Divergence time estimations and contrasting patterns of genetic diversity between Antarctic and southern South America benthic invertebrates

Diversity, abundance and composition of taxonomic groups in the Southern Ocean differ from elsewhere in the planet, since the biogeography in this region refl ects the complex interactions of tectonics, oceanography, climate and biological elements since the Eocene. Several groups of marine benthic organisms exhibit high levels of genetic divergence among provinces in this region, supporting the existence of a vicariance process through plate tectonics, while other groups with high dispersive capacity exhibit recent divergence processes. Moreover, the discovery of nonAntarctic decapod larvae in Antarctic Peninsula suggests that some groups can travel across the Antarctic Circumpolar Current. Here we analyzed levels of genetic divergence in congeneric species of three Southern Ocean’s benthic invertebrate groups with dispersive potential. For this purpose we included in the analyses COI sequences of an echinoid (Sterechinus), a gastropod (Nacella), and a bivalve (Yoldia). Considering the levels of genetic differentiation and assuming the molecular clock hypothesis we estimated the separation of invertebrates from the two continents. We also compared levels of genetic variation between Antarctic and sub-Antarctic species of Nacella and Sterechinus to determine the effect of the Quaternary glacial episodes in the demography of these species. We detected clear genetic differences between Antarctic and sub-Antarctic congeneric species of Sterechinus, Nacella, and Yoldia. According to our results, the installation of an effective barrier between Antarctica and sub-Antarctica occurred almost at the same time (between 3.7 and 5.0 Ma) for these groups of organisms, long after the physical separation of both continents. Genetic comparisons between Antarctic and Sub-Antarctic species of Nacella and Sterechinus detected lower levels of genetic diversity in Antarctic species, suggesting more pronounced effects of the glacial episodes in Antarctica than in South America. These results may refl ect the dramatic effect of the Quaternary glacial cycles on Antarctic population sizes, especially in groups with narrow bathymetric ranges. The present study provides new evidence about the differentiation processes between Antarctic and South American organisms. None of the analyzed genera showed evidence for recurrent gene fl ow across the Antarctic Circumpolar Current since the Mio-Pliocene. Genetic comparisons indicate that Antarctic and Sub-Antarctic species were differentially affected by glacial periods.


INTRODUCTION
The Southern Ocean (SO) includes all waters south of the Polar Front, a well-defi ned circum-Antarctic oceanographic area that marks the nor thernmost extent of cold sur face water (Rintoul et al. 2001, Aronson et al. 2008).Its total area is about 35 million km 2 , of which more than 60 % is covered by ice during the winter maximum and 20 % is covered during the summer minimum (Zwally et al. 2002).Since the Mesozoic this region has undergone major tectonic, oceanographic and climatic changes that operated at dif ferent temporal and spatial scales (Knox 1980, Clarke & Crame 1989, 1992, Zachos et al. 2001, Griffi ths et al. 2009).The geological histor y of the SO is deeply connected to the fragmentation and the dispersion of the continental blocks that formed the supercontinent Gondwana and to the opening of gateways between Antarctica, Australia and South America (Lawver et al. 1993, Barker & Thomas 2004, Pfühl & McCave 2005, Scher & Martin 2006, Torsvik et al. 2008).Reconstructions of paleo-currents of the past 55 Ma have shown that the opening of the Tasman gateway and the Drake Passage shaped past and present oceanographic circulation of the SO (Kennett 1980, Knox 1980, 2007, Woodruff et al. 1989, Stickley et al. 2004, Livermore et al. 2005).The formation of these gateways influenced the initiation of the Antarctic Circumpolar Current (ACC), the major current system transporting more than 130 Sv (1 Sv = 10 6 m 3 seg -1 ) through the Drake Passage (Orsi et al. 1995).This current fl ows around Antarctica and is delimited by the Polar and the sub-Antarctic fronts (Barker et al. 2007).Mackensen ( 2004) recognized three main periods in which major changes affected the circulation of the ACC and the climate of the SO: (1) the Ecocene/Oligocene boundary when the origin of the ACC signifi cantly modifi ed global oceanic circulation.(2) The middle Miocene (~14 Ma) when the re-establishment of an East Antarctic ice sheet influenced mode and levels of Antarctic bottom water formation, generating an intensifi cation of the ACC.(3) The Quaternar y, characterized by the alternation between glacial and interglacial periods.It is likely that the intensity of the ACC has varied in response to processes related to Quaternary glacial cycles (Gersonde et al. 2005, Hassold et al. 2009).In the South Atlantic, Indian and South Pacifi c Oceans, the whole current system can migrate north or south by several degrees of latitude as a response to a change in volume transport (Pudsey & Howe 1998).The positions of the ACC fronts (Polar and Sub-Antarctic) have important implications for the biogeography of the region, they act as a gene fl ow barrier for some species (Shaw et al. 2004, González-Wevar et al. 2010), while for others they constitute important transportation vectors (Beu et al. 1997, Page & Linse 2002, Thorpe et al. 2004, Waters 2008, Fraser et al. 2009, Díaz et al. 2011).
