Ribosomal RNA Genes Challenge the Monophylyof the Hyalospheniidae (Amoebozoa: Arcellinida)

Enrique Laraa,b,c,e, Thierry J. Hegerb,c, Flemming Ekelunda,Mariusz Lamentowiczd, and Edward A.D. Mitchellb,c,1

aTerrestrial Ecology, Biological Institute, University of Copenhagen, Øster Farimagsgade 2D,DK-1353 Copenhagen K, Denmark

bWSL, Swiss Federal Institute for Forest, Snow and Landscape Research, Ecosystem Boundaries Research Unit,Wetlands Research Group, Station 2, CH-1015 Lausanne, Switzerland

cEcole Polytechnique Federale de Lausanne (EPFL), Laboratory of Ecological Systems, Station 2,CH-1015 Lausanne, Switzerland

dDepartment of Biogeography and Palaeoecology, Faculty of Geosciences, Adam Mickiewicz University in Poznan,Dziegielowa 27, 61-680 Poznan, Poland

eUnite d’Ecologie, Systematique et Evolution-CNRS UMR8079 Universite Paris-Sud 11, batiment 360,91405 Orsay Cedex, France

To date only five partial and two complete SSU rRNA gene sequences are available for the lobosetestate amoebae (Arcellinida). Consequently, the phylogenetic relationships among taxa and thedefinition of species are still largely dependant on morphological characters of uncertain value, whichcauses confusion in the phylogeny, taxonomy and the debate on cosmopolitanism of free-livingprotists. Here we present a SSU rRNA-based phylogeny of the Hyalospheniidae including the mostcommon species. Similar to the filose testate amoebae of the order Euglyphida the most basal cladeshave a terminal aperture; the ventral position of the pseudostome appears to be a derived character.Family Hyalospheniidae appears paraphyletic and is separated into three clades: (1) Heleoperasphagni, (2) Heleopera rosea and Argynnia dentistoma and (3) the rest of the species from generaApodera, Hyalosphenia, Porosia and Nebela. Our data support the validity of morphologicalcharacters used to define species among the Hyalospheniidae and even suggest that taxa describedas varieties may deserve the rank of species (e.g. N. penardiana var. minor). Finally our results suggestthat the genera Hyalosphenia and Nebela are paraphyletic, and that Porosia bigibbosa branchesinside the main Nebela clade.

Key words: Amoebozoa; Arcellinida; biogeography; phylogeny; SSUrRNA gene; testate amoebae.

Introduction

The testate lobose amoebae (Arcellinida Kent,1880) make up a group of protozoa presentworldwide mostly in freshwater and mosses, withsome representatives also inhabiting soils andeven marine habitats. The recent finding of some

1

Corresponding author; WSL, Swiss Federal Institute forForest, Snow and Landscape Research, Ecosystem Bound-aries Research Unit, Wetlands Research Group, Station 2,CH-1015 Lausanne, Switzerland, fax +41 21 693 39 13e-mail edward.mitchell@wsl.ch (E.A.D. Mitchell).

Published in Protist 159, issue 2, 165-176, 2008which should be used for any reference to this work

1

vase-shaped microfossils possibly belonging tothe Arcellinida (Porter et al. 2003) has invested thisgroup with a particular importance: these fossilsare generally considered as the oldest undispu-table eukaryotic remains ever found, puttingArcellinida in a central position in dating theapparition of life on earth (Berney and Pawlowski2006). However, in spite of their importance, thephylogenetic position of the Arcellinida within thetree of life has remained unresolved for a long timeand their taxonomy still remains unsatisfactory.The Arcellinida have recently been shown tobelong to the Amoebozoa as a sister-group tothe lobose naked amoebae families Amoebidaeand Hartmannellidae (Nikolaev et al. 2005). Theseven taxa analysed by these authors suggestedthat the Arcellinida are monophyletic.

The family Hyalospheniidae Schultze 1877includes some of the most common, conspicuous,and well-studied lobose testate amoebae. Mem-bers of this family are especially abundant anddiverse in oligotrophic wetland such as Sphagnumpeatlands but some can also be found in othermosses, freshwater habitats, and soils (Mitchellet al. 1999, 2000). It was suggested that the750 Mya old vase-shaped microfossils belongedto this family (Porter and Knoll 2000; Porter et al.2003). Several species have been claimed to havea limited geographic distribution, the most famouscase being certainly Apodera vas (Deflandre 1936;Van Oye 1944). This and other species of theApodera genus have not been recorded in theextra-tropical northern Hemisphere where moststudies on the ecology and taxonomy of testateamoebae took place (Mitchell and Meisterfeld2005). Members of the family Hyalospheniidaesensu Schultze 1877 are characterized by anovoid or pyriform shell, which is laterally com-pressed. The shell may be composed solely ofproteinaceous material (genus Hyalosphenia) or ofsiliceous shell plates recycled from other testateamoebae such as Euglypha, sometimes withagglutinated mineral particles (Meisterfeld 2000).Some species, like Heleopera sphagni and Hyalo-sphenia papilio host photosynthetic endosym-bionts and are thus mixotrophic (Penard 1902).

