Zoologica Scripta Molecular phylogeny of Abyssocladia (Cladorhizidae: Poecilosclerida) and Phelloderma (Phellodermidae: Poecilosclerida) suggests a diversification of chelae microscleres in cladorhizid sponges SERGIO VARGAS, DIRK ERPENBECK, CHRISTIAN GÖCKE, KATHRYN A. HALL, JOHN N. A. HOOPER, DORTE JANUSSEN & GERT WÖRHEIDE Submitted: 21 March 2012 Accepted: 2 July 2012 doi:10.1111/j.1463-6409.2012.00560.x Vargas, S., Erpenbeck, D., Göcke, C., Hall, K. A., Hooper, J. N. A., Janussen, D. & Wörheide, G. (2012) Molecular phylogeny of Abyssocladia (Cladorhizidae: Poecilosclerida) and Phelloderma (Phellodermidae: Poecilosclerida) suggests a diversification of chelae microscleres in cladorhizid sponges. —Zoologica Scripta, 00, 000–000. The taxonomic placement of Abyssocladia Lévi, 1964 (Poecilosclerida) is controversial, having been assigned at various times to three different families (Mycalidae, Cladorhizidae and Phellodermidae) in two different suborders (Mycalina and Myxillina, respectively), since its inception in 1964. It shares the general body plan with the carnivorous sponge family Cladorhizidae (Mycalina), including the lack of an aquiferous system. Nevertheless, it also has chela spicules apparently identical to those in Phelloderma Ridley & Dendy 1886 (Phellodermidae, Myxillina). The ongoing debate on the position of Abyssocladia ultimately reduces to a discussion on the use of chelae morphology to infer phylogenetic relationships within Poecilosclerida. Here, we infer the phylogenetic relationships of the genera Phelloderma and Abyssocladia using two independent molecular markers (28S rDNA and COI), showing that Abyssocladia is not closely related to Phelloderma and belongs in Cladorhizidae. We suggest that despite their complexity, chelae morphology can evolve independently in different poecilosclerid lineages and as such might be potentially misleading as indicator of the phylogenetic history of the group. We also provide the first phylogenetic analysis of the carnivorous sponge family Cladorhizidae and give first insights into the evolution of this feeding mode in Poecilosclerida and, more generally, in Porifera. Corresponding author: Gert Wörheide, Department of Earth- & Environmental Sciences, Palaeontology and Geobiology, Ludwig-Maximilians-Universtität München, Richard-Wagner Str. 10, D-80333 München, Germany; GeoBio-CenterLMU, Richard-Wagner Str. 10, D-80333 München, Germany and Bavarian State Collections of Palaeontology and Geology, Richard-Wagner Str. 10, D-80333 München, Germany. E-mail: woerheide@lmu.de Sergio Vargas, Dirk Erpenbeck Department of Earth- & Environmental Sciences, Palaeontology and Geobiology, Ludwig-Maximilians-Universtität München, Richard-Wagner Str. 10, D-80333 München, Germany. E-mail: s.vargas@lrz.uni-muenchen.de Dirk Erpenbeck, GeoBio-CenterLMU, Richard-Wagner Str. 10, D-80333 München, Germany. E-mail: erpenbeck@lmu.de Christian Göcke, Dorte Janussen, Forschungsinstitut und Naturmuseum Senckenberg, Senckenberganlage 25, D-60325 Frankfurt am Main, Germany. E-mails: christian.goecke@senckenberg.de, dorte.janussen@senckenberg.de Kathryn A. Hall, John N.A. Hooper, Queensland Museum, PO Box 3300, South Brisbane, Qld 4101, Australia. E-mail: kathryn.hall@qm.qld.gov.au, john.hooper@qm.qld.gov.au John N.A. Hooper, Eskitis Institute for Cell and Molecular Therapies, Griffith University, Brisbane, Australia. E-mail: john.hooper@qm.qld.gov.au Introduction Within the species-rich demosponge order Poecilosclerida Topsent, 1928, members of the family Cladorhizidae ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters Dendy, 1922 are well known for their carnivorous feeding mode (Vacelet & Boury-Esnault 1995). Carnivorous sponges passively prey on small crustaceans and other 1 Morphological diversification in carnivorous sponges d S. Vargas et al. micro-invertebrates and have a body-plan that is atypical for sponges because it lacks an aquiferous system, or if present, have one which is highly modified and not used for filter feeding. This remarkable trait, in association with its relatively high species richness (123 spp. in seven genera in Cladorhizidae), has made cladorhizid sponges the subject of special attention among sponge biologists (Vacelet & Duport 2004; Riesgo et al. 2007), especially with regard to their taxonomy (e.g. Lehnert et al. 2005, 2006; Vacelet 2008; Vacelet et al. 2009; Ise & Vacelet 2010). From a systematic perspective, Cladorhizidae is a problematic taxon within Poecilosclerida as it lacks any conclusive synapomorphy. Its high diversity in chela morphology – microscleres with alae at each end of a central shaft (Fig. 1 Boury-Esnault & Rützler 1997) – is at odds with the current classification of poecilosclerid sponges. This classification is largely influenced by the assumption that chelae are homologous in that group and can be used to reconstruct phylogenetic relationships because of their morphological complexity and presumed selective neutrality (Hajdu et al. 1994; Hajdu & Van Soest 1996; van Soest 2002; Vacelet 2007). A number of likely evolutionary innovations that have been used to diagnose cladorhizids and could potentially represent ‘strong’ synapomorphies (e.g. carnivory, the absence or major modification of the aquiferous system, a stipitate symmetrical body shape, a special arrangement of megascleres and ⁄ or microscleres, and the presence of sigmancistra among this last spicule type) and have been questioned because taxa in other poecilosclerid families (e.g. Esperiopsis koltuni Ereskovsky & Willenz 2007 in Esperiopsidae Hentschel, 1923; species of Euchelipluma Topsent, 1909 in Guitarridae Dendy, 1924) also exhibit these characters rendering them either homoplasious or symplesiomorphic (Hajdu & Vacelet 2002; Ereskovsky & Willenz 2007). This interpretation, how- Fig. 1 Chelae microscleres present in representatives of the order Poecilosclerida. From left to right: palmate chelae of Clathria australiensis, arcuate chelae of Hamigera dendyi and anchorate chelae of Crella incrustans. S, shaft, A, alae. Images not at scale. Electron microscope photos: J. N. A. Hooper. 