Palabras clave: barrera oceanográfi ca, COI, Corriente Circumpolar Antártica, planctotrofía, separación en el Plioceno.relict autochthonous fauna; (2) a fauna derived from adjacent deep-water basins; (3) a fauna dispersing from South America along the Scotia Arc; and (4) a fauna which has spread in the opposite direction from Antarctica northwards along the Scotia Arc (Knox & Lowr y 1977, Clarke 2008).Several groups of marine invertebrates that are abundant and diverse in other adjacent regions are scarcely represented or even absent in the SO.Examples include key elements of gastropods, bivalves, decapods and fi shes (Crame 1999, Aronson & Blake 2001).However, other marine groups such as porifera, bryozoa, echinodermata, polychaeta, ascidians, pycnogoniids, amphipods and isopods are highly abundant and diverse, suggesting that major climatic and oceanographic changes in the region did not impede their evolutionar y success (Clarke & Crame 1989, 1992, Clarke & Johnston 1996, Aronson & Blake 2001, Linse et al. 2006, Rogers 2007, Aronson et al. 2008).In a recent revision of the biogeographical patterns of the marine benthic fauna in the Southern Ocean, Griffi ths et al. ( 2009) stated that the regions in the SO differ depending upon the class of organisms being considered.According to their results, some general rules are possible, including high levels of endemism (around 50 %), a single Antarctic Province and a clear distinction between the sub-Antarctic islands infl uenced by South America and those of New Zealand.
High levels of faunal af finities ar e particularly clear between Antarctica and the southern tip of South America, commonly known as the Antarctic-Magellan connection (Arntz 1999, 2005, Brandt et al. 1999, Crame 1999, Arntz et al. 2005, Thatje et al. 2005, Rogers 2007, Aronson et al. 2008).The traditional interpretation for this af finity is that these regions were contiguous until the opening of the Drake Passage and were progressively separated by deep waters from the Eocene/Oligocene (Crame 1999).Marine inver tebrates from dif ferent provinces of the SO such as Euphausia (Patar nello et al. 1996), Af frolittorina and Austrolittorina (Williams et al. 2003, Waters et al. 2007), and fishes (Clarke & Johnston 1996, Waters et al. 2000) exhibit important levels of genetic divergence, supporting vicariance speciation by the plate tectonics hypothesis.Nevertheless, new molecular evidence in other groups of marine invertebrates, especially in those with high oceanic dispersive capacity, suggest more recent divergence processes than those expected under the vicariance hypothesis and provide evidence for the importance of longdistance dispersal in the distribution of the SO marine benthic fauna (Helmuth et al. 1994, O'Foighill et al. 1999, Coyer et al. 2001, Page & Linse 2002, Donald et al. 2005, Gérard et al. 2008, Waters 2008, Fraser et al. 2009, 2010, Díaz et al. 2011).Moreover, recent observations of non-Antarctic anomuran and bachyuran zoea stages in King George, Antarctic Peninsula (Thatje & Fuentes 2003) indicate that some groups can travel across the ACC (Tavares & De Melo 2004, Clarke et al. 2005).Similarly, records of non-Antarctic lithoid crabs in deep water of f the Antarctic continental slope suggests that these crabs may be returning to this region (Thatje & Arntz 2004, Thatje 2005, Thatje et al. 2005).These fi ndings highlight the permeability of the polar front in space and time, raising questions about how organisms got to Antarctica and how often these processes happened in the past.Thatje & Fuentes (2003) suggest that larvae might cross the polar front using eddies or intrusions of Sub-Antarctic water masses through the ACC.Satellite imagery indicates that the ACC, far from being a continuous barrier, has a complex mesoscale structure including eddies over a wide range of scales.Eddies are impor tant transpor t mechanisms across the ACC, where warmcore rings can transport sub-Antarctic plankton to Antarctica, and cold-core rings can carr y Antarctic plankton to warmer waters of the north (Glorioso et al. 2005).