The classification of the Hyalospheniidae sensuSchultze 1877 is based on characters of the testsuch as composition (proteinaceous or aggluti-nated), shape of the aperture (circular, oval oralmost linear) and shape of the shell. Jung (1942)split the very large and widespread genus Nebelainto twelve genera. Meisterfeld (2000) recognizesfour families within the Hyalospheniidae sensuSchultze 1877: Nebelidae (restricted only to the

taxa which formerly were included in the genusNebela), Hyalospheniidae sensu stricto, Lesquer-eusiidae and Heleoperidae. To date, however, theexisting morphological and molecular data has notallowed validation of this classification. In thisstudy, we therefore aimed to provide new mole-cular data on some representative genera andspecies of Hyalospheniidae sensu Schultze 1877in order to clarify the phylogenetic position of themain morphological types.

Testate amoebae, in general, and the Arcellinidain particular, are increasingly used in ecology andpalaeoecology with applications in the reconstruc-tion of past climatic fluctuations over the Holo-cene (Charman 2001). However, both the phylo-geny and taxonomy of this group urgently needsto be updated (Mitchell et al. in press). It has beenshown that, for bioindication purposes, the high-est possible taxonomic resolution should beattained to maximize the ecological indicatorvalue of the assemblages, recorded (Bobrovet al. 2002). Clearly, with a more reliable taxonomythe value of this group of organisms for ecologistsand palaeoecologists would be even higher. Thusour broader goal is both to increase the knowl-edge on the phylogeny and taxonomy of thisgroup and to improve its value for other fields ofresearch.

Results

We obtained a total of 16 SSU rRNA genesequences from 14 different taxa (Figs 1 and 2),which included representatives from three of thefour families proposed by Meisterfeld (2000), andinferred their phylogenetic relationships based onlikelihood and Bayesian analyses.

Structure of the SSU rDNA Gene of theHyalospheniidae

We found a class I intron of approximately 450base pairs starting at position 1200 on the SSUrDNA gene of Schizosaccharomyces pombe(accession file X58056 in GenBank) in seven taxa:N. carinata, N. tincta var. tincta, N. tincta var. major(isolate from Sweden only), Porosia bigibbosa,Nebela tubulosa, Hyalosphenia papilio and Heleo-pera rosea. This intron has been also found inmyxomycetes like Fuligo septica (Fiore-Donnoet al. 2005) and corresponds to position S956according to the nomenclature by Johansen andHaugen (2001).

2

Figure 1. Illustrations of the sequenced Arcellinida taxa. The illustrated individuals correspond to thesequenced individuals with three exceptions for which no illustrations were available: (a) Argynnia dentistoma(from Grindelwald CH and not from Modrava CZ), (b) Heleopera rosea (from Sweden and not from La Chauxd’Abel bog CH), (c) Nebela lageniformis, (d) Apodera vas, (e) Hyalosphenia elegans, (f) H. papilio, (g) Porosiabigibbosa, (h) N. carinata, (i & j) N. penardiana var. minor, (k) N. penardiana (from Chlebowo mire PL and notfrom Grindelwald, CH), (l) N. tubulosa, (m) N. tincta var. major from Ireland, (n) N. tincta var. major fromSweden, (o) N. tincta, (p) N. flabellulum. Scale bars ¼ 50 mm, except for N. penardiana var. minor andH. elegans: Scale bar ¼ 20 mm.

3

Figure 2. Morphological variation within the studied population of Nebela lageniformis from Peru (a—f; scalebar ¼ 50 mm). The original drawings from the descriptions of Nebela lageniformis (g, after Penard 1902 and h,after Deflandre 1936) and Nebela lageniformis var. cordiformis (i, after Heinis 1914) are given for comparison(original drawings lacked scale).

