Amphibia-Reptilia 34 (2013): 579-584

Incongruence between and genetics: three divergent lineages within two of the rare Transcaucasian rat ( hohenackeri)

David Jandzik1,2,∗, Aziz Avcı3, Václav Gvoždík4,5

Abstract. Two morphologically differentiated subspecies with presumably allopatric distribution have traditionally been recognized in the rare Transcaucasian , Zamenis hohenackeri: Z. h. hohenackeri (Caucasus region and E , NW and N ) and Z. h. tauricus (S Anatolia and Levant). Both subspecies are sometimes considered just colour forms of a monotypic . We used sequences of the mitochondrial gene cytochrome b to find out whether both subspecies are also genetically differentiated, or whether their haplotypes form a cluster without clear subdivision. We found that the species forms three divergent lineages: one corresponds to the nominotypic subspecies, another comprises an individual from the region of the type locality of Z. h. tauricus in southern Anatolia and the third, with no available name and separated from the previous by the Amanos (Nur) Mountains, is distributed in the Levant. The pattern of divergence and distribution of the lineages correspond to other species of amphibians and , demonstrating that similar geo-climatic events presumably formed genetic variation and triggered speciation processes in this region.

Keywords: Anatolian Diagonal, Central Anatolian Plateau, , phylogeography, reptiles, Serpentes, systematics.

The Transcaucasian rat snake, Zamenis ho- level (Dalyan, Turkey, voucher of a roadkill ju- henackeri (Strauch, 1873), is a relatively small venile in National Museum in Prague NMP6V slender snake with known distribution in the 71067). Beside the nominotypic subspecies oc- area of the Caucasus, Asia Minor and north- curring in the Caucasus region (type locality: western Levant (fig. 1). It is a highly secretive “Helenendorf” = [Goygol], “Elisabethpol” = terrestrial snake with predominantly crepuscu- [Ganja] in and “Tiflis” = [Tbil- lar and nocturnal activity, which makes it ex- isi] in ; Strauch, 1873), another sub- tremely difficult to encounter and sample in the species, Z. h. tauricus (Werner, 1898) was de- field. It mainly inhabits montane localities at scribed from the Taurus Mountains in south- altitudes between 900 and 3000 m above sea ern Turkey (type locality: Gülek), initially as level (a.s.l.), but in southern parts of the range a species distinct from “Coluber Hohenackeri it can occasionally be found as low as 300 m Strauch” (Werner, 1898). Both subspecies are a.s.l. (Schulz, 1996), or even approaching sea sometimes considered just colour forms of a monotypic species (Nilson and Andrén, 1984; 1 - Department of Zoology, Faculty of Natural Sciences, Bischoff, 1993) – Z. h. hohenackeri has ochre Comenius University in Bratislava, Mlynska dolina B- or brown dorsal spots arranged in two rows sep- 1, 842 15 Bratislava, Slovakia arated by a light vertebral line, while Z. h. tau- 2 - Department of Ecology and Evolutionary Biology (EBIO), University of Colorado, Ramaley N122, Cam- ricus has only one row formed by lateral spots pus Box 334, 80309-0334 Boulder, CO, USA joined together (Venchi and Sindaco, 2006). Ac- 3 - Department of Biology, Faculty of Science and Arts, cording to published data, the ranges of Z. h. Adnan Menderes University, 09010 Aydin, Turkey 4 - Department of Zoology, National Museum, Cirkusova hohenackeri and Z. h. tauricus seem to be sep- 1740, 193 00 Prague, Czech Republic arated from each other by the Central Anato- 5 - Laboratory of Molecular Ecology, Institute of lian Plateau, which partially overlaps with the Physiology and Genetics, Academy of Sciences of the Czech Republic, 277 21 Libechov,ˇ Czech Republic Anatolian Diagonal – mountain ranges passing ∗Corresponding author; across the eastern, central and southern Anato- e-mail: [email protected] lia known to serve as an isolating barrier and