The main objective of this study is to evaluate if the ACC has constituted an effective oceanographic bar rier for lar val dispersal between two Provinces of the Souther n Ocean to estimate since it has operated.For this purpose we selected species of three genera, all characterized by possessing a planktotrophic lar val stage.In or der to discount the possibility of a deep-sea connection after the opening of the Drake Passage, we included groups of organisms with narrow bathymetrical distribution that are restricted to the continental shelves of both provinces.We determined the levels of molecular divergence between congeneric species of broadcasters, marine invertebrates from Antarctic Peninsula and southern South America.First we compared Sterechinus neumayeri (Meissner, 1900), a regular echinoid with a circum-Antarctic distribution with S. agassizii (Mor tensen, 1910) from the Argentinian continental shelf.Second, we compared the Antarctic limpet Nacella concinna (Strebel, 1908) with its Magellanic relative Nacella magellanica (Gmelin, 1791).Finally, we compared the Antarctic bivalve Yoldia eightsi (Jay, 1839) with the Magellanic species Yoldia woodwardi (Hanley, 1860).The information contained in their DNA sequences will permit us to estimate rhythms and trends in the biogeography of marine benthic near-shore organisms in this Region.

METHODS
We analyzed a partial fragment of the mitochondrial gene Cytochrome c Oxidase Subunit I (COI) in congeneric species of Sterechinus (945 bp, Díaz et al. 2011), Nacella (662 bp, González-Wevar et al. 2011a, 2011b) and Yoldia (688 bp) from the Antarctic Peninsula and the Magellanic Province.Specimens of the Antarctic limpet Nacella concinna were collected from fi ve localities along the western Antarctic Peninsula (González-Wevar et al. 2011a) and N. magellanica was collected in three localities along its distribution in the Magellanic Province (González-Wevar et al. 2011b).Sterechinus neumayeri were collected from two localities of the western Antarctic Peninsula (Fildes and Covadonga Bay) and S. agassizi samples were collected from different localities on the Argentinean continental shelf, Magellanic Province (Díaz et al. 2011).Finally, we included in the analyses fi ve individuals of Yoldia eightsi collected in Fildes Bay, King George Island, Antarctica and five specimens from Porvenir Bay, Magellan Strait.The COI gene was amplifi ed in Yoldia using the universal primers described by Folmer et al. (1994).Amplicons were purifi ed and sequenced in both directions by Macrogen (South Korea).Sequences were edited with Proseq 2.91 (Filatov 2002) and aligned with Clustal W (Thompson et al. 1992).
We estimated divergence times between Antarctic and Magellanic lineages considering the number of pairwise differences between species from the provinces and assuming a strict molecular clock hypothesis.Previous to assuming this hypothesis we performed a likelihood ratio test (Felsenstein 1981) using DAMBE (Xia & Xie 2001).Divergence time estimations were made using specifi c mutational rates for each group.In the case of Nacella we used a substitution rate estimated for nacellid limpets (1.0 % per million year, González-Wevar unpublished data).In the case of Sterechinus, we selected two substitution rates 0.51 % and 0.72 % per million year, according to Lee et al. (2004) for Echinidae.For Yoldia we used a substitution rate of 0.95 % per million year (Wares & Cunningham 2001).
We constructed genealogical relationships in all three genera with haplotype networks using the Median-Joining algorithm in Network 4.6 (Röhl 2002, http://www.fluxus-engineering.com).This method allows simple reconstructions of phylogenies based on intraspecifi c genetic data such as mitochondrial DNA variation (Bandelt et al. 1999, Posada & Crandall 2001).