2 ever, relies principally on the assumed ‘correctness’ of the currently accepted taxonomic classification of the order Poecilosclerida, which has not been tested to date and has not been supported by independent evidence, for example, in any molecular phylogeny published to date (Erpenbeck & Wörheide 2007; Wörheide et al. 2012). Vacelet (2007) first recognized the important position of cladorhizid sponges within poecilosclerid systematics; he suggested that the microsclere morphological diversity observed in carnivorous sponges could have resulted from its convergent evolution. Cladorhizidae is the only poecilosclerid family that possesses all basic forms of chela (palmate, anchorate and arcuate; Fig. 1). The diversity in chelae morphology within the groups means that Cladorhizidae can fit the current definition of several poecilosclerid families in the suborders Mycalina Hajdu et al. 1994 and Myxillina Hajdu et al. 1994; suggesting that the subordinal classification of Poecilosclerida requires closer scrutiny (as suggested by Erpenbeck & Wörheide 2007). Abyssocladia provides an exemplary case-in-point to test the ‘puzzling’ problem cladorhizid microscleres represent within the current poecilosclerid classification. The position of Abyssocladia has been controversial. Van Soest & Hajdu (2002) included Abyssocladia within the genus Phelloderma Ridley & Dendy 1886 based on the ‘similar and peculiar shape of their isochelae’, changing thereby its subordinal assignment from the Mycalina to the Myxillina. Abyssocladia was, however, relocated in Cladorhizidae by Vacelet (2006) who considered the shape of a single spicule insufficient to justify the (subordinal) transfer. From an evolutionary perspective, testing the phylogenetic position of Abyssocladia, whether in Cladorhizidae (Mycalina) (with which it shares the lack of an aquiferous system, a stipitate symmetrical body and a carnivorous habit, and the presence of mycalostyles and sigmancistra spicules) or in Phellodermidae (Myxillina) (as indicated by the shape of its arcuate chelae), is not a trivial matter (for a discussion see Lopes et al. 2012). There are important implications for the use of established or traditional morphological characters (e.g. chela microscleres) for the classification of poecilosclerid sponges. If chelae are not good indicators of the evolutionary history of the order, any interpretations of its internal phylogenetic structure based on this character may be inaccurate. Here, we present a phylogenetic analysis of the genera Abyssocladia and Phelloderma aimed to clarify their relative position within the order Poecilosclerida. We sequenced two independent molecular markers (i.e. 28S rDNA and COI) for all non-monotypic cladorhizid genera and a new species of Phelloderma collected in the Southern Ocean and assess alternative hypotheses on the position of Abyssocladia advanced by researchers using morphological ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters S. Vargas et al. observations. We provide the first phylogenetic assessment of the family Cladorhizidae, give insights into the evolution of chelae within Cladorhizidae and discuss implications for the systematics of Poecilosclerida. Materials and methods Specimen collection and identification Most specimens used for the present study were collected during the German ANT XXIV ⁄ 2-SYSTCO I Expedition (2007 ⁄ 2008) to the Antarctic Weddell Sea. Specimens were collected mainly using a big trawl (similar to an Agassiz trawl) and a small Rauschert dredge, but some sponges were also collected using a boxcorer or a multicorer. All the specimens were preliminary sorted on board and fixed in 96% ethanol. Cladorhizid specimens were identified to genus level by Dorte Janussen and Alexander Plotkin and are deposited at Senckenberg Museum and Research Institution, Frankfurt, Germany (see Table 1 for details on the catalogue numbers). A brief description of these specimens including SEM or light microscopy photographs is provided in the Supporting information. Phelloderma was identified by Christian Göcke and Eduardo Hajdu, the specimen is deposited at Senckenberg Museum and Research Institution, Frankfurt, Germany (Table 1) and its detailed description will be published elsewhere (Christian Göcke and Dorte Janussen, in prep.). In addition to these specimens, one specimen of Asbestopluma obae was collected in Antarctica during the New Zealand’s National Institute of Water and Atmospheric Research (NIWA) BioRoss (2004) expedition and was identified by Michelle Kelly (NIWA), and a second specimen of this genus, Asbestopluma hypogea Vacelet & Boury-Esnault, 1996 was collected by Jean Vacelet off Marseille (Table 1). Tissue of specimens of monotypic cladorhizid genera was not available to us at the time of the study, thus these genera were not included in the analysis. Molecular methods We obtained genomic DNA using the DNeasy Blood & Tissue kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. After extraction, the D13-E13 domains of the 28S rDNA and the standard barcoding (Folmer) fragment of the mitochondrial cytochrome oxidase 1 gene regions were amplified using primers NL4F and NL4R (Nichols 2005), and dgLCO1490 and dgHCO2198 (Meyer et al. 2005), respectively. PCR products