Confirmation of European snow vole Chionomys nivalis (Mammalia: Rodentia: Cricetidae) occurrence in Portugal

Abstract

The European snow vole (Chionomys nivalis) is a microtine rodent with a highly fragmented distribution range, mostly associated with the main mountain systems from southern Europe to Turkmenistan. In this paper we confirm the occurrence of the snow vole in Portugal, based on morphological characteristics, biometrics and genetic analysis of two individuals captured in the Montesinho Mountain range (northeastern Portugal). Both mitochondrial and nuclear genetic markers were used to confirm the species identity. The analysis of cytochrome b supports previous conclusions on the phylogeographic structure of the species, revealing the existence of several distinct lineages. Moreover, it shows that the Portuguese specimens are closely related to the other Iberian populations. This finding is of great interest as it adds new information regarding the spatial distribution of the snow vole, by redefining the southwestern limits of the species’ range, and it highlights the need for accurate assessment of regional small mammal population trends and conservation status.

Keywords:

• Rodent
• Chionomys nivalis
• cytochrome-b
• IRBP
• Iberian Peninsula

Introduction

The European snow vole Chionomys nivalis (Martins, 1842) is a rock-dwelling microtine that lives in mountain regions (1000–4700 m; Amori 1999; Luque-Larena et al. 2002; Pérez-Aranda et al. 2007b). The species inhabits stable rocky formations with open slopes and patches of mixed shrubs and grass meadows (LeLouarn & Janeau 1975; Jones & Carter 1980; Janeau & Aulagnier 1997; Luque-Larena et al. 2002), and displays some morphological and behavioural characteristics, such as excellent climbing ability, long tail (nearly half the body length), long vibrissae (up to 60 mm) and well-developed hind feet, which are considered to be an adaptation to those habitats (Figure 1; Janeau & Aulagnier 1997; Luque-Larena et al. 2002; Darvish et al. 2005).

Figure 1. (a) Habitat features of the sampling site; (b) the juvenile female snow vole captured; (c) the adult male snow vole captured (this picture was obtained through infrared beam cells).

The distribution of the European snow vole is wide, ranging from southwestern Europe through southeastern Europe and the Caucasus, to Iran and Turkmenistan (Wilson & Reeder 2005). However, due to the specificity of its ecological requirements, the species distribution is highly fragmented and restricted to mountainous regions including the Pyrenees, the Alps, the Apennines, the Tatra Mountains, the Carpathians and the Balkan Peninsula (Nadachowski 1991; Nappi 2002; Darvish et al. 2005; Pérez-Aranda et al. 2007a). Although considered a widespread species with stable populations and an International Union for Conservation of Nature (IUCN) conservation status of least concern (LC), the snow vole is regionally protected, being included in the Mammal Red Books of Spain, Italy and Ukraine (Pérez-Aranda et al. 2007b; Kryštufek & Amori 2008; Bertolino et al. 2014). Indeed, in a recent study that defined a ranking system for small mammal conservation priorities, the snow vole in Italy was identified as one of those with the highest score, due to its ecological specialisation and potential vulnerability to habitat changes (Bertolino et al. 2014).

Several studies have focused on the evolutionary history of the species through the analysis of mitochondrial and nuclear DNA (Castiglia et al. 2009; Yannic et al. 2012; Bannikova et al. 2013). The first of these, based on mitochondrial DNA, suggested a European origin of the species; however, the later studies, which sampled a wider range of individuals and analysed both mitochondrial and nuclear markers, supported a Middle Eastern origin (Yannic et al. 2012; Bannikova et al. 2013). Nonetheless, they all demonstrate the existence of a mitochondrial phylogeographic structure with several lineages that occupy geographically congruent ranges, mostly associated with previously described subspecies (Castiglia et al. 2009; Yannic et al. 2012; Bannikova et al. 2013). The Iberian Peninsula represents the southwestern limit of the snow vole distribution range and is occupied by one of the lineages that were highly supported in all of the studies. This lineage includes sequences from two populations (the Pyrenees and Central System) of the five main core areas that have been identified in Spain (Pérez-Aranda et al. 2007b). The Iberian populations, particularly those from southern areas, are already considered of conservation concern due to their highly fragmented distribution and low densities (Pérez-Aranda et al. 2007b). There are no historical or recent records of the species’ occurrence in Portugal, despite the presence of suitable habitat in both the northern (Serra do Gerês and Serra de Montesinho) and central (Serra da Estrela) regions. This study describes the first record of Chionomys nivalis in Portugal, which is confirmed by morphologic characteristics, biometrics and genetic analysis, and discusses the importance of this new distributional data in terms of the species’ genetic structure and conservation.

