10 Natl Sci Rev, 2019, Vol. 6, No. 1 PERSPECTIVES

GEOSCIENCES Natatanuran used the Indian Plate to step-stone disperse and radiate across the Indian Ocean Downloaded from https://academic.oup.com/nsr/article-abstract/6/1/10/5090988 by University of Kentucky Libraries user on 29 April 2019 Zhi-Yong Yuan1,2,†, Bao-Lin Zhang1,3,†, Christopher J. Raxworthy4, David W. Weisrock5, Paul M. Hime5,6, Jie-Qiong Jin1,7, Emily M. Lemmon8, Alan R. Lemmon9, Sean D. Holland8, Michelle L. Kortyna8, Wei-Wei Zhou1,7, Min-Sheng Peng1,10,JingChe1,7,11,∗ and Elizabeth Prendini4

Investigating the evolutionary history [6–9]. Previous studies addressed some data set assembled for this group. Sam- of widespread higher taxa, subjected to of these issues from palaeobiogeograph- ples include all major lineages, 85 Natata- multiple tectonic events, can provide ical and evolutionary perspectives, but nura species and 20 outgroup taxa (Sup- evidence for or against various palaeo- were inconclusive due to the selection of plementary Data). The novel evidence geographical models of early Earth taxa that did not include all landmasses or reveals how Natatanuran frogs inter- history [1,2]. Contemporary biotic dis- the limited recovery of evolutionary rela- changed between Laurasia and Gond- tributions have been strongly influenced tionships due to the use of sequence data wana around the Indian Ocean during by events associated with the breakup from only a few genes [10–13]. the Cretaceous–Palaeogene, challenging of Gondwana into present-day Africa, The neobatrachian clade Natatanura recent biogeographical assumptions and Antarctica, Australia, South America, are an ideal group to infer Gondwanan providing new insights into Indian Ocean Madagascar and India, during the Late geological and environmental history due biotic exchanges. Mesozoic and Early Palaeogene [2]. The to their ancient origins (divergence from fragmentation of Gondwana and subse- Afrobatrachia at around 100 Ma), high quent tectonic drift ultimately allowed species diversity (>1500 extant species), biotic exchanges between Laurasia and almost cosmopolitan distribution (ab- Gondwana [2,3], influencing the global sent only from Antarctica), general low distributions of many taxa. terrestrial vagility and poor overwater The relative positions of the post- dispersal capabilities [14]. Previous breakup Gondwanan landmasses during studies suggested the divergence of the Late Cretaceous, especially of the In- Natatanura was characterized by a dian and Australian plates around the historical association with the breakup of Indian Ocean, are highly debated [1]. Gondwanan plates [12]. These frogs are The plate reshuffling was probably ac- thus an appropriate group of organisms companied by the formation of multi- to test hypotheses of Cretaceous– KP ple temporary land bridges and involved Palaeogene biotic exchanges between GR biotic exchange among the plates. Al- Laurasia and Gondwana around the though most models agree that the In- Indian Ocean. However, prior studies dian Plate carried a biotic ‘ferry’ of taxa that have included Natatanura failed to (both plants and ) to Asia after resolve the major nodes in its phylogeny Figure 1. Schematic representation of differ- it broke away from other Gondwanan or suffered from incomplete lineage ent hypotheses regarding land connections and landmasses from about 88 to 55 Ma sampling, which, until now, hampered corridors for dispersal among the landmasses [2,4], both geological and paleontologi- conclusive testing of these hypotheses around the Indian Ocean from 88 to 55 Ma. cal data also support land bridges or mi- [12,15]. (1) Africa and India were reconnected to each nor marine barriers that permitted biotic Here, we integrate phylogenetic, other directly [3,5]; (2) Asia and Madagascar exchanges with other Gondwanan land- biogeographic and molecular dating were linked by India, with possible disper- masses (e.g. Africa, Madagascar; Fig. 1) methods to reconstruct the spatiotem- sal between Asia and Madagascar over