4

Phylogenetic Position of the StudiedSpecies

For each phylogenetic analysis achieved in thiswork, maximum likelihood and Bayesian methodsof tree building gave exactly the same topology.Figure 3 shows the position of the studied taxa inthe phylogenetic tree of the Amoebozoa (selectedclosely related naked amoebae). This tree illus-trates the paraphyly of Hyalospheniidae sensulato. Heleopera sphagni branched at the base ofthe whole arcellinid clade, while Heleopera roseaand Argynnia dentistoma formed another group(maximum likelihood bootstrap ¼ 87; Bayesiananalysis posterior probabilities ¼ 0.63) at the baseof a clade formed by species with a ventralpseudostome (Bullinularia indica, Trigonopyxisarcula, Arcella artocrea and Centropyxis laevigata)and the rest of the Hyalospheniidae. Inside this

group, N. lageniformis and Apodera vas, were themost basal species. We observed a continuum oftest shapes between typical N. lageniformis andN. l. var. cordiformis (Fig. 2) in our studiedpopulation that suggested the possible existenceof more than one taxon. However, sequencingof the whole sample provided only one singlesequence, and the cloning of PCR productsyielded only one type of clone. Besides thesespecies, all other members of genera Nebelaand Hyalosphenia branched together in arobust clade (100/1.00), hereafter referred to as‘‘core Nebelas’’.

To improve resolution, we restrict our analysis toshort branch Hyalospheniidae, allowing us to use830 positions instead of 450. This analysis (Fig. 4)confirms the existence of a clade constituted(for now) of two closely related species, Apoderavas and Nebela lageniformis which branch

Figure 3. Maximum likelihood bootstrap consensus tree (GTR model, 450 sites) of the Arcellinida and relatedtaxa among the Tubulina illustrating the paraphyly of the Hyalospheniidae. Species names in bold representnew sequences. The numbers along the branches represent the bootstraps obtained by the maximumlikelihood method and the posterior probabilities obtained by Bayesian analysis.

5

together, with a good support mostly from max-imum likelihood analysis (77/0.70).

The phylogeny of the core Nebelas is detailled inFigure 5. The species Apodera vas and Nebelalageniformis were used as outgroups in thisanalysis. At the base of the clade, Hyalospheniaelegans formed a rather long branch, and did notgroup with its congeneric H. papilio. Instead, thelatter was linked, in a basal position, to otherrepresentatives of genus Nebela (83/0.76). Theapparent paraphyly of genus Hyalosphenia mightbe due to a long-branch artefact phenomenoncaused by the sequence of Hyalosphenia elegans.

Inside this clade, the taxa P. bigibbosa, Nebelacarinata, N. penardiana var. penardiana andN. penardiana var. minor are grouped together(57/1.00). The two varieties of N. penardianabranched together indeed with strong support(99/1.00), but were nevertheless genetically quitedistinct (10% of sequence difference on thestudied fragment). The taxa N. tincta var. tincta,N. tincta var. major and N. flabellulum branchedtogether robustly (95/0.83), and their sequenceswere very similar (o1% of sequence difference onthe studied fragment). N. tubulosa branched with

this latter group in the two tree constructionmethods; however, its position in the tree washighly unstable, the support of this branch beingvery low (bootstrap value/posterior probabilityo50/ 0.5).

It is noteworthy that the sequences obtainedfrom different samples, from different geographicorigins, from the same species were identical.Also, our sequences from Nebela tincta major andHyalosphenia papilio were identical to the shortersequences presented by Nikolaev et al. (2005),available in Genbank under the accession filesAY848968 and AY848966, respectively. Thesesequences were not included in the analysis,for they are shorter than the sequences studiedhere.

Discussion

Structure of the Fragment of the StudiedSSU rRNA Gene

The presence of a class I intron in the SSU rRNAgene of the some of the studied species, which

Figure 4. Maximum likelihood bootstrap consensus tree (GTR model, 830 sites) of the relatively short branchforming Arcellinida (i.e. excluding Centropyxis laevigata and Heleopera sphagni), illustrating the relatedness ofNebela lageniformis and Apodera vas forming an independent clade separated from the core Nebelas.Species names in bold represent new sequences. The numbers along the branches represent the bootstrapsobtained by the maximum likelihood method and the posterior probabilities obtained by Bayesian analysis.

6

was also present in three of the seven taxa studiedby Nikolaev et al. (2005), is remarkable, but thepattern of presence and absence of this introndoes not seem to have any phylogeneticalsignificance. Indeed, it is present in the basalbranching Heleopera rosea, but is absent in, forinstance, both varieties of Nebela penardiana.

Nikolaev et al. (2005) did not find it in theirsequence from N. tincta var major, and we didnot find it either in the sequence of our isolate ofthis taxon from Ireland, although it was present inour sequence from Sweden. The inconstantpresence of such an intron in the SSU rRNA genein a given taxon has been already documented in

Figure 5. Maximum likelihood bootstrap consensus tree (GTR model, 948 sites) of the core Nebelas.Numbers at the nodes represent successively bootstrap values obtained by maximum likelihood andposterior probabilities obtained by Bayesian analysis. Only bootstrap values above 50 are represented. Thebranch of Hyalosphenia elegans is reduced from half of its size for clarity.

