global eukaryote phylogeny: combined small- and large...

Molecular Phylogenetics and Evolution xxx (2006) xxx–xxxwww.elsevier.com/locate/ympev

ARTICLE IN PRESS

Global eukaryote phylogeny: Combined small- and large-subunit ribosomal DNA trees support monophyly of Rhizaria, Retaria

and Excavata

David Moreira a,¤, Sophie von der Heyden b,1, David Bass b, PuriWcación López-García a, Ema Chao b, Thomas Cavalier-Smith b

a Unité d’Ecologie, Systématique et Evolution, UMR CNRS 8079, Université Paris-Sud, Bâtiment 360, 91405 Orsay Cedex, Franceb Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK

Received 26 July 2006; revised 11 October 2006; accepted 2 November 2006

Abstract

Resolution of the phylogenetic relationships among the major eukaryotic groups is one of the most important problems in evolution-ary biology that is still only partially solved. This task was initially addressed using a single marker, the small-subunit ribosomal DNA(SSU rDNA), although in recent years it has been shown that it does not contain enough phylogenetic information to robustly resolveglobal eukaryotic phylogeny. This has prompted the use of multi-gene analyses, especially in the form of long concatenations of numer-ous conserved protein sequences. However, this approach is severely limited by the small number of taxa for which such a large number ofprotein sequences is available today. We have explored the alternative approach of using only two markers but a large taxonomic sam-pling, by analysing a combination of SSU and large-subunit (LSU) rDNA sequences. This strategy allows also the incorporation ofsequences from non-cultivated protists, e.g., Radiozoa (D radiolaria minus Phaeodarea). We provide the Wrst LSU rRNA sequences forHeliozoa, Apusozoa (both Apusomonadida and Ancyromonadida), Cercozoa and Radiozoa. Our Bayesian and maximum likelihoodanalyses for 91 eukaryotic combined SSU+LSU sequences yielded much stronger support than hitherto for the supergroup Rhizaria(Cercozoa plus Radiozoa plus Foraminifera) and several well-recognised groups and also for other problematic clades, such as theRetaria (Radiozoa plus Foraminifera) and, with more moderate support, the Excavata. Within opisthokonts, the combined tree stronglyconWrms that the Wlose amoebae Nuclearia are sisters to Fungi whereas other Choanozoa are sisters to animals. The position of somebikont taxa, notably Heliozoa and Apusozoa, remains unresolved. However, our combined trees suggest a more deeply diverging positionfor Ancyromonas, and perhaps also Apusomonas, than for other bikonts, suggesting that apusozoan zooXagellates may be central forunderstanding the early evolution of this huge eukaryotic group. Multiple protein sequences will be needed fully to resolve basal bikontphylogeny. Nonetheless, our results suggest that combined SSU+LSU rDNA phylogenies can help to resolve several ambiguous regionsof the eukaryotic tree and identify key taxa for subsequent multi-gene analyses.© 2006 Published by Elsevier Inc.

Keywords: SSU rDNA; LSU rDNA; Eukaryotic phylogeny; Combined analysis; Rhizaria; Retaria; Excavata

1. Introduction

* Corresponding author. Fax: +33 1 6915 4697.E-mail address: [email protected] (D. Moreira).

1 Present address: Department of Botany and Zoology, University ofStellenbosch, Private Bag X1, Matieland 7602, South Africa.

1055-7903/$ - see front matter © 2006 Published by Elsevier Inc.doi:10.1016/j.ympev.2006.11.001

Please cite this article in press as: Moreira, D. et al., Global eukaryotrees support monophyly of Rhizaria, Retaria and Excavata, Mol. P

For nearly two decades, eukaryote phylogeny has beenstudied using SSU rDNA2 as preferred, almost exclusive,

2 Abbreviations used: BP; bootstrap proportion; LSU rDNA; large-sub-unit ribosomal DNA; ML; maximum likelihood; PP; posterior probability;SSU rDNA; small-subunit ribosomal DNA.

te phylogeny: Combined small- and large-subunit ribosomal DNAhylogenet. Evol. (2006), doi:10.1016/j.ympev.2006.11.001

mailto: [email protected]

mailto: [email protected]

2 D. Moreira et al. / Molecular Phylogenetics and Evolution xxx (2006) xxx–xxx

ARTICLE IN PRESS

marker (for a historical overview see Cavalier-Smith, 2002;Philippe and Adoutte, 1998; Roger, 1999). Although someof the earliest eukaryotic molecular phylogenies were con-structed using partial LSU rDNA sequences (Baroin et al.,1988; Perasso et al., 1989), this marker was curiously soonabandoned under the overwhelming impetus given to theuse of SSU rDNA (Sogin, 1991). Nevertheless, in recentyears a number of studies have demonstrated that any indi-vidual marker, including SSU rRNA, has a series of inher-ent problems that can engender artefactual results wheninferring the global phylogeny of eukaryotes. Among themare an insuYcient number of informative sites, mutationalsaturation, and uneven evolutionary rates across sites andtaxa (Philippe and Adoutte, 1998; Philippe et al., 2000;Stiller and Hall, 1999). This has prompted eVorts to apply atotal-evidence approach to the reconstruction of eukaryoticphylogeny by combining diVerent markers into one analy-sis by direct concatenation (Baldauf et al., 2000; Moreiraet al., 2000; Rodríguez-Ezpeleta et al., 2005) or likelihoodsummation (Bapteste et al., 2002). This has provided strongsupport for groups whose monophyly was suspected, suchas Plantae (Moreira et al., 2000; Rodríguez-Ezpeleta et al.,2005) and Conosa (Bapteste et al., 2002), as well as animproved resolution of internal metazoan phylogeny (Phi-lippe et al., 2005).

A major limitation of this approach is the very reducedtaxonomic sampling for which a large collection of genesequences is available, so that many important phyla arenot represented. This problem has been addressed in diVer-ent ways, one of the most eVective being the massivesequencing of expressed sequence tags (EST) from diVerentprotist species (Bapteste et al., 2002; de Koning et al., 2005;Rodríguez-Ezpeleta et al., 2005). Nevertheless, we are stillvery far from a comprehensive representation of eukaryoticdiversity. In addition, this approach is diYcult to apply toorganisms that cannot be maintained in pure, or nearlypure, culture or that resist the extraction of high qualitymRNA. Finally, the computational burden of analysinghuge multi-gene data sets increases substantially with theaddition of more taxa.

The diYcult trade-oV between the merits of a largeamount of sequence information for each taxon and a widetaxonomic sampling is crucial to the quest for a robusteukaryotic phylogeny. From a theoretical point of view, analternative to the analysis of increasingly longer sequencesfor a relatively limited taxon sampling is the use of a largertaxonomic diversity, even if the number of sites remains rel-atively small. In fact, several simulation studies have shownthat in many cases the addition of taxa can be more beneW-cial than the addition of sites to increase the accuracy of thephylogenetic tree, even for complex cases prone to beaVected by artefacts such as long-branch attraction (LBA)(Graybeal, 1998; Hillis, 1996). For the historical reasonmentioned above and its easy retrieval (by PCR and directsequencing), SSU rDNA is the marker with the richest tax-onomic representation available today. It is therefore agood candidate to take part in a combination of few genes


for a wide taxon collection. We have decided to associate itwith a second marker that has been relatively littleexplored, the LSU rDNA, considering that both markersmay reasonably be accommodated within a single model ofsequence evolution for a combined analysis. LSU rDNAoVers the additional advantage of being generally adjacentto SSU rDNA in the genome so that both markers can beeasily retrieved from uncultured protists from metagenomicanalyses (López-García et al., 2002).