were either sequenced directly after EXOSAP-IT (Affymetrix, Santa Clara, California, USA) clean-up or cloned using the TOPO TA Cloning kit (Invitrogen, Carlsbad, California, USA). Positive colonies (at least 8) were picked and boiled in 10 lL HPLC grade water for 5 min to release the DNA; DNA was then amplified using T3 ⁄ T7 ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters d Morphological diversification in carnivorous sponges primers. PCR products of the expected size were excised from a 1.5% agarose gel and sequenced in both directions using the T3 ⁄ T7 primers and the BIGDYE Terminator version 3.1 chemistry (Applied Biosystems, Foster City, California, USA). Sequencing reactions were precipitated using standard ethanol-sodium acetate precipitation and analysed on an ABI 3700 Genetic Analyzer at the Genomic Sequencing Unit of the Department of Biology, LMU München. The resulting chromatograms were visualized and assembled in CodonCode Aligner (Codon Code Corporation, Dedham, Massachusetts, USA). Poriferan origin of all sequences was determined using NCBI BLAST (Johnson et al. 2008). Table 1 contains the EMBL sequence database accession numbers for each sequence generated in this study. Phylogenetic analyses New sequences were manually aligned to existing demosponge (structurally annotated) 28S rDNA and (Folmer) COI (for details see http://www.spongegenetrees.org/ Erpenbeck et al. 2007, 2004, 2008) data matrices. In addition, both data sets were supplemented with poecilosclerid sequences generated for the Sponge Barcoding Project (Table 1; http://www.spongebarcoding.org; Wörheide & Erpenbeck 2007). We included new sequences of the poecilosclerid genera Neofibularia (Desmacellidae) and Rhabderemia (Rhabderemiidae). Previous molecular studies have shown that these genera are not closely related to poecilosclerid sponges bearing chelae microscleres (Erpenbeck et al. 2007). We, therefore, used a comprehensive sample of demosponge genera and families in the analysis of 28S rDNA and COI. As this broad taxon sampling can result in the exclusion of important phylogenetic information (unalignable regions) in the analysis of the 28S rDNA, we ran ML and Bayesian analyses using a restricted data set including only chelae-bearing poecilosclerids and using representatives of the family Clionaidae as outgroup for Poecilosclerida (Morrow et al. 2012). Phylogenetic analyses were performed on each data matrix separately. For COI, the computer programs RAXML 7.2.8 (Stamatakis 2006) and MRBAYES 3.1.2 (Ronquist & Huelsenbeck 2003) were used to infer a bootstrapped (1000 fast bootstrap pseudoreplicates; Stamatakis et al. 2008) maximum likelihood (ML) and a Bayesian phylogenetic tree, respectively. The GTR model of sequence evolution (Tavaré 1986) was used in both analyses, and among-site rate variation was modelled using a discrete Gamma with four rate categories (Yang 1994). The GTR model is the only DNA model available in RAXML, we used this same model for the Bayesian analysis to facilitate comparisons between ML and Bayesian phylogenies and because over-parameterization, normally, does not negatively 3 Morphological diversification in carnivorous sponges d S. Vargas et al. Table 1 New poecilosclerid specimens sampled for this study. Institution: NIWA, National Institute of Water and Atmospheric Research, Wellington, New Zealand; SMF, Senckenberg Museum Frankfurt, Frankfurt am Main, Germany; QM, Queensland Museum, Brisbane, Australia. SBP, Sponge Barcoding Project Location Accession number Suborder Family ID Institution ⁄ Voucher-No. Latitude Longitude 28S DNA COI SBP Number Mycalina Cladorhizidae Mycalidae Abyssocladia sp. Asbestopluma hypogea Asbestopluma obae Cladorhiza sp. Cladorhiza sp. Chondrocladia sp. Neofibularia hartmani Neofibularia irata Mycale mirabilis Microcionidae Clathria abietina SMF ⁄ 11750 – NIWA ⁄ 28893 SMF 11753 SMF ⁄ 11751 SMF ⁄ 11752 QM ⁄ G306606 QM ⁄ G307266 QM ⁄ G300561 QM ⁄ G306269 QM ⁄ G307148 QM ⁄ G305553 QM ⁄ G300508 QM ⁄ G306048 QM ⁄ G306154 QM ⁄ G304846 QM ⁄ G311840 QM ⁄ G306423 QM ⁄ G307278 QM ⁄ G314045 QM ⁄ G305498 QM ⁄ G305135 QM ⁄ G304980 QM ⁄ G300696 QM ⁄ G300541 QM ⁄ G300607 QM ⁄ G312904 QM ⁄ G300820 QM ⁄ G300289 QM ⁄ G306063 QM ⁄ G313145 QM ⁄ G301219 QM ⁄ G304489 QM ⁄ G305838 QM ⁄ G301225 QM ⁄ G304373 QM ⁄ G304056 QM ⁄ G300194 QM ⁄ G305473 QM ⁄ G307601 QM ⁄ G304247 QM ⁄ G305520 QM ⁄ G304718 QM ⁄ G306386 QM ⁄ G313319 QM ⁄ G304254 QM ⁄ G304256 QM ⁄ G300550 SMF ⁄ 11726 – – )71.7675 79.1333 )62.0145 )70.0669 )22.2011 )23.4669 )25.6167 )25.3336 )23.4514 )21.7847 )11.4517 )19.4672 )19.9333 40.3050 )28.4869 7.2842 )23.4669 )23.8939 )21.7847 )14.8192 )22.3347 )22.3000 )12.4000 )20.5667 )10.2183 )10.4167 )11.8000 )19.6853 1.2189 1.1667 4.6667 9.6850 40.4269 )14.5836 )35.0834 )28.3000 )21.9686 )23.4347 )14.4675 )21.7847 )10.2675 7.2514 )15.3619 )14.7006 )14.7006 )28.7839 )64.48 – – 171.1092 4.4833 )2.9833 )3.3267 155.2353 151.9342 113.3833 153.0181 151.9019 152.4511 136.4686 117.5675 118.2167 151.5000 113.5639 134.5014 151.9342 252.4111 152.4503 145.5192 152.7178 166.4167 130.8167 149.0834 148.1025 141.7678 149.2667 118.1014 103.8347 103.7500 119.0167 118.7347 116.7833 145.3356 137.7347 113.8000 152.4675 151.8850 145.4358 152.4503 124.0342 134.3181 137.5550 145.4347 145.4347 113.9681 2.8753 HE611627 HE611628 – HE611629 HE611630 HE611631 – HE611632 – HE611633 – – – HE611634 – – – – – – – – – HE611635 – – – – – HE611636 HE611637 HE611638 HE611639 HE611640 HE611641 – – HE611642 – HE611643 HE611644 – HE611645 – – – – – HE611646 HE611581 HE611582 HE611583 HE611584 HE611585 HE611586 HE611587 HE611588 HE611589 HE611590 HE611591 HE611592 HE611593 HE611594 HE611595 HE611596 HE611597 HE611598 HE611599 