© Koninklijke Brill NV, Leiden, 2013. DOI:10.1163/15685381-00002911 580 Short Notes

Figure 1. Map showing the origin of Z. hohenackeri samples used in this study. The range of Z. hohenackeri in dark grey colour is based on published records reviewed in Schulz (1996) and Sindaco, Venchi and Grieco (2013). It is not clear whether the records from northern central Turkey and western Iran (dark grey circles) represent isolated occurrences or extremes of a continuous distribution. Origin of the samples: Z. h. hohenackeri – 1) Chechenia, Russia, pet trade sample, 2) Boshuri, Georgia, 3) Artvin, Turkey, 4) Trabzon, Turkey (Burbrink and Lawson, 2007); Z. h. tauricus – 5) Adana, Turkey; Z. hohenackeri Southern lineage – 6) Hatay, Turkey, 7) Bcharré (Cedars), . a speciation centre of several animal and plant HZSi1 (5-TATGGGTGRAATGGGAT-3). The aligned and species (see fig. 1; Davis, 1971; Sindaco et al., trimmed sequences used in analyses are 1102 bp in length. The nucleotide sequences of all individual haplotypes of 2000; Kapli et al., 2013; review in Bilgin, 2011). newly obtained Z. hohenackeri (see figs 1, 2), Z. linea- The alleged present pattern of distribution cou- tus (haplotype L1 – Ressutano, Sicily, Italy; L2 – Etna, pled with conflicting opinions on the intraspe- Sicily, Italy) and Z. situla (haplotype S1 – Bacina,ˇ Croa- cific taxonomy raise the question of whether tia) were deposited in the GenBank under accession num- bers KF639740-KF639748. For our analyses we supple- the genetic structure within Z. hohenackeri cor- mented the dataset with Cytb sequences of other species of responds to phenotypically-characterized sub- the Zamenis and related genera (sensu Pyron et al., species, or whether the species represents ge- 2011) retrieved from GenBank: Z. hohenackeri DQ902137, Z. persicus DQ902117, Z. situla DQ902125, Orthriophis netically undifferentiated group. To address this taeniurus DQ902122 (Burbrink and Lawson, 2007); Z. we analysed mitochondrial DNA sequence data longissimus HQ392548, HQ392559-61 (Musilová et al., of Z. hohenackeri from both parts of the range 2011); sauromates AY486931, Rhinechis scalaris AY486932, Coronella austriaca AY486930 (Nagy et al., within the phylogenetic context of the genus Za- 2004). menis. No stop codons were detected in the dataset with the pro- gram DnaSP 5.00 (Librado and Rozas, 2009). The best-fit DNA was collected via oral swabs from live indi- model of sequence evolution was selected by the Akaike viduals, or extracted from liver or muscle pieces from ethanol-preserved or road-killed (see fig. 1). To- information criterion (AIC) as implemented in jModel- tal genomic DNA was extracted from tissue using stan- Test 0.1.1 (Posada, 2008) and MrModeltest 2.3 (Nylander, dard phenol-chlorophorm extraction and the complete mi- 2004), respectively. The phylogeny was reconstructed using tochondrial DNA gene for cytochrome b (Cytb) was am- Maximum likelihood (ML) and Bayesian inference (BI). We plified with primers L14910 and H16064 (Burbrink, Law- employed the program PhyML 3.0 (Guindon et al., 2010) son and Slowinski, 2000) following standard PCR pro- for ML analysis with the best approach combining nearest cedures. The PCR products were sequenced by Macro- neighbour interchanges with subtree pruning and regraft- gen Inc. (Seoul, S. Korea, http://www.macrogen.com) us- ing algorithm and using the TIM2 + G model of sequence ing H16064 primer and newly designed internal sequenc- evolution. Branch support was quantified as bootstrap val- ing primers LZSi1 (5-TTCTCAATYAACGACCC-3)and ues (based on 100 resampled datasets) calculated in PhyML Short Notes 581