7

another amoebozoan, the mycetozoan Fuligoseptica (Fiore-Donno et al. 2005).

Phylogenetic Position of the StudiedSpecies

Based on the available sequence including the 16new ones presented here the Arcellinida remainmonophyletic. However, molecular data is stillmissing from several important taxa (e.g. Crypto-difflugia, Phryganella, etc.). Our results suggestthat a ventral position of the pseudostome is mostprobably a derived character amongst the Arcelli-nida. The most basal clades (Heleopera sphagniand H. rosea/Argynnia dentistoma) have anacrostome opening, like the more derived coreNebelas. This situation is similar to the oneobserved in the Euglyphida, where ventral openedpseudostomes like observed in Trinematidaederived from acrostome forms such as Euglyphi-dae and Assulinidae (Lara et al. 2007).

Molecular data can be useful to assess thephylogenetic value of various morphological cri-teria used in taxonomy. Our results show that thepresence of foreign particles in the test is notnecessarily a reliable criterion for phylogeneticrelationships. For example, Hyalosphenia papilioas well as H. elegans lack foreign material. Still,according to our analysis H. papilio is more closelyrelated to the large Nebela spp., to which itresembles both in shape and size, than toH. elegans. However, as long branches had aneffect on the tree topology this result should beconfirmed by further analyses on more taxa.Nevertheless, test composition sometimes corre-lates with molecular phylogeny, e.g. Argynniadentistoma, which was initially included in thegenus Nebela, differs from Nebela s.str. by theporous cement agglutinating the (rough) particlesthat compose the test (Jung 1942; Vucetich 1974).In accordance with this, molecular phylogenysuggests that Argynnia is only distantly related tomembers of genus Nebela. Conversely, a syna-pomorphy of Nebela s.l. appears to be a test builtfrom an organic amorphous matrix in which preymaterial (mainly from euglyphid amoebae) may beincluded. Interestingly, in the absence of preys,some species like N. collaris are able to formentirely organic tests (MacKinlay 1936). The factof having a test without xenosome is thereforenot a sufficient criterion for defining the genusHyalosphenia.

Morphologically, the genus Heleopera iswell defined by its slit-like terminal positioned

pseudostome, which is generally considered to bean apomorphic character. However, our moleculardata suggest that genus Heleopera is paraphyletic(Fig. 3). This result should be mitigated by the factthat the long branch formed by H. sphagnicertainly has an effect on tree topology. Thenumerous insertions in its SSU rRNA genesequence favour that H. sphagni is a fast-evolvingtaxon. Here again, more data are needed toascertain whether genus Heleopera is actuallyparaphyletic or not.

Inside the core Nebelas, an important criterionfor deep phylogenetic relationships seems to bethe general shape of the shell.

The two species with an elongated tubular neck,Apodera vas and Nebela lageniformis branchtogether at the base of the ‘‘core Nebelas’’ clade,and their relatedness is relatively well-supported,mostly with the ML analysis (77/0.70). The mainmorphological difference between these two taxa(besides the size) is the presence of a constrictedneck in A. vas. Hyalosphenia elegans, with its longneck and characteristic test with many smalldepressions also branches alone. It should benoted that, based on its morphology, the ‘‘variety’’of N. lageniformis studied here (Fig. 2) couldpossibly be assigned to N. l. var. cordiformis(Heinis 1914). This taxon is wider than the type.However, in our sample from Peru, we observeda continuum of test shapes between typicalN. lageniformis and N. l. var. cordiformis (seeFig. 2), but we obtained only one single sequenceand only one type of clone from this population.Until the taxonomic relationships between thesetwo taxa are better understood through a com-bined morphological and molecular study, weprefer to adopt a conservative position andconsider that N. l. var. cordiformis is the productof the phenotypic plasticity of N. lageniformis, andthat these two forms are not separated geneti-cally. It would however not be impossible thatfurther work may show that, contrary to thisconservative view, the different varieties ofNebela lageniformis are genetically distinct, aswe report here for the varieties of Nebela tinctaand N. penardiana (see further).