We determined levels of genetic polymorphism in Nacella and Sterechinus species using standard diversity indices: number of haplotypes (k), number of segregating sites (S), and haplotype diversity (H) for each province using DnaSP 5.00.07 (Librado & Rozas 2009).We also estimated average pairwise sequence differences (Π) and nucleotide diversity (ϖ).We performed mismatch distribution analyses in Nacella and Sterechinus species using pairwise distances as an assessment of population demographic histories.To determine whether the populations have undergone sudden population growth we compared the mismatch distribution of haplotype differences among haplotypes of Nacella and Sterechinus COI sequence data sets with expectations of a sudden expansion model (Rogers & Harpending 1995).The goodness of fi t between the observed and expected mismatch distributions was tested using a parametric bootstrap approach that uses the sum of squared deviations between observed and expected mismatch distributions as a test statistic, as implemented in Arlequin 2.0 (Schneider et al. 2000).

RESULTS
We detected major genetic discontinuities between Antarctic and Magellanic congeneric species of Sterechinus, Nacella and Yoldia.This was shown by the high levels of genetic divergence (> 7.0 %) between species from the two provinces.Sterechinus neumayeri (Antarctica) and S. agassizi (South America) exhibited 7.2 % dif ference, with an average of 56 nucleotide differences between species (Fig. 1A).Similarly, N. concinna (Antarctica) and N. magellanica (Magellanic Province) had 7.7 % difference, with an average of 51.4 nucleotide differences between species (Fig. 1B).Yoldia eighsi from Antarctica showed 7.0 % dif ference with Y. woodwardi from southern South America and an average of 48.2 nucleotide differences between species (Fig. 1C).Divergence time estimations between species from the two provinces of the SO indicate that the analyzed congeneric species were separated by the following mutational times: 28 for Sterechinus, 26.2 for Nacella and 24.1 for Yoldia.Considering specific substitution rates, the separation of the three groups of benthic inver tebrates occur red during the Pliocene (between 5.0 and 3.7 Ma).In the case of Sterechinus the last contact between S. neumayeri and S. agassizi occurred between 4.4 and 5.0 Ma, while the separation between the Antarctic limpet N. concinna from its Magellanic relative N. magellanica took place ∼3.7 Ma.Finally, the separation between Yoldia eightsi from Antarctica and Y. woodwardi from southern South America occurred ∼3.9 Ma.
H a p l o t y p e n e t w o r k s s h o w e d c l e a r differences between Antarctic and Magellanic species of Nacella and Sterechinus (Figs. 1A and 1B).In general, haplotype networks of Magellanic species (S. agasizzi and N. magellanica) showed higher levels of diversity in terms of the number of haplotypes and the extension of the genealogy than their Antarctic relatives (S. neumayeri and N. concinna).These results were further corroborated by the levels of genetic diversity detected in Antarctic and Magellanic species of Nacella and Sterechinus.In both genera, South American species exhibited higher levels of genetic diversity measured as haplotype numbers (k), polymorphic sites (S), and haplotype diversity (H).Similarly, the average number of pairwise dif ference (Π) was six times greater in S. agassizi (1.88) from the Magellanic Province than in S. neumayeri (0.30) from Antarctica and almost three times greater in N. magellanica (2.338) than in the Antarctic limpet N. concinna (0.850, Table 1).We detected clear differences in the shape of the mismatch distributions between Antarctic and sub-Antarctic species of Nacella and Sterechinus that refl ect differences in their demographic histories.For instance pair wise dif ferences of Antarctic species S. neumayeri (Fig. 2A) and N. concinna (Fig. 2C) are characterized by L-shaped distributions, while S. agassizi (Fig. 2B) and N. magellanica (Fig. 2D) from souther n South America showed unimodal and bimodal patterns of distribution, respectively, further supporting clear differences in trends and rhythms of the demographic changes between the provinces of the SO.