With that aim, we have signiWcantly improved thetaxonomic sampling of LSU rDNA, providing 14new sequences, enabling us to include six key eukaryoticgroups on LSU trees for the Wrst time: Cercozoa, Radiozoa(D classical Radiolaria minus Phaeodarea, which appearto be Cercozoa: Polet et al., 2004; Yuasa et al., 2006), cen-trohelid Heliozoa and an unknown micro-heliozoan,Apusomonadida, Ancyromonadida and Jakobea. Maxi-mum likelihood (ML) and, especially, Bayesian analyses ofSSU+LSU rDNA data sets yielded phylogenetic trees withmuch better resolution for a wide diversity of eukaryoticgroups than in the corresponding separate SSU and LSUrDNA analyses. These trees provide support for severalgroups (e.g., Conosa, Excavata) that is comparable to thatproduced by very large amino acid data sets (Bapteste et al.,2002; Hampl et al., 2005). We also retrieved intriguingresults, in particular strong support for the monophyly ofthe Retaria (Radiozoa and Foraminifera) and indicationsof a deep divergence of Ancyromonas (Apusozoa) fromother bikonts. These results suggest that the combination ofboth rDNA sequences may provide a valuable marker forreconstructing global eukaryote phylogeny, with the advan-tage of the usually easy acquisition of new LSU rDNAsequences and the already extremely rich SSU rDNA database. Combined SSU+LSU rDNA trees can be particularlyuseful for identifying key taxa for subsequent EST projectsand multi-gene analyses.

2. Materials and methods

2.1. Protist cultures and DNA extraction

We used culture conditions already described to growthe two centrohelid heliozoa Sphaerastrum fockii, Pterocys-tis sp. (von der Heyden, 2004) and an unknown marine“microheliozoan” (Cavalier-Smith and Chao, 2003a), thetwo cryptomonads Cryptomonas (DChilomonas) parame-cium and Goniomonas sp. ATCC 50108 (von der Heydenet al., 2004), the four cercozoans Paracercomonas marinaCavalier-Smith and Bass ATCC 50344 (formerly misidenti-Wed as Cercomonas longicauda: see Karpov et al., 2006),Cercomonas sp. strain C-59, Dimorpha-like sp. ATCC 50522and Thaumatomastix sp. ATCC 50250 (Bass and Cavalier-Smith, 2004; Cavalier-Smith and Chao, 2003b), and the twoapusozoans Apusomonas proboscidea and Ancyromonas‘sigmoides’ ATCC 50267 (Cavalier-Smith and Chao,2003c). Cells were collected and DNA puriWed as describedpreviously (Cavalier-Smith and Chao, 1995).


D. Moreira et al. / Molecular Phylogenetics and Evolution xxx (2006) xxx–xxx 3

ARTICLE IN PRESS

2.2. Environmental samples and fosmid libraries

A metagenomic library of large DNA fragments(»35 kbp) was constructed from sea plankton collectedfrom mesopelagic (500 m) Antarctic water, after Wltrationto retain only the 0.2–5 �m biomass fraction. Details on thislibrary and the screening for fosmid clones with SSUrDNA genes are in López-García et al. (2002). The twoclones (KW16 and HA2) containing SSU rDNA genesrelated to radiolaria were examined also for the presence ofLSU rDNA genes (see below).

2.3. LSU rDNA ampliWcation and sequencing

For both the protist DNA and the environmentalfosmid clones, the LSU rDNA genes were ampliWed byPCR using speciWc primers 28S-1F (ACCCGCTGAATTTAAGCAT) and 28S-4R (TTCTGACTTAGAGGCGTTCAG). PCRs were carried out under the following condi-tions: 30 cycles (denaturation at 94 °C for 15 s, annealing at55 °C for 30 s, extension at 72 °C for 2 min) preceded by2 min denaturation at 94 °C, and followed by 10 min exten-sion at 72 °C. The amplicons were directly sequenced usingthe same primers and a set of internal primers, including28S-568F (TTGAAACACGGACCAAGGAG), 28S-2F(GCAGATCTTGGTGGTAG), 28S-1611R (CTTGGASACCTGMTGCGG), and 28S-3R (CACCTTGGAGACCTGCT). All sequences have been submitted to GenBankunder Accession Nos. AY752987–AY752993 andDQ386164–DQ386169.

2.4. Reclinomonas americana LSU rDNA sequence assembly

A set of partial LSU rDNA sequences (»400–800 bp) ofthe jakobid R. americana was generously provided by Drs.B. Franz Lang, Gertraud Burger and Michael W. Gray.These sequences were obtained through the Protist ESTProgram (PEP: http://megasun.bch.umontreal.ca/pepdb/pep_main.html). The partial sequences were assembled intoa nearly complete LSU rDNA sequence manually.

2.5. Data set construction

The 14 new sequences were added to a data set of LSUrDNA sequences (»350) retrieved from GenBank (http://ncbi.nlm.nih.gov/) and aligned using the ED program of thepackage MUST (Philippe, 1993). A SSU+LSU rDNA dataset containing 110 taxa was constructed by concatenationof the aligned SSU and LSU rDNA sequences from eachspecies. When both markers were not available for the samespecies, we concatenated the sequences from two closelyrelated species (see Supplementary material Table 1).

2.6. Phylogenetic analysis

Preliminary phylogenetic analyses of the individual LSUand concatenated SSU+LSU rDNA data sets including all


available sequences were done by neighbour-joining (NJ)using MUST (Philippe, 1993). This helped to identify theslowest evolving species (those with the shortest branches)within each taxonomic group, which were subsequentlyanalysed in more detail using maximum likelihood (ML)and Bayesian methods. For some groups (in particularConosa and Excavata), most species showed long branches,so that we were constrained to include them in our analyses.The conserved positions of the alignment used for phyloge-netic reconstruction were chosen using GBLOCKS (Castre-sana, 2000). For both the ML and Bayesian methods weapplied a general time reversible (GTR) model of nucleo-tide substitution, with a six-category discrete approxima-tion of a � distribution plus invariable sites (GTR+�+Imodel), which was selected using MODELTEST (Posadaand Crandall, 1998) as the best of 56 testable substitutionmodels. ML trees were reconstructed with PHYML v. 2.4.4(Guindon and Gascuel, 2003) and bootstrapped using 1000replicates. Bayesian trees were inferred using MrBAYES 3(Ronquist and Huelsenbeck, 2003). All the Markov chainMonte Carlo searches were run with 4 chains for 1,000,000generations, with trees being sampled every 100 genera-tions. The Wrst 2000 trees were discarded as “burnin”, keep-ing only trees generated well after those parametersstabilised. Stabilisation of the chain parameters (tree likeli-hood, � shape parameter and proportion of invariant sites)was studied with the program TRACER (Rambaut andDrummond, 2003). Additional Bayesian trees were alsoreconstructed using other model speciWcations: GTR+�+Iwith allowance for covarions, GTR+�+I with intersite rateautocorrelation, and GTR+�+I with covarions and inter-site rate autocorrelation. For each data set and model spec-iWcation three independent runs were carried out and thetree topologies obtained were compared to check whetherthey were congruent or not using the Approximately Unbi-ased (AU) test (Shimodaira, 2002) implemented in the pro-gram CONSEL (Shimodaira and Hasegawa, 2001).Sequence alignments are available upon request.