Materials and methods

In October 2014, live trapping was performed during six consecutive days in Lama Grande, Montesinho Mountain (41°59′00″N, 6°47′24″W; Bragança, northeastern Portugal) at an elevation of 1370 m. The habitat is heterogeneous and includes stable and extensive patches of large granite rocks, together with shrubs and low herbaceous cover (Figure 1; Carvalho 2005). A total of 27 Sherman traps (23 × 7.5 × 8.5 cm) were set, baited with fresh apples and carrots, and with hydrophobous cotton as bedding material. The two snow voles captured were weighed and sexed, and the reproductive status and standard biometric measurements were recorded (Barnett 1992; Blanco 1998; Gurnell & Flowerdew 2006). The animals were released immediately at the trapping location.

Tissue samples from the snow voles captured (1 mm ear punch) were collected to confirm the species identity by molecular analysis. DNA was extracted using an Easy Spin^® Genomic DNA Minipreps Tissue Kit (Citomed, Lisbon, Portugal) following the manufacturer’s instructions. The amplification and sequencing of one mitochondrial gene [cytochrome b (cyt-b)] and one nuclear gene [interphotoreceptor retinoid-binding protein (IRBP)] were then performed, according to the procedures described in Barbosa et al. (2013). The sequences obtained were visually inspected and assembled in BioEdit version 7.2.4 (Hall 1999) and were subsequently matched to the reference data set compiled by Barbosa et al. (2013) for species identification. In addition, 41 cyt-b snow vole sequences and three IRBP sequences were retrieved from GenBank (Table I) and aligned with the sequences obtained from the specimens captured, using the CLUSTAL W algorithm (Thompson et al. 1994) as implemented in BioEdit (Hall 1999). The cyt-b 855-bp alignment was reduced into haplotypes with DnaSP 5.10 (Librado & Rozas 2009). A median-joining network was then constructed for the cyt-b alignment, to represent the evolutionary relationships among haplotypes, with the algorithm implemented in the software Network 4.6.1.3 (Bandelt et al. 1999; http://fluxus-engineering.com).

Table I. Locality and accession numbers of the Chionomys nivalis cyt-b and interphotoreceptor retinoid-binding protein (IRBP) sequences retrieved from GenBank and used in the analysis. Number of sequences is given in parentheses.

