Objections to the proposed reclassification of Eubacterium rectale as Agathobacter rectalis

We disagree with a recent proposal by Rosero et al. (2016) to transfer Eubacterium rectale into a newly proposed genus, Agathobacter. The role of E. rectale as one of the most abundant species in the human intestine means that any reclassification could have a major impact on human microbiota research and the supporting evidence therefore needs to be assessed critically. The species E. rectale as defined by Moore & Holdeman Moore (1986) included four subgroups, but many of the original human colonic E. rectale isolates (Moore & Holdeman, 1974) would now be classified within four newly defined species of Roseburia (Duncan et al. 2006). This has led to a narrower definition of the species E. rectale based on the type strain (ATCC 33656) and neotype strain A1-86 (5DSM 17629) (Duncan & Flint, 2008). Nevertheless, E. rectale and species of the genus Roseburia show considerable similarities in their 16S rRNA gene sequences, genome organization and physiology (Aminov et al., 2006; Louis & Flint 2009; Neville et al., 2013; Sheridan et al., 2016). It is unfortunate that Rosero et al. (2016) (in their Table 1) give an erroneous figure for the mol% DNA G+C content of Roseburia intestinalis, which should be 42.6 mol%, a value very close to those of E. rectale strains and of other species of the genus Roseburia (Duncan et al., 2006). It is also unfortunate that not all type strains of species of the genus Roseburia were included in their phylogenetic trees. Phylogenetic tree topologies are of course notoriously sensitive to the number and type of sequences used in tree reconstruction, but we find that the 16S rRNA gene sequence of Agathobacter ruminis is as distant from that of E. rectale as it is from that of Roseburia faecis (both 0.034 base substitutions per site). We also note that most of the trees presented by Rosero et al. (2016) either do not support reclassification of E. rectale as a member of the genus Agathobacter (Figs S3, S4 and S7) or show that species of the genus Agathobacter, E. rectale and species of the genus Roseburia could be split into one, two or three genera (Figs S5 and S6) depending on which node is used. Furthermore, when considering additional criteria for taxonomic proposals we note that neither substrate utilization nor lipid profiles reported by Rosero et al. (2016) appear to indicate any closer similarity between E. rectale and A. ruminis than between E. rectale and R. intestinalis.