Geographic Pattern of Genetic Variation in the European

MEC_1618.fm Page 2337 Tuesday, October 22, 2002 6:09 PM

Molecular Ecology (2002) 11, 2337–2347

BlackwellGeographic Science, Ltd pattern of genetic variation in the European globeflower Trollius europaeus L. (Ranunculaceae) inferred from amplified fragment length polymorphism markers

LAURENCE DESPRES, SANDRINE LORIOT and MYRIAM GAUDEUL Laboratoire de Biologie des Populations d’Altitude, CNRS-UMR 5553, Université J. Fourier, BP 53–38041 Grenoble Cedex 09, France

Abstract The distribution of genetic variation and the phylogenetic relationships between 18 populations of the arctic-alpine plant Trollius europaeus were analysed in three main regions (Alps, Pyrenees and Fennoscandia) by using dominant AFLP markers. Analysis of molecular variance revealed that most of the genetic variability was found within populations (64%), although variation among regions (17%) and among populations within regions (19%) was highly significant (P < 0.001). Accordingly, the global fixation index

FST averaged over loci was high (0.39). The among-population differentiation indicates restricted gene flow, congruent with limited dispersal of specific globeflower’s pollinating flies (Chiastocheta spp.). Within-population diversity levels were significantly higher

in the Alps (mean Nei’s expected heterozygosity HE = 0.229) than in the Pyrenees (HE = 0.197) or in Fennoscandia (HE = 0.158). This finding is congruent with the species- richness of the associated flies, which is maximum in the Alps. We discuss the processes involved in shaping observed patterns of genetic diversity within and among T. europaeus populations. Genetic drift is the major factor acting on the small Pyrenean populations at the southern edge of T. europaeus distribution, while large Fennoscandian populations result probably from a founder effect followed by demographic expansion. The Alpine populations represent moderately fragmented relics of large southern ancestral populations. The patterns of genetic variability observed in the host plant support the hypothesis of sympatric speciation in associated flies, rather than recurrent allopatric speciations. Keywords: amplified fragment length polymorphism markers (AFLP); genetic variation; habitat fragmentation; population differentiation; postglacial recolonization; sympatric speciation, Trollius europaeus Received 5 March 2002; revision received 24 June 2002; accepted 19 July 2002

European plant species were shown to be restricted to Introduction three main southern ice-free refugia: one in Portugal– The genetic structure of a species is both influenced by its Spain, one in Italy and one in the Balkans (Taberlet et al. past history and by current gene flow. The Quaternary cold 1998; Hewitt 2001). Most present plant species distri- periods in Europe have heavily influenced the distribution butions in Europe result from a northward recolonization of plant species with many range contractions–expansions from those southern refugia after the last glaciation, about in direct relation with climatic variations (Hewitt 1996; 12 000 years ago (the ‘tabula rasa hypothesis’, Nordal 1987; Comes & Kadereit 1998). The ice sheets of the Northern Birks 1996). Under this hypothesis, the southern part hemisphere began to grow about 2.5 My ago and the of Europe should present the highest genetic diversity major climatic oscillations occurred during the last 700 ky by contrast with the northern part recently recolonized (Webb & Bartlein 1992). During glaciations, many temperate (founder effect). This picture is likely to be quite different for cold-tolerant plants which may have survived in ice- Correspondence: Laurence Després. Fax: + 33 4 76 51 42 79; E- free northern, alpine or eastern refugia during Pleistocene mail: [email protected] glaciations (Lagercrantz & Ryman 1990; Abbott et al. 1995;