Country	Locality	GenBank Acc. No. (n)		References
		cyt-b	IRBP
Andorra	Andorra	JX457754 (1)	JX457672 (1)	Barbosa et al. (2013)
Andorra	Andorra		JX457673 (5)	Barbosa et al. (2013)
Spain	Girona	AY513848 (1)		Jaarola et al. (2004)
Spain	Girona	JX457753 (1)		Barbosa et al. (2013)
Spain	Sierra de Gredos	AM392367 (1)		Galewski et al. (2006)
Spain	Sorpe	JX457752 (1)	JX457671 (1)	Barbosa et al. (2013)
Italy	Duchessa	GQ150802 (1)		Castiglia et al. (2009)
Italy	Gran Sasso	GQ150799 (1)		Castiglia et al. (2009)
Italy	Marta Alpi Liguri	GQ150794 (1)		Castiglia et al. (2009)
Italy	Marta Alpi Liguri	GQ150800 (1)		Castiglia et al. (2009)
Italy	Trento	AY513845 (1)		Jaarola et al. (2004)
Italy	Trento	AY513846 (1)		Jaarola et al. (2004)
Italy	Val Masino	GQ150795 (1)		Castiglia et al. (2009)
Italy	Val Masino	GQ150796 (1)		Castiglia et al. (2009)
Italy	Valle d’Aosta	GQ150797 (1)		Castiglia et al. (2009)
Italy	Valle d’Aosta	GQ150798 (1)		Castiglia et al. (2009)
Italy	Valle d’Aosta	GQ150801 (1)		Castiglia et al. (2009)
Switzerland	Derborence	GU954316 (1)		Fink et al. (2010)
Switzerland	Derborence	GU954317 (1)		Fink et al. (2010)
Slovenia	Mt Sneznik	GQ150792 (1)		Castiglia et al. (2009)
Slovenia	Mt Sneznik	GQ150793 (1)		Castiglia et al. (2009)
Slovakia	West Tatra Mts	AY513847 (1)		Jaarola et al. (2004)
Macedonia	Mt Pelister	GQ150791 (4)		Castiglia et al. (2009)
Bulgaria	Vitosha Mountain	CQ352464 (1)		Bannikova et al. (2013)
Romania	Eastern Carpatians, Breskul Mountain	CQ352465 (1)		Bannikova et al. (2013)
Adygea	Caucasian Reserve, Abago Mountain	JX440342 (2)		Bannikova et al. (2013)
Georgia	Pass of Tskhra-tskharo, Transcaucasia	HQ901803 (2)		Yannic et al. (2012)
Georgia	Pass of Tskhra-tskharo, Transcaucasia	HQ901799 (1)		Yannic et al. (2012)
Georgia	Pass of Tskhra-tskharo, Transcaucasia	HQ901800 (1)		Yannic et al. (2012)
Georgia	Pass of Tskhra-tskharo, Transcaucasia	HQ901801 (1)		Yannic et al. (2012)
Georgia	Bacuriany, Caucasus	HQ901802 (2)		Yannic et al. (2012)
Turkey	Ciglikara	GQ150786 (1)		Castiglia et al. (2009)
Turkey	Ciglikara	GQ150787 (1)		Castiglia et al. (2009)
Turkey	Ciglikara	GQ150788 (1)		Castiglia et al. (2009)
Turkey	Central Taurus	JN244707 (1)		Bannikova et al. (2013)
Syria	Saleh	AY513849 (1)		Jaarola et al. (2004)
Israel	Mt Hermon	GQ150789 (1)		Castiglia et al. (2009)
Israel	Mt Hermon	GQ150790 (1)		Castiglia et al. (2009)
Turkmenistan	Ashabad, Kopet-Dag	HQ901804 (1)		Yannic et al. (2012)
Turkmenistan	Ashabad, Kopet-Dag	HQ901805 (1)		Yannic et al. (2012)
Turkmenistan	Ashabad, Kopet-Dag	HQ901806 (1)		Yannic et al. (2012)
Turkmenistan	Ashabad, Kopet-Dag	HQ901807 (1)		Yannic et al. (2012)

A Bayesian coalescent genealogy of the C. nivalis cyt-b sequences was also inferred using BEAST version 1.8.0, inputting through BEAUti version 1.8.0 (Drummond & Rambaut 2007). A two-partition (first and second codon position linked; third position separate) HKY (Hasegawa, Kishino and Yano) (+Γ) nucleotide substitution model was used for this, as recommended for protein coding sequence data (Shapiro et al. 2006). Three replicate runs of 100 million generations were performed using a constant-size population model and with a sampling frequency that provided a total of 10,000 samples for each run. Tracer version 1.5 (Rambaut & Drummond 2007) was used to assess convergence of the runs. The results were then combined, after the removal of 10% burn-in, and a representative genealogy was identified from the posterior sample of trees.

Results and discussion

The two snow voles captured during the trapping session, one juvenile female and one adult male, exhibit morphological characteristics (Figure 1) and biometric measurements within the range described for the species (Table II; Luque-Larena et al. 2004; Darvish et al. 2005; Pérez-Aranda et al. 2007b; Metcheva et al. 2008). Although the comparison with previous studies of snow voles from Iran and Bulgaria shows that the tail of the captured male is relatively long, the length is compatible with the range of tail size that has been found in Spanish populations (Table II).

