Evolutionary Diversification of the Lizard Genus Bassiana (Scincidae) across Southern Australia

Background Relatively recent (Plio-Pleistocene) climatic variations had strong impacts on the fauna and flora of temperate-zone North America and Europe; genetic analyses suggest that many lineages were restricted to unglaciated refuges during this time, and have expanded their ranges since then. Temperate-zone Australia experienced less severe glaciation, suggesting that patterns of genetic structure among species may reflect older (aridity-driven) divergence events rather than Plio-Pleistocene (thermally-mediated) divergences. The lizard genus Bassiana (Squamata, Scincidae) contains three species that occur across a wide area of southern Australia (including Tasmania), rendering them ideally-suited to studies on the impact of past climatic fluctuations. Methodology/Principal Findings We performed molecular phylogenetic and dating analyses using two partial mitochondrial genes (ND2 and ND4) of 97 samples of Bassiana spp. Our results reveal a pattern of diversification beginning in the Middle Miocene, with intraspecific diversification arising from 5.7 to 1.7 million years ago in the Upper Miocene-Lower Pleistocene. Conclusions/Significance In contrast to the temperate-zone Northern Hemisphere biota, patterns of evolutionary diversification within southern Australian taxa appear to reflect geologically ancient events, mostly relating to east-west discontinuities imposed by aridity rather than (as is the case in Europe and North America) relatively recent recolonisation of northern regions from unglaciated refugia to the south.


Introduction
The geographic distribution of genetic diversity within a biological lineage can tell us a great deal about the processes that have affected those organisms over evolutionary time. For example, major genetic discontinuities can reveal places where barriers to dispersal occurred, often reflecting areas unsuitable for the species' persistence at some time in the past [1]. Information on the timing of such events, also obtainable from genetic data, can help to identify the nature of processes (e.g., climatic variation, sea-level fluctuation, volcanic activity) that created barriers to gene flow.
Palynological data suggest that the impact of climatic variation on the vegetation was weaker in the Southern Hemisphere than in the North, because large ice sheets were rare and full glacial conditions did not persist during interglacial periods; thus, forested or partially forested areas existed in mesic habitats of southeastern and southwestern Australia even during the Last Glacial Maximum [2]. Nonetheless, the distributions of particular vegetation types presumably expanded and contracted during the glacial fluctuations of the Upper Pliocene and Pleistocene [3,4]. Also, extreme arid conditions rendered large regions of inland Australia unsuitable for vegetation, leading to the formation of active sand and clay dune systems. Clearly, these modifications would have had a substantial impact on the distributions of all types of fauna in arid temperate areas [5].
Although the impacts of Pleistocene climatic variations on the fauna of southern Australia remain unclear, the past climatic history of southern Australia is relatively well-understood. From the Lower-Middle Miocene (23 mya [million years ago]), Australia was warm and wet, even in areas now covered by desert [6]. Further south in Antarctica, ice was absent until 14 mya [7]. In the Upper Miocene, the disappearance of these stable, warm, wet conditions caused a contraction of rainforest, an expansion of xerophyllous plant species such as Eucalyptus and Casuarinaceae, and the retreat of the southern marine basin (e.g., the Murray and Gippsland basins in southeastern Australia: [5,8]. Although warm and wet conditions (and thus, rainforest) returned in the Lower Pliocene (523 mya), climatic fluctuations (cool-dry to warm-wet) and a global cooling-drying trend through the Pliocene favoured the expansion of woodlands, sclerophyllous forest and grasslands (but still wetter than today; [5,[8][9][10][11][12]). The beginning of the Pleistocene marked a transition to cold and dry conditions, and the lower sea level (to more than 120 m lower; [13,14]) periodically connected continental islands to the mainland (e.g. Tasmania; [15]). Pleistocene climatic fluctuations reached a maximum amplitude over the last 0.4 Myr, with temperatures 10 to 5uC lower during the last Glacial Maximum 21,000 years ago than present-day conditions and/or fully arid conditions (LGM; [5,8,16]). However, glaciation was limited to Tasmania and the southeastern highlands [8].
These climatic fluctuations were likely to exert stronger effects on the distribution and abundance of some species than others. For example, terrestrial ectotherms (such as reptiles) depend upon ambient thermal heterogeneity for behavioural thermoregulation, and hence are sensitive to local climatic conditions (e.g. [17,18]). Consequently, reptiles provide excellent model systems with which to study the impact of past climatic variations, because the genetic signatures of past events are likely to be discernable in these taxa. In the present study, we performed molecular phylogenetic and dating analyses on a widespread genus of lizards (Bassiana, Scincidae, Squamata) from southern Australia (Fig. 1), to document patterns of genetic diversification within this group and to clarify potential impacts of past climatic variations and geographic barriers on gene flow in this taxon.