3. Results and discussion

3.1. The eukaryotic LSU rDNA tree

We have substantially improved the broad-scale taxo-nomic sampling of the LSU rDNA with 14 new sequences,especially for groups of uncertain phylogenetic aYnity:Cryptomonada, Centrohelida, Ancyromonadida, Apuso-monadida, Cercozoa and Radiozoa. Bayesian analysis of aLSU rDNA data set (91 taxa, 1513 sites) yields a tree thatsupports with strong posterior probability (PPD1) themonophyly of major clades such as opisthokonts, Conosaand Rhizaria (Fig. 1). In general, support for the diVerentgroups is stronger in the LSU rDNA tree than in the corre-sponding SSU rDNA tree (see Supplementary materialFig. 1). This is particularly noticeable for Conosa (PPD1vs. 0.63) and Heterokonta (PPD 1 vs. 0.72), whereas othergroups, such as Excavata, are polyphyletic in the SSU


http://megasun.bch.umontreal.ca/pepdb/pep_main.html



http://ncbi.nlm.nih.gov/




ARTICLE IN PRESS


Fig. 1. Unrooted Bayesian phylogenetic tree constructed with an LSU rDNA data set (91 taxa, 1513 sites; GTR+�+I plus covarion model). Newsequences obtained in this work are highlighted in bold. Several long branches were shortened to one fourth of their actual length (labelled 1/4). Numbersat nodes are posterior probabilities and ML bootstrap proportions (only those higher than 0.50 and 50%, respectively, are shown). Other evidence (Rich-ards and Cavalier-Smith, 2005) indicates that the root of the eukaryote tree lies between unikonts (opisthokonts plus Amoebozoa) and bikonts (all othereukaryotes, including kingdoms Plantae and Chromista). However, as unikonts do not appear monophyletic on this tree or Fig. 2, both are arbitrarilydrawn as if rooted between opisthokonts and other eukaryotes.



ARTICLE IN PRESS

rDNA tree but monophyletic in the LSU tree. We do notWnd any case where SSU rDNA provides much strongersupport than LSU rDNA for a given node. In all our trees,posterior probabilities are higher than the correspondingML bootstrap proportions (ML BP), in agreement with thegeneral trend observed in the comparisons between thesetwo measures of statistical support (Douady et al., 2003;Erixon et al., 2003; Suzuki et al., 2002).

The SSU and LSU rDNA trees diVer in the position ofseveral taxa, but all of these concern weakly supportednodes (PP < 0.80). In fact, an approximately unbiased (AU)test (Shimodaira, 2002) shows that the two tree topologiesare not signiWcantly diVerent (PD 0.78). To maximise thephylogenetic information, we therefore also analysed a con-catenation of both markers. We considered that both couldbe accommodated in a single concatenated analysis withthe same substitution model, GTR+�+I, since the twoexhibit similar values for several model parameters: similarnucleotide frequencies, � parameter values of the � distri-bution of 0.53 and 0.54, and proportion of invariant sites of0.13 and 0.16 for the SSU and LSU rDNA, respectively.That homogeneity could help circumvent problemsencountered when markers with signiWcantly diVerent sub-stitution patterns are analysed together (Bapteste et al.,2002; Pupko et al., 2002).

3.2. The eukaryotic SSU+LSU rDNA tree

Concatenation of SSU and LSU rDNA sequencesallows the use of over 2500 well conserved sites for phyloge-netic reconstruction, excluding all columns that contained agap in more than one taxon (the exact number depends onthe set of species analysed). For a few taxa (Didymium sp.,rotaliid foraminifer, Fucus sp., Mallomonas sp., Goniomonassp. and Spathidium sp.), we concatenated SSU and LSUrDNA sequences from two closely related species becauseboth gene sequences were not available for exactly the samespecies. To study the impact of the taxonomic sampling, weconstructed phylogenetic trees using a variety of speciescombinations. Fig. 2 shows an unrooted Bayesian tree withthe largest representation of eukaryotic phyla available forthese markers (91 taxa, 2574 sites). Several deep nodes ofthis SSU+LSU rDNA tree show only moderate support,which indicates that the concatenation of the two markersis not enough to obtain a global eukaryotic phylogeny witha good resolution of all nodes. Nevertheless, as in the indi-vidual LSU rDNA tree, various eukaryotic supergroups aresupported by strong posterior probabilities (PPD 1), in par-ticular opisthokonts, Conosa and Rhizaria.

The internal phylogeny of opisthokonts shows severalinteresting relationships. The nucleariid Wlose amoebaNuclearia simplex branches robustly as sister taxon to theFungi (PPD 1, ML BPD95). This result conWrms, with alarger taxonomic sampling, a previous analysis of opistho-kont phylogeny also based on concatenated SSU+LSUrDNA sequences (Medina et al., 2003) and refuted a puta-tive aYnity with choanoXagellates based on 18S rRNA


alone (Cavalier-Smith, 2000). The sisterhood of fungi andnucleariid amoebae also agrees with a recent multi-proteinsequence analysis (elongation factor 1�, actin, �-tubulinand heat shock protein 70) of opisthokont internal relation-ships (Steenkamp et al., 2005). Our phylogenetic trees alsoagree with this multi-protein analysis for the position ofIchthyosporea, which emerge within a clade also includingMetazoa (DAnimalia) and ChoanoXagellatea (Steenkampet al., 2005). This position is also retrieved by phylogeneticanalyses of a data set of concatenated proteins encoded bymitochondrial DNA (Lang et al., 2002). In our concate-nated SSU+LSU rDNA trees Ichthyosporea (representedby Ichthyophonus hoferi) branch robustly (PPD 1, MLBPD99) at that position (Fig. 2). This relationship is alsoretrieved using ML tree reconstruction, and it is notaVected by the use of diVerent combinations of metazoansequences, nor even by the inclusion of the fast evolvingsequences characteristic of bilaterian animals (not shown).

A second species frequently classiWed as an ichthyospo-rean, Capsaspora owczarzaki (Hertel et al., 2002), branchesstrongly (PPD 1, ML BPD 86) as sister to choanoXagellatesin our concatenated analysis. This agrees with a recentanalysis of combined SSU+LSU rDNA+actin sequences(Ruiz-Trillo et al., 2004), although in that analysis the cladecomprising C. owczarzaki, one ichthyosporean and onechoanoXagellate was sister to animals only when thesequence evolution model took covarions into account. Asmentioned above, our analyses retrieve a more stable topol-ogy, with C. owczarzaki always grouping with choanoXa-gellates in a clade sister to animals independently of thesequence evolution model applied. This increased stabilitycould be due to the inclusion of a second choanoXagellatespecies in our analyses. Moreover, AU tests carried outwith the SSU+LSU data set show that topologies where theIchthyosporea, C. owczarzaki and the ChoanoXagellateaare constrained to branch as sisters of N. simplex + Fungiare signiWcantly worse (P < 0.05). The aYnity of C. owczar-zaki with choanoXagellates appears to be supported mostlyby the LSU rDNA since, in individual analyses (Fig. 1), thismarker supports this relationship strongly (PPD0.98, MLBPD87), whereas the SSU rDNA yields trees where C.owczarzaki is sister of I. hoferi with weaker support(PPD 0.73, ML BPD 54; Supplementary material Fig. 1).Increasing the sampling of LSU rDNA sequences for othernucleariid and ichthyosporean species, as well as for otheropisthokonts of uncertain aYnity, such as ministeriids, cor-allochytreans and Eccrinales (Cafaro, 2005; Cavalier-Smithand Chao, 2003c; Steenkamp et al., 2005), will help test thispossible reorganisation of opisthokont phylogeny.

Conosa, one of the two subphyla of phylum Amoebozoa(Cavalier-Smith, 2004), is another group well supported bythe SSU+LSU rDNA analyses (Fig. 2). The two conosan sub-groups Archamoebae (represented by species of the generaPhreatamoeba and Entamoeba) and Mycetozoa (Dictyoste-lium, Didymium and Physarum) emerge in all Bayesian trees assisters with strong support (PPD1), which is weaker for theML analyses (ML BPD65). Inspection of the bootstrapped



ARTICLE IN PRESS

ML trees reveals that this is due to the branching of the fora- results obtained by analysing the genes separately. One case
miniferan sequence within the Conosa in a proportion ofthese trees (not shown). Good support is also found for thesisterhood of Alveolata and Heterokonta (Fig. 2), both by theBayesian (PPD0.97) and ML analyses (BPD81).
In contrast, for a number of groups the concatenatedSSU+LSU rDNA data set did not improve signiWcantly the


concerned Plantae, which encompass the three primaryphotosynthetic eukaryotic groups: Glaucophyta, Rhodo-phyta and Viridaeplantae (Moreira et al., 2000; Rodríguez-Ezpeleta et al., 2005). Plantae emerge mixed with somechromistan groups in our trees; thus neither Plantae norChromista is recovered as holophyletic. One possible

Fig. 2. Unrooted Bayesian phylogenetic tree constructed with a SSU+LSU rDNA data set (91 taxa, 2574 sites; GTR+�+I plus covarion model). Taxa forwhich the SSU and LSU rDNA sequences belonged to diVerent but related species are labelled as ‘composite’ (see Supplementary Material Table 1 fordetails). Several long branches were shortened to one fourth of their actual length (labelled 1/4). Triangles indicate the alternative positions for the speciesA. sigmoides that were examined using the AU test. Numbers at nodes are posterior probabilities and ML bootstrap proportions (only those higher than0.50 and 50%, respectively, are shown).