HE611600 HE611601 HE611602 HE611603 HE611604 HE611605 HE611606 HE611607 HE611608 HE611609 HE611610 HE611611 HE611612 HE611613 – – HE611614 HE611615 – HE611616 HE611617 HE611618 HE611619 HE611620 HE611621 HE611622 HE611623 HE611624 HE611625 HE611626 – – – – – – SBP642 SBP650 SBP572 SBP646 SBP645 SBP731 SBP756 SBP745 SBP738 SBP741 SBP1022 SBP643 SBP641 SBP1023 SBP739 SBP742 SBP743 SBP652 SBP570 SBP579 SBP1026 SBP653 SBP682 SBP736 SBP1027 SBP723 SBP688 SBP730 SBP729 SBP696 – – SBP732 SBP648 SBP700 SBP737 SBP687 SBP647 SBP1025 SBP690 SBP701 SBP571 – Desmacellidae Microcionina Clathria Clathria Clathria Clathria cervicornis cancellaria reinwardti kylista Clathria conectens Acarnidae Myxillina Rhabderemiidae Coelosphaeridae Crambeidae Clathria rugosa Paracornulum dubium Paracornulum sp. 819 Rhabderemia sorokinae Lissodendoryx sp. 489 Monanchora sp. 994 Monanchora sp. 0605 Monanchora clathrata Hymedesmiidae Crella sp. 4778 Crella incrustans Crella spinulata Phorbas fictitioides Iotrochotidae Phorbas sp. 1539 Iotrochota baculifera Crellidae Iotrochota coccinea Phellodermidae Iotrochota acerata Phelloderma sp. affect Bayesian phylogenetic inference (Huelsenbeck & Rannala 2004). For the Bayesian analysis, two independent runs with one cold and five heated Metropolis4 Coupled Monte-Carlo Markov Chain (MCMCMC) chains each were set to sample trees every 500 generations for a total of 10 000 000 generations using the default prior and ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters S. Vargas et al. temperature settings available in MRBAYES 3.1.2. After completion, 25% of the samples were discarded as burn-in and a 50% majority-rule consensus tree was calculated. Chain convergence was assessed using AWTY (Wilgenbusch et al. 2004). With the 28S rDNA data set, we inferred a bootstrapped (1000 fast pseudo-replicates) ML and a Bayesian phylogeny with the programs RAXML 7.2.8 and PHASE 2.0 (http://www.bioinf.manchester.ac.uk/resources/phase/), respectively. For the analysis of this data set, we used the second most general structural model available (i.e. S7A see Savill et al. 2001) in PHASE for helices (paired-sites) and the GTR + GAMMA model for the unpaired sites of the RNA molecule. Structural annotation of the 28S rDNA data set was carried out following Erpenbeck et al. (2004) & Erpenbeck et al. (2007). For the Bayesian analysis, two independent PHASE 2.0 Monte-Carlo Markov Chain (MCMC) runs were set to sample every 500 generations for a total of 10 000 000 generations; prior to each run, 500,000 generations were discarded as burn-in. To evaluate the probability of alternative topological arrangements for Abyssocladia and Phelloderma as well as other groupings proposed in the taxonomic literature, trees sampled during the COI Bayesian analyses were imported into PAUP*4.0 (Swofford 2003) and filtered using constraints corresponding to the different morphology-based hypotheses. In a Bayesian context, the posterior probability of a bipartition is the frequency with which the bipartition occurs in the trees sampled from the posterior distribution during MCMC (Lewis 2001). Results COI phylogeny The inferred COI Bayesian phylogeny (Supporting information) supported the monophyly of chelae-bearing poecilosclerids and, from this order, the exclusion of several genera that do not bear chelae, such as Rhabderemia Topsent, 1890 (Rhabderemiidae, Microcionina) and the desmacellid genera Biemna Gray, 1867 and Neofibularia Hechtel, 1965 (Mycalina). In contrast, the ML phylogeny (Supporting information) inferred a polyphyletic chelaebearing Poecilosclerida with representatives of the genera Crambe Vosmaer, 1880 and Monanchora Carter, 1883 forming a poorly supported clade with sequences attributed to species of Niphates Duchassaing & Michelotti, 1864 (Haplosclerida) and Scopalina Schmidt, 1862 (Halichondrida). In general, both COI phylogenies (Fig. 2) showed at least some level of disagreement with the current classification of poecilosclerid sponges (sensu Hooper & van Soest 2002), with most suborders – Suborder Latrunculina Kelly & Samaai, 2002 was not included in the present analysis – not recovered as monophyletic in either ML or Bayesian trees. Suborder Myxillina was polyphyletic in all trees sampled in the Bayesian MCMCMC, comprising three inde- ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters d Morphological diversification in carnivorous sponges pendent clades. Relationships among chelae-bearing Microcionina Hajdu et al. 1994 were unresolved in the Bayesian consensus tree but the monophyly of these taxa received low posterior probability (P = 0.2061). Within Mycalina, species of Mycale (Arenochalina) and Cladorhizidae were not related to members of Podospongiidae de Laubenfels, 1936, which was polyphyletic. Abyssocladia was included within Cladorhizidae, which formed the sister clade of Mycale (Arenochalina) (with high support in both the ML and Bayesian trees). Phelloderma (Myxillina) was included in a clade with Crella, Phorbas, Strongylacidon and a specimen labelled as Timea sp. (Supporting information). This result was also obtained after removal of Timea sp. (Fig. 2), which likely represents a contamination or a misidentification deposited in GenBank. Finally, a topology consistent with the monophyly of Abyssocladia + Phelloderma was not found among the trees (N = 15 001) sampled during one arbitrarily chosen chain of the MCMCMC of the Bayesian analysis. 