed Pelagibacteraceae family.

Figure 1. Nucleotide and amino acid composition of the dataset used in this study.

A) Bar graph displaying the percentage of A+T in the concatenation of the nucleotide sequences. B) Reduced dimensionality plot showing the main principal components of the global amino acid compositions. Dots are coloured as in Figure 1A. The group of overlapping dots at the bottom right corner contains the species with the lowest A+T %. The variances that explain the two first axes are respectively 92.3% and 6.3%.

Figure 2. Phylogeny based on 24 mitochondrial/bacterial proteins (6,542 amino acid positions) inferred by Bayesian Inference with the WAG + F (left) or the CAT mixture (right) model.

Numbers indicate posterior probability values. Branches without values are supported by posterior probabilities of 1.0. The scale bar denotes the estimated number of amino acid substitution per site. See supplementary material for complete species names and proteins used.

Figure 3. Phylogeny based on 24 mitochondrial/bacterial proteins (6,542 amino acid positions) inferred by Bayesian Inference with the GTR + F (left) or the CAT mixture (right) model on the Dayhoff recoded dataset.

(see Methods for details). Branches without values are supported by posterior probabilities of 1.0. The scale bar denotes the estimated number of amino acid substitution per site. See supplementary material for complete species names and proteins used.

Figure 4. Phylogeny based on 24 mitochondrial/bacterial proteins (6,542 amino acid positions) inferred by Bayesian Inference with the WAG/GTR + F (left) or the CAT mixture (right) model on the non recoded and Dayhoff recoded dataset.

(see Methods for details). Branches without values are supported by posterior probabilities of 1.0. When at least one dataset gives posterior probability <1, both values are shown, standard coding on the left and Dayhoff coding on the right. The scale bar denotes the estimated number of amino acid substitution per site. See supplementary material for complete species names and proteins used.