Surviving the Messinian Salinity Crisis? Divergence patterns in the genus Dendropoma (Gastropoda: Vermetidae) in the Mediterranean Sea

Highlights

• Coalescent-based phylogenetic analysis confirmed four well-supported clusters.

• Clusters are delimited along a Mediterranean west–east axis.

• The estimates obtained indicated two main stages of diversification.

• The data suggested that the different lineages predated the MSC.

• Within-lineage diversification events occurred in late Pliocene–early Pleistocene.

Abstract

Four genetically distinct clades were recently described under the name Dendropoma petraeum, a Mediterranean endemic vermetid gastropod. The aim of this work is to date the processes that drove to the diversification within this taxon and to relate them to the corresponding historical events occurred in the Mediterranean Sea. Sequences from mitochondrial and nuclear markers were obtained from specimens collected in 29 localities spanning over 4000 km across the entire distribution range of D. petraeum species complex. The phylogenetic and coalescent-based analyses confirmed the four well-supported and largely differentiated lineages of D. petraeum, clearly delimited geographically along a west–east axis within the Mediterranean Sea: Western, Tyrrhenian–Sicilian, Ionian–Aegean and Levantine lineages. Divergence time estimates, obtained using a range of known substitution rates for other marine gastropods, indicated two main stages of diversification. In the first period (between 9.5 and 4.5 mya), the ancestral D. petraeum diverged into the current four lineages. The most recent period occurred between 3.72 and 0.66 mya in the late Pliocene–early Pleistocene, and included the main within-lineage diversification events. Therefore, if the divergence time between the major lineages of Dendropoma in the Mediterranean actually predated or coincided with the Messinian Salinity Crisis, then they should have survived to this dramatic period within the Mediterranean, as supported by Bayes Factors model comparison. Conversely, if the divergence started after the crisis, congruent with the idea that no true marine organism survived the Messinian Salinity Crisis, then our results indicate substitution rates of Dendropoma much higher than usual (5.16% per million years for COI, 3.04% for 16S). More recent climate changes seem to have conditioned the demographic history of each lineage differently. While Western and Tyrrhenian–Sicilian lineages both underwent an increase in their effective population sizes from 1.5 to 0.6 mya coinciding with a long interglacial period, the Ionian–Aegean and Levantine lineages showed constant effective population sizes since 2–2.5 mya, suggesting that these eastern lineages might represent small and relict populations surviving the subsequent Quaternary glaciations in isolated refugia.

Introduction

Species distribution is influenced mainly by geological events and by the dispersal capacity of each species. Due to a lack of basic data, however, the complex processes that led to today’s biodiversity patterns of coastal marine taxa are not fully understood. Comparative molecular studies can provide considerable insight into marine biodiversification processes by assigning the genetic structuring of taxa to an inferred historical context (Grosberg and Cunningham, 2001). Phylogeography explicitly integrates micro- and macro-evolution, relating ecology to evolution, current distribution to historical events, the physical environment to genetic structure, and patterns of variation within species to patterns of variation across species (Avise, 2000). It therefore provides a framework to explain and integrate patterns of marine biodiversity at intra- and supra-specific levels (Dawson, 2005).

The Mediterranean Sea has experienced dramatic changes in its configuration and climate over the last ten million years (Taviani, 2002). For instance, the formation of species or major phylogenetic lineages within species in the Mediterranean has been strongly influenced by modifications to the coastline, intensified during the Pleistocene by repeated glaciations. Hence, the combination of different events, such as the opening and closure of the Strait of Gibraltar, advancing and retreating glaciation and changes in current patterns, have made the Mediterranean region a notably dynamic hotspot of diversity (Pannacciulli et al., 1997).