28S rDNA analysis Despite the lesser number of Poecilosclerida samples that were available for the 28 rDNA analysis, the results from the 28S rDNA corroborated those from COI (Fig. 3). Poecilosclerid genera lacking chelae (i.e. Rhabderemia and Neofibularia in the present data set) were not recovered as being related to chelae-bearing genera in either the ML or Bayesian phylogenies (Supporting information). Chelaebearing taxa (Fig. 3) belonging to different suborders formed clades that contradicted the currently hypothesized subordinal classification of Poecilosclerida. For example, Coelocarteria Burton, 1934 (Mycalina) was not related to Cladorhizidae + Mycale (Arenochalina), but rather to representatives of Clathria (Microcionina). Regarding the phylogeny of Abyssocladia and Phelloderma, these two genera were not indicated as closely related in the 28S rDNA ML or Bayesian phylogenetic trees. In accordance with the COI results, Abyssocladia was included within the family Cladorhizidae which, together with Mycale (Arenochalina), formed a highly supported clade. Phelloderma, in contrast, was related to representatives of Lissodendoryx, Crella and Phorbas Duchassaing & Michelotti, 1864 (Myxillina); all these genera formed a highly supported monophylum. These results were independent of taxonomic sampling or outgroup choice as the analysis of either the full data set (see Materials and Methods) or a reduced data set (Fig. 3) excluding sequences labelled as Geodia spp., which have been shown to be contaminations (Cárdenas et al. 2010), and Tethya sp., which can be either a contamination or a misidentification deposited in GenBank, resulted in a topology congruent with the inclusion of Abyssocladia within a monophyletic Cladorhizidae, far from Phelloderma. 5 Morphological diversification in carnivorous sponges d S. Vargas et al. Fig. 2 Phylogenetic hypothesis of chelaebearing poecilosclerid sponges based on partial COI sequences. The topology is based on the results of the Bayesian analysis. Support values are given only for branches with posterior probability ‡0.95. Above the branches an asterisk indicates posterior probability ‡0.95 (left asterisk), Maximum Likelihood bootstrap proportion ‡70% (right asterisk). Vertical bars indicate subordinal membership: Myc, Mycalina; Mic, Microcionina; Myx, Myxillina. The gray boxes correspond to the genera belonging to the families Cladorhizidae and Phellodermidae. The genera Abyssocladia and Phelloderma are in bold face. SBP# refer to the Sponge Barcoding Project reference numbers, accession numbers follow the Sponge Gene Tree Server formats. The complete Bayesian and Maximum Likelihood phylogenetic trees with support values and branch lengths are provided as Supporting information. 6 ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters S. Vargas et al. d Morphological diversification in carnivorous sponges Fig. 3 Phylogenetic hypothesis of chelae-bearing poecilosclerid sponges inferred from structurally annotated partial 28S rDNA sequences. The topology is based on the results of the Bayesian analysis. Support values are given only for branches with posterior probability ‡0.95. Above the branches an asterisk indicate posterior probability ‡0.95 (left asterisk), Maximum Likelihood bootstrap proportion ‡70% (right asterisk). Vertical bars indicate subordinal membership: Myc, Mycalina; Mic, Microcionina; Myx, Myxillina. Highlighted in gray genera belonging to the families Cladorhizidae and Phellodermidae. The genera Abyssocladia and Phelloderma are in bold face. SBP# refer to the Sponge Barcoding Project reference numbers, accession numbers follow the Sponge Gene Tree Server formats. The complete Bayesian and Maximum Likelihood phylogenetic trees with support values and branch lengths are provided as Supporting information. Discussion The current classification of poecilosclerid sponges is based principally on the assumption that the morphology of ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters chela microscleres reflects the phylogenetic history (cf. van Soest 2002). Nevertheless, chelae morphology can be homoplasic – as is the case of the macro- and microscleres 7 Morphological diversification in carnivorous sponges d S. Vargas et al. of other sponge groups (Cárdenas et al. 2011) – and their presence ⁄ absence has been demonstrated to be environmentally plastic in at least one poecilosclerid genus (see Maldonado et al. 1999). The alternative taxonomic positions proposed for Abyssocladia (see Introduction and Vacelet 2006; van Soest & Hajdu 2002) represent an almost perfect case study on the diversity of interpretations of chelae significance for the systematics of Poecilosclerida. Based on the morphology of chelae, Abyssocladia should belong in Phellodermidae (van Soest & Hajdu 2002). However, the lack of an aquiferous system, the overall body shape, and the spicule complement and skeletal organization suggest affinities of the genus to Cladorhizidae (Vacelet 2006). Here, we have shown that Phelloderma and Abyssocladia are not closely related, and that Abyssocladia belongs in Cladorhizidae. This family was recovered as monophyletic, with high bootstrap support and posterior probability in our analysis of COI. In the 28S rRNA ML and Bayesian phylogenies, Cladorhizidae received low bootstrap support and posterior probabilities. It is worth noting, however, that the uncertainty regarding the monophyly of Cladorhizidae in these analyses was caused by the unstable position of Mycale (Arenochalina) in the Bayesian phylogeny and its inclusion within Cladorhizidae in the ML analysis and was not caused by the exclusion of any cladorhizid genera from the family. Additionally, the analysis of a data set restricted to chelae-bearing poecilosclerids (see Supporting information) resulted in a monophyletic Cladorhizidae sister to Mycale (Arenochalina) in agreement with the COI results. With respect to the relationships between Abyssocladia