Phylogenetics analyses
The 97 samples showed 92 different haplotypes of 1412 bp (combined dataset; GenBank accession numbers GU811882-GU812071), containing 414 variable sites, of which 353 were parsimony-informative (excluding outgroups). Because trees from the BA, ML and MP analyses showed very similar relationships, only the ML tree is shown (with ML, MP, and BA support values: Fig. 2).
The genus Bassiana, as well as the three species within it, formed monophyletic units (support of 100, 100 and 1.0 for MP, ML, and BA analyses except for Bassiana platynota where BA analyses provided only poorly support for its monophyly; Fig. 2). However, the phylogenetic relationships between the three species remain unclear.
Major phylogenetic divisions were evident within each of the three Bassiana species. Two well-supported lineages corresponding to different geographic locations were found within B. trilineata The three species are well-separated genetically. Mean K2P genetic distances between species were 11.1% for duperreyi-platynota, 12.3% for duperreyi-trilineata, and 12.8% for platynota-trilineata. The genetic distance between the two lineages of B. trilineata was 2.9%, whereas the genetic distances between the lineages of B. duperreyi varied from 2.3 (XII-XIII) to 5.3% (VIII-XII), and between the lineages of B. platynota from 3.3 (V-VI) to 4.7% (III-IV). Genetic distances within lineages varied from 0.3 (IV) to 1.1% (XII) in B. duperreyi, from 0.8 (VI) to 3.1% (IV; excluding lineage III represented by one sample) in B. platynota, and from 0.5 (II) to 0.8% (I) in B. trilineata.

