View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Magazine R665 Correspondence are not secondarily derived from use a uniquely derived fusion of the the bikonts, and that bikonts are a mitochondrial cytochrome oxidase clade, not a paraphyletic group. Do 1 and 2 genes shared by other Amoebozoa also have Dictyostelium and Acanthamoeba The root of the separate DHFR and TS genes? [8] to show that they are part of a eukaryote tree Using our new sequences from broader amoebozoan clade and Hartmannella, we searched that neither the eukaryote nor the pinpointed ongoing genome sequencing opisthokont root can lie within this projects of different Entamoeba amoebozoan subclade. species and Dictyostelium An internal gene duplication to Alexandra Stechmann and discoideum without success [4]. form an N-terminal catalytic part Thomas Cavalier-Smith Our inability to amplify the and C-terminal regulatory part of bifunctional gene from two other phosphofructokinase (PFK) also Determining the precise position of Amoebozoa, Phalansterium and cleanly partitions eukaryotes into the root of the eukaryote Phreatamoeba, suggests that all unikonts and bikonts. This evolutionary tree is very important Amoebozoa lack the gene fusion. duplication/fusion is a shared for understanding cell evolution [1], Another derived gene fusion, of derived state of unikonts, so far but has been a major challenge, three of the six enzymes in the found only in animals, fungi and because of systematic biases in pyrimidine synthesis pathway [7], Dictyostelium [9,10]. If it originated gene sequence trees [2,3]. provides the first really compelling in their last common ancestor it However, one can deduce the support for Amoebozoa being would be a unikont synapomorphy position indirectly, by using derived sisters to opisthokonts rather than like the triple gene fusion. Bikonts genic properties to exclude the to bikonts. In eubacteria and have very different PFK possibility that the root lies within archaebacteria all six enzymes of paralogues, lacking both this clades that share them [4]. A this pathway are separately duplicated PFK and its derived gene fusion between translated, as are the two subunits unduplicated orthologue [11], dihydrofolate reductase (DHFR) of the first of these, carbamoyl- widespread in bacteria but and thymidylate synthase (TS) phosphate synthetase II (CPSII) [7]. apparently absent in genes strongly indicated that the CPSII had clearly undergone a actinobacteria or archaebacteria. root is not amongst the eukaryote fusion between its two subunits in As the ancestor of eukaryotes [1] groups ancestrally with two cilia the ancestral eukaryote, as they is likely to be a transient (bikonts [1]). Bikonts share a fused are fused in the trypanosomatids intermediate between these two DHFR-TS gene that is clearly and Sporozoa. In plants (at least groups, this paralogue probably derived [4]. However, the precise angiosperms), this fused enzyme entered eukaryotes either by position of the root of the has been replaced by unfused lateral gene transfer into unikonts, eukaryote tree remained uncertain, paralogues from the cyanobacterial which would make it a unikont because of the unclear status of ancestor of chloroplasts. In synapomorphy, or, more likely, via the Amoebozoa [4]. We have animals, Fungi and Dictyostelium, the ancestral mitochondrial sequenced the region upstream of however, the first three enzymes symbiosis, which would make its the TS gene in the amoebozoan are fused into a multienzyme loss a bikont synapomorphy. Hartmannella cantabrigiensis and protein [7]. No bikonts are known Thus, this unikont gene find that the TS and DHFR genes to have this three-gene fusion, duplication implies that unikonts are separate. As the two genes are which is hence a shared derived or bikonts (or both) are encoded on opposite strands they character that unites opisthokonts holophyletic. If the PFK must be translated separately and and Amoebozoa into a single clade duplication continues to prove thus be unfused (inset in Figure 1). and thus implies that the root of absent from all bikonts with more Also the choanozoan the tree does not lie within them. extensive taxon sampling, we Corallochytrium limacisporum We previously designated this would have two shared derived contains only non-coding DNA clade unikonts [6], because we characters uniting the unikonts, as upstream of the TS gene, but not a inferred that its common ancestor well as two independent shared DHFR gene. Thus, all major was uniciliate and probably derived characters uniting the opisthokont groups lack fused unicentriolar (unikont [1]). bikonts, all together supporting a DHFR-TS genes. Until now, this As the DHFR-TS fusion can be position of the root