ARTICLE IN PRESS

source of problems could be the biased nucleotide composi-tion of the rhodophyte sequences, as shown by a chi-squaretest (Schmidt et al., 2002), and their faster evolutionary rate(longer branches) compared with several of the glauco-phytes and green plants. The probably spurious attractionbetween glaucophytes and cryptophytes, which has alreadybeen noticed in several SSU rDNA analyses (e.g., Cavalier-Smith and Chao, 2003a; Nikolaev et al., 2004), appears alsoto be important in hampering the inference of monophy-letic Plantae and Chromista. This attraction seems mostlydue to the SSU part of the SSU+LSU rDNA concatenation(see Supplementary material Fig. 1), since the LSU rDNAalone did not support any robust location for these twogroups (Fig. 1).

We observe a similarly inconstant position for centroh-elid heliozoa, which very often branch with weak supportclose to cryptophytes and glaucophytes (Fig. 2) or as anindependent lineage. Even the monophyly of Heliozoa (cen-trohelids plus the microheliozoan) remains uncertain, sincethe undetermined marine “microheliozoan” (which is prob-ably not a centrohelid) branches far from the two centrohe-lids (Pterocystis sp. and Sphaerastrum fockii) in manysingle-gene trees (Supplementary material Fig. 2), despitebeing consistently together on the concatenated trees. Itappears that the two heliozoan branches group togetherweakly on 18S rRNA trees when long branch taxa areexcluded and more nucleotides used (Cavalier-Smith andChao, 2003c; Nikolaev et al., 2004), but are separate, withthe microheliozoan weakly grouping instead with red algae,when long branch taxa and fewer nucleotides are included(Supplementary Fig. 1 and Cavalier-Smith, 2004). Theirgrouping together as a weakly supported heliozoan cladeon the concatenated trees that include long branch taxa(Fig. 2), and with distinctly stronger support when longbranch taxa are excluded and over 900 more nucleotidesused (Supplementary Fig. 2), suggests that the relationshipis real, but additional gene sequences are required to dem-onstrate it convincingly. Neither our combined SSU+LSUrDNA analyses nor the Wrst protein phylogenies includingcentrohelid sequences (Nikolaev et al., 2004; Sakaguchiet al., 2005) have resolved the phylogenetic position of thisgroup within bikonts.

Albeit with only moderate support (PPD0.72, MLBPD67), our combined tree did reproducibly recover thesister relationship between Rhodophyta and Viridaeplan-tae, which is only occasionally found on 18S rRNA treesand was absent from a recent tree using 143 nuclear pro-teins (Rodríguez-Ezpeleta et al., 2005), probably because ofthe very long rhodophyte branch, but is clearly supportedby a shared gene replacement of fructose bis-phosphatealdolase (Patron et al., 2004).

We also used the SSU+LSU rDNA data set to investi-gate the phylogeny of Excavata, represented in our analysesby jakobid, diplomonad, Trichomonas and euglenozoansequences. This was a diYcult analysis because the longbranches of most excavates were prone to be attracted bythose of Amoebozoa. Data sets excluding Amoebozoa sup-


port the monophyly of Excavata with strong PP (1) butweak ML BP (56) (not shown). When Amoebozoa areincluded, excavates still emerge as a monophyletic groupthough with weaker support (PPD0.87, ML BPD 17)(Fig. 2). This suggests that the SSU+LSU rDNA data setcontains just enough phylogenetic information to supportthe monophyly of excavates even if other fast evolvinggroups are included. However, only the Bayesian analysis isable to retrieve this result, whereas the ML phylogenetictrees yield a mix of all the long branching amoebozoa andexcavate sequences within an unsupported (ML BPD 28)group (not shown). Within excavates, our analyses stronglysupport (PPD1, ML BPD85) the monophyly of the phy-lum Metamonada (Cavalier-Smith, 2003a), represented byseveral diplomonads and the parabasalian Trichomonasvaginalis (Fig. 2). This agrees with a recent multi-gene (sixnuclear protein-coding genes) phylogeny, which also con-Wrms the inclusion of Carpediemonas within Metamonada(Simpson et al., 2006). In our SSU+LSU rDNA trees met-amonads branch weakly (PPD 0.64, ML BPD 32) as sistersof Euglenozoa, and the jakobid Reclinomonas americanaoccupies a sister position to all the other excavates,whereas, in that multi-gene analysis, Euglenozoa were sis-ters of Reclinomonas and Heterolobosea (Naegleria) withquite robust support (ML BPD85). Interestingly, the sup-port in our analyses for the monophyly of all Excavata iscomparable to that found in another recent multi-markeranalysis (SSU rDNA + 8 protein genes) which, in contrastto our data sets, did not include any other long-branchingtaxa that might interfere with the long branches of the exca-vate species (Hampl et al., 2005).

3.3. The monophyly of Rhizaria and the position of Foraminifera: evidence for the Retaria?

The determination of new LSU rDNA sequences fromseveral cercozoan species and two environmental clonesrelated to the radiozoan groups Taxopodida and Polycysti-nea allows testing the reliability of grouping these clades,together with Foraminifera, within the infrakingdomRhizaria (Cavalier-Smith, 2002; Nikolaev et al., 2004). Inthe recent past, SSU rDNA analyses have provided contra-dictory results and variable support for this supergroup,especially because of the extremely fast evolutionary rate ofthe foraminiferan sequences that induced LBA artefacts.Thus, some ML analyses (which did not include radiolariansequences) were able to retrieve a relationship betweenForaminifera and Cercozoa only when other fast-evolvingtaxa were removed (Berney and Pawlowski, 2003). Never-theless, more recent analyses (both including radiolariansequences) did retrieve the monophyly of Rhizaria, even inthe presence of several fast-evolving groups, but with onlymoderate (ML BPD72; Nikolaev et al., 2004) or very low(distance ME BPD20: Cavalier-Smith and Chao, 2003c)bootstrap values.

Our Bayesian global eukaryotic phylogenetic treesconstructed using both the LSU and SSU+LSU rDNAs



ARTICLE IN PRESS

provide much stronger support than hitherto for theholophyly of Rhizaria (PP 1). They also support(PPD0.92–0.97) a clade comprising Foraminifera andRadiozoa (represented in our analyses by environmentalpolycystinean- and taxopodid-related sequences) within theRhizaria (Figs. 1 and 2). The monophyly of Rhizaria andthe Foraminifera/Radiozoa grouping are also both reason-ably supported by our SSU+LSU rDNA ML tree (78% and75%). This Foraminifera/Radiozoa clade contradicts someprevious single-gene trees that suggested that the closestrelatives of Foraminifera might be species of the endomy-xan cercozoan genus Gromia, which are marine Wlosetestate amoebae. Both SSU rDNA and RNA polymeraseII phylogenies retrieved a relationship with Gromia,although with only moderate bootstrap support (Berneyand Pawlowski, 2003; Longet et al., 2003). However,radiozoan sequences were not included in those analyses;when Radiozoa were included, Foraminifera grouped withpolycystines in SSU rRNA distance analyses with strong(Cavalier-Smith and Chao, 2003b) or weak support(Cavalier-Smith and Chao, 2003c). A more recent workincorporating several radiozoan sequences in a supertreeanalysis combining SSU rDNA and actin phylogenies(Nikolaev et al., 2004) also retrieved sisterhood of Forami-nifera and Gromia. However, that was essentially due to theresults of the SSU rDNA analysis, which were incompatiblewith those of the actin phylogeny because Foraminiferashowed two deeply divergent actin paralogues, one relatedto Gromia and the other to the polycystine Radiolaria.

