Landscape heterogeneity and ecological niche isolation shape the distribution of spatial genetic variation in Iranian brown bears, Ursus arctos (Carnivora: Ursidae)
Mohammad Reza Ashrafzadeh, Rasoul Khosravi, Mohsen Ahmadi, Mohammad Kabolia
Mohammad-Reza Ashrafzadeh
Department of Environmental Sciences, Faculty of Natural Resources, University of Tehran, Karaj, Iran
Department of Fisheries and Environmental Sciences, Faculty of Natural Resources and Earth Sciences, Shahrekord University, Shahrekord, Iran
Mohammad-Reza Ashrafzadeh
Department of Environmental Sciences, Faculty of Natural Resources, University of Tehran, Karaj, Iran
Department of Fisheries and Environmental Sciences, Faculty of Natural Resources and Earth Sciences, Shahrekord University, Shahrekord, Iran
Mohsen Ahmadi
Department of Desert Regions Management, School of Agriculture, Shiraz University, Shiraz, Iran
Swiss Federal Research Institute WSL, 8903, Birmensdorf, Switzerland
Mohammad Kaboli
View the author's ORCID recordDepartment of Environmental Sciences, Faculty of Natural Resources, University of Tehran, Ka
Introduction
Unlike other carnivores, the brown bear (Ursus arctos) shows a widespread phylogeographic structure according to the mitochon- drial genome (Calvignac et al., 2008; Davison et al., 2011). Brown bears in the Middle East, including Iran, have lost much of their historical range and are now at risk of extinction (Calvignac et al.,
2009, 2008). Currently, almost all existing populations in Iran are confined to the Alborz and Zagros Mountains surrounded by areas of dense human settlements and road networks. Development and extension of urban areas may have significant consequences on genetic diversity of populations and migration of individu als between demes. The results of mtDNA shows that Iranian bears form a new lineage, which includes two to three distinct geographic sub-clades (Ashrafzadeh et al., 2016). It is unknown whether gene flow occurs between putatively isolated populations. Moreover, knowledge is scarce on other population attributes such as intraspecific diversity, patterns of spatial genetic composition, and landscape connectivity among populations, which hampers progress in developing action plans for the conservation of the brown bear. Hence, it is vital to understand the variety of con- straints influencing population genetic structure of bears in order to ensure gene flow among the remaining populations, especially in areas where it has been interrupted or reduced as a result of anthropogenic land use.
Geographic distances and landscape elements differently influ- ence individual movements and lead to genetic differentiations through isolation by distance (IBD, Wright, 1943), isolation by resistance (IBR, McRae, 2006), and isolation-by-barriers (IBB). IBD and IBR affect random mating among individuals, migration, and drift, and tend to lead to a loss of genetic diversity within local genetic neighborhoods (Landguth et al., 2010). As IBD assumes movement in straight lines, this approach explains only a small portion of the observed genetic differentiation and fails to account for landscape characteristics that might influence movement. Landscape genetics addresses this gap by adopting the approach of “isolation-by- resistance” (IBR) which considers the effects of landscape features on the distribution of genetic variation in an explicitly spatial man- ner using inter-deme least-cost paths or circuit-based analyses (McRae and Beier, 2007; Spear et al., 2010). Although geographic and landscape resistance between demes are among the com- ponents with potential influence on gene flow and population connectivity (Crispo et al., 2006; Lee and Mitchell-Olds, 2011), environmental heterogeneity and local adaptations are considered as other drivers of genetic differentiation (Thorpe et al., 2005; Wang and Bradburd, 2014). Isolation-by-environment (IBE) is a more recent approach predicting the relationship between spatial genetic dissimilarity and environmental heterogeneity (Wang and Bradburd, 2014; Wang and Summers, 2010).
While a large amount of literature has focused on exploring the effects of IBD, IBR and IBE on genetic differentiations of species with limited dispersal abilities (e.g. Cushman et al., 2006; Wasserman et al., 2010), little is known about dynamics of population genetics of large-bodied species, such as the brown bear (e.g. Mateo-Sánchez et al., 2015), in heterogeneous landscapes. Therefore, it is still poorly understood whether large carnivores with high dispersal abilities show clinal population structures and to what extent iso- lation by geographic distance or landscape resistance is responsible for their genetic differentiations.