Table II. Age, sex, number of specimens analysed in each study (N) and biometric measures of Chionomys nivalis (standard errors are given in parentheses when available). According to body weight and hair colour, two age classes, juveniles and adults (body mass > 35 g), were defined (LeLouarn & Janeau 1975; Luque-Larena et al. 2002; Pérez-Aranda et al. 2007b).

Country	Age	Sex	N	Body weight (g)	Head-body length (mm)	Tail length (mm)	Hind foot length (mm)	Ear length (mm)	References
Portugal	A	M	1	49.0	122.4	72.8	22.0	17.3	This study
	J	F	1	27.2	103.1	53.7	19.5	14.5	This study
Spain	A	M	111	50.3 (0.9)	–	–	–	–	Luque-Larena et al. (2004)
	A	F	85	45.4 (0.9)	–	–	–	–
	–	–	NI	35.0–70.0	92.0–140.0	45.0–74.0	17.5–25.0	12.5–21.0	Pérez-Aranda et al. (2007b)
Iran	–	–	4	–	104.0–116.0	58.0–60.0	18.0–19.0	12.0–13.0	Darvish et al. (2005)
Bulgaria	A	M	11	54.8 (11.7)	143.4 (19.6)	61.5 (4.5)	21.6 (1.6)	14.4 (2.8)	Metcheva et al. (2008)
	A	F	13	50.5 (12.2)	128.2 (17.0)	63.4 (9.1)	20.6 (1.3)	14.5 (3.0)

A – adult; J – juvenile; M – male; F – female; NI – no information.

Genetic analysis of both mitochondrial and nuclear markers confirmed the species identity. The 1143-bp sequences obtained for cyt-b corresponded to the same haplotype (GenBank accession number KT175903) and showed 11 nucleotide substitutions (10 transitions and one transversion) from the most closely related haplotype, which came from Sierra de Gredos in Spain (AM392367). Two different 1040-bp sequences were obtained for the IRBP gene (GenBank accession numbers KT175904 and KT175905), and these had three polymorphisms in relation to the published data (two transitions and one transversion; JX457671-JX457673, Barbosa et al. 2013).

The analysis of the evolutionary relationships of cyt-b haplotypes revealed several geographic clusters, in accordance with previous studies on the species’ phylogeography (Figure 2; Castiglia et al. 2009; Yannic et al. 2012; Bannikova et al. 2013). In the phylogenetic analysis three main clades were retrieved with high support (Figure 2), concordant with the results obtained by Yannic et al. (2012). Moreover, there is a haplotype from Turkey that does not cluster with any of the main groups (JN244707; Bannikova et al. 2013), showing a divergence of about 9% to these clades. This particular specimen has already been reported by the previous authors as divergent both at the mitochondrial and nuclear levels (Bannikova et al. 2013). Considering the three main clades, the most divergent includes the specimens from Turkmenistan (C. nivalis dementievi) and has an uncorrected p-distance of 3.5 to 4.0% from the other clades. The other two clades, which correspond to the Near East and western European specimens, are more closely related (uncorrected p-distance of 1.8% ± 0.34) and each includes several highly supported lineages [Near East: Turkey, Israel, Syria and Georgia; western Europe: Iberian Peninsula, Italy (two lineages), Slovakia, Slovenia, Macedonia, Bulgaria and Romania; see Figure 2]. The haplotype from Lama Grande clustered with the remaining Iberian haplotypes with high support (Figure 2). Moreover, although the records available for the IRBP gene are limited, as they only include a small number of specimens from the Iberian Peninsula (Sorpe and Andorra; Barbosa et al. 2013), there is some evidence of differentiation between the western Iberian and Pyrenean populations. This overall pattern is consistent with a Middle Eastern origin of the species, as proposed by other authors (Yannic et al. 2012; Bannikova et al. 2013), and suggests that there is a certain degree of divergence between fragmented populations.

