Evolutionary history and taxonomic position for cryptic species may be clarified by using molecular and cytogenetic methods. The subterranean rodent, the Alay mole vole Ellobius alaicus Vorontsov et al., 1969 is one of three sibling species constituting the subgenus Ellobius Fischer, 1814, all of which lost the Y chromosome and obtained isomorphic XX sex chromosomes in both males and females. E. alaicus is evaluated by IUCN as a data deficient species because their distribution, biology, and genetics are almost unknown. We revealed specific karyotypic variability (2n = 52–48) in E. alaicus due to different Robertsonian translocations (Rbs). Two variants of hybrids (2n = 53, different Rbs) with E. tancrei Blasius, 1884 were found at the Northern slopes of the Alay Ridge and in the Naryn district, Kyrgyzstan. We described the sudden change in chromosome numbers from 2n = 50 to 48 and specific karyotype structure for mole voles, which inhabit the entrance to the Alay Valley (Tajikistan), and revealed their affiliation as E. alaicus by cytochrome b and fragments of nuclear XIST and Rspo1 genes sequencing. To date, it is possible to expand the range of E. alaicus from the Alay Valley (South Kyrgyzstan) up to the Ferghana Ridge and the Naryn Basin, Tien Shan at the north-east and to the Pamir-Alay Mountains (Tajikistan) at the west. The closeness of E. tancrei and E. alaicus is supported, whereas specific chromosome and molecular changes, as well as geographic distribution, verified the species status for E. alaicus. The case of Ellobius species accented an unevenness in rates of chromosome and nucleotide changes along with morphological similarity, which is emblematic for cryptic species.

speciation, hybridization, chromosome painting, cytochrome b gene, nuclear XIST and Rspo1 genes, Robertsonian translocations, synaptonemal complex, Ellobius

An origin of species due to chromosome changes is still debatable (King 1993, Castiglia 2014, Dobigny et al. 2017). The problem of chromosomal speciation is closely connected with the phenomenon of sibling species. Mole voles of the genus Ellobius Fischer, 1814, and some other rodents, such as Mus Linnaeus, 1758, Nannomys Peters, 1876 (Gropp et al. 1972, Capanna et al. 1976, Capanna and Castiglia 2004, Veyrunes et al. 2010, Garagna et al. 2014), and subterranean Spalax Guldenstaedt, 1770, Fukomys Kock et al., 2006, Ctenomys Blainville, 1826 etc. (Wahrman et al. 1969, Nevo et al. 2000, Van Daele et al. 2007, Deuve et al. 2008, Kryštufek et al. 2012, Buschiazzo et al. 2018), demonstrate a broad chromosome variability at the species and intraspecies levels without morphological differences (Lyapunova et al. 1980). The lack of clear morphological characters, by which specimens can be easily distinguished in museum collections, as well as in nature, makes such species problematic for study and protection. New molecular methods, especially DNA sequencing and cross-species chromosome painting, can be a precise approach for studying the most intriguing groups (Graphodatsky et al. 2011).

The genus Ellobius divides into two subgenera: Bramus Pomel, 1892 and Ellobius Fischer, 1814 (Musser, Carleton 2005). The subgenus Bramus includes two species: E. fuscocapillus Blyth, 1843 (2n = 36, XX♀–XY♂), and E. lutescens Thomas, 1897 (2n = 17, X0♀-X0♂) (Matthey 1953, Vorontsov et al. 1969, Lyapunova, Vorontsov 1978). Species of the subgenus Ellobius (E. talpinus Pallas, 1770, E. tancrei Blasius, 1884, and E. alaicus Vorontsov et al. 1969) are cryptic ones, indistinguishable by morphological features (Yakimenko and Vorontsov 1982), the main diagnostic features are distant karyotypes. E. talpinus, E. tancrei, and E. alaicus are unique in mammals. Along with autosomal changes, the species lost the Y chromosome, the Sry gene, and obtained isomorphic XX chromosomes in both males and females (Lyapunova and Vorontsov 1978, Vorontsov et al. 1980, Kolomiets et al. 1991, Just et al. 1995, Romanenko et al. 2007, Bakloushinskaya et al. 2012, Bakloushinskaya and Matveevsky 2018). The study of E. lutescens and E. talpinus whole genomes was not able to reveal any sex determining factors (Mulugeta et al. 2016). The first signs of sex chromosomes heteromorphism in E. talpinus and E. tancrei were observed in the meiotic behaviour of XX chromosomes in males (Kolomiets et al. 1991 2010, Matveevsky et al. 2016 2017).