Another source of information was provided by theextremely well conserved protein polyubiquitin. One- ortwo-amino acid insertions have been found in theubiquitin monomer-monomer junctions of the poly-ubiquitin sequences of Cercozoa (including Gromia) andForaminifera (Archibald et al., 2003; Bass et al., 2005).These insertions appear to be absent in the rest ofeukaryotes studied, including the polycystine Radiozoaand the Acantharea (Bass et al., 2005). Given theextraordinary degree of evolutionary conservation ofpolyubiquitin, the insertions shared by Foraminifera andCercozoa were considered a potentially good characterfor the sisterhood of these two groups to the exclusion ofRadiozoa (Bass et al., 2005). However, the case of Phae-odarea shows that these insertions may be not as stableas previously thought. Although Phaeodarea were con-sidered for a long time to be members of the Radiolaria(Haeckel, 1887), recent phylogenetic analyses based onSSU rDNA sequence comparison have robustly shownthat Phaeodarea belong in Cercozoa within the subphy-lum Filosa (Polet et al., 2004; Yuasa et al., 2006). Never-theless, their polyubiquitin sequences lack the insertionsbetween the ubiquitin monomers found in other Cerco-zoa (Bass et al., 2005). If Phaeodarea really branch withinCercozoa, they have lost the insertion, which could there-fore also be the case for Radiozoa, making this characterunreliable for excluding their possible closer relationshipto Foraminifera.


We have studied this problem using a SSU+LSU rDNAdata set that included a partial 5�-end LSU rDNA sequence(1507 bp) of Gromia oviformis (Pawlowski et al., 1994).Addition of this sequence to our concatenated SSU+LSUrDNA data set reduced the number of conserved sites use-ful for phylogenetic inference (to a Wnal number of »2200,depending on the set of species retained), but the concate-nation was still signiWcantly larger than the separated SSUand LSU rDNA data sets. As in the global eukaryotic phy-logeny (Fig. 2), a Bayesian phylogenetic tree of all rhizariansequences plus several slow evolving outgroup speciesretrieves with strong statistical support (PPD1, MLBPD 100) the monophyly of Foraminifera plus Radiozoa(Fig. 3A). G. oviformis emerges with equal support at thebase of the other cercozoan sequences. The same result isobtained when two additional partial foraminiferanSSU+LSU rDNA sequences are included, despite an evensmaller number of usable sites (2077; Fig. 3B). DiVerentcombinations of outgroup sequences yield the same results(not shown). However, if Radiozoa are removed from thedata sets, Foraminifera turn out to branch with moderatesupport (PPD 0.87, ML BPD78) as sisters of G. oviformis(Fig. 3C). Overall these results favour Foraminifera beingmore closely related to Radiozoa than to Gromia, and sug-gest that the placement of Foraminifera within the end-omyxan Cercozoa, seen in the absence of radiozoansequences, is artefactual.

A speciWc relationship between Foraminifera andRadiozoa was already advanced by Cavalier-Smith(1999), who erected an infrakingdom Retaria to accom-modate these two groups (Cavalier-Smith, 1999); laterRetaria was reduced in rank to phylum (Cavalier-Smith,2002), making Radiozoa and Foraminifera subphyla(Cavalier-Smith, 2003a). However, although our resultssupport this, the conXicting evidence in the literaturewith diVerent taxon samples makes it desirable to verifythe monophyly of Retaria using sequences of several con-served proteins from diVerent members of the Rhizaria,and to determine whether Radiozoa are sisters of Foram-inifera (Fig. 3A) or ancestral to them (Fig. 3B). This rela-tionship should also be tested by new LSU rRNAsequences from other deeply branching cercozoangroups, such as Haplosporidia and Phytomyxea (Niko-laev et al., 2004; Reece et al., 2004). If retarian mono-phyly were conWrmed, this would imply that the aminoacid insertion in the polyubiquitin sequence was reversedin Radiozoa. However, since all available radiolarianpolyubiquitin sequences have been obtained from speci-mens directly collected from seawater samples, it cannotbe excluded that they actually belonged to undetectedcontaminant organisms (Bass et al., 2005). The possibilitythat the root of the Rhizaria lies between the monophy-letic phyla Retaria and Cercozoa (with Gromia a deeplybranching member of the endomyxan Cercozoa) suggeststhat the ancestral rhizarian might even have beenendowed with an organic and/or mineralised test in itstrophic phase.



ARTICLE IN PRESS

3.4. The positions of Ancyromonadida and Apusomonadida group together (Cavalier-Smith, 2003a,b; Nikolaev et al.,
(Apusozoa) remain unresolved
All our SSU+LSU rDNA analyses place the zooXagel-late Ancyromonas with the opisthokonts and Conosa, oftenbranching weakly between these two groups (Fig. 2). Thisagrees with the Wrst SSU rDNA phylogenies reconstructedfor this organism (Atkins et al., 2000; Cavalier-Smith,2000). Support for such separation of Ancyromonas fromother bikonts was strong in the Bayesian analyses (PPD1)but weaker in the ML trees (ML BPD 71). The latterappears to be mainly due to the attraction of Ancyromonastowards Apusomonas proboscidea observed in several MLtrees, also discussed by Atkins et al. (2000). Cavalier-Smith(1998a,b) placed Ancyromonas (Ancyromonadida: Cava-lier-Smith, 1998a) and Apusomonadida (Apusomonas plusAmastigomonas) in Apusozoa, later grouping both orderswith Hemimastigida (e.g., Hemimastix, Spironema, forwhich sequence data are not yet available) as the classThecomonadea within a more narrowly deWned phylumApusozoa, which now also includes the class Diphyllatea(Cavalier-Smith, 2000, 2003a,b). Although on most earlySSU rDNA trees Ancyromonas and Apusomonadida wereseparate (but often close) deep branches among the bikonts(Cavalier-Smith, 2000, 2002, 2004), on some well sampledgamma-corrected ML and distance trees they do weakly


2006). It has therefore been unclear whether Thecomona-dea and Apusozoa are monophyletic or not (Atkins et al.,2000; Cavalier-Smith and Chao, 2003a). To test that furtherwe analysed an additional sequence data set excluding alllong-branch species, which allowed maximising the numberof conserved positions useful for phylogenetic reconstruc-tion (905 additional sites for a total of 3479). Bayesian anal-yses of this data set also support strongly (PPD 1) theproximity of Ancyromonas to the opisthokonts and no rela-tionship with A. proboscidea (Supplementary materialFig. 2). In our Bayesian trees, A. proboscidea branches withmoderate support (PP 70.87) within a group containingRhizaria, Alveolata and Heterokonta (Fig. 2 and Supple-mentary Fig. 2). In contrast, the corresponding ML trees(not shown) join Ancyromonas and A. proboscidea in aweak monophyletic group (ML BP < 53) sister to the opi-sthokonts. This leaves open the possibility that these twoXagellates are indeed related (Cavalier-Smith, 1998a,b,2000, 2003a,b), but additional sequence data are necessaryto solve this question. In fact, an AU test carried out on thewidest SSU+LSU rDNA data set (91 taxa, Fig. 2) does notreject the possibility that these two species form a mono-phyletic group at the base of the bikonts (PD0.086). Wehave also tested topologies where Ancyromonas alone, orwith A. proboscidea, branches as sister to Conosa,

Fig. 3. Bayesian phylogenetic trees of the Rhizaria constructed with SSU+LSU rDNA data sets: (A) 9 rhizarian taxa, 2189 sites; (B) 11 rhizarian taxa,2077 sites; (C) 7 rhizarian taxa, 2186 sites. GTR+�+I plus covarion model. Taxa for which the SSU and LSU rDNA sequences belonged to diVerent butrelated species are labelled as ‘composite’ (see Supplementary Material Table 1 for details). Several long branches were shortened to one sixth of theiractual length (labelled 1/6). Numbers at nodes are posterior probabilities and ML bootstrap proportions. All the trees were rooted using nine slowly evolv-ing sequences (Cyanophora paradoxa, Gnetum gnemon, Saccharomyces cerevisiae, Suberites Wcus, Prymnesium patelliferum, Perkinsus andrewsi, Prorocen-trum micans, Skeletonema pseudocostatum and Pelagomonas calceolata) as outgroup.



