*3.6. General Aspects*

Traditionally and especially due to the pioneering work of Carl Woese [25], the 16S rRNA gene sequence has been established as the basic tool for the analysis of bacterial phylogenies. Though the sequence information contained in this molecule is of limited size (approx. 1400 nt), it is considered as particularly conservative in evolutionary terms. In consequence, 16S rRNA gene sequences still are used as a backbone for phylogenetic considerations, although limitations are to be expected due to the comparable small sequence information and restricted resolution. Further limitations may be due to multiple changes in individual sequence positions and insertions/deletions over time, which could blur the phylogenetic roots in particular of the deep branching lineages.

In this context, it was quite remarkable that some of those species/groups that revealed the most obvious differences between the PS tree and RNA tree also showed the deepest branching points within the PS tree. In fact, a number of those species and branches that were not congruen<sup>t</sup> with respect to RNA and PS phylogeny had statistically poorly supported positions in either one or both of the trees. This was especially relevant for most of the *Alphaproteobacteria*, which appeared—with the remarkable exception of the *Rhodobacterales—*to be the most ancient group of phototrophic *Proteobacteria*.

Despite the uncertainty in the resolution of the very deeply branching lineages, the transfer of photosynthesis genes could explain several of the discrepancies between the PS and RNA trees. Such mechanisms have been postulated earlier [26,29], and genetic exchange could have occurred repeatedly in the early ages of photosynthesis in *Proteobacteria*. Examples of such possible exchange events during early diversification of the phototrophic purple bacteria may be found in *Gemmatimonas phototrophica*, the *Rhodobium*/*Hoeflea* group, and the *Betaproteobacteria* with different events of *Rhodocyclus* and *Rubrivivax* and their relatives.

The situation was different within the *Rhodobacterales*, which is a well-resolved group with a clear distinction from other *Proteobacteria*. It is the younges<sup>t</sup> diversification within the phototrophic *Proteobacteria*. For this group, lateral gene transfer has been demonstrated [41]. It could, in fact, be shown that the photosynthetic gene cluster in several genomes, including *Sulfitobacter* and *Roseobacter* species, is located on a plasmid, which enforces the genetic exchange of the whole cluster [41]. Several of such exchanges could explain the divergences between photosynthesis phylogeny and RNA phylogeny among the *Rhodobacterales* [41]. Despite the established gene transfer inside the *Rhodobacterales*, it appears highly unlikely that the whole group received the photosynthetic gene cluster by lateral transfer from an external donor. The long phylogenetic distance to most other photosynthesis systems and the basically good correlation of RNA and PS phylogeny in regard to relations of *Sphingomonadales*, *Brevundimonas,* and *Rhodobacterales* precludes the transfer from any other known phototrophic lineage. As we have no knowledge of the existence of similar gene transfer agents in other phototrophic bacteria, this kind of genetic exchange of the complete photosynthetic gene cluster could be a late acquisition and unique to the *Rhodobacterales*. However, it would be interesting to study the situation in *Erythrobacter marinus,* which could have similarly received its photosynthesis genes from a relative of the *Sphingomonadaceae* branch.