3.0. (Guindon et al., 2010). BI was performed with Mr- no taxonomic significance (Nilson and Andrén, Bayes 3.1.2 (Huelsenbeck and Ronquist, 2001; Ronquist 1984; Bischoff, 1993). The Caucasian lineage and Huelsenbeck, 2003). The analysis was set with par- titions for each codon position within Cytb (1st/2nd/3rd corresponds to the nominotypic subspecies Z. h. codon position: GTR + I + G/HKY + G/GTR + I + G) hohenackeri. Another subspecies, Z. h. tauricus, with parameters optimized during the run. Two indepen- was previously thought to comprise all popula- dent BI analyses were run each with four coupled chains tions from southern Anatolia and north-western that were run for six million generations. Parameter and tree samples were saved every 100 generations and the sampled Levant, which are presumably geographically trees, after discarding the first 1/10 of trees as the burn-in, separated from Z. h. hohenackeri by the Central were used to construct a 50% majority-rule consensus tree. Anatolian Plateau. However, our study showed The posterior probabilities were calculated as the frequency of samples recovering any particular clade (Huelsenbeck that this part of the range is inhabited by at and Ronquist, 2001). Uncorrected p-distances among the least two distinct lineages (figs 1, 2). One of sequences were calculated using PAUP* 4.0b10 (Swofford, them could be assigned to the subspecies Z. 2003) and then manually averaged between the taxa. h. tauricus as it contains a sample collected in Among the 1102 bp-long fragments of the the region of the type locality of this taxon. mtDNA gene Cytb, 372 sites are variable and The Southern lineage, seemingly with no avail- 279 of them are parsimony informative in the able name, is represented by samples from ex- dataset comprising all species of the genus Za- treme southern-Turkish locality Hatay and from menis with all outgroups. Both ML and BI Mt. Makmel in Lebanon. All three lineages produced trees with the same topology within are relatively deeply divergent from each other. the genus Zamenis, however with low statisti- The genetic distance between Z. h. hohenack- cal support for most branches (fig. 2). Zamenis eri and other two lineages is more than 5% (in hohenackeri forms a highly supported mono- p-distance), which is equivalent to the genetic phyletic clade in both analyses. It comprises distance between many closely related snake three divergent lineages: one formed by speci- species, such as e.g. species of the crotaline mens from the Caucasus and north-eastern Ana- genus Lachesis, vipers V. ursinii and V. renardi, tolia, representing thus the nominotypic sub- cobras Naja siamensis and N. mandalayensis or species Z. h. hohenackeri, a second corresponds species of the elapid genus Acanthophis (Zamu- to Z. h. tauricus from the Adana Province in dio and Greene, 1997; Slowinski and Wüster, southern Turkey, and a third, which we refer to 2000; Wüster et al., 2005; Gvoždík et al., 2012). as “Southern lineage” throughout the text, con- Comparison of divergences among the species sists of two haplotypes from the Amanos (Nur) of the genus Zamenis, however shows that the and Lebanon Mountains in north-western Lev- genetic distances between the lineages of Z. ho- ant. The genetic distance between the Z. h. ho- henackeri are considerably smaller than those henackeri lineage and the two remaining lin- between any other two Zamenis species (from eages is 5.1% and 5.2%, respectively, in p- 8.7% between Z. longissimus and Z. lineatus distance, while the distance between Z. h. tauri- to 14.5% between Z. hohenackeri and Z. lin- cus and the Southern lineage is 4.3%. The over- eatus; data not