and Phelloderma, a topology compatible with an Abyssocladia + Phelloderma clade was not sampled during the Bayesian analysis of COI or 28S rRNA sequences, indicating that the posterior probability of a tree including this clade is negligible in comparison with that of the optimal and near-optimal trees – in other words, the hypothesized clade is artificial. Importantly, this result was not sensitive to taxon sampling or outgroup choice as revealed by the analysis of different matrices with diverse taxonomic composition and by the close phylogenetic relationship between Abyssocladia and Cladorhizidae found in trees showing different sister group relationships for Poecilosclerida. From a morphological perspective, the phylogenetic results should not be surprising as these genera differ greatly in their body shape; Abyssocladia is stalked with a spherical or disc-shaped body (Vacelet 2006), while Phelloderma is subglobular with a cork-like cortex, papillae and a general skeletal organization resembling the family Suberitidae (Ridley & Dendy 1886; van Soest & Hajdu 2002). Regarding the relationships of Cladorhizidae, this family was shown as sister to the mycalid Mycale (Arenochalina) as 8 noted above, however, this sister group relationship should be treated with caution pending a more complete taxonomic sampling within ‘Mycalina’, and generally within (chelae-bearing) Poecilosclerida. From an evolutionary perspective, our results indicate that a diverse complement of chelae have been independently acquired in the cladorhizid lineage. Owing to their functional role in the capture of prey (Vacelet & Duport 2004), it is likely that cladorhizid chelae are acted upon by strong selective pressures. Thus, it is probable that the carnivorous habit of cladorhizid sponges has led to the diversity of chelae forms observed among modern representatives of this family (see Introduction and Price et al. 2010 for an example). More speculatively, innovations in chelae morphology within cladorhizid genera might have had an impact on the speciation rates within the family, as judged by its relatively high species richness (see Introduction) and species-specific chelae morphologies. At present, testing this hypothesis is not possible because of the sparse intra-genus sampling available. A more thorough sampling of the family and of other carnivorous sponges currently classified in different poecilosclerid families (Esperiopsidae and Guitarridae) will be required to establish a correlation (if any) between carnivory, chelae morphological diversity and speciation rates in Cladorhizidae, and carnivorous sponges in general. Our findings have broader implications on the biology of Poecilosclerida beyond the phylogenetic position of Abyssocladia and Phelloderma. Several hypotheses regarding the evolution of a carnivorous feeding mode in Poecilosclerida have been advanced by sponge biologists (see Vacelet 2007 for a review). The phylogenetic analyses presented in this study suggest that a carnivorous feeding mode and the morphological innovations associated with it evolved once in poecilosclerid sponges. However, species currently classified in two other genera of non-cladorhizid poecilosclerids, Euchelipluma (Guitarridae) and Esperiopsis (Esperiopsidae), have been associated with a carnivorous feeding mode. The phylogenetic position of these two poecilosclerid genera remains to be determined to corroborate the monophyly of all carnivorous sponges suggested by their general body plan and the shared presence of sigmancistra spicules. A broader analysis of this peculiar sponge group will further clarify the evolution of carnivory in the diverse order Poecilosclerida and in sponges in general. Conclusions We have presented a phylogeny of the genera Abyssocladia and Phelloderma using independent molecular data sets, clarified their relative position within the order Poecilosclerida and evaluated hypotheses concerning their morphological characters. Our molecular data support the ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters S. Vargas et al. proposal of Vacelet (2006) that Abyssocladia is a cladorhizid, and that this family is monophyletic. Conversely, Phelloderma formed a clade with sponges now classified in the suborder Myxillina, and as such, Phelloderma and Abyssocladia are only distantly related, contrary to previous morphological hypotheses based on chelae morphologies (van Soest & Hajdu 2002). This result implies the independent acquisition of a diverse chelae complement in Cladorhizidae, likely resulting from an evolutionary diversification related to the carnivorous habit of cladorhizid sponges. These findings have significant implications for the systematics of the Poecilosclerida, as foreseen by Vacelet (2006), and in particular the alleged taxonomic importance of chela morphotypes for higher taxonomy. The present analyses provide only a small contribution to a reevaluation of Poecilosclerida using two independent molecular markers, but highlight the significance and growing potential of molecular-based approaches to reciprocally illuminate morphological evidence and to better resolve the evolutionary history of Poecilosclerida – the most diverse order of Porifera. Acknowledgements We thank Annamarie Gabrenya and Astrid Schuster for assistance and support in the laboratory. Jean Vacelet kindly provided the specimen of Asbestopluma hypogea and reviewed the manuscript. We thank Alexander Plotkin for his help identifying some of the specimens deposited at SMF and for doing SEM photographs of most of the material. Eduardo Hajdu helped with the identification of the new species of Phelloderma. Constructive comments of past and