Our knowledge of species’ phylogeography in the Mediterranean Sea is mainly based on organisms with a high dispersal capacity (such as fish, planktonic invertebrates or those with planktotrophic larvae) (Patarnello et al., 2007 for a review). This bias may be hindering our understanding of evolutionary patterns within the Mediterranean Sea, since past and contemporary population connectivity may obscure the signal of historical events (Provant and Bennett, 2008). Such phylogeographic studies on organisms with a high dispersal potential have established a common set of processes resulting from fragmentation into glacial refugia, range expansions via postglacial water masses, and/or secondary contact zones among formerly disjunct lineages. However, species with reduced dispersal are likely to have faced ice scouring, local extinctions and isolation in refugia, which may indirectly result in loss of local genetic variability and increased differentiation among populations by promoting the effects of local selection, inbreeding, and drift. Moreover, shallow subtidal/intertidal organisms capable only of short distance dispersal are likely to have faced the many challenges of historical sea level fluctuations, constantly reshaping coastlines and displacing their suitable habitats (Provant and Bennett, 2008). Hence, the study of such species with limited dispersal capacity may reveal different historical patterns of refuges and migration routes, and may thus offer a more complete understanding of relationships between the evolutionary history and contemporary distribution of genetic variation.

In a prior molecular analysis, Calvo et al. (2009) showed that at least four genetically distinct clades exist under the name Dendropoma petraeum. These four cryptic clades seem geographically well-delimited and are highly divergent at both the levels of mitochondrial and nuclear genes. In effect, this may be interpreted as indicative of differentiation at the species-level. This cryptic radiation is characterized by genetic speciation with little or no ecological diversification (‘nonadaptive radiation’ sensu Gittenberger, 1991) and with low morphological differentiation (‘morphostatic radiation’, sensu Davis, 1992). The four cryptic species of the D. petraeum complex are ecological equivalents (occupying the same type of habitat and playing the same ecological role in the different areas they inhabit) (Calvo et al., 2009). A posteriori searching for biological/morphological differences between these cryptic species has revealed some almost unappreciable morphological differences but significant changes in their mode of intracapsular larval development (Calvo et al., 2009, Usvyatsov and Galil, 2012).

Both ‘nonadaptive radiation’ and ‘morphostatic radiation’ require allopatric speciation. Actually, the allopatric distribution of these cryptic species supports predominantly vicariance-based cladogenetic patterns coupled to restricted post-speciation range expansions. Several lines of indirect evidence support the hypothesis that the allopatric distribution of these four clades persisted over evolutionary time scales in the absence of secondary contacts. Vicariance plays a key role in generating coastal biodiversity and dispersal explains many of the deviations from the phylogeographic hypotheses, but vicariant hypothesis is valid only if temporal congruence between vicariant events and cladogenesis can be established. Additionally, estimating the timing of divergence events is one of the central goals of contemporary evolutionary biology.

The high genetic divergence between the four clades obtained in our previous study (Calvo et al., 2009) could be indicative of an ancient radiation that might have predated the Messinian Salinity Crisis (MSC) between 5.9 and 5.3 Mya (Krijgsman et al., 1999). However, the generally more accepted hypothesis poses that virtually no stenohaline marine organism could have survived to the extreme dryness and salinity levels that existed during this epoch in the Mediterranean basin, and consequently the current fauna and flora need to have subsequently colonized the Mediterranean from the near Atlantic populations (Taviani, 2002). There are nevertheless some data supporting species survival during this period in Mediterranean refugia, probably existing in the deepest areas or near large river delta zones (Por, 2009), but this issue is still under debate.

The main objectives of the present study were: (1) to date the observed divergence events within D. petraeum complex; (2) to relate the vicariance processes driving differentiation to major climatic, oceanographic and geological changes occurring in the Mediterranean Sea; and (3) to infer associated demographic changes within this species-complex in the Mediterranean. Our ultimate goal was therefore to address the questions: (1) could an Atlantic ancestor of Dendropoma have colonized the Mediterranean in an eastward direction after the Messinian or did some Dendropoma species survive the MSC?, and (2) when and where did major cladogenetic events take place among the Mediterranean clades of Dendropoma?