Conservation of genetic uniqueness in remaining populations of red squirrels (Sciurus vulgaris L.) in the South of England

Abstract The Eurasian red squirrel (Sciurus vulgaris) is an emblematic species for conservation, and its decline in the British Isles exemplifies the impact that alien introductions can have on native ecosystems. Indeed, red squirrels in this region have declined dramatically over the last 60 years due to the spread of squirrelpox virus following the introduction of the gray squirrel (Sciurus carolinensis). Currently, red squirrel populations in Britain are fragmented and need to be closely monitored in order to assess their viability and the effectiveness of conservation efforts. The situation is even more dramatic in the South of England, where S. vulgaris survives only on islands (Brownsea Island, Furzey Island, and the Isle of Wight). Using the D‐loop, we investigated the genetic diversity and putative ancestry of the squirrels from Southern England and compared them to a European dataset composed of 1,016 samples from 54 populations. We found that our three populations were more closely related to other squirrels from the British Isles than squirrels from Europe, showed low genetic diversity, and also harbored several private haplotypes. Our study demonstrates how genetically unique the Southern English populations are in comparison with squirrels from the continental European range. We report the presence of four private haplotypes, suggesting that these populations may potentially harbor distinct genetic lineages. Our results emphasize the importance of preserving these isolated red squirrel populations for the conservation of the species.

The replacement of the native squirrel in much of the UK, and the role of squirrelpox virus (SQPV) in this process, is a wellknown example of disease-mediated invasion (Bosch & Lurz, 2012;Tompkins, White, & Boots, 2003) and the risks associated with release of non-native species. Combined effects of disease and competition have enabled the gray squirrel to replace its native congener with native strongholds remaining in the north of the country and isolated populations in the south (Gurnell et al., 2006(Gurnell et al., , 2004Tompkins et al., 2003). Recent evidence also suggests that genetic diversity in S. vulgaris may be lower in UK populations, compared with European congeners, with potential implications for their susceptibility to disease (Ballingall et al., 2016).
The conservation genetics of S. vulgaris presents interesting challenges for a number of reasons. It has been classified into up to 42 subspecies on the basis of morphological differences including coat color and body size (Shorten, 1954), and the number of estimated subspecies has varied (Lurz et al., 2005). Sidorowicz (1971) suggested a classification into 17 subspecies mapped into geographic subregions but only a few subspecies have been supported by molecular data. Grill et al. (2009) (Wauters et al., 2017). The 17 subspecies classification included a British subspecies S. v. leucourus which has been noted as far back as the 18th century on the basis of their white or "bleached" tails (Shorten, 1954). However, there is scant evidence that such a subspecies is still present in the UK and uncertainty over whether it was a true subspecies, as color coat is thought to be a poor species marker (Lowe & Gardiner, 1983) and specimens suitable for an in-depth morphological study and molecular confirmation have not been identified (Hale et al., 2004).
The population structure of S. vulgaris in Britain is unlikely to be straightforward as it has experienced dramatic declines and recoveries over several centuries. In the 15-16th century, and again the 18th century, deforestation in Scotland resulted in squirrels coming close to extinction in that region, except possibly the far north. This was followed by several successful reintroductions and afforestation, with a subsequent recovery of the red squirrel until foresters considered the species a pest by the late 19th century (Shorten, 1954). A history of translocations of continental S. vulgaris to the British Isles during these reintroductions (Lowe & Gardiner, 1983;Shorten, 1954) adds another level of complexity to the challenges of conservation genetics of this species (Hale et al., 2004). Indeed, Hale et al. (2004) found that the majority of the British S. vulgaris had a continental origin with many animals carrying a Scandinavian haplotype. Although Barratt et al. (1999) examination of mtDNA from a range of British sites indicated no clear population structure and concluded that translocations between regions could be advised, subsequently, Hale et al. (2001) found significant genetic differences between some British regions.
Likewise, Finnegan, Edwards, and Rochford (2008) found evidence for significant differences among Irish red squirrel populations and suggested that these should be treated as separate conservation management units.
Although the red squirrel is now largely limited to the north of Britain, there are small populations remaining on islands off the south coast of England. These isolated populations may harbor unique genetic variation which needs to be accounted for in conservation management. Using mitochondrial data from a wide range of European samples, this study aims to infer the possible origin as well as the conservation value of the isolated populations of S. vulgaris in the South of England currently living on three islands: the Isle of Wight, and two islands in Poole Harbour, Dorset: Brownsea Island and Furzey Island.