It should be noted that A. vas is one of the‘‘flagship species’’ of microorganisms with alimited distribution while N. lageniformis is con-sidered to be cosmopolitain (Mitchell and Meis-terfeld 2005). The two taxa differ only in 3% in theportion of the SSU rRNA gene examined herebut they nevertheless are certainly different fromeach other both from the genetic and themorphological points of view. Smith and Wilkinson

8

(1987) suggested that cosmopolitan speciesevolved prior to the break-up of the supercontinent of Pangea, while the taxa specific tothe southern hemisphere has evolved during theJurassic era in Gondwanaland. Although thelack of a reliable fossil record makes it impossibleto date the separation between A. vas andN. lageniformis, we believe that this event is morerecent than the separation of Laurasia andGondwana because of (1) their similar morphologyand SSU rRNA gene sequence and (2) the putativeextremely long evolutionary history of the Arcelli-nida (Porter et al. 2003). As Apodera vas has notyet been recorded in North America or in Eurasiawhere most of the studies on testate amoebaehave been done, it can be assumed that it evolvedsomewhere in the former Gondwana, and hasstill not reached similar habitats in the northerncontinents.

The rest of the studied species, referred to as‘‘core Nebelas’’ clustered together with a strongsupport. These species all had a shorter neckor a test generally tapering from the broadestpart to the pseudostome. At the base of this clade,the mixotrophic Hyalosphenia papilio brancheswith a moderate support; notably, it is the onlyspecies in the ‘‘core Nebelas’’ investigated inthis study never to incorporate xenosomes(mineral elements from external origin). Possiblythis more transparent (but amber coloured) shellallows more light to reach its photosyntheticendosymbionts.

Besides H. papilio, all members of this cladebelong to genus Nebela, with the exception ofPorosia bigibbosa. Porosia was suggested byJung (1942), to include forms with two invaginatedpores on the test. The ecology of P. bigibbosadiffers greatly from the other studied members ofthe Nebela s.l. genus, being abundant in forestlitter and rare or exceptional in mosses andSphagnum (Todorov 2002). In spite of thesemorphological and ecological differences, itsphylogenetic position does not support its classi-fication inside the monotypic genus Porosia.

The two varieties of Nebela penardiana studiedhere robustly cluster together with high statisticalsupport; however, they appear distantly related inthe tree. Morphologically, they can be easilydifferentiated; besides the size difference, thenominal variety being more pear-shaped thanN. p. var. minor. In addition, it seems that theminor variety is restricted to tropical regions,where it can be relatively abundant (Gauthier-Lievre 1957). Based on our results, it might seemadequate to give the specific status to this variety.

However, we feel that more data would berequired to revise the taxonomy.

Nebela tincta var. tincta, N. tincta var. major, andN. flabellulum make up a very distinctive clade.Some of these relatively smaller and roundedforms are not restricted to the wet and acidicSphagnum environment. Genetic data suggestthat the three taxa presented here are very closelyrelated. Interestingly, while the first two taxa areconsidered as subspecies of N. tincta, and areincluded in the ‘‘N. tincta major-bohemica-collarisspecies complex’’ (Charman et al. 2000; Warner1987), previous studies did not relate N. flabellu-lum to this group. Our molecular data furthersuggests that N. tincta tincta is even more closelyrelated to N. flabellulum than it is to N. tinctamajor (76/0.96). While the two subspecies ofN. tincta share an identical morphology and aredifferentiated only by size, N. tincta tincta andN. flabellulum have the same cell length, butclearly differ in width, the latter having a muchbroader test and almost no neck (Fig. 1, Table 1).

Methods

Origin of the samples, identification: A total of fourteenspecies and subspecies were obtained from Sphagnum, othermosses, and forest litter collected from a broad range ofsampling locations (Table 1, Figs 1 and 2). Two species,Nebela carinata and N. tincta major, were isolated from twodifferent locations, as a very first step towards a study of theirgeographical variability. The amoebae were extracted fromthe mosses by sieving and back sieving using appropriatemesh sizes (300 mm down to 10 mm), picked individually with anarrow diameter pipette under the dissecting or invertedmicroscope and isolated for DNA extraction.

Scanning electron microscopy imaging: For SEM obser-vations, the tests were rinsed with demineralized water andthen kept during 1 week in a desiccator. The samples werecoated with gold in a Bal-Tec SCD005 sputter, or in a Jeol 4x.Samples were observed either with a PHILIPS XL30 FEGmicroscope at a tension of 5 kV or with a Jeol 840a at atension 20 kV.

