The Phylum Porifera (sponges) is currently divided into three extant classes – the Hexactinellida, Demospongiae, and Calcarea – and one fossil class – the Archaeocyatha [1
]. Hexactinellids differ from the canonical sponge body plan in lacking discrete motile cells (see [1
] for a detailed review). The tissue of hexactinellids forms a continuous multinucleate syncytium. Cellular components exist, but all "cells" are connected by cytoplasmic bridges to one-another and to the syncytium. Choanocytes are branched; collar-flagella units ("collar bodies") form as enucleate buds, several arising from a single nucleated choanoblast [2
]. The distinct tissue organization was considered important enough to warrant separation of hexactinellids from other Porifera at the subphylum level – Symplasma for the Hexactinellida and Cellularia for the Demospongiae and Calcarea [5
]. In no other animal is the tissue so inclusively connected in a giant syncytium, making the hexactinellid body construction unique among Porifera, as well as all Metazoa.
The Porifera has long been considered the earliest branching group of the metazoan crown, based on both morphological and molecular evidence, although the precise relationships of the lower metazoan phyla remains uncertain [6
]. One on-going approach to resolve the overall metazoan phylogeny, as well as the problematic class-level relationships among the sponges, has been comparative analysis of mitochondrial genomes (mtDNA). To date, this effort has seen the sequencing and description of five complete demosponge mtDNAs [GenBank: NC_006894
], and most recently, that of two partial hexactinellid sponge mtDNAs [GenBank: EF537576
]. Phylogenetic estimations based on concatenated mitochondrial protein sequences have been sensitive to taxon sampling, outgroup choice and algorithm implementation. These trees reveal artifacts likely due to variable rates of molecular evolution, such that placozoans, cnidarians, and demosponges are recovered as a monophyletic clade, with hexactinellids the sister group of bilaterians [12
]. The tree topologies may become more stable, and consistent with plausible hypotheses of the evolution of morphological traits, as the mtDNA of additional taxa are added to the data set.
While sequence-based mtDNA comparisons of the lower metazoa have not been phylogenetically definitive, the efforts have dispelled some commonly held myths and revealed some general characteristics of animal mtDNA. Placozoan mtDNA, for example, is twice as large as most animal molecules, ranging from 32 to 43 kilobases, and retains various traits of the non-metazoan outgroups such as substantial intergenic space, introns and large open reading frames less commonly found in other animal mtDNA [12
]. On the other hand, demosponge mtDNA is more typical of other animals – compact molecules, between 18 and 25 kb, with few or no introns, little intergenic space, and coding for twelve to fourteen respiratory chain proteins and two ribosomal RNAs [13
]. Two partial hexactinellid mitochondrial genomes, those of Sympagella nux
(Order Lyssacinosida, Family Caulophacidae) and Iphiteon panicea
(Order Hexactinosida, Family Dacylocalycidae), have recently been reported [14
]. These genomes were found to have similar protein-coding gene content as the published demosponges, but had several features such as tRNA structure and content more similar to that of bilaterians [14
]. The current paper reports on the complete mtDNA sequence of the hexactinellid sponge Aphrocallistes vastus
(Order Hexactinosida, Family Aphrocallistidae), compares it to previously published poriferan mtDNAs, and highlights two translational frameshifts, a phenomenon that is unique to the hexactinellids among reported lower metazoan mitochondrial genomes.