In this study, we aimed to fill the current knowledge gap regard- ing population genetic diversity, spatial genetic structure, and effects of geographic distance, landscape features, and environ- ment dissimilarity on genetic composition of the Iranian brown bear by adopting an individual-based landscape genetics approach. First, we present the population genetic structure of the brown bear throughout its range in Iran by analyzing microsatellite loci. Then, we use a combination of individual-based landscape genetics approach and population-based ecological niche modeling (ENM) to test hypotheses regarding the effects of landscape attributes (i.e. IBD, IBR, and IBE) and niche divergence on gene flow (i.e. population level niche comparisons). These hypotheses were organized into four frameworks incorporating geographic distance, landscape- resistance, environment dissimilarity, and niche modeling.
Material and methods
Study area and genetic sampling
We sampled all recognized locations of the brown bear through- out the species’ range in Iran, between 2014 and 2016. Within Iranian borders, the Alborz Mountains, stretching from the north- west to the northeast, and the Zagros Mountains, stretching from the northwest to the southwest, prevent moisture bearing sys- tems to pass through to the south and the east of the country, respectively. Hence, deciduous forest habitats are more abundant in the Alborz and Zagros Mountains in the study area. The area is a fragmented mosaic of land-cover and land-use practices domi- nated by agricultural fields, highways and human settlements that may potentially limit migrations between brown bear populations especially in the Alborz Mountains.
The remaining populations of brown bear have been largely con- fined to Protected Areas (PAs) and No-Hunting Areas (NHAs). A total of 66 tissue samples were collected opportunistically from a variety of sources including carcasses of bears that died from poaching, hunting, disease, and road accidents. Locations of indi- vidual samples were recorded using a global positioning system (Fig. 1). All samples were preserved in 96% ethanol prior to DNA extraction. This study was permitted by the Iranian Department of Environment under permit number 91/52621.
DNA extraction and microsatellite amplification
All samples were subjected to whole genomic DNA extraction using the QIAamp DNA MicroKit (Qiagen) and also DNA Extraction Tissue Kit (Bioneer). We amplified nineteen dinucleotide species- transferred microsatellites with dye-labelled forward primers for each DNA extraction (Supporting information, Table S1). The PCR conditions for each primer were optimized in a total volume of
15 L using 1.5 l of DNA template, 1.5 L 10X PCR Buffer, 1.5 mm MGCL2, 0.5 U Taq DNA polymerase, 0.2 of each dNTP, 0.064 m of forward primer and 0.16 m of reverse primer, and 0.12 m M13 (Schuelke, 2000). PCR was carried out in 35 cycles as follows: initial denaturation (5 min at 95 ◦C), denaturation (40 s at 94 ◦C), annealing (30 s at 53 to 62.7 ◦C), extension (30 s at 72 ◦C), and lastly, a final extension step (5 min at 72 ◦C; Table S1). Two (homozygote) and three (heterozygote) PCR per locus were conducted to minimize genotyping errors resulting from degraded DNA (Bellemain and Taberlet, 2004). Fragment analysis was carried out using ABI3730xl (Macrogen Inc., South Korea). Alleles were scored using Geneious R9 (9.0.5; Biomatters Ltd.). All DNA extractions and PCRs were car- ried out at the Agriculture and Natural Resources laboratory of University of Tehran and also Natural History Museum laboratory of the Iranian Department of Environment (DoE). Both steps were monitored for contamination using negative controls.
Population genetic characteristics
MICRO-CHECKER 2.2.3 (Van Oosterhout et al., 2004) was used to statistically test for genotyping errors, large allele dropouts, and scoring errors (1000 randomizations). We calculated diver- sity statistics for each genetically different group, including the mean number of alleles across all loci (A), allelic richness (Ar, Kalinowski, 2004), expected heterozygosity (HE), and observed het- erozygosity (HO) using FSTAT 2.9.3 (Goudet, 2001). We tested for site-specific deviations from Hardy-Weinberg equilibrium (HWE) for each locus and linkage disequilibrium across all pairs of loci using GENEPOP (Raymond and Rousset, 1995). Significance for tests was assessed using the Markov chain method (100 batches,
1000 iterations and 1000 dememorisation steps). We used sequen- tial Bonferroni correction to control for multiple comparisons. The inbreeding coefficient (FIS) was tested by GENEPOP. A Two Phased Mutation Model (TPM) with 90% Stepwise Mutation Model (SMM) and 10,000 iterations was used to detect recent population bot- tlenecks in the program BOTTLENECK (Piry et al., 1999). Wilcoxon signed-rank test along with a mode-shift was used to assess sigificant heterozygote deficiency in samples. We also calculated the mean number of alleles in each recognized population divided by the allelic size range (M-ratio) with ARLEQUIN 3.11 to test for bot- tleneck events that could have occurred over longer time periods (Wall et al., 2014). M-ratios lower than the critical value of 0.68 indicate a population bottleneck (Garza and Williamson, 2001). Inference of population genetic structure and identification of dispersal events