Bayerxenia, , Alderslade, 2001

Reinicke Heike Wägele Ralph Tollrian Florian Leese, Kristina Stemmer Ingo Burghardt Christoph Mayer Götz B., 2013, Morphological and genetic analyses of xeniid soft coral diversity (Octocorallia; Alcyonacea), Organisms Diversity & Evolution (New York, N. Y.) 13 (2), pp. 135-150 : 146-147

Bayerxenia

Bayerxenia sp. 2 , identified as a distinct and well-defined species on the basis of various morphological characters, is paraphyletic in the SRP54 analysis with two distinct and strongly supported groups, but remains unresolved in the ND6/ND3 analysis. Specimens of Bayerxenia sp. 2 occurred in dense aggregations of distinguishable and sometimes fused colonies and were collected from three different locations along the same beach on Lizard Island (Figure S6 E). The red-brownish colonies had a size of about 4 cm and the autozooids were pulsating. Siphonozooids were visible between the dense, feathery autozooids. At the same locality, also colonies of Xenia sp. 3 were found. Both species formed huge and dense intermingling colonies (Figure S6 F). Only Bayerxenia sp. 2 was colonised by the slug Phyllodesmium lizardensis, Burghardt et al. (2008b) (Figure S4 C). Affeld et al. (2009) showed that two new secondary metabolites (sesquiterpenes) were only present in the dimorphic xeniid and the associated slug, but not in the sympatric species Xenia sp. 3 . In their study, the dimorphic xeniid was still assigned to the genus Heteroxenia because the sclerite structure of their material was not analysed until now in this study.

It is astonishing that in the SRP54 gene analysis some of the specimens identified as Bayerxenia sp. 2 from Lizard Island group together with a morphotype classified as Bayerxenia sp. 1 from Bali (SRP54 clade 5). Differences between these Bayerxenia sp. 2 specimens are 8.3 % uncorrected pairwise distances for SRP54. Specimens assigned to Bayerxenia sp. 1 can be distinguished from Bayerxenia sp. 2 by larger sclerites, longer anthocodiae and a smaller number of pinnule rows with similarities to Heteroxenia pinnata , described for the Philippine Sea by Roxas (1933). This discrepancy can potentially be explained by ancestral polymorphisms together with incomplete lineage sorting. Alternatively, members of Bayerxenia sp. 2 may belong to two different species (clade 5 and clade 7 in SRP54, Fig. 3 ), and the morphological characters used for the delimitation of species may be phenotypically plastic and thus of limited use. However, both morphospecies can be assigned unambiguously to the genus Bayerxenia because of the distinct sclerite surface structures with triangular corpusculars as described by Alderslade (2001). This also applies to Bayerxenia sp. 3 , which we initially determined as a Xenia species based on the lack of siphonozooids. The analysis of the sclerites revealed the typical triangular corpusculars of Bayerxenia (Fig. 2h), and both genetic analyses confirmed its assignment to this genus. Achituv and Benayahu (1990) have shown that siphonozooids are not present throughout the whole life cycle in dimorphic xeniid species. This clearly shows that the absence of siphonozooids is an ambiguous character and ontogenetic variability has to be taken into consideration in the process of species identification. Similar results were obtained for two further specimens identified preliminarily as members of the genus Xenia because of the absence of siphonozooids ( Xenia sp. 5 and sp. 6). They clearly group within the dimorphic clade ( Heteroxenia / Bayerxenia , Fig. 3 ), but the two genetic analyses are not congruent in the assignment to a certain genus. Whereas SRP54 indicates a closer relationship of these two species to one of the Bayerxenia clades (clade 5), the ND3/ ND6 analysis shows no resolution. We assume that the Xenia sequence taken from GenBank (AF530512) also represents a misidentification.

In this study, Asterospicularia is resolved as the sister group to the dimorphic clade in the SRP54 analysis but shows no particular affiliation to any xeniid genus in the less resolved tree of the ND6/ND3 analysis. Its grouping within the Xeniidae analysed in this study confirms the rejection of the monotypic family Asterospiculariidae and the placement of the monogeneric Asterospicularia within the family Xeniidae ( Alderslade 2001) .

In summary, resolution was much higher for the fast-evolving nuclear marker SRP54 compared to the mitochondrial gene fragment ND6/ND3. Furthermore, several incongruencies have been found between morphological and genetic characters: On the one hand, the molecular genetic data cannot confirm the validity of some distinct morphospecies, whereas, on the other hand, the morphologically identical specimens of Xenia sp. 3 revealed high sequence divergence in the fast-evolving SRP54 gene, indicative of overlooked or cryptic species. A reasonable explanation for these incongruencies between genetic markers and morphological characters could lie in the properties of the very fast-evolving gene SRP54 and its heterozygote nature. But they could also be the result of the usage of polymorphic morphological characters as diagnostic features, which hence are of limited use for species classification. Similar observations were made by Concepcion et al. (2008) for the octocoral genus Carijoa . Furthermore, sampling of differing ontogenetic stages may lead to a misidentification. We need more information on the ontogeny, life cycle and environmentally induced changes in the morphology of Xeniidae to re-evaluate diagnostic characters used for discriminating species of the family Xeniidae and also other phenotypic plastic octocorals. So far, it cannot be determined whether the genetically distinct clades in this study represent cryptic species that have not been recognised previously or whether they represent other, already described species whose morphs are difficult to distinguish or have been synonymised ( McFadden et al. 2006). At this point it is also interesting to note that the specimens within each of the eight clades always originated from the same sampling site ( Table 1), indicating a clear genetic differentiation with respect to geographic location.

Our results give evidence that SRP54 is a suitable marker for phylogenetic analyses on the generic and species levels within Xeniidae , whereas ND6/ND3 probably will contribute more to the generic and higher taxa levels. It is also evident that SRP54 is a good marker for discriminating several xeniid species, whereas morphological characters showed limitations and therefore have to be re-evaluated. Even though ND6/ND3 is less variable than SRP54, it was sufficiently variable for genus and even species delimitation in some cases.

The mitochondrial gene msh1 (MutS homolog 1) has been in the focus with regard to octocoral phylogenies ( McFadden et al. 2006). But according to McFadden et al. (2006), this marker shows a low genetic divergence among xeniid species. For the octocoral Narella, Baco and Cairns (2012) showed that 83 % of the species within this group could be resolved by a combination of COI and msh1 along with the ND2 marker. Future studies should focus on comparing existing and finding additional markers for studying biodiversity and evolution in the xeniid genera.

Figures

Fig. 3 Bayesian phylogenetic tree of Ovabunda, Xenia, Bayerxenia and Heteroxenia. Node labels indicate Bayesian posterior probabilities and maximum likelihood bootstrap support values (in italics) based on