The northern mole vole, E. talpinus, with 2n = NF = 54 (Ivanov 1967, Romanenko et al. 2007), has no described chromosomal variability, but significant mtDNA variability was revealed recently along its wide range (Bogdanov et al. 2015). The eastern mole vole, E. tancrei has stable 2n = 54, NF = 56 in most of its range, and demonstrates enormous karyotype variability (2n = 54-30) in the Pamir-Alay region (Vorontsov and Radzhabli 1967, Lyapunova et al. 1984 2010, Bakloushinskaya et al. 2013). The third species was described first as a chromosomal form of E. talpinus sensu lato (a chromosomal form of E. tancrei from the modern point of view) with one pair of large metacentric chromosomes and small submetacentrics, specific 2n = 52, NF = 56 (Vorontsov and Radzhabli 1967), and later it was designated as the Alay mole vole E. alaicus (Vorontsov et al. 1969, Lyapunova and Vorontsov 1978). The Alay Valley, the terra typica of the Alay mole vole, extending appr. 180 km from Tajikistan in the west to China in the east between two mountain systems: the Tien Shan and the Pamir. Range of the species was limited to the Alay Valley and the Northern slopes of the Alay Ridge, Tien-Shan (Kyrgyzstan). E. alaicus was listed by IUCN as data deficient species; cytogenetic data are scarce, no molecular study has been made ever (Gerrie and Kennerley 2016).

We studied the G-band structure of the E. alaicus karyotype previously and described a morphological homology for one pair of large metacentrics of the species to the Robertsonian metacentrics of E. tancrei from the Pamir-Alay (Bakloushinskaya 2003). We also discovered different forms of E. alaicus and their hybrids with E. tancrei with 2n = 50-53 from other parts of the Inner Tien-Shan (Lyapunova et al. 1985, Bakloushinskaya, Lyapunova 2003). But the study was incomplete, and application of modern cytogenetical and molecular techniques is required to confirm the karyotype structure, validity of E. alaicus as a species and its distribution.

The main objectives of this study were to reveal the chromosomal variability in E. alaicus and prove species affiliations for mole voles from adjacent to the Alay Valley territories of the Inner Tien-Shan and the Pamir-Alay Mountains. To bring a phylogenetic framework to the delimiting species, we examined the phylogeny of the subgenus Ellobius using the mitochondrial DNA marker, complete cytochrome b gene, cytb, and two nuclear DNA markers, fragments of the XIST (X-inactive specific transcript) and Rspo1 (R-spondin 1) genes.

We analyzed karyotypes or cytb structure, or both, of 116 specimens of E. alaicus and E. tancrei mole voles from 27 localities across the Alay Valley and adjacent territories, as well as 7 E. talpinus specimens from 6 localities of Russia (Fig. 1, Table 1). Fragments of the XIST and Rspo1 genes were studied for nine specimens of three species.

Figure 1.

The geographic location of studied populations of the mole voles E. alaicus (dark triangles) and E. tancrei (dark spots). Localities are numbered as in Table 1. Localities 23–27 are outside the map.

A few studies dealt with Ellobius molecular phylogeny before. Conroy and Cook (1999) studied cytb of two species, E. fuscocapillus and E. tancrei, and their position in the Arvicolinae tree appeared to be unstable in different models. Data on variations of short fragments of nuclear genes (partial LCAT and exon 10 GHR) in E. talpinus and E. tancrei contradicted the conventional view that Ellobius is an ancient group because of simplicity of rooted molars and the peculiar structure of the skull (Abramson et al. 2009). Fabre et al. (2012) re-analyzed these data among others for comparative meta-analyses of the rodent diversity and phylogeny without special attention to Ellobiini. Nevertheless, the genus Ellobius appears to be a young group; its morphological characters indicate adaptation to subterranean life and provide no phylogenetic signal. E. talpinus and E. tancrei separated not earlier than the latest Pliocene and Early Pleistocene between ca. 2.1–1.0 Ma (Abramson et al. 2009). The phyletic lineage leading to the recent E. talpinus includes at least two chronospecies: Late Pliocene-Early Pleistocene E. kujalnikensis and early Middle Pleistocene E. melitopoliensis; E. talpinus was recognized from the late Middle Pleistocene (Tesakov 2009). There are no such data for E. tancrei and E. alaicus.

