Phylogenetic relationships of heteroscleromorph demosponges and the affinity of the genus Myceliospongia (Demospongiae incertae sedis)

Abstract

Class Demospongiae - the largest in the phylum Porifera (Sponges) - encompasses 7,581 accepted species across the three recognized subclasses: Keratosa, Verongimorpha, and Heteroscleromorpha. The latter subclass contains the majority of demosponge species and was previously subdivided into subclasses Heteroscleromorpha sensu stricto and Haploscleromorpha. The current classification of demosponges is the result of nearly three decades of molecular studies that culminated in a formal proposal of a revised taxonomy (Morrow and Cardenas, 2015). However, because most of the molecular work utilized partial sequences of nuclear rRNA genes, this classification scheme needs to be tested by additional molecular markers. Here we used sequences and gene order data from complete or nearly complete mitochondrial genomes of 117 demosponges (including 60 new sequences determined for this study and 6 assembled from public sources) and three additional partial mt-genomes to test the phylogenetic relationships within demosponges in general and Heteroscleromorpha sensu stricto in particular. We also investigated the phylogenetic position of Myceliospongia araneosa - a highly unusual demosponge without spicules and spongin fibers, currently classified as Demospongiae incertae sedis.\n\nOur results support the sub-class relationship within demosponges and reveal four main clades in Heteroscleromorpha sensu stricto: Clade 1 composed of Spongillida, Sphaerocladina, and Scopalinida; Clade 2 composed of Axinellida, Biemnida, Bubarida; Clade 3 composed of Tetractinellida and \"Rhizomorina\" lithistids; and Clade 4 composed of Agelasida, Polymastida, Clionaida, Suberitida, Poecilosclerida, and Tethyida. The four clades appear to be natural lineages that unite previously defined taxonomic orders. Therefore, if those clades are to be systematically interpreted, they will have the rank of superorders (hence S1-S4). We inferred the following relationships among the newly defined clades: (S1(S2(S3+S4))). Analysis of molecular data from Myceliospongia araneosa - first from this species/genus - placed it in S3 as a sister group to Microscleroderma sp. and Leiodermatium sp. (\"Rhizomorina\").\n\nMolecular clock analysis indicated that the origin of the Heteroscleromorpha sensu stricto as well as the basal split in this group between S1 and the rest of the superorder go back to Cambrian, while the divergences among the three other superorders occurred in Ordovician (with the 95% standard variation from Late Cambrian to Early Silurian). Furthermore most of the proposed orders within Heteroscleromorpha appear to have middle Paleozoic origin, while crown groups within order date mostly to Paleozoic to Mesozoic transition. We propose that these molecular clock estimates can be used to readjust ranks for some of the higher taxa within Heteroscleromorpha.\n\nIn addition to phylogenetic information, we found several unusual mtgenomic features among the sampled species, broadening our understanding of mitochondrial genome evolution in this group and animals in general. In particular, we found mitochondrial introns within cox2 (first in animals) and rnl (first in sponges).