MEC_1618.fm Page 2338 Tuesday, October 22, 2002 6:09 PM

2338 L. DESPRES, S. LORIOT and M . GAUDEUL

Tremblay & Schoen 1999; Vendramin et al. 2000), resulting the same; our expectation is therefore to find the highest in the maintenance of high genetic diversity in nordic and/ plant genetic diversity in the Alps. Our main objective is to or alpine populations. Much fewer phylogeographical determine the evolutionary processes responsible for the analyses had been undertaken on these taxa, so that their observed pattern of genetic variation in T. europaeus, and to glacial refugia are still poorly known, but are likely to be infer those driving Chiastocheta spp. radiation. The time- found at higher latitude/altitude than those described for scales involved in Chiastocheta radiation and in T. europaeus temperate taxa. Moreover, not only past history, but also phylogeography are different: Chiastocheta speciation current balance between gene flow and genetic drift shapes events occurred several times during the last 2 million the observed genetic patterns of species: genetic drift in years (Després et al. 2002), whereas the present study of T. small populations at the edge of the species range will lead europaeus genetic structure considers only the plant history to reduced within population genetic diversity and high since the end of the last glacial period (12 000 years ago). between population differentiation, while populations However, the last 2 million years have been characterized admixture can lead to increased genetic diversity (Walter by several range contraction–expansions comparable to & Epperson 2001). Discriminating between patterns of the last one, in concordance with cyclical climatic changes genetic variation caused by gene flow/drift balance from of similar amplitude. Recurrent allopatric speciations may those caused by common ancestry requires analysing have occurred in the Chiastocheta genus during fragmenta- genetic variability at both spatial and temporal scales: tion periods of the host-plant range, followed by remixture within vs. among populations and regions throughout of allopatrically differentiated species. Alternatively, Europe, and past history (phylogeography) vs. recent sympatric speciation may have occurred in ancestral large history (drift and gene flow). host-plant southern populations throughout disruptive The European globeflower Trollius europaeus L. (Ranun- selection on the date of oviposition (Després & Jaeger 1999; culaceae) is a perennial arctic-alpine species growing in Ferdy et al. 2002), followed by local species extinction moist and cold meadows. The species occurs at low during fragmentation of the host-plant populations and/ altitude in northern Europe (i.e Fennoscandia and Russia) or species loss during range expansion. The finding of where populations are usually large and more or less well-differentiated host-plant lineages originating from continuously distributed. By contrast, in southern Europe, different geographical regions and coexisting in the Alps T. europaeus is restricted to mountains above 700 m and would support the first hypothesis, whereas the finding that forms much more fragmented populations, especially at Pyrenean and Alpine populations represent fragmented the southern edges of its distribution, where populations relics of large southern, low altitudinal, interconnected are usually small and patchily distributed. During the last populations would support the second hypothesis. glaciation, glaciers covered most of the present range of the In this study, we analyse the genetic diversity of the species. The pollination ecology of T. europaeus has been globeflower from the north to the south of its geographical studied extensively as it is one of the rare reported cases distribution. Nuclear and chloroplast sequencing failed of a plant being pollinated exclusively by a seed parasite to reveal nucleotide variability among T. europaeus (Pellmyr 1989; Jaeger & Després 1998; Jaeger et al. 2000, populations (Després et al. submitted). For this study, 2001). Pollinators are small anthomyiid flies (genus we therefore chose the amplified fragment length poly- Chiastocheta) whose larvae develop by eating a fraction of morphism (AFLP) technique, validated as efficient for the plant’s seeds. Several Chiastocheta species may coexist biogeographical questions (Weising et al. 1995; Mueller & in T. europaeus populations with up to six species observed Wolfenbarger 1999). The AFLP technique, described first in many Alpine populations (Després & Jaeger 1999). by Vos et al. (1995), offers the advantage of generating a These species differ in the date of oviposition. By contrast large number of markers, spanning the whole genome with Alpine populations, a maximum of five species were without requiring any prior knowledge. The major flaw described throughout the whole northern range (Pellmyr of this marker is that it is a dominant marker, therefore 1992; Després et al. 2002), and only three have been precluding any inference of the within-population genetic observed so far in the five Pyrenean populations sampled structure. in the present study (L.D., unpublished data). Moreover, We first studied the within- and among-population the highest level of mtDNA sequence polymorphism genetic diversity in three main geographical regions: Fenno- within Chiastocheta species was observed in the Alps scandia (i.e northern part of the T. europaeus distribution (Després et al. 2002). This high level of biodiversity (both in area), Alps and the Pyrenees (i.e southern part of the terms of species number and within species variability) species distribution area). We then inferred the phylogen- found in the globeflower flies in the Alps may reflect etic relationships between these populations in relation to the history of T. europaeus populations in Europe. Indeed, their geographical location (phylogeography), and finally because of the intimacy of the relationship linking the plant proposed a scenario describing the history of T. europaeus and the insect, their evolutionary history is likely to be populations during the last postglaciation period.

MEC_1618.fm Page 2339 Tuesday, October 22, 2002 6:09 PM

GENETIC DIVERSITY IN THE EUROPEAN GLOBEFLOWER 2339

Fig. 1 Geographic location of the 18 studied populations of Trollius europaeus.