Here, for the first time, we demonstrated data on molecular, mitochondrial (cytb) and nuclear (XIST and Rspo1 fragments) specificity of E. alaicus. The cytb variability in the subgenus Ellobius, which we demonstrated here, is comparable and even higher than in Ctenomys, subterranean rodents with numerous species-specific chromosome changes (Buschiazzo et al. 2018). In Ctenomys genetic distances, calculated on cytb gene, range from 0 to 2.28%, whereas 2n varies from 41 to 70, and autosomal fundamental numbers (NFa) from 72 to 84. Nevertheless, cytb appears to be more informative for phylogenetic reconstructions compared to nuclear markers. Published data on partial sequences of XIST and Sox9 revealed no differences for E. talpinus and E. tancrei (Just et al. 2007, Bagheri-Fam et al. 2012). Our data on fragments of Eif2s3x and Eif2s3y for E. talpinus, E. tancrei, and E. alaicus also reveal no changes in the exonic part of the genes (Matveevsky et al. 2017). The cryptic Ellobius species are rather young ones, so this may be why nuclear DNA markers were insufficient. However, our new data on XIST and Rspo1 variability demonstrated apparent clustering for all species of the Ellobius subgenus despite interspecific genetic distances were rather low and relatively high difference of E. tancrei specimens from Tajikistan and Uzbekistan, as nuclear markers of the latest (specimen 25159) could not be assigned to any of the two clades.

Figure 6.

Chromosome synapsis in pachytene spermatocytes of E. alaicus, 27024, ♂ (2n = 48, NF = 56), locality #9’. Axial SC elements were identified using anti-SYCP3 antibodies (green), anti-CREST for kinetochores (red). Numbers of SC correspond to chromosome numbers in the karyotype (see Fig. 5). Scale bar: 10 μm.

Originally, E. alaicus was described as a species with specific karyotype structure, including a pair of very large bi-armed chromosomes (Vorontsov et al. 1969, Lyapunova, Vorontsov 1978). Now we proved, that this Rb(2.11) metacentric is the same as in the E. tancrei forms with 2n = 30 and 2n = 48 from the northern bank of the Surkhob River (Bakloushinskaya et al. 2010 2013), but not the Rb(2.18) as in the form with 2n = 50 from the opposite bank of the river. Moreover, translocations Rb(1.3), Rb(4.9), and Rb(3.10) were revealed in the Alay mole voles only. Thus, the Alay mole vole generated a distinctive Robertsonian variability with special structure that highlights genetic distinctness of this species compared to E. tancrei. No specimens with 2n = 52 and a single pair of Rb(2.11) were found among over 400 studied E. tancrei with Rb translocations (Bakloushinskaya, Lyapunova 2003). Probably, the translocation Rb(2.11) originated independently in E. alaicus and E. tancrei. The results of the phylogenetic analyses support this assumption because both ML and BI trees demonstrated distant positions for E. alaicus and E. tancrei specimens carrying Rb(2.11). Their relationships were established indirectly through Uzbekistan and South-West Tajikistan populations of E. tancrei, which have no any Robertsonian translocations (Fig. 7).

Figure 7.

Trees of the subgenus Ellobius inferred from complete mitochondrial cytb gene sequences (1143 bp) of 53 specimens a a tree was got by using the Maximum Likelihood method based on the Tamura-Nei model, bootstrap support is listed above main branches. Only values greater than 70 percent are shown b Bayesian inference tree was made in MrBayes ver. 3.2 (Ronquist et al. 2012), posterior probabilities >0.75 are given above nodes. E. tancrei with 2Rb(2.11) were marked by black spots in both trees.