ARTICLE IN PRESS

Excavata, Viridaeplantae, Cryptomonada, Haptophyta,Heliozoa, Glaucophyta, Rhizaria, Rhodophyta, Hetero-konta or Alveolata. All of them are rejected (P < 0.05), sup-porting the deep divergence of Ancyromonas (alone or withApusomonas) from all these bikonts in our trees.

This tendency of Ancyromonas to group with opi-sthokonts and conosans is intriguing (see also Nikolaevet al. (2006) where Thecomonadea as a whole weaklyoccupy this position on the SSU rRNA tree, but otherputative Apusozoa—Diphylleia and Breviata—groupweakly with the excavates). Several lines of evidence suggestthat Amoebozoa are the closest relatives of opisthokonts(Richards and Cavalier-Smith, 2005; Stechmann and Cava-lier-Smith, 2003). Although we do not retrieve this result inour SSU+LSU rDNA trees, because Ancyromonas appearseven closer to opisthokonts, Amoebozoa are the secondclosest group to opisthokonts, and statistical support forthe relative positions of Ancyromonas and Amoebozoa withrespect to opisthokonts is weak (PPD0.73, ML BP < 54).Thus our analyses are compatible with the very probablesisterhood of Amoebozoa and opisthokonts. If we acceptthat the root of the eukaryotic tree is betweenopisthokonts + Amoebozoa (i.e., unikonts) and the rest ofthe phyla (bikonts), one could argue that the aYnity ofAncyromonas (or Ancyromonas plus apusomonads: Niko-laev et al., 2006) towards the opisthokonts and conosansmight arise because this lineage is really sister to all otherbikonts. Because of its very deep branching near the uni-kont-bikont divergence, Ancyromonas may be a veryimportant model for the study of early evolution of eukary-otes as a whole or bikonts in particular. We stress that ithas not yet been established whether Ancyromonas pos-sesses the DHFR-TS fusion gene characteristic of bikonts(Richards and Cavalier-Smith, 2005; Stechmann and Cava-lier-Smith, 2003) or the myosin II gene characteristic of uni-konts (Richards and Cavalier-Smith, 2005). We are activelypursuing this question to clarify its position by these molec-ular cladistic markers. Also important for clarifying thebasal branching of bikonts and the evolutionary aYnitiesand proper circumscription of Apusozoa will be sequencesfor other Ancyromonas species and for Micronuclearia, anon-Xagellate protozoan with cortical and mitochondrialultrastructural resemblances to it (Mikrjukov and Mylni-kov, 2001), and LSU rRNA sequences for Breviata, whichSSU rRNA trees weakly group with apusomonads orAncyromonas, or with Diphylleia and/or loukozoan exca-vates (Cavalier-Smith, 2003a, 2004; Cavalier-Smith andChao, 2003c; Nikolaev et al., 2006; Walker et al., 2006).

4. Conclusions

For several decades, inference of global eukaryotic phy-logeny has been mainly based on single-marker analysesthat, in most cases, have proved to be aVected by a numberof tree reconstruction artefacts (Philippe and Adoutte,1998). This promoted the preferential use of a multi-markerapproach using increasingly larger multi-gene data sets


(e.g., Baldauf et al., 2000; Hampl et al., 2005; Moreira et al.,2000; Steenkamp et al., 2005), especially thanks to the con-struction of the Wrst expressed sequence tag (EST) databases for several protist species (Bapteste et al., 2002; Rod-ríguez-Ezpeleta et al., 2005) and the Wrst phylogenetic infer-ences made possible by the complete sequencing of a fewprotist genomes (e.g., Eichinger et al., 2005). However,despite great eVorts, especially for massive EST sequencing,this approach is yielding results at a still relatively slowpace so that, today, we have a huge number of sequencesbut for only a very small number of protist species (Rodrí-guez-Ezpeleta et al., 2005). In addition to this multi-geneway, a complementary approach to the challenge of recon-structing the evolutionary history of eukaryotes could bethe use of fewer markers but a much richer taxonomic sam-pling (Graybeal, 1998). Obviously, SSU rDNA oVers thebest sampling, but we have shown that it can be eYcientlycombined with LSU rDNA that, like its small subunitcounterpart, has the advantage of being easy to amplify byPCR and to sequence, even from uncultured organisms(López-García et al., 2002). Hence, with relatively modesteVort it should be possible to gather a SSU+LSU rDNAdatabase with a rich taxonomic sampling very rapidly. Itwill certainly not resolve the entire eukaryotic phylogeny(see, for example, the ambiguous results for centrohelidheliozoa) but it can provide a more reliable and more com-prehensive backbone that could be useful, among otherthings, to identify key species for EST projects.

Acknowledgments

DM and PLG thank the French Centre National de laRecherche ScientiWque (CNRS), section ‘Dynamique dela Biodiversité’ for an ATIP grant. TCS thanks the UKNatural Environment Research Council (NERC) forresearch grants and a Professorial Fellowship. S.H. andD.B. thank NERC for research studentships. T.C.S. andD.M. thank the Canadian Institute for AdvancedResearch Evolutionary Biology Program for Fellowshipand Associateship support. We thank Franz Lang, Gert-raud Burger and Michael Gray for providing Reclino-monas sequences.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.ympev.2006.11.001.

References

Archibald, J.M., Longet, D., Pawlowski, J., Keeling, P.J., 2003. A novelpolyubiquitin structure in cercozoa and foraminifera: evidence for anew eukaryotic supergroup. Mol. Biol. Evol. 20, 62–66.

Atkins, M.S., McArthur, A.G., Teske, A.P., 2000. Ancyromonadida: a newphylogenetic lineage among the protozoa closely related to the com-mon ancestor of metazoans, fungi, and choanoXagellates (Opi-sthokonta). J. Mol. Evol. 51, 278–285.


http://dx.doi.org/10.1016/j.ympev.2006.11.001




ARTICLE IN PRESS

Baldauf, S.L., Roger, A.J., Wenk-Siefert, I., Doolittle, W.F., 2000. A king-dom-level phylogeny of eukaryotes based on combined protein data.Science 290, 972–977.

Bapteste, E., Brinkmann, H., Lee, J.A., Moore, D.V., Sensen, C.W., Gor-don, P., DuruXe, L., Gaasterland, T., Lopez, P., Muller, M., Philippe, H.,2002. The analysis of 100 genes supports the grouping of three highlydivergent amoebae: Dictyostelium, Entamoeba, and Mastigamoeba.Proc. Natl. Acad. Sci. USA 99, 1414–1419.

Baroin, A., Perasso, R., Qu, L.H., Brugerolle, G., Bachellerie, J.P., Adoutte,A., 1988. Partial phylogeny of the unicellular eukaryotes based onrapid sequencing of a portion of 28S ribosomal RNA. Proc. Natl.Acad. Sci. USA 85, 3474–3478.

Bass, D., Cavalier-Smith, T., 2004. Phylum-speciWc environmental DNAanalysis reveals remarkably high global biodiversity of Cercozoa (Pro-tozoa). Int. J. Syst. Evol. Microbiol. 54, 2393–2404.