DISCUSSION
We l l -s u p p o r t e d r e s u l t s d e r i v e d f r o m m o l e c u l a r d a t a c a n p o t e n t i a l l y r e v e a l important information about biogeographical and phylogeographical patterns, systematic r elationships, conser vation issues and divergence time estimations.In this study, our molecular analyses established clear dif ferences between congeneric species of Sterechinus, Nacella and Yoldia.For each genus, species from Antarctic Peninsula and Southern South America constitute distinct Evolutionary Signifi cant Units that were separated several million years ago.Based on these results, the ACC appears as an old and effi cient barrier for these genera.It is important to note that the separation of congeneric species in three genera belonging to echinoids, gastropods and bivalves occurred in a brief evolutionary time measured in terms of mutational steps (between 28 and 24.1).These results suggest that the installation of an effective barrier for faunal interchange between Antarctica and souther n South America occur red almost at the same time among these broadcastspawning invertebrates.Considering lineagespecific substitution rates, divergence time estimations suggest that the separation of these groups occurred during the Pliocene between 3.7 and 5.0 Ma.According to these divergence time estimates, the separation between Antarctic and South American taxa started long after the physical separation of both continents, estimated between 41 Ma (Livermore et al. 2005) and 23.9 Ma (Eagles & Livermore 2002, Pfühl & McCave 2005, Scher & Mar tin 2006, Barker et al. 2007, Lyle et al. 2007).This separation seems to be more related to climatic and oceanographic processes during the end of the Miocene, including an increase of d 18 O values associated with polar cooling, major growth of ice sheets in eastern Antarctica and main changes in ocean circulation (Woodruf f & Savin 1989, Flower & Kennett 1994, Shevenell et al. 2004, Mackensen 2004).In fact, during the late Miocene an intense pattern of thermal zonation in the oceans has been described.This event might be responsible for an intensification of the ACC resulting in the differentiation of Antarctic and Sub-Antarctic fauna (Crame 1999).It is important to note that the taxonomy of Yoldia is still unclear.According to Rabarts & Whybrow (1979) Yoldia in the Magellanic Province includes two dif ferent species, Y. eightsi and Y. woodwardi, both distributed in the Falkland Islands and Tierra del Fuego.We considered the Magellanic specimens of Yoldia as Y. woodwardi because of the observed dif ferences with Y. eightsi.However, Dell (1964), Villarroel & Stuardo (1998) and Huber (pers.comm.2011) synonymized Y. eightsi and Y. woodwardi.In spite of these uncertainties in the taxonomy of Yoldia, the genetic divergence between Antarctic and sub-Antarctic specimens of Yoldia indicate that they do not constitute the same Evolutionary Signifi cant Unit and should be considered as separate entities.
Our estimated times of the split between Antar ctic and Sub-Antar ctic species of Sterechinus, Nacella and Yoldia are congruent with other molecular studies in dif ferent taxonomic groups.For example, the separation b e t w e e n A n t a r c t i c a n d S u b -A n t a r c t i c notothenioid fishes (Patagonotothen and Lepidonotothen) ranged from 9 Ma (Bargelloni et al. 2000a) to 7.1-6.1 Ma (Stankovic et al. 2002).Divergence analyses of the Antarctic Genetic diversity indices in species of Nacella and Sterechinus from the Antarctic Peninsula and southern South America.n = number of sampled individuals; k = number of haplotypes; S = polymorphic sites; H = haplotype diversity; Π = average number of nucleotide differences; ϖ = nucleotide diversity.