present members of the Molecular Geo- and Palaeobiology Lab, LMU München greatly improved the project. This study was possible thanks to the funding of the German Science Foundation (DFG), through grants Wo896 ⁄ 9-1,2 to G. Wörheide and JA1063 ⁄ 14-1,2 to D. Janussen, respectively. We also want to thank the NIWA Invertebrate Collection for providing access to samples from the TAN0402 cruise of the Ministry of Fisheries, NIWA under the BioRoss biodiversity survey of the western Ross Sea and Balleny Islands, and to Michelle Kelly (NIWA) for identifying the specimens. The Sponge Barcoding Project thanks the Marine Barcode of Life initiative (MarBol) funded by the Alfred P. Sloan Foundation and the GeoBio-CenterLMU for financing the subsampling and extraction of 17.000 sponge specimens of the Queensland Museum. S. Vargas is indebted to N. Villalobos Trigueros and to M. Vargas Villalobos for their constant support during the course of the study. Constructive comments from Manuel Maldonado and three anonymous reviewers greatly improved the manuscript, we thank them. ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters d Morphological diversification in carnivorous sponges References Boury-Esnault, N. & Rützler, K. (1997). Thesaurus of sponge morphology. Smithsonian Contributions to Zoology, 596, 1–55. Cárdenas, P., Rapp, H. T., Schander, C. & Tendal, O. S. (2010). Molecular taxonomy and phylogeny of the Geodiidae (Porifera, Demospongiae, Astrophorida) a combining phylogenetic and Linnaean classification. Zoologica Scripta, 39, 89–106. Cárdenas, P., Xavier, J. R., Reveillaud, J., Schander, C. & Rapp, H. T. (2011). Molecular phylogeny of the Astrophorida (Porifera, Demospongiae) reveals an unexpected high level of spicule homoplasy. PLoS ONE, 6, e18318. Ereskovsky, A. V. & Willenz, P. (2007). Esperiopsis koltuni sp. nov. (Demospongiae: Poecilosclerida: Esperiopsidae), a carnivorous sponge from deep water of the sea of Okhotsk (North Pacific). Journal of the Marine Biology Association of the U.K., 87, 1–8. Erpenbeck, D., McCormack, G., Breeuwer, J. & van Soest, R. (2004). Order level differences in the structure of partial LSU across demosponges (Porifera): new insights into an old taxon. Molecular Phylogenetics and Evolution, 32, 388–395. Erpenbeck, D., Duran, S., Rützler, K., Paul, V., Hooper, J. N. A. & Wörheide, G. (2007). Towards a DNA taxonomy of Caribbean Demosponges: a gene tree reconstructed from partial mitochondrial COI gene sequences supports previous rDNA phylogenies and provides a new perspective on the systematics of Demospongiae. Journal of the Marine Biology Association of the U.K., 87, 1–8. Erpenbeck, D., Nichols, S. A., Voigt, O., Dohrmann, M., Degnan, B. M., Hooper, J. N. A., Wörheide, G. (2007). Phylogenetic analyses under secondary structure-specific substitution models outperform traditional approaches: Case studies with diploblast LSU. J Mol Evol, 64, 543–557. Erpenbeck, D., Voigt, O., Gültas, M. & Wörheide, G. (2008). The sponge genetree server-providing a phylogenetic backbone for poriferan evolutionary studies. Zootaxa, 1939, 58–60. Hajdu, E. & Vacelet, J. (2002). Cladorhizidae. In J. N. A. Hooper & R. W. M. van Soest (Eds) Systema Porifera: A Guide to the Classification of Sponges. (pp. 636–641). New York: Kluwer Academic ⁄ Plenum Publishers. Hajdu, E. & Van Soest, R. W. M. (1996). Choosing among poriferan morphological characters within the cladistic paradigm. Bulletin van het Koninklijk Belgisch Instituut voor Natuurwetenschappen. Biologie= Bulletin de l’Institut Royal des Sciences Naturelles de Belgique. Biologie, 66, 81–88. Hajdu, E., Van Soest, R. & Hooper, J. (1994). Proposal for a phylogenetic subordinal classification of poecilosclerid sponges. In R. W. M. van Soest, T. M. G. van Kempen & J. C. Braekman (Eds) Sponges in Time and Space: Biology, Chemistry, Paleontology. (pp. 123–139). Amsterdam: A. A. Balkema. Hooper, J. N. A. & van Soest, R. W. M. (2002). Poecilosclerida. In J. Hooper & R. W. M. van Soest (Eds) Systema Porifera: A Guide to the Classification of Sponges. (pp. 403–408). New York: Kluwer Academics ⁄ Plenum Publishers. Huelsenbeck, J. & Rannala, B. (2004). Frequentist properties of bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Systematic Biology, 53, 904–913. Ise, Y. & Vacelet, J. (2010). New carnivorous sponges of the genus Abyssocladia (Demospongiae, Poecilosclerida, Cladorhizidae) 9 Morphological diversification in carnivorous sponges d S. Vargas et al. from Myojin Knoll, Izu-Ogasawara arc, southern Japan. Zoological Sciences, 27, 888–894. Johnson, M., Zaretskaya, I., Raytselis, Y., Merezhuk, Y., Mcginnis, S. & Madden, T. L. (2008). NCB Blast: a better web interface. Nucleic Acids Research, 36, W5–W9. Lehnert, H., Watling, L. & Stone, R. (2005). Cladorhiza corona sp. nov.(Porifera: Demospongiae: Cladorhizidae) from the Aleutian islands (Alaska). Journal of the Marine Biology Association of the U.K., 85, 1359–1366. Lehnert, H., Stone, R. & Heimler, W. (2006). New species of deep-sea demosponges (Porifera) from the Aleutian islands (Alaska, USA). Zootaxa, 1250, 1–35. Lewis, P. (2001). Phylogenetic systematics turns over a new leaf. Trends in Ecology and Evolution, 16, 30–37. Lopes, D. A., Bravo, A. & Hajdu, E. (2012). New carnivorous sponges (Cladorhizidae: Poecilosclerida: Demospongiae) from off Diego Ramı́rez archipelago (South Chile), with comments on taxonomy and biogeography of the family. Invertebrate Systematics, 25, 407–443. Maldonado, M., Carmona, M., Uriz, M. & Cruzado, A. (1999). Decline in mesozoic reef-building sponges explained by silicon limitation. Nature, 401, 785–788. Meyer, C., Geller, J. & Paulay, G. (2005). Fine scale endemism on coral reefs: archipelagic differentiation in turbinid gastropods. Evolution, 59, 113–125. Morrow, C. C., Picton, B. E., Erpenbeck, D., Boury-Esnault, N., Maggs, C. A. & Allcock, A. L. (2012). Congruence between nuclear and mitochondrial genes in Demospongiae: a new hypothesis for relationships within the G4 clade (Porifera: Demospongiae). Molecular Phylogenetics and Evolution, 62, 174– 190. Nichols, S. A. (2005). An evaluation of support for order-level monophyly and interrelationships within the class Demospongiae using partial data from the large subunit rDNA and cytochrome oxidase subunit I. Molecular Phylogenetics and Evolution, 34, 81–96. Price, S. A., Wainwright, P. C., Bellwood, D. R., Kazancioglu, E., Collar, D. C. & Near, T. J. (2010). Functional innovations and morphological diversification in parrotfish. Evolution, 64, 3057– 3068. Ridley, S. O. & Dendy, A. (1886). Preliminary report on the Monaxonida collected by h.M.S. ‘Challenger’. Annals and Magazine of Natural History, 5, 325–351, , 470-493. Riesgo, A., Taylor, C. & Leys, S. P. (2007). Reproduction in a carnivorous sponge: the significance of the absence of an aquiferous system to the sponge body plan. Evolution and Development, 9, 618–631. Ronquist, F. & Huelsenbeck, J. (2003). Mrbayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics, 19, 1572–1574. Savill, N., Hoyle, D. & Higgs, P. (2001). RNA sequence evolution with secondary structure constraints: comparison of substitution rate models using maximum-likelihood methods. Genetics, 157, 399–411. van Soest, R. W. M. (2002). Myxillina. In J. N. A. Hooper & R. W. M. van Soest (Eds) Systema Porifera: A Guide to the Classification of Sponges. (pp. 515–520). New York: Kluwer Academics ⁄ Plenum Publishers. van Soest, R. W. M. & Hajdu, E. (2002). Phellodermidae. In J. N. A. Hooper & R. W. M. van Soest (Eds) Systema Porifera: A 10 Guide to the Classification of Sponges. (pp. 621–624). New York: Kluwer Academic ⁄ Plenum Publishers. Stamatakis, A. (2006). RAxML-VI-HPC: maximum likelihoodbased phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 22, 2688–2690. Stamatakis, A., Hoover, P. & Rougemont, J. (2008). A rapid bootstrap algorithm for the raxml web servers. Systematic Biology, 57, 758–771. Swofford, D. L. (2003). PAUP*: Phylogenetic Analysis Using Parsimony (* and other methods), Version 4. Sunderland: Sinauer Associates. Tavaré, S. (1986). Some probabilistic and statistical problems in the analysis of DNA sequences. Lectures on Mathematics in the Life Sciences, 17, 57–86. Vacelet, J. (2006). New carnivorous sponges (Porifera, Poecilosclerida) collected from manned submersibles in the deep pacific. Zoological Journal of the Linnean Society, 148, 553–584. Vacelet, J. (2007). Diversity and evolution of deep-sea carnivorous sponges. In M. Custodio, G. Lôbo-Hajdu, E. Hajdu & G. Muricy (Eds) Porifera Research: Biodiversity, Innovation and Sustainability. (pp. 107–115). Rio de Janeiro: Museu Nacional. Vacelet, J. (2008). A new genus of carnivorous sponges (Porifera: Poecilosclerida, Cladorhizidae) from the deep NE Pacific, and remarks on the genus Neocladia. Zootaxa, 1752, 57–65. Vacelet, J. & Boury-Esnault, N. (1995). Carnivorous sponges. Nature, 373, 333–335. Vacelet, J. & Duport, E. (2004). Prey capture and digestion in the carnivorous sponge Asbestopluma hypogea (Porifera : Demospongiae). Zoomorphology, 123, 179–190. Vacelet, J., Kelly, M. & Schlacher-Hoenlinger, M. (2009). Two new species of Chondrocladia (Demospongiae: Cladorhizidae) with a new spicule type from the deep South Pacific, and a discussion of the genus Meliiderma. Zootaxa, 2073, 57–68. Wilgenbusch, J. C., Warren, D. L. & Swofford, D. L. (2004). AWTY: A system for graphical exploration of MCMC convergence in Bayesian phylogenetic inference. http:// ceb.csit.fsu.edu/awty. Wörheide, G. & Erpenbeck, D. (2007). DNA taxonomy of sponges—progress and perspectives. Journal of the Marine Biology Association of the U.K., 87, 1–6. Wörheide, G., Dohrmann, M., Erpenbeck, D., Larroux, C., Maldonado, M., Voigt, O., Borchiellini, C. & Lavrov, D. V. (2012). Deep phylogeny and evolution of sponges (Phylum Porifera). In: M. A. Becerro, J. M. Uriz, M. Maldonado & J. Turon (Eds). Advances in Marine Biology, 61, 1–78. Yang, Z. (1994). Maximum-likelihood phylogenetic estimation from DNA-sequences with variable rates over sites—approximate methods. Journal of Molecular Evolution, 39, 306–314. Supporting information Additional Supporting Information may be found in the online version of this article: Fig. S1. COI Bayesian Tree inferred with MRBAYES excluding Timea sp. from GenBank. Fig. S2. COI ML Tree inferred with RAXML excluding Timea sp. from GenBank. Fig. S3. LSU Bayesian phylogeny inferred using PHASE 2.0 and the 7A model of sequence evolution. ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters S. Vargas et al. Fig. S4. LSU phylogeny inferred with RAXML and the 7SA model of sequence evolution. Fig. S5. LSU phylogeny inferred with PHASE 2.0 and the 7SA model of sequence evolution. This analysis excludes specimens of Geodia and Tethya which represent contaminants or missidentified specimens. Fig. S6. The analysis excludes specimens of Geodia and Tethya that represents missidentifications or contaminants. Data S1. Genus assignment of Cladorhizidae used in: Vargas et al. Molecular phylogeny of Abyssocladia (Clad- ª 2012 The Authors d Zoologica Scripta ª 2012 The Norwegian Academy of Science and Letters d Morphological diversification in carnivorous sponges orhizidae: Poecilosclerida) and Phelloderma (Phellodermidae: Poecilosclerida) suggest a diversification of chelae microscleres in cladorhizid sponges. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. 11