was collected and dried immediately in silica gel. Genetic Materials and methods analyses were performed on 10 individuals per population, resulting in a total of 180 individuals genotyped. Plant material Trollius europaeus L. is an arctic-alpine, early-flowering AFLP protocol perennial. This diploid plant (2n = 16) is an obligate outcrosser depending on Chiastocheta flies for its pollina- DNA extraction was performed with the Dneasy Plant Mini tion throughout its geographical range (Pellmyr 1989; Kit (Qiagen, Hilden, Germany) according to the manufac- Jaeger & Després 1998; Hemborg & Després 1999). The turer’s protocol, using about 20 mg of dried leaf material. plant is 0.2–1 m high and usually produces a single yellow DNA concentration was determined by fluorometry with globose flower composed of several multi-ovulate carpels. the Picogreen® DNA quantification Kit (Molecular Probes, The closed shape of the flower does not allow visitors Leiden, Netherlands) and ranged from 15 to 25 ng/µL. The other than Chiastocheta flies. Males and females feed, rest AFLP method was performed as described by Gaudeul and mate inside the flowers, pollinating them passively. et al. (2000): EcoRI adapters were 5′-CTCGTAGACTGCGT- Females deposit eggs on carpels and larvae develop on ACC-3′ and 5′-AATTGGTACGCAGTCTAC-3′, and MseI seeds. Mature follicules open at the end of the summer, adapters 5′-GACGATGAGTCCCTGAG-3′ and 5′- liberating seeds and pupae that fall down in the soil to TACTCAGGACTCAT-3′. For the preselective amplification germinate/emerge the following spring. Only the under- by PCR, parameters were as follows: 2 min at 72 °C, 30 ground taproot persists over winter, and there is no cycles of 30 s denaturing at 94 °C, 30 s annealing at 56 °C, evidence of asexual multiplication. Adult Chiastocheta flies and 2 min extension at 72 °C, ending with 10 min at 72 °C do not survive during winter. for complete extension, using EcoRI primer E.A (5′- Eighteen populations of T. europaeus were chosen in GACTGCGTACCAATTCA-3′) and MseI primer M.C order to cover a broad geographical range (Fig. 1). Five (5′-GATGAGTCCTGAGTAAC-3′). For the selective populations in each of three main regions were sampled: amplification, PCR parameters were: 10 min at 95 °C Fennoscandia, Alps and the Pyrenees. In addition, three followed by 36 cycles of 30 s denaturing at 94 °C, 30 s more isolated populations were sampled in Germany, annealing, and 1 min extension at 72 °C, ending with Poland and Romanian Carpathians. The population sizes 10 min at 72 °C for complete extension. Annealing was and elevations were distributed widely ranging, respect- initiated at a temperature of 65 °C, which was then reduced ively, from 30 to thousands of individuals, and from 0 to by 0.7 °C for the next 12 cycles and maintained at 56 °C for 2500 m (Table 1). Thirty individuals per population were the subsequent 23 cycles. Three selective PCR primer pairs sampled randomly and for each individual, leaf material were selected over 16 tested for the quality of the produced

MEC_1618.fm Page 2340 Tuesday, October 22, 2002 6:09 PM

2340 L. DESPRES, S. LORIOT and M . GAUDEUL

Table 1 Name, location, altitude, size and collector name of the 18 studied populations

Approximate Population Region–country Longitude/latitude Elevation (m) number of plants Collector

1 Areche Alps — France 6°34′ E/45°40′ N 1700 > 1000 L. Després 2 Menee Alps — France 5°31′ E/44°42′ N 1400 > 1000 L. Després 3 Fournel Alps — France 6°53′ E/44°79′ N 1500 500 M. Gaudeul 4 Cherlieu Alps — France 5°46′ E/45°18′ N 950 > 1000 L. Després 5 Galibier Alps — France 6°24′ E/45°04′ N 2300 > 1000 L. Després 6 Fulda Germany 9°39′ E/51°25′ N 800 < 500 J. Johannesen 7 Pradella Pyrenees — France 2°01′ E/42°32′ N 1850 50 L. Després 8 Eynes Pyrenees — France 2°10′ E/42°30′ N 2000 500 L. Després 9 Gazies Pyrenees — France 2°79′ W/42°91′ N 1500 100 F. d’Amico 10 Gabardères Pyrenees — France 2°79′ W/42°85′ N 1680 50 F. d’Amico 11 Puymorens Pyrenees — France 1°49′ E/42°32′ N 1850 500 L. Després 12 Kvaloeya Fennoscandia — Norway 18°43′ E/69°39′ N 180 500 G. Yoccoz 13 Ringassoeya Fennoscandia — Norway 19°00′ E/69°50′ N 0 500 G. Yoccoz 14 Rhindunjira Fennoscandia — Sweden 18°49′ E/68°21′ N 400 > 1000 A. Hemborg 15 Slattatjakka Fennoscandia — Sweden 18°46′ E/68°20′ N 670 > 1000 A. Hemborg 16 Oulanka Fennoscandia — Finland 28°59′ E/66°26′ N 200 > 1000 P. Siikamäki 17 Biebrza Poland 22°28′ E/53°25′ N 150 < 500 A. Wrodlewska 18 Piatra Fontanele Romania 24°50′ E/47°12′ N 1000 500 I. Chintauan

