Molecular phylogeny of Demospongiae: implications for classification and scenarios of character evolution

An analysis of the phylogenetic relationships of the 13 orders of Demospongiae, based on 18S and C1, D1 and C2 domains of 28S rRNA (for, respectively, 26 and 32 taxa) has been performed. The class Demospongiae as traditionally defined is not found to be monophyletic. Instead, a clade comprising all demosponges except Homoscleromorpha is well-supported, and we define phylogenetically the name Demospongiae in this more restricted sense to preclude the possibility of drastic alterations of the meaning of Demospongiae in the future, depending on the position of Homoscleromorpha. Within this clade Demospongiae s.s., ceractinomorphs and tetractinomorphs are polyphyletic, implying homoplastic evolution of characters such as reproductive strategies (viviparity vs. oviparity) and skeleton architecture (reticulate vs. radiate). The topology derived from our molecular data provides a basis for proposing a new classification of Demospongiae s.s., and suggests a reverse polarity of some characters, with respect to traditional conceptions: viviparity, presence of monaxon spicules and of spongin appear to be ancestral, whereas oviparity, and presence of tetraxon spicules appear as derived characters.     

Molecular systematics of sponges (Porifera)

The first application of molecular systematics to sponges was in the 1980s, using allozyme divergence to dis-criminate between conspecific and congeneric sponge populations. Since this time, a fairly large database has been accumulated and, although the first findings seemed to indicate that sponge species were genetically more divergent than those of other marine invertebrates, a recent review of the available dataset indicates that levels of interspecific gene identities in most sponges fall within the normal range found between species of other invertebrates. Nevertheless, some sponge genera have species that are extremely divergent from each other, suggesting a possible polyphyly of these genera. In the 1990s, molecular studies comparing sequences of ribosomal RNA have been used to reappraise the phylogenetic relationships among sponge genera, families, orders and classes. Both the 18S small subunit and the 28S large subunit rRNA genes have been sequenced (41 complete or partial and 75 partial sequences, respectively). Sequences of 18S rRNA show good support for Porifera being true Metazoa, but they are not informative for resolving relationships among genera, families or orders. 28S rRNA domains D1 and D2 appear to be more informative for the terminal nodes and provide resolution for internal topologies in sufficiently closely related species, but the deep nodes between orders or classes cannot be resolved using this molecule. Recently, a more conserved gene, Hsp70, has been used to try to resolve the relationships in the deep nodes. Metazoan monophyly is very well supported. Nevertheless, the divergence between the three classes of Porifera, as well as the divergence between Porifera, Cnidaria and Ctenophora, is not resolved. Research is in progress using other genes such as those of the homeodomain, the tyrosine kinase domain, and those coding for the aggregation factor. For the moment the dataset for these genes is too restricted to resolve the phylogenetic relationships of these phyla. However, whichever the genes, the phylogenies obtained suggest that Porifera could be paraphyletic and that the phylogenetic relationships of most of the families and orders of the Demospongiae have to be reassessed. The Calcarea and Hexactinellida are still to be studied at the molecular level.     

Reassessment of homology of morphological characters in tetractinellid sponges based on molecular data

In sponges, as in other taxa with simple organization, the evaluation and use of morphological characters is difficult. Phylogenetic analysis of the first 850 nucleotides from the 5' end of the 28S rRNA gene is used here to assess the homology of spicules used in the classification of the subclass Tetractinellida. A single well-supported MP tree was obtained. The monophyly of the nine Tetractinellida species studied confirms the tetraxon megasclere as a morphological synapomorphy for the Tetractinellida. Two species are reallocated, Penares helleri as a Geodiidae, now thought to have lost sterraster microscleres, and Stryphnus mucronatus to the Streptosclerophorida. SEM micrographs of Stryphnus microscleres show that the morphology of the sanidasters is compatible with the hypothesis that they are homologous with streptoscleres and confirm this reallocation. Two other synapomorphies are confirmed within the tetractinellid clade, the simultaneous presence of tetraxon megasclere and aster-type microsclere (Astrophorida) and the loss of the streptosclere and persistence of the euaster s.s. microscleres (Euastrophorida) evidenced by the reallocation of Stryphnus mucronatus. The streptosclere microscleres cannot be evaluated in terms of homology because Streptosclerophorida may be paraphyletic (although these nodes are not supported by reliable bootstrap proportions) contrary to the currently accepted classification.