| Study sites and sample collection
Brownsea includes about 200 ha of mixed woodland and approximately 150-200 squirrels (Thain & Hodder, 2015). Furzey is a 13 ha island approximately 300 m from Brownsea with six hectares of woodland dominated by Pinus sylvestris  and it is home to a population of around 30 red squirrels (Thain & Hodder, 2015). In 2009, eight samples of plucked hairs were collected from squirrels livetrapped on Furzey Island as part of conservation monitoring and one cadaver was collected on Brownsea ( Figure 1). Twenty additional plucked hair samples from livetrapped squirrels were collected in 2016 as part of a squirrel leprosy research project on Brownsea Island. Hair was plucked from the base of the tail.
The Isle of Wight, with over 3,600 ha of woodland, is home to the largest remaining population of the red squirrel in southern England estimated as 3,300 squirrels assuming 1.1 squirrels per hectare (Pope & Grogan, 2003).
Red squirrel tissue samples from the Isle of Wight were collected during routine postmortem examinations undertaken by Wight Squirrel Project. DNA was extracted at the Moredun Research Institute using conditions described in Simpson et al. (2015). Twenty-five of those samples were used in the present study.
All the sequences generated in the present study were submitted to GenBank: accession number MK234640-MK234695 and MK258734-MK258755.

| Phylogenetic analysis
The Brownsea Island, Furzey Island, and Isle of Wight sequences were aligned to previously published data and used the British populations as defined by Hale et al. (2004). The final alignment has a length of 238 bp with 72 informative variants from 1,016 samples from across Europe (see references in Table 1). The numbers of haplotypes, haplotype diversity, nucleotide diversity, and neutrality tests were calculated using DNAsp (Librado & Rozas, 2009). F ST and AMOVA calculations were performed using Arlequin ver. 3.5.2.2. (Excoffier & Lischer, 2010). A median-joining haplotype network was constructed in PopART (Leigh & Bryant, 2015), and the Mantel test was calculated using R software and Ade4 package (Dray & Dufour, 2007).

| Phylogenetic tree
A phylogenetic tree was generated with MrBayes (Ronquist et al., 2012) using the sequences generated in the present study as well as all the sequences available from S. vulgaris from Europe (Table 1). Ogden et al., 2005) were not used because the D-loop fragment sequences in their study did not correspond to those used in the rest of the studies used in our analysis. The generation number was set at 600,000 MCMC with 25% of burn-in. A sequence from Sciurus lis (AB249880) was used as an out-group. The substitution model HKY + G was chosen using jModelTest (Darriba, Taboada, Doallo, & Posada, 2012). The tree was visualized using FigTree v1.4 (http:// tree.bio.ed.ac.uk/softw are/figtr ee/).

| Migrate-n analysis
The potential introduction pattern of the S. vulgaris was investigated using Migrate-n (Beerli, 2009). The transition/transversion rate was found to be 7.2920 for the Brownsea/Furzey dataset and 2.7591 for the Isle of Wight dataset using jModelTest (Darriba et al., 2012) and π, nucleotide diversity.
TA B L E 1 (Continued) was used for the Migrate analysis. The parameters for the Migrate-n analysis were set following 500,000 generations with a 25% burn-in and with 10 concurrent chains per run. Convergence of all the parameters was not always obtained; however, each migrate-n run was replicated three times independently and Bayes factor compared to ensure that the parameter space was explored in the same way by all three analyses. All the models tested are described in S1 and S2.

| Population differentiation
Pairwise F ST statistics were calculated across all 54 European populations available from GenBank ( Figure 4). Interestingly, F ST values between most of the populations and Italy were found to be low (between 0 and 0.39 with a mean of 0.17, SD = 0.129- Figure 4- Table   S1). The F ST between the Isle of Wight, Brownsea and Furzey, and Dorset was particularly high (F ST > 0.7 for the three pairwise comparisons- Table S1). As expected, geographically close populations had a lower F ST than populations further apart (Figure 4)

| Phylogeography of the red squirrels
A Bayesian phylogenetic tree was calculated using a 238 bp D-loop fragment from 1,016 red squirrel sequences from all across Europe (see references in Table 1) using S. lis (AB249880) as an out-group.
A total of 216 haplotypes was found in the dataset (Figure 3). Small clades were found in South West England. Brownsea and Furzey islands clustered in Clade 1 and 2 (Figure 3

| Putative origin of the S. vulgaris on Brownsea Island, Furzey Island, and the Isle of Wight
The colonization hypotheses for each island were investigated using Migrate-n. The hypothesized source regions were proposed using the clustering of the phylogenetic tree ( Figure 3) as well as the F ST matrix (Figure 4). Eight putative origins were tested for Brownsea Island and Furzey Island (S1). Model 8, with a Northern English origin for the Furzey red squirrels and a North West English origin for the Brownsea red squirrels, found to be most likely (  (Table 3).

