16.5 Brief Comms Mx

brief communications

2. Hinnebusch, B. J., Fischer, E. R. & Schwan, T. G. J. Infect. Dis. with genomic data, such as phylogenetic fast-evolving) positions, and is thus proba- 178, 1406–1415 (1998). trees based on multiple genes3,4. This sug- bly artefactual. The artefactual attraction of 3. Hinnebusch, B. J., Perry, R. D. & Schwan, T. G. Science 273, 8 367–370 (1996). gests that rRNA is rarely transferred and the long branches of Archaea and hyper- 4. Costerton, J. W., Lewandowski, Z., Caldwell, D. E., indicates that species phylogeny can be thermophilic Bacteria, and the high guanine Korber, D. R. & Lappin-Scott, H. M. Annu. Rev. Microbiol. 49, deduced by using this marker. and cytosine content of rRNA in most 711–745 (1995). 9 5. Avery, L. Genetics 133, 897–917 (1993). However, rRNA phylogenies can be Archaea, Aquificales and Thermotogales , 6. Brubaker, R. R. Clin. Microbiol. Rev. 4, 309–324 (1991). seriously affected by artefacts of tree could explain why hyperthermophiles are 7. Jones, H. A., Lillard, J. W. Jr & Perry, R. D. Microbiology 145, construction5. As the most slowly evolving often found to branch early in Bacteria. 2117–2128 (1999). positions are less prone to confounding The emergence of hyperthermophilic 8. Gerke, C., Kraft, A., Sussmuth, R., Schweitzer, O. & Gotz, F. J. Biol. Chem. 273, 18586–18593 (1998). factors such as multiple substitutions, they Bacteria among mesophiles suggests a 9. Heilmann, C. et al. Mol. Microbiol. 20, 1083–1091 (1996). probably retain an ancient phylogenetic secondary adaptation to life at very high 10.Hinnebusch, B. J. J. Mol. Med. 75, 645–652 (1997). signal6. Phylogenies constructed by exam- temperature for these organisms, which Competing financial interests: none declared. ining these positions are less affected by may have been facilitated by massive gene artefacts. We used the ‘slow–fast’ method7 transfer from the Archaea10. The most con- to identify these positions, which has vincing support for this idea is provided by proved efficient in analysing eukaryotic the enzyme reverse gyrase, which is specific Phylogeny phylogeny based on rRNA5. to hyperthermophiles11. This enzyme, which There are several discrepancies between introduces positive supercoils in DNA and A non-hyperthermophilic the standard phylogeny2 and the one we is thought to stabilize it at high tempera- ancestor for Bacteria have deduced using the most conserved tures, has been independently acquired by positions (Fig. 1). Our version shows Aquifex and Thermotoga from Archaea11. he first phyla that emerge in the tree of hyperthermophilic Bacteria (Aquificales Consistent with the ancestral guanine and life based on ribosomal RNA (rRNA) and Thermotogales) to be monophyletic, cytosine content of rRNA9, our results indi- Tsequences are hyperthermophilic, which is consistent with large-scale stud- cate that the most recent common ancestor which led to the hypothesis that the univer- ies3,4, instead of paraphyletic as in classical of Bacteria was not hyperthermophilic. sal ancestor, and possibly the original living versions. More important, our phylogeny We find that Planctomycetales are the organism, was hyperthermophilic1. Here we shows the late emergence of hyperthermo- first branching bacterial group (Fig. 1). reanalyse the bacterial phylogeny based on philes, as they are clustered with Fusobac- This is a significant, although understud- rRNA using a more reliable approach, and teria within a vast multifurcation containing ied, division of Bacteria, whose members find that hyperthermophilic bacteria (such almost all the phyla. share several peculiar traits, such as a bud- as Aquificales and Thermotogales) do not Surprisingly, Planctomycetales emerge at ding mode of reproduction and the lack of emerge first, suggesting that the Bacteria the base of the Bacteria. Whereas the sup- peptidoglycan in their cell walls12. The had a non-hyperthermophilic ancestor. It port for the early branching of hyperther- most intriguing feature of this group is the seems that Planctomycetales, a phylum with mophilic Bacteria is low (bootstrap value existence of a single or double membrane numerous peculiarities, could be the first about 20%), slowly evolving positions (Fig. around the chromosome in Gemmata emerging bacterial group. 1) provide reasonable, albeit inconclusive, sp. and Pirellula sp., respectively, which rRNA is a useful tool for investigating support for the early emergence of Plancto- has been compared to the eukaryotic the universal phylogeny of life, particularly mycetales (bootstrap value about 70%). The nucleus12. However, evolutionary homology for uncultured organisms2. The phylogeny basal position of hyperthermophilic phyla is of these structures with the eukaryotic of Bacteria based on rRNA is congruent found only when examining noisy (that is, nucleus has not been proved. An early emergence of Planctomycetales, as inferred Figure 1 Prokaryotic phylo- from our work, must be confirmed by a Green sulphur Bacteria genetic tree based on conserved careful phylogenetic analysis of genome Thermus/Deinococcus group positions in ribosomal RNA. The sequences from several apparently early- Cyanobacteria ‘slow–fast’ method7 was used branching Bacteria. If our finding is Low-G+C Gram-positive Bacteria to estimate the rate of evolution verified, the origin of Bacteria should be at each position as the sum of seriously reconsidered. High-G+C Gram-positive Bacteria the number of substitutions Céline Brochier, Hervé Philippe Green non-sulphur Bacteria within 19 predefined phyla. Phylogénie, Bioinformatique et Génome, UMR ε-Proteobacteria Alignments were constructed 7622 CNRS, Université Pierre et Marie Curie, δ-Proteobacteria for various thresholds (0–15) 9 quai St Bernard, 75005 Paris, France α-Proteobacteria and the most parsimonious was e-mail : [email protected] inferred using PAUP 4b8 (align- β-Proteobacteria 1. Stetter, K. O. Ciba Found. Symp. 202, 1–10 (1996). ments and trees are available 2. Pace, N. R. Science 276, 734–740 (1997). γ-Proteobacteria from the authors). The relative 3. Brochier, C., Bapteste, E., Moreira, D. & Philippe, H. Fusobacteria positions of Planctomycetales, Trends Genet. 18, 1–5 (2002). 4. Daubin, V. & Gouy, M. Genome Inform. Ser. Workshop Genome Aquificales and Thermotogales Thermotogales Inform. 12, 155–164 (2001). remain very similar for all 5. Philippe, H., Germot, A. & Moreira, D. Curr. Opin. Genet. Dev. Aquificales thresholds from 1 to 10. The 10, 596–601 (2000). Spirochaetes phylogeny shown here is based 6. Felsenstein, J. J. Mol. Evol. 53, 447–555 (2001). Cytophagales/Fusobacteria/ 7. Brinkmann, H. & Philippe, H. Mol. Biol. Evol. 16, 817–825 on the 751 positions with a Bacteroidaceae (1999). Planctomycetales threshold of fewer than 5 sub- 8. Felsenstein, J. Syst. Zool. 27, 401–410 (1978). Crenarchaeota stitutions, which represents 45% 9. Galtier, N., Tourasse, N. & Gouy, M. Science 283, 220–221 of informative positions and is (1999). Euryarchaeota 10.Nelson, K. E. et al. Nature 399, 323–329 (1999). thus a valid compromise bet- 11.Forterre, P., Bouthier de la Tour, C., Philippe, H. & Duguet, M. 1 change per nucleotide ween the quality and the quanti- Trends Genet. 16, 152–154 (2000). ty of phylogenetic information. 12.Fuerst, J. A. Microbiology 141, 1493–1506 (1995).