Bass, D., Moreira, D., Lopez-Garcia, P., Polet, S., Chao, E.E., von der Hey-den, S., Pawlowski, J., Cavalier-Smith, T., 2005. Polyubiquitin insertionsand the phylogeny of Cercozoa and Rhizaria. Protist 156, 149–161.

Berney, C., Pawlowski, J., 2003. Revised small subunit rRNA analysis pro-vides further evidence that Foraminifera are related to Cercozoa. J.Mol. Evol. 57 (Suppl. 1), S120–S127.

Cafaro, M.J., 2005. Eccrinales (Trichomycetes) are not fungi, but a clade ofprotists at the early divergence of animals and fungi. Mol. Phylogenet.Evol. 35, 21–34.

Castresana, J., 2000. Selection of conserved blocks from multiple align-ments for their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552.

Cavalier-Smith, T., 1998a. Neomonada and the origin of animals andfungi. In: Coombs, G.H., Vickerman, K., Sleigh, M.A., Warren, A.(Eds.), Evolutionary relationships among Protozoa. Chapman andHall, London, pp. 375–407.

Cavalier-Smith, T., 1998b. A revised six-kingdom system of life. Biol. Rev.Camb. Philos. Soc. 73, 203–266.

Cavalier-Smith, T., 1999. Principles of protein and lipid targeting in second-ary symbiogenesis: euglenoid, dinoXagellate, and sporozoan plastid ori-gin and the eukaryote family tree. J. Eukaryot. Microbiol. 46, 347–366.

Cavalier-Smith, T., 2000. Fleagellate megaevolution. In: Leadbeater,B.S.C., Green, J.C. (Eds.), The Xagellates: Unity, diversity and evolu-tion. Taylor and Francis Limited, London, pp. 361–390.

Cavalier-Smith, T., 2002. The phagotrophic origin of eukaryotes and phy-logenetic classiWcation of Protozoa. Int. J. Syst. Evol. Microbiol. 52,297–354.

Cavalier-Smith, T., 2003a. The excavate protozoan phyla MetamonadaGrassé emend. (Anaeromonadea, Parabasalia, Carpediemonas,Eopharyngia) and Loukozoa emend. (Jakobea, Malawimonas): theirevolutionary aYnities and new higher taxa. Int. J. Syst. Evol. Micro-biol. 53, 1741–1758.

Cavalier-Smith, T., 2003b. Protist phylogeny and the high-level classiWca-tion of Protozoa. Eur. J. Protistol. 39, 338–348.

Cavalier-Smith, T., 2004. Only six kingdoms of life. Proc. Roy. Soc. Lond.B 271, 1251–1262.

Cavalier-Smith, T., Chao, E.E., 1995. The opalozoan Apusomonas isrelated to the common ancestor of animals, fungi and choanoXagel-lates. Proc. Roy. Soc. Lond. B. 261, 1–6.

Cavalier-Smith, T., Chao, E.E., 2003a. Molecular phylogeny of centrohelidheliozoa, a novel lineage of bikont eukaryotes that arose by ciliary loss.J. Mol. Evol. 56, 387–396.

Cavalier-Smith, T., Chao, E.E., 2003b. Phylogeny and classiWcation of phy-lum Cercozoa (Protozoa). Protist 154, 341–358.

Cavalier-Smith, T., Chao, E.E., 2003c. Phylogeny of Choanozoa, Apuso-zoa, and other protozoa and early eukaryote megaevolution. J. Mol.Evol. 56, 540–563.

de Koning, A.P., Tartar, A., Boucias, D.G., Keeling, P.J., 2005. Expressedsequence tag (EST) survey of the highly adapted green algal parasite,Helicosporidium. Protist 156, 181–190.

Douady, C.J., Delsuc, F., Boucher, Y., Doolittle, W.F., Douzery, E.J., 2003.Comparison of Bayesian and maximum likelihood bootstrap measuresof phylogenetic reliability. Mol. Biol. Evol. 20, 248–254.


Eichinger, L., Pachebat, J.A., Glockner, G., Rajandream, M.A., Sucgang,R., Berriman, M., Song, J., Olsen, R., Szafranski, K., Xu, Q., Tunggal,B., Kummerfeld, S., Madera, M., Konfortov, B.A., Rivero, F., Bankier,A.T., Lehmann, R., Hamlin, N., Davies, R., Gaudet, P., Fey, P., Pilcher,K., Chen, G., Saunders, D., Sodergren, E., Davis, P., Kerhornou, A.,Nie, X., Hall, N., Anjard, C., Hemphill, L., Bason, N., Farbrother, P.,Desany, B., Just, E., Morio, T., Rost, R., Churcher, C., Cooper, J., Hay-dock, S., van Driessche, N., Cronin, A., Goodhead, I., Muzny, D., Mou-rier, T., Pain, A., Lu, M., Harper, D., Lindsay, R., Hauser, H., James, K.,Quiles, M., Madan Babu, M., Saito, T., Buchrieser, C., Wardroper, A.,Felder, M., Thangavelu, M., Johnson, D., Knights, A., Loulseged, H.,Mungall, K., Oliver, K., Price, C., Quail, M.A., Urushihara, H., Hernan-dez, J., Rabbinowitsch, E., SteVen, D., Sanders, M., Ma, J., Kohara, Y.,Sharp, S., Simmonds, M., Spiegler, S., Tivey, A., Sugano, S., White, B.,Walker, D., Woodward, J., Winckler, T., Tanaka, Y., Shaulsky, G.,Schleicher, M., Weinstock, G., Rosenthal, A., Cox, E.C., Chisholm,R.L., Gibbs, R., Loomis, W.F., Platzer, M., Kay, R.R., Williams, J.,Dear, P.H., Noegel, A.A., Barrell, B., Kuspa, A., 2005. The genome ofthe social amoeba Dictyostelium discoideum. Nature 435, 43–57.

Erixon, P., Svennblad, B., Britton, T., Oxelman, B., 2003. Reliability ofBayesian posterior probabilities and bootstrap frequencies in phyloge-netics. Syst. Biol. 52, 665–673.

Graybeal, A., 1998. Is it better to add taxa or characters to a diYcult phy-logenetic problem? Syst. Biol. 47, 9–17.

Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm toestimate large phylogenies by maximum likelihood. Syst. Biol. 52, 696–704.

Haeckel, E., 1887. Report on radiolaria collected by H.M.S. Challengerduring the years 1873–1876. Rep Sci Res Voyage H.M.S. Challenger1873–76 18, 1–1803.

Hampl, V., Horner, D.S., Dyal, P., Kulda, J., Flegr, J., Foster, P.G., Embley,T.M., 2005. Inference of the phylogenetic position of oxymonads basedon nine genes: support for Metamonada and Excavata. Mol. Biol.Evol. 22, 2508–2518.

Hertel, L.A., Bayne, C.J., Loker, E.S., 2002. The symbiont Capsasporaowczarzaki, nov. gen. nov. sp., isolated from three strains of the pulmo-nate snail Biomphalaria glabrata is related to members of the Meso-mycetozoea. Int. J. Parasitol. 32, 1183–1191.

Hillis, D.M., 1996. Inferring complex phylogenies. Nature 383, 130–131.Karpov, S.A., Bass, D., Mylnikov, A.P., Cavalier-Smith, T., 2006. Molecu-

lar phylogeny of Cercomonadidae and kinetid patterns of Cercomonasand Eocercomonas gen. nov (Cercomonadida, Cercozoa). Protist 157,125–158.

Lang, B.F., O’Kelly, C., Nerad, T., Gray, M.W., Burger, G., 2002. The clos-est unicellular relatives of animals. Curr. Biol. 12, 1773–1778.

Longet, D., Archibald, J.M., Keeling, P.J., Pawlowski, J., 2003. Foraminif-era and Cercozoa share a common origin according to RNA polymer-ase II phylogenies. Int. J. Syst. Evol. Microbiol. 53, 1735–1739.