Dating analyses
The phylogenetic split between the three Bassiana species

Discussion
Our analyses reveal strong biogeographic structure within Bassiana in southern Australia, with an initial diversification of the genus occurring in the Miocene (16.229.7 mya; fig. 2), at about the same time that climatic conditions in southern Australia became less stable and more arid [5]. Diversification within each of the three Bassiana species also began long ago, probably from the Upper Miocene (5.7 Myr) to the Lower Pleistocene (1.6 Myr). Most of this intraspecific diversification occurred during the Pliocene (Fig. 2), a period marked by a global cooling and drying trend [9][10][11]. Plausibly, those climatic changes rendered many parts of Australia unsuitable for these lizards, splitting formerly continuous populations into at least 13 major genetic lineages, each in a different geographic region.
Although mainland Australia and Tasmania currently are separated by Bass Strait (average depth of 60 m: [20]), variations in sea level created land bridges (and thus, opportunities for terrestrial dispersal of lizards) several times over the last 10 Myr. At least two such events occurred during the last 0.5 Myr, with sea levels up to 150 m lower than at present [15]. The considerable antiquity of these phylogenetic divergences within Bassiana also is evident from the diversification of haplotypes within intraspecific lineages (except lineage XIII of B. duperreyi, represented by only one sample). Many of these divergences appear to date back to the Upper Pliocene through to the Lower and Middle Pleistocene. Any impact of the last Glacial Maximum (21,000 years ago; [16]) would be apparent in genetic homogeneity over large areas (reflecting recent recolonisation) and/or the elimination of the genetic signatures of more ancient divergence events (because of extirpation of populations over wide areas). Neither of these patterns is apparent, suggesting that these recent climatic fluctuations had no substantial impact on present genetic diversity within lineages. Conditions over most of southern Australia appear to have remained suitable for population persistence, allowing endemic lineages of these lizards to persist.
In summary, lizards of the genus Bassiana have evolved in southern Australia over a period of at least 10 million years (interspecific divergences), and most of the major intraspecific phylogenetic divergences among this lineage occurred at least 5.7 million years ago to 1.7 million years ago. Climatic fluctuations during the Upper Pleistocene (including the Last Glacial Maximum) do not appear to have substantially affected the extent of areas within lizards could persist, with the result that the genetic signature of ancient divergence events (between and within lineages) remains clearly expressed in modern-day populations in the form of strong geographically-associated genetic structure.
These conclusions accord well with the results of phylogenetic analyses of other reptile and amphibian lineages within southern Australia. Some widespread taxa show little geographic genetic structure, presumably reflecting recolonisation events after older populations were wiped out by relatively recent (Pleistocene) sealevel fluctuations (tigersnake Notechis scutatus; [21]). More commonly, however, reptile and amphibian lineages in southern Australia are highly structured genetically, and reflect the results of ancient diversifications. For example, geckos of the Diplodactylus vittatus complex began to diverge about 20-25 mya, in the Lower Miocene and Pliocene [22]). Unsurprisingly, different lineages have reacted to historical climatic fluctuations in different ways. For example, a deep divergence between mainland Australia and Tasmania is evident in some lineages (e.g., the skinks Lerista bougainvilli [23] and Bassiana duperreyi [current study], and the froglet Crinia signifera [24]) whereas animals in the two areas are similar genetically in other taxa (e.g., the frogs Limnodynastes peronii and L. tasmaniensis [25], the skink Egernia whitii [26], and the snake Notechis scutatus [21]). Plausibly, such interspecific differences in spatial genetic structure reflect interspecific differences in factors such as habitat requirements, past distributions and dispersal capacities.
These consistently strong patterns of genetic structure suggest that much of the modern reptile and amphibian fauna of southern Australia bears a strong stamp of divergence events that occurred many million of years ago, and were caused by climatic events that long preceded Pleistocene climatic fluctuations (i.e. in the Miocene and Pliocene). Phylogenetic studies on arid adapted-species from central Australia also suggest a pattern of diversification dating from the mid-Miocene, driven by increasing aridity at that time rather than by thermal fluctuations during the Pleistocene (see [24] for a detailed review). This pattern offers a striking contrast with studies on Northern Hemisphere taxa, where thermal fluctuations (glacial periods) during the Pleistocene eliminated animals from extensive areas, erasing the genetic signature of earlier endemic radiations (e.g. [29][30][31][32]). Consistent with these studies, a metaanalysis [33] revealed that species of reptiles and amphibians from temperate-zone areas of the Northern Hemisphere tend to be younger (in term of the age of earliest intraspecific diversification) than are taxa found at similar latitudes in the Southern Hemisphere. This long-retained genetic evidence of ancient aridity-driven phylogenetic divergence within presently wideranging ectotherms in southern Australia thus provides strong support for the hypothesis that Pleistocene climatic fluctuations in southern Australia were not sufficiently intense to eradicate reptiles and amphibians from most areas.
We analysed a total of 97 samples of Bassiana from Australian museum collections (Museum Victoria, Australian National Wildlife Collection CSIRO, and South Australian Museum) or collected recently in the field. The study was approved by the University of Sydney Animal Care and Ethics Committee (L04/7-2007/3/4665), and samples were collected in ACT and NSW with the permits LT2009355 and S10826, respectively (see Table S1 for more details). Small tail clips or liver (for samples from museum collections) were used to isolate total cellular DNA, and tissues were placed in 200 mL of 5% Chelex containing 0.2 mg/mL of proteinase K, incubated overnight at 56uC, boiled at 100uC for 10 min, and centrifuged at 13,300 g for 10 min. The supernatant, containing purified DNA, was removed and stored at -20uC.