shown). The lowest interspe- all high genetic diversity within Z. hohenackeri cific genetic distance is almost twice higher is accentuated by the fact that each of the seven than the distance between Z. h. hohenackeri sampled localities contained a different haplo- and the other two Z. hohenackeri lineages. We type. also lack information from independent mark- The existence of three divergent phylogenetic ers such as nuclear DNA and/or detailed mor- lineages within Z. hohenackeri is in contrast to phological comparison which are essential to the previous taxonomy of this species, which confirm the level of evolutionary divergence. either recognized only two subspecies, or con- Few differences exist in colouration and mor- sidered these groups merely colour forms with phology between Z. h. hohenackeri and popula- 582 Short Notes

Figure 2. Maximum likelihood haplotype tree of 1102 bp fragment of Cytb mtDNA showing the phylogenetic relationships within the genus Zamenis, with emphasis on Z. hohenackeri. The substitution model TIM2 + G was selected under AIC. The statistical support of the branches is expressed as percentage bootstrap values/Bayesian posterior probabilities. Letters/numbers accompanying the terminal clade names are haplotype names; (GB) – sequences obtained from GenBank. Short Notes 583 tions from the southern parts of the range (pre- that other populations phylogenetically related viously referred to as Z. h. tauricus, however to Z. h. tauricus could occur in this area – a containing both mtDNA lineages revealed in similar phylogeographic pattern is known in the this study), but the morphological differentia- treefrog Hyla savignyi (Gvoždík et al., 2010). tion has never been subject to a detailed anal- However, the rat from this region show ysis, nor are any differences known between Z. phenotypic characteristics of Z. h. hohenackeri h. tauricus and the populations corresponding (Afrasiab and Mohamad, 2011), suggesting they to the Southern lineage. Based on the relatively are presumably related to this subspecies. high genetic distance between Z. h. tauricus and Divergence between Z. h. tauricus from the Southern lineage (4.3% in p-distances), the southern Anatolia and the Southern lineage Southern lineage presumably represents a sepa- from Levant, which are geographically sepa- rate subspecies. This, however, needs to be con- rated by the Amanos (Nur) Mountains, corre- firmed by more detailed sampling and indepen- sponds to the pattern observed in the urodelan dent lines of evidence first. Taking all this into amphibian Ommatotriton vittatus (Litvinchuk et account, at this time we refrain from drawing al., 2005), treefrog Hyla savignyi (Gvoždík et any taxonomic conclusions from our results. al., 2010), and in the vipers of the Montivipera Although the knowledge of the distribution xanthina complex (Stümpel and Joger, 2009). of this secretive snake species remains incom- Thus it seems that the split described here re- plete, the data published so far suggest that the flects a more general pattern of distribution of recent range of Z. hohenackeri is fragmented. If amphibians and reptiles in this region, which this is true, it presumably represents remnants could be potentially attributed to the fact that of historically continuous range in the Cauca- the genetic variation of herpetofauna has been sus, Anatolia and Levant. Based on the genetic formed by similar geo-climatic factors. distances between the three lineages and esti- mated mutation rate of Cytb (2% of sequence divergence per Myr; Johns and Avise, 1998), Acknowledgements. We thank P. Boruvka, ˚ D. Jablonski, S. Moeller, P. Vlcekˇ and O. Zahradnícekˇ for providing the basal split within the species occurred about