Such a tree would suggest that proteases within the groups 3b/3d developed before the proteases of group 3a and 4, which seems far-fetched since proteases of group 3a and 4 type cleaves hydrogenases that are deeper branched then the 3b/3d hydrogenases. We therefore suggest that the placement of HOX-specific proteases (3d) and the scattered

result of 3b proteases in the phylogenetic tree may be the result of horizontal gene transfer (HGT). HGT is today seen as a major force in evolution and has occurred numerous times between archaea and bacteria [30–33]. Within prokaryotes almost no gene family is untouched by HGT [34] and there are also numerous cases of HGT within cyanobacteria [35]. [NiFe]-hydrogenases have not been spared from this mechanism and an archaeal SB273005 price organism is believed to be the origin of the Ech- hydrogenase in Thermotoga maritima [36]. By comparing the phylogenetic tree of hydrogenases and

their specific protease and assuming that the [NiFe]-hydrogenase and its specific protease have evolved together the most likely scenario is that an early group 3 [NiFe]-hydrogenase with or without its specific protease was transferred, most probably from an archaeal organism to a bacterial. If we assume that the LOXO-101 ic50 type 3 hydrogenase and the protease transferred together then this indicates that most likely the root of the tree should be placed between group 3a and 4 (point Z; Figure 1) and that the protease transferred is the ancestor of all type 1, 2 and 3d proteases (Figure 8). If we assume the opposite, (that the hydrogenase transferred alone), then the root should instead be placed between type 1/2/3d and type 3a/4 proteases (point Y; Figure 1) and the transferred hydrogenase must have incorporated an already existing type 1 protease to its maturation process. The scattered impression of type 1 and 3b proteases from the less robust phylogenetic tree with additional

hydrogenase specific proteases (Additional file 1) could be the result e.g. older phylum branching off close to the HGT point, poor Epigenetics inhibitor resolution of the phylogenetic tree or by additional oxyclozanide HGT and so does not contradict our proposed theory of HGT. Rooting the tree with an outgroup; germination protease (GPR), the closest relative to the [NiFe]-hydrogenase specific proteases, (data not shown) placed the root between group 3a and 4 suggest that the first scenario, a root between group 3a and 4, is more plausible (point Z; Figure 1). However, all attempts at rooting the tree resulted in very unstable phylogenetic trees. When considering both GPR endopeptidase function (bacterial spoluration) and taxonomic location (bacterial phylum of firmicutes only) it is plausible that the [NiFe]-hydrogenase specific proteases are instead the ancestor of GPR, making any tree with GPR as outgroup unreliable.