India [3,5]. There is also debate about whether poral diversification of Natatanura (see and the Seychelles plateau [5]; (3) Antarctic– Australia– and Madagascar were Antarctica–Australia–New Guinea was Supplementary Data). Results resolve connected to: (i) the Indian Plate via the connected by the Gunnerus Ridge (GR) [6,7]; the evolutionary history of Natatanuran (4) Antarctic–Australia–New Guinea and India Kerguelen Plateau (KP) land bridge and frogs (Supplementary Figs 1–3), based were connected by the Kerguelen Plateau (KP) (ii) Madagascar via the Gunnerus Ridge on molecular data from 376 nuclear loci, [13]. Paleo-reconstructions are modified from (GR) land bridge in the Late Cretaceous representing by far the largest molecular Briggs [5] and Bossuyt et al. [12]. PERSPECTIVES Yuan et al. 11

STEPPING-STONE ROLE OF THE from the Indian landmass [18] and that scenario will also be recovered for other INDIAN PLATE FOR BIOTIC the Indian Plate would have been well nonvolant organisms with Indian Ocean EXCHANGE BETWEEN AFRICA, placed to minimize oceanic dispersal dis- distributions. ASIA AND MADAGASCAR tances between Asia and Madagascar, the Indian Plate could have served as a step- Using the traditional Gondwana and ping stone for long-distance dispersal, as

Laurasia model, it has been commonly Downloaded from https://academic.oup.com/nsr/article-abstract/6/1/10/5090988 by University of Kentucky Libraries user on 29 April 2019 suggested previously [13,15]. DISPERSALS WITH assumed that the Indian Plate was an Geological and paleontological ev- ∼ AUSTRALIA–NEW GUINEA isolated island between 88–55 Ma, and idence has previously challenged the the Indian landmass served as an ‘ark’ ‘Indian-biotic Ark’ model. For example, AND ASIA to transport lineages from various biotic Briggs [5] proposed that India was in The Antarctica–Australia–New Guinea groups ‘Out-of-India’ into Asia, after close proximity to other landmasses Plate has been proposed to have been India broke away and drifted northward during its journey northwards because a connected to the Indian Plate by the KP from Gondwana in the Late Cretaceous significant endemic biota did not evolve land bridge, or connected to Madagascar [4]. According to this model, there was on the Indian Plate during this period. by the GR land bridge (Fig. 1)inthe no biotic exchange (which would have Ali and Aitchison [8] proposed the exis- Late Cretaceous [6,7], although these required crossing marine barriers) be- tence of a palaeogeographic connection land bridges have been disputed due to a tween India and its nearby landmass after between Madagascar and India in the lack of evidence that they were sub-aerial ∼ it separated from Madagascar 88 Ma, Late Cretaceous, which may have been during this period [8,9]. Concerning and dispersal events only resumed after formed by the Seychelles–Mascarene Natatanura diversification, we find no India’s collision with Asia in the early Plateau. More recently, Chatterjee et al. sister relationships between frogs from ∼ Eocene ( 65–55 Ma, [16]). ‘Rafting’ [3] argued that biotic links were possibly Antarctica–Australia–New Guinea and of the flora and fauna on the Indian re-established between India and Africa either India or Madagascar (Fig. 2), and Plate enabled unidirectional migration of during the Late Cretaceous, during In- thus find no support for biota exchanges Gondwanan taxa into Asia. However, our dia’s collision with the Kohistan–Ladakh among these landmasses via a KP and phylogenomic results reject this model. Arc along the Indus Suture in the Late GR land bridge. In addition, we date We do not find any periods between 88 Cretaceous. Sister relationships and Natatanura dispersal into Australia–New and 55 Ma, when there was no biotic divergence dates for some vertebrate Guinea to be much later than these exchange occurring between Africa fossil groups (Supplementary Data) also hypothetical Late Cretaceous land and India, India and Asia, or India and support stepping-stone biotic exchange bridges. Madagascar. In contrast, our ancestral via the Indian Plate, consistently with Bossuyt et al.