Earlier (Lyapunova et al. 1990) we obtained the experimental hybrids of E. alaicus, 2n = 52, 2 Rb(2.11) (#3, Fig. 1, Table 1) and E. tancrei with 2n = 50, 2 Rb(2.18), 2 Rb(5.9) from the left bank of the Surkhob River (#18, Fig. 1, Table 1). In meiosis, during pachytene I, chains of chromosomes were described (Lyapunova et al. 1990). Now we can explain the results by the partial, monobrachial homology of Rbs involved in the meiotic chains: Rb(2.11) of E. alaicus and Rb(2.18) of E. tancrei, 2n = 50. Complex chains in meiotic prophase I led to the reduction of fertility in hybrids or even sterility. It might be a possible post-copulation mechanism for reproductive isolation. Here, we demonstrated, that the synapsis and behaviour of E. alaicus (2n = 48) meiotic chromosomes were very similar to E. tancrei and E. talpinus ones (Kolomiets et al. 1991, 2010, Bakloushinskaya et al. 2012, Matveevsky et al. 2016, 2017). Isomorphic sex chromosomes exhibit a functional heteromorphism in the meiotic prophase I in all three species, that is a unique case for mammals.

Therefore, characteristic nucleotide substitutions in mitochondrial and nuclear genes, distinct Rbs variability and independent origin of typical for E. alaicus translocation Rb(2.11) support the species status of the Alay mole vole notwithstanding the closeness to E. tancrei.

The discovery of different heterozygous animals with 2n = 53 and two different Rb translocations raised the question of natural hybridization and mechanisms of genome stability. Animals that carried 1 Rb(2.11) with a high probability were hybrids of E. alaicus, 2n = 52 and E. tancrei, 2n = 54. For the second variant, 2n = 53 and 1 Rb(1.3), two scenarios are possible. The first is the existence of an unknown form (or species) with 2n = 52, 2 Rb(1.3), which hybridized with E. tancrei, 2n = 52, so hybrids of the first generation or backcrosses had 2n = 53, 1 Rb(1.3). Another possibility is that they were remote hybrids of E. alaicus with 2n = 50, 2 Rb(2.11), 2 Rb(1.3) (as animals from the Lake Chatyr-Kel’ vicinities, #4 or Naryn district, #6) and E. tancrei, 2n = 54. In that case, hybrids might have lost the Rb(2.11) in numerous generations under meiotic drive (de Villena and Sapienza 2001, Lindholm et al. 2016). Sociality described in mole voles (Smorkatcheva and Lukhtanov 2014, Smorkatcheva and Kuprina 2018) and underground lifestyle could accelerate the fixation of mutations in disjunct populations.

As we mentioned previously (Bogdanov et al. 2015), the differentiation of wide-ranging steppe species E. talpinus has occurred because of isolation due to geographic barriers, for example, large rivers such as the Volga River and the Irtysh River. E. tancrei and E. alaicus inhabit mountainous steppes and alpine meadows. Mountain ranges might be the most important geographic barriers for the spreading of mole voles because the animals do not inhabit mountains higher than 3500–4000 m above sea level. In the Tien Shan, the Pamir and the Pamir-Alay a distribution of mole voles should be sporadic because suitable habitats are mosaic. The complex orography of the regions may be a main source for geographical separation and ensuing fixation of the chromosomal forms (Bush et al. 1977). The situation is further complicated by the rapid change in the landscape due to neotectonic activity. The Alay Valley is an asymmetric intra-montane sedimentary basin with an average elevation of 2700 m, which formed in response to the convergence between India and Eurasia during the late Cenozoic (Coutand et al. 2002). The Pamir continues to move northward with a large fraction absorbed near the Alay Valley. The highest observed rate of the North-South convergence is between 10 and 15 mm/year as derived from Global Positioning System (GPS) measurements (Zubovich et al. 2016). The Pamir-Tien Shan region accommodates a high deformation over a short distance and is capable of producing magnitude 7 earthquakes in nearly decadal repeat times (Storchak et al. 2013). The last large seismic event was the 2008 magnitude 6.6 Nura earthquake with an epicenter just east of the Alay Valley (Sippl et al. 2014). Large earthquakes, which appeared to be in the Tien Shan and the Pamir, can trigger landslides (Havenith et al. 2003). Mudflows and landslides may quickly separate habitats of subterranean mole voles (Vorontsov and Lyapunova 1984). All three E. alaicus forms (2n = 52, 50 and 48) live in valleys, which are bordered by the mountain ranges. The evident pathways for mole voles spreading are the river banks in canyons crossing the ridges. Mole voles have a complex system of burrows, with at least three horizontal levels and numerous vertical connecting tunnels. But sometimes, most often at night, the animals run out onto the surface and move quickly over the ground. They probably can use human-made bridges, which are often destroyed by flows; new bridges may open a new route for mole voles. The suggestion was inspired when bursts of variations in chromosome numbers in mole voles from the opposite banks of the Vakhsh River were discovered at places close to bridges (Lyapunova et al. 1980 1984). In some cases, as when mole voles inhabit opposite banks of the Kyzyl-Suu River in a deep canyon (localities # 8, 9, Figs 1, 9, 10), we can only explain how animals cross a mountain river if we assume that they use human-made bridges.