Nucleic acids extraction and amplification: DNA wasextracted using a guanidine thiocyanate protocol (Chomc-zynski and Sacchi 1987). Partial SSU sequences wereobtained by PCR. DNA was amplified with the universaleukaryotic primers EUK516F (GGAGGGCAAGTCTGGT) andEUK1643R (GACGGGCGGTGTGTACA), in a total volume of25 ml with an amplification profile consisting in a 35 cyclesprogram of 1 min at 94 1C, 1 min at 50 1C and 1 min 30 at 72 1Cwith a final elongation of 10 min at 72 1C. The PCR mixcontained 5% DMSO and 10 mM TMA. Afterwards, a semi-nested protocol was used to obtain specific amplification oftestate amoebae DNA, using the following primers incombination with the universal primers: Arcell1F (GAAA-GTGGTGCATGGCCGTTT) with EUK1643R and Hyalo2R(GCATTTCATTGTAACGCGC) with EUK516. In both cases,PCR products from the first amplification were diluted 250times and used as a template for the second amplification.

9

For the first pair of primers, a first denaturation step of 5 minwas applied, followed by a 30 cycles program consisting of1 min at 94 1C, 1 min at 60 1C and 1 min at 72 1C for theamplification of a product of about 400 bp. For the second pairof primers, an identical protocol was applied except that theannealing temperature was of 58 1C and the elongation time at72 1C was of 1 min 30, and the product was around 950 bp(1400 with the intron).

Sequencing: The PCR products where purified with theNucleoFasts 96 PCR Clean Up kit from Macherey-Nagel (Duren,Germany) and sequenced at MWG Biotech (Martinsried,Germany). Products where sequenced using each time bothforward and reverse primer, as well as the sequencing primer

Hyalo3R (CAATACAAGTGGCCCCAAC) which served to se-quence the 50 end of the second fragment (as described above).

Alignment and phylogenetic analysis: The SSU rRNAgene sequences obtained in this study where alignedmanually using the BioEdit software (Hall 1999). We performedthree phylogenetic analyses using three different datasets:(1) A first analysis including 38 sequences from the taxonTubulina including our Hyalospheniidae (450 sites out of 550positions), rooted with the most basal members of theTubulina, the group of Echinamoeba and Hartmannellavermiformis (Fahrni et al. 2003). (2) A second analysisincluding all relatively short branch forming Arcellinids (i.e.without Centropyxis laevigata and Heleopera sphagni), with

Table 1. List of sequenced taxa sampling location and morphometric measurements.

Taxon Sampling location Co-ordinates Altitude(m)

Shell dimensions�

Length (mm) Breadth (mm)

Apodera vas Wet moss fromBlepharidophyllum-dominated Mire, MarionIsland (ZA)

461530 S, 371510 E 27 16274 8572

Argynniadentistoma

Poor fen with Molinia,Modrava, (CZ)

491010 N, 131260 E 980 9673 6471

Helopera rosea Sphagnum, La Chauxd0Abel Peatland (CH)

471100 N, 061560 E 1006 10772 7772

Hyalospheniaelegans

Sphagnum, CachotPeatland (CH)

471500 N, 061400 E 1050 9976 5174

Hyalospheniapapilio

Sphagnum,Ryggmossen Peatland(SE)

601000 N, 171150 E 58 12379 75717

Nebela carinata Sphagnum, Bog pool,Ryggmossen Peatland(SE)

601000 N, 171150 E 58 17777 12477

Nebela carinata Sphagnum, Praz-RodetBog, (CH)

461330 N, 061100 E 1041 179710 12975

Nebela flabellulum Sphagnum, Pitsligobog, Scotland (GB)

571360 N, 021090 W 110 7874 8872

Nebelalageniformis

Moss, summit ofHuayna Pichu (PE)

131090 S, 721320 W 2300 11675 6874

Nebela penardianavar. penardiana

Brown moss, AlpineRich Fen, Grindelwald(CH)

461400 N, 081010 E 2343 14478 12475

Nebela penardianavar. minor

Moss, rainforest nearVolcan Arenal (CR)

101280 N, 841150 W 1000 8174 4173

Nebela tincta var.major

Sphagnum, Poor Fen‘‘Scragh Bog’’ (IE)

531350 N, 071220 W 98 11775 8274

Nebela tincta var.major

Sphagnum,Ryggmossen Peatland(SE)

601000 N, 171150 E 58 11875 8273

Nebela tincta var.tincta

Sphagnum,Ryggmossen Peatland(SE)

601000 N, 171150 E 58 8174 12475

Nebela tubulosa Sphagnum, Poor Fen,Granges-Narboz (FR)

461520 N, 061160 E 823 215713(N ¼ 3)

120717(N ¼ 3)

Porosia bigibbosa Forest litter under Piceaabies, near MarchairuzPass (CH)

461330 N, 061150 E 1300 15474 9975

�Average and standard deviation. N ¼ 10, unless otherwise stated.