| D ISCUSS I ON
Our analysis of S. vulgaris from southern English island populations, in the context of a European dataset of S. vulgaris, provided insight into the population differentiation of the species across Europe.
We were able to corroborate the findings of previous phylogenetic studies (e.g., Grill et al., 2009) which also showed no evidence for a phylogeographic pattern in Europe. In contrast, our results showed high population differentiation within Britain, differing from continental Europe which followed a pattern of isolation by distance.
More interestingly, several private haplotypes were found in the three isolated populations from southern England representing unique lineages which could be valuable for the conservation of the species.  (Lowe & Gardiner, 1983;Shorten, 1954). Those translocations could explain the high population differentiation found in Britain.

| Origin of S. vulgaris on Brownsea and Furzey islands
Our results indicated that S. vulgaris can migrate between Brownsea and Furzey or that the populations have a common origin, as haplotypes are shared between squirrels on the two islands.

Migration between those islands is feasible as Brownsea and Furzey
Islands are around 300 m apart, well within the ability of this species to swim (Bosch & Lurz, 2012) and evidence exists of an individual successfully crossing the greater distance from these islands to a peninsula on the mainland  Harbour about 600 m from Furzey island; however, this was an unsuccessful translocation . Therefore, it is not likely that this translocation has contributed to the populations of the squirrels on Brownsea or Furzey.
F I G U R E 4 Pairwise F ST calculated for the 54 populations of red squirrels

| Origin of S. vulgaris on the Isle of Wight
The Isle of Wight is home to the largest remaining population of the red squirrels in southern England. The population has been estimated as 3,300 squirrels (Pope & Grogan, 2003). We found that the S. vulgaris population on this island was more genetically diverse than Brownsea or Furzey islands. The result was expected as the population on the Isle of Wight is much larger (<3,000) than on Brownsea (<300) and Furzey (~30). Furthermore, many studies highlight the positive correlation between island area and genetic diversity (Cheylan, Granjon, Granjon, & Britton-Davidian, 1998;Jenkins, Yannic, Yannic, Schaefer, Conolly, & Lecomte, 2018;White & Searle, 2007). Indeed, the haplotype diversity on the Isle of Wight is similar to the one found in the Parc de Sceaux, an urban park close to Paris in France (Table 1).
This result is encouraging as Rézouki et al. (2014) demonstrated that this population of S. vulgaris, despite being an "urban island," was viable and self-sustaining. However, it was also found that migration of red squirrels to the park was possible via ecological corridors and forested habitats in the urban environment (Rézouki et al., 2014

| CON CLUS ION
The preservation of island population genetic diversity may be crucial for the conservation of the locally adapted individuals. The three islands studied are more than 250 km away from the main S. vulgaris populations in the UK and represent the only remnant populations of Southern England. Our analysis confirmed a British origin of these populations as well as lineages of S. vulgaris that appear to be unique to the islands and, therefore, reinforces the importance of preserving these S. vulgaris populations for the conservation of the species.

ACK N OWLED G M ENTS
Many thanks to Ade Parvin, Perenco for funding and access permission on Furzey Island. We also would like to thank National The authors would like to thank Dr. Emiliano Mori and two anonymous reviews for their help in improving the manuscript.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R S ' CO NTR I B UTI O N S
KHH, AKS, AM, PL, HB, and REK conducted the fieldwork for this investigation. EAH, OGR, EC, WJL, and JF conducted the laboratory work. EAH and MBS conducted the data analysis. EAH and KHH conceived the study and wrote the final manuscript. All authors were involved in writing and data interpretation and read and approved the final manuscript.

DATA ACCE SS I B I LIT Y
All the sequences generated in the present study were submitted to GenBank: accession number MK234640-MK234695 and MK258734-MK258755.