Heredity (2003) 91, 125–135 & 2003 Nature Publishing Group All rights reserved 0018-067X/03 $25.00 www.nature.com/hdy Spatial patterns of mitochondrial and nuclear gene pools in chamois (Rupicapra r. rupicapra) from the Eastern Alps H Schaschl, D Kaulfus, S Hammer1 and F Suchentrunk Research Institute of Wildlife Ecology, Veterinary Medicine University of Vienna, Savoyenstrasse 1, A-1160 Vienna, Austria We have assessed the variability of maternally (mtDNA) and variability as a result of immigration of chamois from different biparentally (allozymes) inherited genes of 443 chamois Pleistocene refugia surrounding the Alps after the withdrawal (Rupicapra r. rupicapra) from 19 regional samples in the of glaciers, rather than from topographic barriers to gene Eastern Alps, to estimate the degree and patterns of spatial flow, such as Alpine valleys, extended glaciers or woodlands. gene pool differentiation, and their possible causes. Based However, this striking geographical structuring of the on a total mtDNA-RFLP approach with 16 hexanucleotide- maternal genome was not paralleled by allelic variation at recognizing restriction endonucleases, we found marked 33 allozyme loci, which were used as nuclear DNA markers. substructuring of the maternal gene pool into four phylogeo- Wright’s hierarchical F-statistics revealed that only p0.45% of graphic groups. A hierarchical AMOVA revealed that 67.09% the explained allozymic diversity was because of partitioning of the variance was partitioned among these four mtDNA- among the four mtDNA-phylogroups. We conclude that this phylogroups, whereas only 8.04% were because of partition- discordance of spatial patterns of nuclear and mtDNA gene ing among regional samples within the populations, and pools results from a phylogeographic background and sex- 24.86% due to partitioning among individuals within regional specific dispersal, with higher levels of philopatry in females. samples. We interpreted this spatial pattern of mtDNA Heredity (2003) 91, 125–135. doi:10.1038/sj.hdy.6800290 Keywords: mtDNA; RFLP; allozymes; population structure; chamois; Rupicapra rupicapra Introduction Chartreuse massif in the French Alps (Shackleton and the IUCN/SSC Caprinae Specialist group, 1997). Gene pool Chamois (Rupicapra spp.) are group-living ungulates that divergence between these two subspecies as assessed by preferentially inhabit alpine pastures and rocky areas in multilocus allozyme electrophoresis is apparently higher diverse mountain regions of Europe and the Middle East. than among local populations of R. r. rupricapra (Pem- Several morphological characteristics place chamois berton et al, 1989), and allozyme studies also indicate somewhat separate from most of the other fossil or somewhat reduced gene flow among several regional extant forms of the Rupicaprini, and fossil remains populations of chamois from the Eastern Alps (Miller suggest an origin of chamois in eastern Europe or south- and Hartl, 1986). Here, we study the degree of gene pool west Asia (Lovari, 1987). Two species of chamois are substructuring of chamois from the Eastern Alps. Most considered to occur in Europe, Rupicapra rupicapra and R. parts of this region were covered by a 2000–3000 m thick pyrenaica (Masini and Lovari, 1988). While Pyrenean ice sheet during long periods of the late Pleistocene, and chamois (R. pyrenaica) are disjunctly distributed in south- deglaciation started only c. 15–12 kyr BP (Van Husen, west Europe (Pyrenees, Cantabrian Mountains) and the 1987). Obviously, post-Pleistocene spread of many tree southern Apennine mountain range in Italy (Shackleton species into central Europe occurred within a few and the IUCN/SSC Caprinae Specialist group, 1997), millennia of the end of the Pleistocene and at the native populations of R. rupicapra roam the Alps, the beginning of the Holocene, which led to rapid reforesta- Tatra massif (Slovak Republic), the Carpathian moun- tion of Alpine valleys and basins after the retreat of the tains (Romania), and various mountain massifs in the glaciers (Lang, 1994; Taberlet et al, 1998). These environ- Balkans, Asia Minor, and the Caucasus (eg, Lovari, 1987). mental changes likely provided only a short period of Within the Alps, chamois are more or less continuously time for colonization of the Alps by chamois, since distributed in higher altitudes, and two subspecies have chamois prefer rocky, unwooded terrain rather than been described: Rupicapra r. rupicapra, ranging over most extended woodlands, where large predators, such as of the Alpine regions, and R. r. carthusiana, from the wolf, lynx, and bear are present. Fossil evidence (eg Do¨ppes, 1997; Galik, 1997) indicates that chamois were roaming central Europe during late-glacial times. Colo- Correspondence: H Schaschl. Institute of Zoology, ZLS, Regent’s Park, nization of the present range of chamois in the Eastern London NW1 4RY, UK. E-mail: [email protected] 1Current address: AG General Genetics, Institute of Medical Biology, Alps might have resulted from only one major cohesive Waehringerstrasse 10, A-1090 