10

Argynnia dentistoma and Heleopera rosea as outgroups (using830 sites out of 1050 positions). (3) A third analysis restrictedto ‘‘core Nebelas’’, using Apodera vas and Nebela lageniformisas outgroups. The number of unambiguously aligned positionsthat were used in this phylogenetic analysis was 948 sites outof a total of about 1050 (1500 with the intron). In all analyses,we performed a maximum likelihood analysis using thesoftware PHYML_v 2.4.4 (Guindon and Gascuel 2003) usingthe GTR+I+G model. The reliability of internal nodes wasestimated by bootstrapping (1000 replicates). In addition, theresulting tree was compared to the one obtained by Bayesiananalysis. These analyses were performed with the softwareMrBayes v. 3.1.2 (Huelsenbeck and Ronquist 2001). Themodel chosen here was the GTR model of substitution(Lanave et al. 1984; Rodriguez et al. 1990), the number ofinvariable sites being estimated, and a gamma-shapeddistribution of variable sites with five rate categories. Thevalues of alpha and the proportions of invariable sites, wererespectively, 0.467 and 0.269, for the general Tubulina tree,0.170 and 0.816 for the short-branching arcellinids, and 0.744and 0.705 for the ‘‘core Nebelas’’ tree. For Bayesian analyses,four simultaneous chains were run for 1,000,000 generations,and 10,000 trees were sampled, and trees corresponding tothe likelihood estimates previous to stabilization of thelikelihood plot where discarded as burn-in: 500 were removedfrom the first analysis (general Tubulina) and 100 for thesecond and the third (short-branching arcellinids and ‘‘coreNebelas’’). Posterior probabilities at all nodes were estimatedfrom the remaining trees.

Acknowledgements

This work was funded by Swiss NSF project no.205321-109709/1 and EU project RECIPE.RECIPE was partly supported by the EuropeanCommission, Directorate I, under the programme‘‘Energy, Environment and Sustainable Develop-ment’’ (no. EVK2-2002-00269) and partly, for theSwiss partners, by the State Secretariat forEducation and Research, Switzerland. We thankAnita Maric for practical help in the lab andFabienne Bobard and Marco Cantoni at the CIME(EPFL) for help with the SEM. We thank Marten FløJørgensen (University of Copenhagen) and SylvainDubey (University of Lausanne) for the help withthe phylogenetic analyses, Ralf Meisterfeld foruseful comments on taxonomy, Jan Pawlowskiand David Moreira for fruitful discussion onphylogeny, and also Herve Philippe and ananonymous reviewer whose criticism substantiallyimproved this manuscript. We also would like toacknowledge the following people for providingsamples: Niek Gremmen (Data-Analyse Ecologie,NL) for the samples from Marion Island, SylvainBuhler for the sample from Peru, Martine Rebetez(WSL) for the mosses from Costa Rica, NicolasDerungs for the cells of Porosia bigibbosa andRebekka Arz (Macaulay Institute-UK) for thesample from Scotland.

References

Berney C, Pawlowski J (2006) A molecular time-scale foreukaryote evolution recalibrated with the continuous micro-fossil record. Proc R Soc Lond B 273: 1867—1872

Bobrov AA, Charman DJ, Warner BG (2002) Ecology oftestate amoebae from oligotrophic peatlands: specific fea-tures of polytypic and polymorphic species. Biol Bull RussAcad Sci 29: 605—617

Charman DJ (2001) Biostratigraphic and palaeoenvironmen-tal applications of testate amoebae. Quart Sci Rev 20:1753—1764

Charman DJ, Hendon D, Woodland WA (2000) TheIdentification of Testate Amoebae (Protozoa: Rhizopoda) inPeats. Quaternary Research Association, London

Chomczynski P, Sacchi N (1987) Single-step method of RNAisolation by acid guanidinium thiocyanate-phenol-chloroformextraction. Anal Biochem 162: 156—159

Deflandre G (1936) Etude monographique sur le genre NebelaLeidy. Ann Protistol 5: 201—286

Fahrni JF, Bolivar I, Berney C, Nassonova E, Smirnov A,Pawlowski J (2003) Phylogeny of lobose amoebae based onactin and small-subunit ribosomal RNA genes. Mol Biol Evol20: 1881—1886

Fiore-Donno A-M, Berney C, Pawlowski J, Baldauf SL(2005) Higher-order phylogeny of plasmodial slime molds(Myxogastria) based on elongation factor 1-A and smallsubunit rRNA gene sequences. J Eukaryot Microbiol 52:201—210

Gauthier-Lievre L (1957) Additions aux Nebela d’Afrique. BullSoc Nat Afr Nord 48: 494—525