[12] suggested that reconstruction suggests Natatanura our results. However, we did not find Australia–New Guinea acted as a raft, en- originated in Africa and then dispersed any studies of nonvolant extant groups abling Gondwanan Natatanura frogs to ∼ to Asia through India 75.6–72.8 Ma that provide substantial evidence for colonize Southeast Asia, although most (Fig. 2 and Supplementary Fig. 3). It using the Indian Plate as a stepping-stone of their basal relationships were not is unlikely that frogs could cross a large route among these three plates from 88 well resolved. Our genomic-based esti- saltwater barrier, although a few extant to 55 Ma. Although several plant and mates of phylogeny, divergence times species may have made more modest groups exhibit sister relationships and biogeographic reconstruction cast Ptychadena oceanic dispersals (e.g. between India (or Asia) and Madagascar doubt on this dispersal route. We found mascareniensis ,[17]). Briggs [5] and (or Africa) (e.g. [19,20]), their deep strong support that the two Australia– et al Chaterjee .[3] suggested a geograph- divergences (>88 Ma) attribute this New Guinea clades, Cornufer and Papu- ical model in which there were ‘corridors’ relationship to ancient vicariance coin- rana, were embedded within Asian clades or ‘landspans’ that reconnected Africa ciding with the breakup of Gondwana. of and Ranidae, re- ∼ ∼ and India from 75 to 60 Ma. Our Nevertheless, other taxa with younger spectively. Ancestral state analysis sug- topologies and divergence time estimates divergences, and distributions that gests Asian origins, followed by migration are consistent with this scenario. A ter- thus cannot be attributed to ancient into Australia–New Guinea for Cornufer restrial route possibly existed from Africa Gondwanan vicariance, will represent (30.2 Ma, 95% highest posterior density to Asia via India, which would have good candidates for testing ‘Indian (HPD): 21.3–40.0 Ma) and Papurana allowed frogs to disperse among these stepping-stone hypotheses’ in the future (14.9 Ma, 95% HPD: 10.4–19.5 Ma) in- landmasses. Moreover, Malagasy man- (e.g. microhylid frogs, [21]). We provide dependently (Fig. 2 and Supplementary tellid frogs are phylogenetically deeply the first case of extant taxa that appears to Fig. 3). These dispersals could have oc- nested within the larger Asian clade and have taken advantage of the Indian Plate curred after the Australia–New Guinea the ancestral reconstruction supports as a stepping-stone route between Africa, Plate first collided with Sundaland in this clade originating from Asia and dis- Asia and Madagascar, although the exact the Early Miocene [22]. Tectonic colli- persing to Madagascar (Fig. 2). Taking positions of land bridges or traversable sion and extensive island formations in into account that the oldest known rha- ocean channels still remain unclear. Wallacea provided a direct colonization Indorana prasadi cophorid fossils ( )are And we predict that this geographic route between Asia and , and 12 Natl Sci Rev, 2019, Vol. 6, No. 1 PERSPECTIVES Downloaded from https://academic.oup.com/nsr/article-abstract/6/1/10/5090988 by University of Kentucky Libraries user on 29 April 2019

Figure 2. Ancestral-area estimations for the species of Ranoidea, using the DEC+J model in BioGeoBEARS. Circles at nodes represent the set of possible ancestral areas and the colours reflect biogeographic designations (see area code key). Clades of interest are numbered in boxes. Models show the stepping-stone role of the Indian Plate to biotic exchange between India and Africa (I) and among India, Asia and Madagascar (II). Paleo reconstructions are modified from Chatterjee et al.[3] and Briggs [5].