samples, A. Chudárková, P. Chudárková, L. Loudová and 2-2.5 Mya, therefore much later than the up- J. Šmíd for help in the field, and the great Tyler Square lifting of the Central Anatolian Plateau (Mid- for language check. We are grateful to all reviewers for critical reading and commenting the manuscript. Part of this dle Miocene, ca. 12 Mya; Hrbek et al., 2004) study was supported by the Ministry of Culture of the Czech possibly separating the northern and southern Republic (DKRVO 2013/14, National Museum, 00023272). parts of the recent species range. It seems thus plausible that Z. hohenackeri, which represents a snake species capable of living at a relatively References broad altitudinal range, inhabited the continu- Afrasiab, S.R., Mohamad, S.I. (2011): First record of the ous range during the climatic optima in Pliocene rat snake, Zamenis hohenackeri (Strauch, 1873), from or Pleistocene warm periods, before the basal north-eastern Iraq with notes on other colubrid species divergence took place. Although our data sug- (Reptilia: Colubridae). Zool. 54: 19-22. Bilgin, R. (2011): Back to the suture: the distribution of gest that both parts of the recent species range intraspecific genetic diversity in and around Anatolia. are allopatrically inhabited by descendants of Int. J. Mol. Sci. 12: 4080-4103. the basal split (Z. h. hohenackeri in the north- Bischoff, W. (1993): Elaphe hohenackeri (Strauch, 1873), Transkaukasische Kletternatter. In: Handbuch der Rep- ern part, and Z. h. tauricus with the Southern tilien und Amphibien Europas, Band 3/I., Schlangen lineage in the southern part), due to the gap in (Serpentes) I, p. 317-329. Böhme, W., Ed., Aula-Verlag, our sampling from regions of SE Turkey, NW Wiesbaden. Burbrink, F.T., Lawson, R. (2007): How and when did Old Iran, N Iraq, and S Transcaucasia (, World rat snakes disperse into the New World? Mol. Azerbaijan), we cannot exclude the possibility Phylogenet. Evol. 43: 173-189. 584 Short Notes

Burbrink, F.T., Lawson, R., Slowinski, J.B. (2000): MtDNA Nylander, J.A.A. (2004): MrModeltest v2. Program Dis- phylogeography of the North American rat snake tributed by the Author. Evolutionary Biology Centre, (Elaphe obsoleta): a critique of the subspecies concept. Uppsala University, Uppsala. Evolution 54: 2107-2118. Posada, D. (2008): JModelTest: phylogenetic model averag- Davis, P.H. (1971): Distribution patterns in Anatolia with ing. Mol. Phylogenet. Evol. 25: 1253-1256. particular reference to endemism. In: Plant Life of Pyron, R.A., Burbrink, F.T., Colli, G.R., Montes de Oca, South-West Asia, p. 15-27. Davis, P.H., Harper, P.C., A.N., Vitt, L.J., Kuczynski, C.A., Wiens, J.J. (2011): Hedge, I.C., Eds, Botanical Society of Edinburgh, Ed- The phylogeny of advanced snakes (Colubroidea), with inburgh. discovery of a new subfamily and comparison of support Guindon, S., Dufayard, J.-F., Lefort, V., Anisimova, M., methods for likelihood trees. Mol. Phylogenet. Evol. 58: Hordijk, W., Gascuel, O. (2010): New algorithms and 329-342. methods to estimate maximum-likelihood phylogenies: Ronquist, F., Huelsenbeck, J.P. (2003): MRBAYES 3: assessing the performance of PhyML 3.0. Syst. Biol. 59: Bayesian phylogenetic inference under mixed models. 307-321. Bioinformatics 19: 1572-1574. Gvoždík, V., Moravec, J., Klütsch, K., Kotlík, P. (2010): Schulz, K.