López-García, P., Rodríguez-Valera, F., Moreira, D., 2002. Towards themonophyly of Haeckel’s Radiolaria: 18S rRNA environmental datasupport the sisterhood of Polycystinea and Acantharea. Mol. Biol.Evol. 19, 118–121.

Medina, M., Collins, A.G., Taylor, J.W., Valentine, J.W., Lipps, J.H., Ama-ral-Zettler, L., Sogin, M.L., 2003. Phylogeny of Opisthokonta and theevolution of multicellularity and complexity in Fungi and Metazoa.Int. J. Astrobiol. 2, 203–211.

Mikrjukov, K.A., Mylnikov, A.P., 2001. A study of the Wne structure and themitosis of a lamellicristate amoeba, Micronuclearia podoventralis gen. etsp. nov. (Nucleariidae, Rotosphaerida). Eur. J. Protistol. 37, 15–24.

Moreira, D., Le Guyader, H., Philippe, H., 2000. The origin of red algaeand the evolution of chloroplasts. Nature 405, 69–72.

Nikolaev, S.I., Berney, C., Fahrni, J.F., Bolivar, I., Polet, S., Mylnikov, A.P.,Aleshin, V.V., Petrov, N.B., Pawlowski, J., 2004. The twilight of Helio-zoa and rise of Rhizaria, an emerging supergroup of amoeboid eukary-otes. Proc. Natl. Acad. Sci. USA 101, 8066–8071.

Nikolaev, S.I., Berney, C., Petrov, N.B., Mylnikov, A.P., Fahrni, J.F., Paw-lowski, J., 2006. Phylogenetic position of Multicilia marina and the evo-lution of Amoebozoa. Int. J. Syst. Evol. Microbiol. 56, 1449–1458.



ARTICLE IN PRESS

Patron, N.J., Rogers, M.B., Keeling, P.J., 2004. Gene replacement of fruc-tose-1,6-bisphosphate aldolase supports the hypothesis of a single pho-tosynthetic ancestor of chromalveolates. Eukaryot. Cell. 3, 1169–1175.

Pawlowski, J., Bolivar, I., Guiard-MaYa, J., Gouy, M., 1994. Phylogeneticposition of foraminifera inferred from LSU rRNA gene sequences.Mol. Biol. Evol. 11, 929–938.

Perasso, R., Baroin, A., Qu, L.H., Bachellerie, J.P., Adoutte, A., 1989. Ori-gin of the algae. Nature 339, 142–144.

Philippe, H., 1993. MUST, a computer package of Management Utilitiesfor Sequences and Trees. Nucleic Acids Res. 21, 5264–5272.

Philippe, H., Adoutte, A., 1998. The molecular phylogeny of Eukaryota:solid facts and uncertainties. In: Coombs, G., Vickerman, K., Sleigh,M., Warren, A. (Eds.), Evolutionary relationships among Protozoa.Chapman and Hall, London, pp. 25–26.

Philippe, H., Lopez, P., Brinkmann, H., Budin, K., Germot, A., Laurent, J.,Moreira, D., Muller, M., Le Guyader, H., 2000. Early-branching orfast-evolving eukaryotes? An answer based on slowly evolving posi-tions. Proc. Roy. Soc. Lond. B Biol. Sci. 267, 1213–1221.

Philippe, H., Lartillot, N., Brinkmann, H., 2005. Multigene analyses ofbilaterian animals corroborate the monophyly of Ecdysozoa, Lopho-trochozoa, and Protostomia. Mol. Biol. Evol. 22, 1246–1253.

Polet, S., Berney, C., Fahrni, J., Pawlowski, J., 2004. Small-subunit ribo-somal RNA gene sequences of Phaeodarea challenge the monophyly ofHaeckel’s Radiolaria. Protist 155, 53–63.

Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model ofDNA substitution. Bioinformatics 14, 817–818.

Pupko, T., Huchon, D., Cao, Y., Okada, N., Hasegawa, M., 2002. Combin-ing multiple data sets in a likelihood analysis: which models are thebest? Mol. Biol. Evol. 19, 2294–2307.

Rambaut, A., Drummond, A., 2003. TRACER: A program for analysingresults from Bayesian MCMC programs. Version 1.2, Oxford.

Reece, K.S., Siddall, M.E., Stokes, N.A., Burreson, E.M., 2004. Molecularphylogeny of the Haplosporidia based on two independent genesequences. J. Parasitol. 90, 1111–1122.

Richards, T.A., Cavalier-Smith, T., 2005. Myosin domain evolution andthe primary divergence of eukaryotes. Nature 436, 1113–1118.

Rodríguez-Ezpeleta, N., Brinkmann, H., Burey, S.C., Roure, B., Burger, G.,LoVelhardt, W., Bohnert, H.J., Philippe, H., Lang, B.F., 2005. Mono-phyly of primary photosynthetic eukaryotes: green plants, red algae,and glaucophytes. Curr. Biol. 15, 1325–1330.

Roger, A.J., 1999. Reconstructing early events in eukaryotic evolution.Am. Nat. 154, S146–S163.


Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogeneticinference under mixed models. Bioinformatics 19, 1572–1574.

Ruiz-Trillo, I., Inagaki, Y., Davis, L.A., Sperstad, S., Landfald, B., Roger,A.J., 2004. Capsaspora owczarzaki is an independent opisthokont line-age. Curr. Biol. 14, R946–R947.

Sakaguchi, M., Nakayama, T., Hashimoto, T., Inouye, I., 2005. Phylogenyof the centrohelida inferred from SSU rRNA, tubulins, and actin genes.J. Mol. Evol. 61, 765–775.

Schmidt, H., Strimmer, K., Vingron, M., von Haeseler, A., 2002. TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartetsand parallel computing. Bioinformatics 18, 502–504.

Shimodaira, H., 2002. An approximately unbiased test of phylogenetic treeselection. Syst. Biol. 51, 492–508.

Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the conW-dence of phylogenetic tree selection. Bioinformatics 17, 1246–1247.

Simpson, A.G., Inagaki, Y., Roger, A.J., 2006. Comprehensive multigenephylogenies of excavate protists reveal the evolutionary positions of“primitive” eukaryotes. Mol. Biol. Evol. 23, 615–625.

Sogin, M.L., 1991. Early evolution and the origin of eukaryotes. Curr.Opin. Genet. Dev. 1, 457–463.

Stechmann, A., Cavalier-Smith, T., 2003. The root of the eukaryote treepinpointed. Curr. Biol. 13, R665–R666.

Steenkamp, E.T., Wright, J., Baldauf, S.L., 2005. The protistan origins ofanimals and fungi. Mol. Biol. Evol. 23, 93–106.

Stiller, J., Hall, B., 1999. Long-branch attraction and the rDNA model ofearly eukaryotic evolution. Mol. Biol. Evol. 16, 1270–1279.

Suzuki, Y., Glazko, G.V., Nei, M., 2002. Overcredibility of molecular phy-logenies obtained by Bayesian phylogenetics. Proc. Natl. Acad. Sci.USA 99, 16138–16143.

von der Heyden, S., 2004. Testing ubiquitous dispersal and freshwater/marine divergence in free-living protist groups. DPhil Thesis, Univer-sity of Oxford.

von der Heyden, S., Chao, E., Cavalier-Smith, T., 2004. Genetic diversity ofgoniomonads: an ancient divergence between freshwater and marinespecies. Eur. J. Phycol. 39, 343–350.

Walker, G., Dacks, J.B., Embley, T.M., 2006. Ultrastructural description ofBreviata anathema, n. gen., n. sp., the organism previously studied as“Mastigamoeba invertens”. J. Eukaryot. Microbiol. 53, 65–78.

Yuasa, T., Takahashi, O., Dolven, J.K., Mayama, S., Matsuoka, A., Honda,D., Bjørklund, K.R., 2006. Phylogenetic position of the small solitaryphaeodarians (Radiolaria) based on 18S rRNA sequences by single cellPCR analysis. Marine Micropal. 59, 104–114.


global eukaryote phylogeny: combined small- and large...

Documents