Vienna, Austria. population or several isolated late-Pleistocene source Received 12 June 2002; accepted 12 February 2003 populations with more or less distinctly differentiated mtDNA and allozyme variability in chamois H Schaschl et al 126 gene pools. In addition to different colonization routes the Eastern Alps (Figure 1) were collected during regular along mountain ranges, drift effects in pioneer popula- hunts in several provinces of Austria (Carinthia, Lower tions (founder effects) and later on during possibly Austria, Salzburg, Styria, Upper Austria), Germany dramatic population crashes caused by sarcoptic mange (Berchtesgaden Nationalpark), Slovenia (Kamnik re- epidemics (see Rossi et al, 1995) might have produced a gion), and northern Italy (South Tyrol, Texel-Gruppe). distinct substructuring of the extant gene pool in the All samples were taken immediately after death and study region. In this study, we examine the hypothesis of frozen at À201C, normally within 12 h. distinct phylogeographic structuring in chamois from the Eastern Alps as a consequence of postglacial immigra- tion from different refugia by using an RFLP approach to mtDNA-RFLP analysis the total mtDNA. However, the maternally inherited mtDNA of 308 chamois was isolated from liver cells mitochondrial DNA (mtDNA) might overemphasize a following the procedures as described in Hartl et al possible phylogeographic structuring, because of the (1993). RFLP analysis was carried out by digesting the higher level of philopatry in female chamois as com- entire mtDNA with 16 hexanucleotide-recognizing type pared to males (Loison et al, 1999). Therefore, we use a II restriction endonucleases (from Roche). The mtDNA multilocus allozyme approach to test for increased gene fragments obtained were separated electrophoretically flow among populations in the nuclear DNA, and expect and then visualized under UV light according to less geographical structuring with this marker system. Hammer et al (1995). The DNA-markers Lambda phage DNA digested with HindIII and Lambda phage DNA digested with a combination of HindIII and EcoRI served Materials and methods as size standards and were applied to determine the fragment lengths of the restriction segments. Further- Study regions and sampling more, we assumed that the length of the entire mtDNA Liver and kidney samples of 443 chamois (R. r. rupicapra) of chamois was approximately the same as the mtDNA from 19 populations from different mountain ranges in genome of cattle, which consists of 16 340 nucleotides c n KIT HUN HAGn* OHGn* TENn* KALn* RR5 RR9 RR1 RR9 RR9 RR8 RR9 RR9 RR1 RR8 RR8 RR8 c TAUa n* RR9 SBS RR5 RR9 RR5 RR1 RR8 TAUbc RR9 e* RR12 RR14 HBS RR8 RR6 RR5 RR1 RR7 SCHc SUEc RR8 RR1 RR10 RR5 RR1 RR5 0 50km e* RR12 KOR RR11 RR15 RR13 RR5 RR7 RR6 RR12 RR1 RR7 RR2 c RR1 RR1 KZK RR1 RR2 RR1 RR5 RR5 KAN s GAI c NOB c KAW s SAUc Figure 1 Study region in the Eastern Alps showing the 19 populations of chamois. The location of the Eastern Alps within Europe is given in the inset. For acronyms of populations, see Table 1. Frequency distribution of mtDNA haplotypes (RR1–RR15) is given for each population. Lower case letters attached to acronyms allocate single populations to one of the four distinct mtDNA-phylogroups found in this study: c ¼ ‘Central mtDNA-phylogroup’, e ¼ ‘Eastern mtDNA-phylogroup’, n ¼ ‘Northern mtDNA-phylogroup’, s ¼ ‘Southern mtDNA-phy- logroup’. Asterisks attached to upper case letters denote geographically marginal populations within the distribution range of chamois. Heredity mtDNA and allozyme variability in chamois H Schaschl et al 127 (Anderson et al, 1982). Fragment lengths shorter than 200 direct side-by-side comparisons of migrating allozymes basepairs (bp) could not be detected by UV visualization. were carried out including marker samples on all gels. Haplotypes were defined based on the composition of Interpretation of band patterns in terms of allelic restriction band patterns, and labelled with ‘RR’ (R. variation was based on the principles outlined in Harris rupicapra) and a serial number. RR1, RR2, RR5, and RR6 and Hopkinson (1976), and Rothe (1994). In numbering detected in this study were previously found by isozymes and designation of alleles we followed Miller Hammer et al (1995) in the Alpine chamois (haplotypes and Hartl (1986). The following enzyme systems were 7 and 8 of Hammer et al (1995), however, were screened (abbreviation, EC number of enzymes, and haplotypes of Pyrenean chamois, different from ‘RR7’ corresponding loci scored in parentheses): a-glyceropho- and ‘RR8’ of our study). sphate dehydrogenase (GDC, 1.1.1.8, Gdc), sorbitol dehydrogenase (SDH, 1.1.1.14, Sdh), lactate dehydrogen- Statistical treatment of mtDNA-RFLP data ase (LDH, 1.1.1.27, Ldh-1,-2), malate dehydrogenase (MOR, 1.1.1.37, Mor-1,-2), malic enzyme (MOD, Based on the presence or absence of restriction sites, 1.1.1.40, Mod-1,-2), isocitrate dehydrogenase (IDH, pairwise genetic distances between haplotypes were 1.1.1.42, Idh-1,-2), 6-phosphogluconate dehydrogenase, calculated after Nei and Li (1979).