Guindon S, Gascuel O (2003) A simple, fast, and accuratealgorithm to estimate large phylogenies by maximum like-lihood. Syst Biol 52: 696—704

Hall TA ed (1999) BioEdit: a user-friendly biological sequencealignment editor and analysis program for Windows 95/98/NT.Nucleic Acids Symp Ser 41: 95—98

Heinis F (1914) Die Moosfauna Columbiens: in: Voyaged’exploration scientifique en Colombie. Mem Soc Neuch SciNat 5: 675—730

Huelsenbeck JP, Ronquist F (2001) MrBayes: Bayesianinference of phylogenetic trees. Bioinformatics 17:754—755

Johansen J, Haugen P (2001) A new nomenclature of group Iintrons in ribosomal DNA. RNA 7: 935—936

Jung W (1942) Illustrierte Thekamoben-Bestimmungstabellen.I. Die Systematik der Nebelinen. Arch Protistkd 95:357—390

Lanave C, Preparata G, Saccone C, Serio G (1984) A newmethod for calculating evolutionary substitution rates. J MolEvol 20: 86—93

Lara E, Heger TJ, Mitchell EAD, Meisterfeld R, Ekelund F(2007) SSU rRNA reveals a sequential increase in shellcomplexity among the euglyphid testate amoebae (Rhizaria:Euglyphida). Protist 158: 229—237

11

MacKinlay RB (1936) Observations on Nebela collaris Leidy(pro parte), a testate amoeba of moorland waters. Part 1. J RMicrosc Soc 56: 307—325

Meisterfeld R (2000) Order Arcellinida Kent, 1880. In Lee JJ,Leedale GF, Bradbury P (eds) The Illustrated Guide to theProtozoa, Vol 2, 2nd edn. Society of Protozoologists,Lawrence, Kansas, USA, pp 827—860

Mitchell EAD, Meisterfeld R (2005) Taxonomic confusionblurs the debate on cosmopolitanism versus local endemismof free-living protists. Protist 156: 263—267

Mitchell EAD, Charman DJ, Warner BG (in press) Testateamoebae analysis in ecological and paleoecological studiesof wetlands: past, present and future. Biodiv Conserv,doi:10.1007/s10531-007-9221-3

Mitchell EAD, Buttler AJ, Warner BG, Gobat JM (1999)Ecology of testate amoebae (Protozoa: Rhizopoda) in Sphag-num peatlands in the Jura mountains, Switzerland andFrance. Ecoscience 6: 565—576

Mitchell EAD, Buttler A, Grosvernier P, Rydin H, AlbinssonC, Greenup AL, Heijmans MMPD, Hoosbeek MR, SaarinenT (2000) Relationships among testate amoebae (Protozoa),vegetation and water chemistry in five Sphagnum-dominatedpeatlands in Europe. New Phytol 145: 95—106

Nikolaev SI, Mitchell EAD, Petrov NB, Berney C, Fahrni J,Pawlowski J (2005) The testate lobose amoebae (orderArcellinida Kent, 1880) finally find their home within Amoebo-zoa. Protist 156: 191—202

Penard E (1902) Les Rhizopodes du bassin du Leman.Kundig, Geneve

Porter SM, Knoll AH (2000) Testate amoebae in theNeoproterozoic Era: evidence from vase-shaped microfossilsin the Chuar Group, Grand Canyon. Paleobiology 26:360—385

Porter SM, Meisterfeld R, Knoll AH (2003) Vase-shapedmicrofossils from the Neoproterozoic Chuar Group, GrandCanyon: a classification guided by modern testate amoebae.J Paleontol 77: 409

Rodriguez F, Oliver JL, Marin A, Medina JR (1990) Thegeneral stochastic model of nucleotide substitution. J TheorBiol 142: 485—501

Smith HG, Wilkinson DM (1987) Biogeography of testaterhizopods in the southern temperate and antarctic zones.Comite Nat. Franc-ais. Rech Antarctiques 58: 83—96

Todorov M (2002) Morphology, biometry and ecology ofNebela bigibbosa Penard, 1890 (Protozoa: Rhizopoda). ActaProtozool 41: 239—244

Van Oye P (1944) Au sujet de la distribution geographique desRhizopodes. Biol Jahrb 11: 83—91

Vucetich MC (1974) Comentarios criticos sobre ArgynniaJung, 1942 (Rhizopoda, Testacea). Neotropica 20: 126—128

Warner BG (1987) Abundance and diversity of testate amoebae(Rhizopoda, Testacea) in Sphagnum peatlands in SouthwesternOntario, Canada. Arch Protistenkd 133: 173—189

12