-D. (1996): Eine Monographie der Schlangengat- Phylogeography of the Middle Eastern tree frogs (Hyla, tung Elaphe Fitzinger. Bushmaster Publications. Hylidae, Amphibia) as inferred from nuclear and mito- Sindaco, R., Venchi, A., Carpaneto, G.M., Bologna, M.A. chondrial DNA variation, with a description of a new (2000): The reptiles of Anatolia: a checklist and zoogeo- species. Mol. Phylogenet. Evol. 55: 1146-1166. graphical analysis. Biogeographia 21: 441-554. Gvoždík, V., Jandzik, D., Cordos, B., Rehák, I., Kotlík, P. Sindaco, R., Venchi, A., Grieco, C. (2013): The Reptiles of (2012): A mitochondrial DNA phylogeny of the endan- the Western Palearctic, Volume 2, Annotated Checklist gered vipers of the Vipera ursinii complex. Mol. Phylo- and Distributional Atlas of the Snakes of Europe, North genet. Evol. 62: 1019-1024. Africa, Middle East and Central Asia, with an Update to Hrbek, T., Stölting, K.N., Bardakci, F., Küçük, F., Volume 1. Edizioni Belvedere. Wildekamp, R.H., Meyer, A. (2004): Plate tectonics and Slowinski, J.B., Wüster, W. (2000): A new cobra (Elapidae: biogeographical patterns of the Pseudophoxinus (Pisces: Naja) from Myanmar (Burma). Herpetologica 56: 257- Cypriniformes) species complex of central Anatolia, 270. Turkey. Mol. Phylogenet. Evol. 23: 297-308. Strauch, A. (1873): Die Schlangen des Russischen Reichs, Huelsenbeck, J.P., Ronquist, F. (2001): MRBAYES: in systematischer und zoogeographischer Beziehung. Bayesian inference of phylogeny. Bioinformatics 17: Mém. Acad. Sci Pétersbourg, Ser. 7 21: 1-288. 754-755. Stümpel, N., Joger, U. (2009): Recent advances in phy- Johns, G.C., Avise, J.C. (1998): A comparative summary logeny and taxonomy of Near and Middle Eastern of genetic distances in the vertebrates from the mito- vipers – an update. ZooKeys 31: 179-191. chondrial cytochrome b gene. Mol. Biol. Evol. 15: 1481- Swofford, D.L. (2003): PAUP*. Phylogenetic Analysis Us- 1490. ing Parsimony (* and Other Methods). Version 4. Sin- Kapli, P., Botoni, D., Ilgaz, C., Kumluta¸s, Y., Avcı, A., auer Associates, Sunderland, USA. Rastegar-Pouyani, N., Fathinia, B., Lymberakis, P., Ah- Venchi, A., Sindaco, R. (2006): Annotated checklist of the madzedeh, F., Polakakis, N. (2013): Molecular phy- reptiles of the Mediterranean countries, with keys to logeny and historical biogeography of the Anatolian species identification. Part 2 – Snakes (Reptilia, Ser- lizard Apathya (, Laceridae). Mol. Phylogenet. pentes). Ann. Mus. Civ. Stor. Nat. “G. Doria” 98: 259- Evol. 66: 992-1001. 364. Librado, P., Rozas, J. (2009): DnaSP v5: a software for Werner, F. (1898): Über einige neue Reptilien und einen comprehensive analysis of DNA polymorphism data. neuen Frosch aus dem cilicischen Taurus. Zool. Anz. 21: Bioinformatics 25: 1451-1452. 217-223. Litvinchuk, S.N., Zuiderwijk, A., Borkin, L.J., Rosanov, Wüster, W., Dumbrell, A.J., Hay, C., Pook, C.E., Williams, J.M. (2005): Taxonomic status of Triturus vittatus (Am- A.J., Fry, B.G. (2005): Snakes across the Strait: trans- phibia: Salamandridae) in western Turkey: trunk ver- Torresian phylogeographic relationships in three ge- tebrae count, genome size and allozyme data. Amph.- nera of Australasian snakes (Serpentes: Elapidae: Acan- Rept. 26: 305-323. thophis, Oxyuranus,andPseudechis). Mol. Phylogenet. Musilová, R., Zavadil, V., Marková, S., Kotlík, P. (2011): Evol. 34: 1-14. Relics of the Europe’s warm past: Phylogeography of Zamudio, K.R., Greene, H.W. (1997): Phylogeography of the . Mol. Phylogenet. Evol. 57: 1245- the bushmaster (Lachesis muta: Viperidae): implications 1252. for neotropical biogeography, systematics, and conser- Nagy, Z.T., Lawson, R., Joger, U., Wink, M. (2004): vation. Biol. J. Linn. Soc. 62: 421-442. Molecular systematics of racers, whipsnakes and rela- tives (Reptilia: Colubridae) using molecular and nuclear markers. J. Zool. Syst. Evol. Res. 42: 223-233. Nilson, G., Andrén, C. (1984): A taxonomic account of Submitted: April 8, 2013. Final revision received: July 11, the Iranian ratsnakes of the Elaphe longissima species- 2013. Accepted: August 16, 2013. group. Amph.-Rept. 5: 157-171. Associated Editor: Sylvain Ursenbacher.