precisely had only been inferred from the used to exclude the root of the tree between unikonts and bikonts. fact that we were not able to detect from the bikonts and the pyrimidine Our pinpointing the root of the the fusion gene in Corallochytrium, pathway three-gene fusion eukaryote tree between the the fact that Choanozoa are sisters excludes it from the unikonts, the unikonts and bikonts has to animals [5] and from the root of the tree lies precisely numerous important evolutionary absence of the fusion gene in both between these two clades. Strictly implications. It shows, contrary to animals and Fungi. speaking, we cannot yet firmly rule earlier ideas [3], that there are no Given the improbability of out the possibility that Amoebozoa extant eukaryotes that branched reversal of the fusion in the bikont are paraphyletic and that only off from the tree prior to the ancestor [4], this provides the best some of them are sisters of divergence of the ancestors of evidence to date that Amoebozoa unikonts. However, we can already animals and plants. For the first Current Biology Vol 13 No 17 R666 BIKONTS UNIKONTS CABOZOA Kingdom PLANTAE chromalveolates Kingdom Kingdom ANIMALIA FUNGI Excavata Viridaeplantae Rhodophyta Kingdom CHROMISTA (green plants) Choanozoa G core * R Alveolata Rhizaria Glaucophyta opisthokonts ? Retaria EF1- alpha insertion (Foraminifera CHLOROPLAST plastid GAPDH Heliozoa Radiolaria) replacement Amoebozoa ? Cercozoa* carbamoyl phosphate synthase/ Ancyromonadida DHO/ACT polyubiquitin Apusomonadida three-gene fusion insertion Apusozoa TS DHFR-TS fusion gene BamHI BamHI 480 bp 332 bp DHFR MITOCHONDRION ancestral heterotrophic aerobic uniciliate eukaryote Hartmannella ‘ clone 5’ Kingdom BACTERIA Current Biology Figure 1. Eukaryote phylogeny integrating ultrastructure, sequence trees, gene fusions and molecular cladistic markers. The unikont topology is established, but the branching order of the six bikont groups remains uncertain. The single enslavement [12] of a red alga (R) to create chromalveolates is supported by a plastid glyceraldehyde phosphate dehydrogenase (GAPDH) replacement [13]. Whether there was a single enslavement of a green alga (G) to form cabozoa or two separate enslavements (asterisks) to form Cercozoa and Excavata is uncertain [12], as is the position of Heliozoa [14]. Polyubiquitin [15] and EF-1α [16] insertions strongly support the clades core Rhizaria and opisthokonts. The inset shows the BamHI restriction fragment from H. cantabrigiensis that was sequenced and analysed in this study, spanning the DHFR and the amino terminus of the TS gene (red, introns are green). The length of the noncoding regions upstream and downstream of the DHFR gene from one of the clones is indicated. time it enables us to reconstruct 4. Stechmann, A. and Cavalier-Smith, horizontal gene transfer and the phenotype of the last common T. (2002). Rooting the eukaryote tree phospho-donor change in the ancestor (cenancestor) of by using a derived gene fusion. evolution of the Science 297, 89–91. phosphofructokinase. Gene, in eukaryotes [1]. Any homologous 5. Cavalier-Smith, T. and Chao, E.E.-Y. press. character present in two taxa on (2003). Phylogeny of Choanozoa, 12. Cavalier-Smith, T. (2003). Genomic opposite sides of the fundamental Apusozoa and other protozoa and reduction and evolution of novel unikont/bikont divide must have early eukaryote megaevolution. J. genetic membranes and protein- Mol. Evol. 56, 540–563. targeting machinery in eukaryote- been in the cenancestor, in the 6. Stechmann, A. and Cavalier-Smith, eukaryote chimaeras (meta-algae). absence of lateral transfer. T. (2003). Phylogenetic analysis of Philos. Trans. R. Soc. London Ser. B Nevertheless, it is highly desirable eukaryotes using heat-shock 358, 109–134. that other genic characters of protein Hsp90. J. Mol. Evol., in 13. Fast, N.M., Kissinger, J.C., Roos, press. D.S. and Keeling, P.J. (2001). comparable phylogenetic 7. Nara, T., Hashimoto, T. and Aoki, T. Nuclear encoded, plastid-targeted decisiveness are found and used (2000). Evolutionary implications of genes suggest a single common to further test our conclusion. the mosaic pyrimidine-biosynthetic origin for apicomplexan and pathway in eukaryotes. Gene 257, dinoflagellate plastids. Mol. Biol. 209–222. Evol. 18, 418–426. Supplemental Data 8. Gray, M.W., Lang, B.F., Cedergren, 14. Cavalier-Smith, T. and Chao, E.E.-Y. Supplemental data are available at R., Golding, G.B., Lemieux, C., (2003). Molecular phylogeny of http://images.cellpress.com/sup- Sankoff, D., Turmel, M., Brossard, centrohelid Heliozoa, a novel lineage of bikont eukaryotes that mat/supmatin.htm. N., Delage, E., Littlejohn, T.G., et al. (1998). Genome structure and gene arose by