MULTIPLE DISPERSALS FROM are suggested to have triggered much of light on this issue, although those that do the biotic interchange between the re- show similar patterns with evidence ASIATOAFRICA gions. For example, Miocene dispersals of immigrations [26]. Additional studies Reconnections between Gondwana between these two regions are known with increased taxon sampling should and Laurasia-origin landmasses in the in plants [23], birds [24] and mammals help to identify the common time peri- Neogene allowed extensive biotic inter- [25]. However, few dated phylogenies ods and directions of dispersals taken by changes between Africa and Eurasia [3]. of herpetofauna exist that directly shed these species. These biotas could have migrated across PERSPECTIVES Yuan et al. 13 the western margin of the Mediterranean FUNDING 10Sino-Africa Joint Research Center, Chinese Academy of Sciences, China Sea or through the Afro-Arabian to This work was supported by the programs of the 11 Eurasian land bridge [27]. Our results Strategic Priority Research Program of the Chinese Center for Excellence in Animal Evolution and suggest three groups of ranoid frogs Academy of Sciences (CAS) (XDA20050201), Genetics, Chinese Academy of Sciences, China ∗ dispersed independently from Asia to the National Natural Science Foundation of Corresponding author. Africa. The dispersals of Ranidae (Amni- China (NSFC) (31672268, 31622052), the E-mail: [email protected] † rana) and Rhacophoridae (Chiromantis) International Partnership Program of Chinese Equally contributed to this work. Downloaded from https://academic.oup.com/nsr/article-abstract/6/1/10/5090988 by University of Kentucky Libraries user on 29 April 2019 appear to have occurred in a similar Academy of Sciences (152453KYSB20170033) time period: ∼21.6 and ∼20.6 Ma, and the Animal Branch of the Germplasm Bank of respectively (Fig. 2 and Supplementary Wild Species, CAS (Large Research Infrastructure Fig. 3). This period is consistent with Funding) to J.C.; the National Science Foun- REFERENCES dation (USA) BIO-DEB 1021247 to E.S.P. and collision of the Afro-Arabian Plate with 1. Upchurch P. Trends Ecol Evol 2008; 23: C.J.R.; BIO-DEB 135500 to D.W.W.; BIO-GRFP Eurasia during the mid-Burdigalian 229–36. ∼ 3048109801 to P.M.H.; BIO-DEB 1021299 to ( 19–21 Ma) causing the emergence K.M. Kjer; BIO-DEB 1120516 to E.M.L.; and 2. Li J, Li Y and Klaus S et al. Proc Natl Acad Sci USA of a terrestrial corridor, called ‘the NSF-IIP to A.R.L. and E.M.L.; the Strategic 2013; 110: 441–3446. Gomphotherium Land Bridge’ [28]. This Priority Research Program, CAS (XDPB020406), 3. Chatterjee S, Goswami A and Scotese CR. Gond- land bridge later became disconnected the National Key Research and Development wana Research 2013; 23: 238–67. intermittently, but it appears to have Program of China (2017YFC0505202), South- 4. Scotese CR. Atlas of Earth History. Arlington: Uni- been continuously present since ∼15 Ma east Asia Biodiversity Research Institute, CAS versity of Texas, 2001. ago, triggering mammals [29], reptiles (Y4ZK111B01: 2017CASSEABRIQG002) to 5. Briggs JC. J Biogeogr 2003; 30: 381–8. W.-W.Z.; the Sino-Africa Joint Research Center, [30], invertebrate [31] and possibly also 6. Case JA. J Vert Paleont 2002; 22: 42A. CAS (SAJC201611) to M.-S.P.; the funding from frogs (our results) to exchange between 7. Hay WW, DeConto RM and Wold CN et al. Al- the State Key Laboratory of Genetic Resources and Africa and Eurasia. Interestingly, our ternative global Cretaceous paleogeography. In: Evolution (GREKF18-15) and NSFC (31501843, results also support another colonization 31702008) to Z.-Y.Y.; J.C and M.-S.P are supported Barrera E and Johnson CC (eds). Evolution of from Asia to Africa by a lineage of by the Youth Innovation Promotion Association the Cretaceous Ocean Climate System. Boulder Hoplobatrachus Dicroglossidae ( ), which CAS. CO: Geological Society of America Special Paper, occurred much later (∼12.7 Ma). This 1999, 1–48. implies that habitats in the North Africa Zhi-Yong Yuan1,2,†, Bao-Lin Zhang1,3,†, 8. Ali JR and Aitchison JC. Earth-Sci Rev 2008; 88: and Afro-Arabian plates were suitable for Christopher J. Raxworthy4, David W. Weisrock5, 145–66. dispersal during the middle Paul M. Hime5,6, Jie-Qiong Jin1,7, Emily 