Figure 8.

Molecular phylogenetic analysis of three Ellobius species based on variability of XIST and Rspo1 genes fragments (1652 bp in total) and constructed by using the Maximum Likelihood method and the Jukes-Cantor model. Bootstrap support is listed for main branches. Only values over 70 percent are shown.

Figure 9.

Ellobius alaicus, locality #8. Photo by I. Bakloushinskaya.

Despite a complex relief of the region, the geographical barriers are not as strong as genomic ones. We revealed no signs of hybridization in neighbor populations of E. alaicus and E. tancrei yet, i.e. between E. alaicus (2n = 48, locality #8, Fig. 1) and E. tancrei (2n = 30, locality # 16, Fig. 1) or E. alaicus (2n = 48, locality #12, Fig. 1) and E. tancrei (2n = 54, locality # 13, Fig. 1). There are no geographical barriers preventing active contact between these populations in about ten or even few kilometers. In such cases, the assumption that genomic (chromosomal) reorganization in mammals is often rapid (Vorontsov, Lyapunova 1989, Bakloushinskaya 2016, Dobigny et al. 2017) seems plausible, if one considers that polymorphism for isolation traits segregates within populations with different genetic compositions and ecological settings. If we assume that loci, which may contribute to a reproductive barrier, are dispersed throughout the genome, and intragenomic interactions that arise from genetic pathways can maintain species-specific differences (Lindtke and Buerkle 2015, Payseur and Rieseberg 2016), we can consider speciation starting with chromosome changes as a reliable and fast way of speciation.

Figure 10.

Habitat of E. alaicus, the Kyzyl-Suu River Valley, locality # 8. Photo by I. Bakloushinskaya.

The study of E. alaicus demonstrates that the difficulty of species delimitation due to lack of morphological differences might be resolved by using chromosomal and molecular markers.

We assumed, that the independent emergence of Robertsonian translocation Rb(2.11) was crucial for the divergence of ancestors of E. alaicus and E. tancrei, which both developed specific karyotypic variability, more extensive in E. tancrei (2n = 54-30) but distinct due to non-homological (except Rb(2.11)) translocations in E. alaicus (2n = 52–48). Notwithstanding, the closeness of species, which was demonstrated here by studying mitochondrial DNA (cytb) and fragments of two nuclear genes, determines the possibility of sporadic hybridization at the zones of species contacts. Using different cytogenetic methods, G-banding and chromosome painting, along with by cytb, XIST, and Rspo1 genes sequencing allowed us to expand the range of E. alaicus from the terra typica, the Alay Valley (South Kyrgyzstan) up to the Ferghana Ridge and the Naryn Basin, Tien Shan at the north-east and to the Pamir-Alay Mountains (Tajikistan) at the west.

We are sincerely grateful to our colleagues for organizing expeditions and providing material, and, especially, V.S. Lebedev, curator of the myomorph rodents collection, Zoological Museum of Moscow State University; I. Pavlinov, A. Zykov, A. Esipov, E. Bykova, S. Ivnitsky, and E. Elina. N. Mugue and D. Schepetov are particularly thanked for consultations and DNA sequencing. This study was supported in part by the Russian Foundation for Basic Researches N 17-04-00618, N 17-00-00146, IDB RAS Government basic research program N 0108-2019-0007, IMCB SB RAS Government basic research program N 0310-2019-0002, and VIGG RAS State Assignment Contract N 0112-2019-0002.