Figure 2. Phylogenetic analysis of the Chionomys nivalis cyt-b gene 855-bp, considering the sequences retrieved from GenBank (see Table I) and the sequences obtained from the captured individuals. (A) Maximum clade credibility tree based on Bayesian coalescence genealogy of the 49 C. nivalis cyt-b sequences with BEAST (Table II). Values on branches indicate posterior probability support for the nodes (≥ 0.8). (B) Median-joining network of the C. nivalis cyt-b haplotypes. Each circle represents one haplotype, and circle size is proportional to the number of sequences of each haplotype. The number of nucleotide differences between haplotypes is indicated on branches. Letters indicate country codes: Adygea (RU-AD), Andorra (AD), Bulgaria (BG), Georgia (GE), Israel (IL), Italy (IT), Macedonia (MK), Portugal (PT; in dark grey), Romania (RO), Slovakia (SK), Slovenia (SI), Spain (ES), Syria (SY), Switzerland (CH), Turkey (TR) and Turkmenistan (TM).

This newly detected population in the highest areas of Montesinho Mountain may be spatially connected with the populations already recorded in the contiguous region of Sanabria (Spain). Although this new finding might be seen as evidence for the recent expansion of the species’ distribution range in the northern part of the Iberian Peninsula, it is highly likely that it actually represents the discovery of an existing but previously undetected population in northeastern Portugal. In fact, although information about the distribution of small mammals in Portugal has been gradually updated (e.g. Cruz et al. 2002; Vale-Gonçalves & Cabral 2014), the available records of species distributions are insufficient to support a consistent national mammal atlas, and the spatial gaps subsist for the range of several species (Vale-Gonçalves & Cabral 2014). Therefore, this new record of snow vole occurrence highlights the need for accurate regional surveys of small mammal distribution and population trends to facilitate the definition of species conservation status, as previously emphasised by Bertolino et al. (2014) for Italy. In Portugal, additional sampling effort is needed to delimit the species’ distribution and determine its conservation status. We recommend as the Portuguese common name the designation of ‘Rato-das-neves’, which is a strict translation of the English name.

Overall, our finding contributes new information regarding the southwestern limit of the snow vole distribution in particular, and regarding Portuguese small mammal diversity as a whole, while emphasising the need for further studies on the genetic structure of the species. Our analysis of both mitochondrial and nuclear data, including that from this newly detected population in Portugal, reinforces previous conclusions on the species’ phylogeographic structure (Castiglia et al. 2009; Yannic et al. 2012; Bannikova et al. 2013), and highlights the existence of several divergent and geographically distinct lineages. Although the divergence between these lineages is within the limit of intraspecific differentiation (5% as suggested by Baker & Bradley 2006), it may indicate that these constitute potential management units for conservation, due to the level of current fragmentation and isolation and their expected vulnerability to habitat change (Bertolino et al. 2014). However, the combination of a more comprehensive sampling design, throughout the species range, with additional genetic analysis, using molecular markers that are widespread in the genome (e.g. microsatellites or SNPs (Single Nucleotide Polymorphisms)), will be crucial to accurately define its distribution and to assess the level of gene flow between the lineages and clarify the level of divergence and isolation of these populations.

Ethical standards

Animals were trapped and ethically handled according to Portuguese law and under permit no. 587/2014/CAPT.

Acknowledgements

This study was supported by funding from several ecological monitoring projects of the Laboratory of Applied Ecology (University of Trás-os-Montes and Alto Douro, UTAD), including the grants BGCT/LEA/402/UTAD/2014 (Hélia Marisa Vale-Gonçalves) and BGCT/LEA/403/UTAD/2014 (Paulo Barros), and by the Portuguese Foundation for Science and Technology (FCT) through the project PEst-OE/AGR/UI4033/2014. Joana Paupério has a postdoctoral grant funded by the project “Genomics and Evolutionary Biology” co-financed by the North Portugal Regional Operational Programme 2007/2013 (ON.2 – O Novo Norte), under the National Strategic Reference Framework, through the European Regional Development Fund. We are grateful to C. M. Ferreira and F. M. S. Martins for collaboration in the genetic analysis, and to P. C. Alves for early discussions of the manuscript. We also thank J. S. Herman for reviewing the English, and the reviewers for valuable comments.