The root of the eukaryote tree pinpointed

Determining the precise position of the root of the eukaryote evolutionary tree is very important for understanding cell evolution [1], but has been a major challenge, because of systematic biases in gene sequence trees [2,3]. However, one can deduce the position indirectly, by using derived genic properties to exclude the possibility that the root lies within clades that share them [4]. A derived gene fusion between dihydrofolate reductase (DHFR) and thymidylate synthase (TS) genes strongly indicated that the root is not amongst the eukaryote groups ancestrally with two cilia (bikonts [1]). Bikonts share a fused DHFR-TS gene that is clearly derived [4]. However, the precise position of the root of the eukaryote tree remained uncertain, because of the unclear status of the Amoebozoa [4]. We have sequenced the region upstream of the TS gene in the amoebozoan Hartmannella cantabrigiensis and find that the TS and DHFR genes are separate. As the two genes are encoded on opposite strands they must be translated separately and thus be unfused (inset in Figure 1). Also the choanozoan Corallochytrium limacisporum contains only non-coding DNA upstream of the TS gene, but not a DHFR gene. Thus, all major opisthokont groups lack fused DHFR-TS genes. Until now, this had only been inferred from the fact that we were not able to detect the fusion gene in Corallochytrium, the fact that Choanozoa are sisters to animals [5] and from the absence of the fusion gene in both animals and Fungi. Given the improbability of reversal of the fusion in the bikont ancestor [4], this provides the best evidence to date that Amoebozoa are not secondarily derived from the bikonts, and that bikonts are a clade, not a paraphyletic group. Do other Amoebozoa also have separate DHFR and TS genes? Using our new sequences from Hartmannella, we searched ongoing genome sequencing projects of different Entamoeba species and Dictyostelium discoideum without success [4]. Our inability to amplify the bifunctional gene from two other Amoebozoa, Phalansterium and Phreatamoeba, suggests that all Amoebozoa lack the gene fusion. Another derived gene fusion, of three of the six enzymes in the pyrimidine synthesis pathway [7], provides the first really compelling support for Amoebozoa being sisters to opisthokonts rather than to bikonts. In eubacteria and archaebacteria all six enzymes of this pathway are separately translated, as are the two subunits of the first of these, carbamoyl-phosphate synthetase II (CPSII) [7]. CPSII had clearly undergone …