Phylogenetic analyses
We aligned sequences using BioEdit [40] and by eye. Concordance of the two genes (ND2 and ND4) used to construct the dataset was evaluated using the partition-homogeneity test with the software PAUP* [41]. As these tests did not reveal significance incongruence between the two genes (P = 0.20), we conducted phylogenetic analyses on concatenated sequences (593 bp for ND2+819 bp for ND4); all codon positions were used. For the combined data set, sequences of species from the infraorder Scincomorpha were used to root the trees, as follows: one species of Cordylidae (Cordylus warreni, AB079613), one of Xantusiidae (Lepidophyma flavimaculatum, AB162908), two of Lacertidae (Lacerta viridis, AM176577; Takydromus takydromoides, AB080237), and one of Scincidae (Eumeces egregius, AB016606) from Albert et al. [42] and as in Dubey & Shine [43]. All the non-Bassiana species were used as outgroups.
For Maximum likelihood (ML) analyses, jModelTest 0.1.1 [44,45] was used to select models of DNA substitution. The TrN+I+G model [46] best fitted the dataset using a Bayesian Information Criterion (BIC; [47] Schwarz 1978; Lset base = (0.3559 0.3272 0.0877 0.2292) nst = 6 rmat = (1.0000 13.9692 1.0000 1.0000 7.6755 1.0000) rates = gamma shape = 0.4430 pinvar = 0.1490). ML heuristic searches and bootstrap analyses (1000 replicates) were performed with phyml [44]. Bayesian analyses (BA) were conducted with the GTR model (nst = 6), using MrBayes ver. 3.1.2.1 [48]. Two independent runs were performed, each consisting of four parallel Markov chain Monte Carlo (MCMC) chains of 3 million generations, allowing a good convergence of the independent runs (the average standard deviation of split frequencies being lower than 0.01). Trees were sampled every 100 generations. Burn-in was assessed by comparing the mean and variance of log likelihoods, both using the program Tracer ver. 1.4 [49] and by eye. Tree parameters reached stationarity after a burn-in period of 300,000 generations. Optimal trees were then sampled every 100 generations to obtain the final consensus BA tree and associated posterior probabilities.
Finally, we used Paup* 4.0b10 [40] to perform maximum parsimony (MP) analyses using 100 random additions of sequences followed by tree bisection and reconnection (TBR) branchswapping, and retaining at most 100 trees at each replicate. Branch support was estimated using 1000 bootstrap replicates with the same heuristic settings.

Molecular dating
We estimated divergence times with Beast 1.4 [50] using a coalescent tree prior (adequate to study intraspecific diversification: [51]. Because the fossil record for Australian skinks is limited [52], we used the combination of three different secondary calibration points from the study by Albert et al. [42] on the phylogeny of squamate reptiles based on two different molecular dating methods (BA: Bayesian, PL: penalized likelihood) and as in the recent study of Dubey & Shine [43] on water skinks (see this  Table S1. doi:10.1371/journal.pone.0012982.g003 study for details of methods and calibration points). Additional simulations were run with the same dataset and the same models, but strictly based on two different rates of divergence (1.3% and 2.3% Myr 21 ) derived from other reliable studies on reptiles [53][54][55][56].
Preliminary analyses were performed with an uncorrelated lognormal relaxed clock to test if a strict molecular clock can be rejected for our dataset (ucld.stdev parameter .1 with a frequency histogram not abutting 0). Because in our simulations (BA and PL based calibration, and the two divergence rates 1.3% and 2.3% Myr 21 ) the ''ucld.stdev'' parameters were all ,0.3 with a frequency histogram abutting 0, we chose a strict molecular clock for the analyses [51]. The analyses were performed with two independent chains and 20 million generations and chains were sampled every 1000 generations with a burn-in of 2 million generations.