9. Ali JR and Aitchison JC. J Biogeogr 2009; 36: Miocene. As a consequence of shrinkage M. Lemmon8,AlanR.Lemmon9, Sean D. Holland8, 1778–84. of the Tethys Sea, desert conditions Michelle L. Kortyna8, Wei-Wei Zhou1,7, 10. Samonds KE, Godfrey LR and Ali JR et al. PLoS expanded across North Africa in the late Min-Sheng Peng1,10,JingChe1,7,11,∗ and ONE 2013; 8:4. Miocene (∼7 Ma), marking the origin of Elizabeth Prendini4 11. Verma O, Khosla A and Goin FJ et al. New the Sahara Desert, and also the deserts of 1State Key Laboratory of Genetic Resources and Mexico Mus Nat Hist Sci Bull 2016; 71: the Middle East and the Arabian Penin- Evolution, Kunming Institute of Zoology, Chinese 317–30. sula [32], which subsequently hindered Academy of Sciences, China 12. Bossuyt F, Brown RM and Hillis DM et al. Syst the migration of most mesic-adapted 2Key Laboratory for Forest Resources Conservation Biol 2006; 55: 579–94. species between Africa and Eurasia. and Utilization in the Southwest Mountains of 13. Crottini A, Madsen O and Poux C et al. Proc Natl China, Ministry of Education, Southwest Forestry Acad Sci USA 2012; 109: 5358–63. SUPPLEMENTARY DATA University, China 14. University of California, Berkeley. Amphib- 3Kunming College of Life Science, University of iaWeb: Information on Amphibian Biology Supplementary data are available at NSR online. Chinese Academy of Sciences, China and Conservation. https://amphibiaweb.org/ 4Department of Herpetology, American Museum /index.html (15 February 2018, date ACKNOWLEDGEMENTS of Natural History, USA last accessed). ). 5Department of Biology, University of Kentucky, 15. Feng YJ, Blackburn DC and Liang D et al. Proc Natl We extend our sincere gratitude to Robert W. Mur- USA Acad Sci USA 2017; 29: E5864–70. phy, Amy Lathrop, Abigail Cramer, Alan Resetar, 6Biodiversity Institute, University of Kansas, 16. Aitchison JC, Ali JR and Davis AM. J Geophys Res Jimmy A. McGuire, Carol Spencer, Robert Drewes, USA 2007; 112: B05423. Jens Vindum, Bryan L. Stuart, Zoe Davids, Aaron 7 Bauer, Raj Patel and John Smith for the use of tis- Southeast Asia Biodiversity Research Institute, 17. Vences M, Kosuch J and Rodel M et al. J Biogeogr sues held in their collections. We are grateful to Chinese Academy of Sciences, Myanmar 2004; 31: 593–601. 8 Alyssa Bigelow Hassinger, Hannah Ralicki, Kirby Department of Biological Science, Florida State 18. Folie A, Rana RS and Rose KD et al. Acta Palaeon- Birch and Ameer Jalal at the Florida State Univer- University, USA tol Pol 2013; 58: 511–24. sity Center for Anchored Phylogenomics for assis- 9Department of Scientific Computing, Florida State 19. Kamei RG, San Mauro D and Gower DJ et al. Proc tance with molecular data collection and bioinfor- University, USA Biol Sci 2012; 279: 2396–401. matics analysis. 14 Natl Sci Rev, 2019, Vol. 6, No. 1 PERSPECTIVES

20. Thomas N, Bruhl JJ and Ford A et al. J Biogeogr 25. Lecompte E, Aplin K and Denys C et al. BMC Evol 30. Tamar K, Carranza S and Sindaco R et al. Mol Phy- 2014; 41: 894–904. Biol 2008; 8: 199–220. logen Evol 2016; 103: 6–18. 21. van der Meijden A, Vences M and Hoegg S et al. 26. Byrne M, Steane DA and Joseph L et al. J Bio- 31. Harzhauser M, Kroh A and Mandic O et al. Zool Mol Phylogen Evol 2007; 44: 1017–30. geogr 2011; 38: 1635–56. Anz 2007; 24: 241–56. 22. Hall R. J Asian Earth Sci 2002; 20: 353– 27. Duggen S, Hoernle K and van den Bogaard P et al. 32. Zhang Z, Ramstein G and Schuster M et al. Nature 431. Nature 2003; 422: 602–6. 2014; 513: 401–4.

23. Sniderman JMK and Jordan GJ. J Biogeogr 2011; 28. Bosworth W, Huchon P and McClay K. J Afr Earth National Science Review Downloaded from https://academic.oup.com/nsr/article-abstract/6/1/10/5090988 by University of Kentucky Libraries user on 29 April 2019 38: 1445–55. Sci 2005; 43: 334–78. 6: 10–14, 2019 24. Moyle RG, Oliveros CH and Andersen MJ et al. 29. Antoine PO, Welcomme J and Marivaux L et al. J doi: 10.1093/nsr/nwy092 Nat Comms 2016; 7:12709. Vertebr Paleontol 2003; 23: 977–80. Advance access publication 5 September 2018