Índices de diversidad genética en especies de Nacella y Sterechinus de Península Antártica y el Sur de Sudamérica.n = número de individuos muestreados; k = número de haplotipos; S = sitios polimórfi cos; H = diversidad haplotípica; Π = número promedio de diferencias de nucleótidos; ϖ = diversidad nucleotídica.2007) are also comparable to those estimated in this study.In summary, these studies between Antarctic and Sub-Antarctic invertebrate taxa indicate that, in spite of the substitution rate selected, the separation of marine benthic taxa of the regions occurred near the transition between Miocene and Pliocene and therefore long after the geographical separation of these provinces of the SO.Climate change is one of the major forces driving population extinctions, par ticularly near the limit of a species' range (Hewitt 1996).Paleontological and palynological records have demonstrated that many species have undergone significant latitudinal shifts in response to the advances and retreats of Quaternary glacial ice sheets, and particularly to the recent Last Glacial Maximum (LGM) between 23.000-18.000years ago (Webb & Bar tlein 1992, Hewitt 2000, 2004, Provan & Bennett 2008).Especially at higher latitudes, ice sheet advances and retreats, surrounding permafrost, lower global temperature, sea level variations, and reduced water availability caused major changes in the distribution of species (Bennett 1997, Huybrechts 2002).During the LGM, many species went extinct over large parts of their range, some dispersed to new template habitats, others sur vived in lower latitude refugia and subsequently expanded their range through interglacial recolonization (Hewitt 2000(Hewitt , 2004)).Relatively little is known about the biotic effects of recent glaciations in the Southern Hemisphere, where ice-sensitive benthic biota of Antarctic and sub-Antarctic regions would have endured significant processes of extinction and recolonization as sea ice covered the SO (Fraser et al. 2009).In fact, contemporary ice scouring is known to purge much of the shallow water benthos in Antarctica (Gutt 2001, Barnes & Conlan 2007), and the coastlines within the LGM sea ice limits (Fraser et al. 2009).Genetic comparisons of standard diversity indices in Antarctic and Sub-Antarctic species of Nacella and Sterechinus suggest a more pronounced ef fect of the Quater nar y glacial cycles in Antarctica than in southern South America.In general, even when we included at least twice the number of individuals of the Antarctic species N. concinna and S. neumayeri, they exhibited lower levels of genetic diversity (k, S, H, Π and ϖ) in comparison to the Magellanic species (N.magellanica and S. agassizi).These results could refl ect the dramatic effect of the Quaternary glacial periods on population sizes, especially for species with narrow bathymetric ranges.In this respect, the results obtained in N. concinna and S. neumayeri contrast with other studies in Antarctic invertebrates that showed higher levels of genetic diversity (Mahon et al. 2008, Thornhill et al. 2008, Krabbe et al. 2009, Wilson et al. 2009, Goodall-Copestake et al. 2010, Arango et al. 2011).However, most of these species have large bathymetrical ranges that could have helped them to prevent the drastic demographic impact during Quaternary glacial cycles (Brey & Clarke 1993, Brey et al. 1996).During glacial maxima the ice edge advanced across all land and the continental shelf, bulldozing the sur viving fauna to the deep continental margin (Grobe & Mackensen 1992, Barnes & Conlan 2007).Shallower continental shelf and terrestrial environments thus had to be repeatedly re-colonized during interglacial periods of ice retreat.Considering the bathymetrical range of Nacella and Sterechinus, the extension of glacial continental ice-sheets over main par t of the Antarctic Peninsula should have drastically reduced their habitats to isolated ice-free glacial areas (Poulin et al. 2002, Thatje et al. 2005).However, there is still no evidence of such ice-free marine areas in Antarctic Peninsula (Bar nes & Conlan 2007).Another possibility includes range contraction of both species to ice-free areas in the northern extreme of their distributions to Antarctic islands of the Scotia Arc such as the South Georgia and Signy Islands.Then during interglacial periods both species could have re-colonized the Antarctic Peninsula through lar val dispersal.According to this, founder ef fects could constitute a plausible explanation for the low genetic diversity detected in these species.This scenario is also supported by star-like genealogies with ver y short branches (Figs. 1A and 1B), and marked L-shaped distributions of pairwise differences in Antarctic Nacella and Sterechinus (Figs. 2A  and 2C).

Species
Finally, the present study gives new insight into the patterns of genetic dif ferentiation between the Antarctic Peninsula and southern South America marine near-shore invertebrates with high dispersive capacity.According to our results, in organisms with free-living larvae such as Sterechinus, Nacella and Yoldia, the divergence between Antarctic and Sub-Antarctic lineages was initiated long after the physical separation of Antarctica and South America.After the Miocene-Pliocene transition the intensification of the ACC constituted an effi cient barrier for gene fl ow across these Provinces.None of the analyzed genera exhibited evidence for recurrent gene fl ow across the APF for several Ma.In this respect, the recent discovery of decapod larvae in Antarctic Peninsula probably represents exceptional events related to present rapid climatic changes or to an increase in marine human traffi c (scientifi c and tourism) between South America and the Antarctic Peninsula.Moreover, genetic diversity analyses indicate that Antarctic and Sub-Antarctic taxa have been differentially impacted by glacial periods, and this situation was more evident in Antarctic species with narrow bathymetrical ranges.