bands (i.e. even distribution of bands with relatively is an obligate outcrosser throughout its geographical homogeneous intensity): E.ATC/M.CAG, E.ATC/M.CAT distribution. Only polymorphic markers were taken into and E.AGA/M.CTC. Products were separated by account in all calculations. electrophoresis for 6 h on a 5% polyacrylamide gel Mean genetic diversity within populations was esti- (automated sequencer ABI 377TM Perkin-Elmer). AFLP mated using popgene Version 1.31 (Yeh et al. 1997) in three patterns were then visualized with Genescan Analysis® ways: (i) the percentage of polymorphic loci out of all 3.1 (Perkin-Elmer): a fluorescent peak corresponds to the polymorphic loci (P%) (ii) Nei’s (1978) unbiased expected

presence of an amplified restriction fragment. Polymorphic heterozygosity (HE) and (iii) Shannon’s index of phenotypic peaks were checked individually and a presence/absence diversity (IS & Lewontin 1972). Estimates of HE and IS were (i.e. 1/0) matrix was manually constructed. Reproducibility obtained by averaging across loci. Genetic differentiation of each primer pair was checked by carrying out two times among populations and regions was calculated as the

the whole AFLP protocol for three individuals chosen unbiased FST estimator of Weir & Cockerham (1984), and randomly. its 95% confidence interval was obtained by bootstrapping 1000 replicates over loci using tfpga Version 1.3 (Miller 1997). To evaluate among-population or among-region Data analysis differentiation, total genetic diversity was also partitioned Independence of markers was assessed first by calculating among regions, among populations within regions, and Σ − their pairwise linkage index DA,B = 1/n i|V(A,i) V(B,i)| within populations by carrying out a hierarchical analysis where V(A,i) is the allele value of individual i for the of molecular variance (amova) on euclidian pairwise Φ marker A and n, the total number of individuals analysed. distances ( ST) using arlequin Version 2.000 (Schneider If marker pairs had values of D = 0.01 or D = 0.99, one was et al. 2000). Finally, Fisher’s exact tests were performed discarded from the data set to avoid redundancy. D-values on marker frequencies at each locus between all pairs were calculated using the program DDM (disequilibrium of populations/regions (Raymond & Rousset 1995): these between dominant markers) version 0.1 (P. Berthier, pers. pairwise tests determine if significant differences in marker comm.). Statistical analyses of AFLP patterns were based frequencies exist between groups of individuals. In order on the following assumptions: (i) AFLP markers behave to test for a correlation between genetic (Nei 1978) and as diploid, dominant markers with alleles either present geographical distances (in km) among populations, Mantel (amplified) or absent (nonamplified); (ii) comigrating tests (Mantel 1967) were performed using tfpga (1000 fragments represent homologous loci; and (iii) populations permutations). To determine the phylogenetic relationships are at the Hardy–Weinberg equilibrium (HWE). This among the populations, a neighbour-joining dendrogram assumption appears justified because the globeflower based on Nei’s distance was constructed, and bootstrap

GENETIC DIVERSITY IN THE EUROPEAN GLOBEFLOWER 2341

values were obtained by resampling with replacement supported by generally high bootstrap values (range 72– over loci (2000 replicates) using the program njbs (J.-M. 99%), except for the Romanian population, different from Cornuet, pers. comm.). all other populations. The two other groups present less robust nodes (range 24– 89%). We therefore defined three groups of populations Results corresponding to three geographical regions: the Pyrenees, the Alps (including the German population Fulda) and AFLP polymorphism Fennoscandia. These three groups are comparable both in Using three primer pairs, 128 scorable polymorphic frag- sample size (respectively five, six and five populations ments were generated. All the 180 individuals genotyped sampled) and in geographical area (populations within a presented different profiles. When all the 180 genotypes region are less than 700 km apart). Because the sampling were considered, a complete or nearly complete linkage was too restricted in eastern Europe (only two populations disequilibrium (D ≤ 0.01) was found for 57 pairs of markers, sampled), the Polish and the Romanian populations were due probably to physical linkage among these markers. excluded from the grouped populations analyses. To avoid redundant information, 11 markers were discarded before subsequent analyses, leading to a total of Within-population and within-region genetic diversity 117 independent locus markers: 46 for the primer pair

E.ATC/M.CAG, 47 for E.ATC/M.CAT and 24 for E.AGA/ The percentage of polymorphic loci (P%) and the Nei (HE) M.CTC. The mean linkage disequilibrium over all pairs of and Shannon (I) indices for each population were strongly loci was D = 0.44. correlated (Table 3, pairwise Spearman’s rank correlation s s coefficient P% vs. HE: r = 0.812; HE vs. I: r = 0.998; P% vs. I: r s = 0.814; all P < 0.001). The mean percentage of poly- Genetic distance analysis morphic loci was significantly higher within the Alpine The neighbour-joining dendrogram based on Nei’s populations (60.78 ± 2.21%) compared to the Pyrenean (1978) unbiased genetic distances between all pairwise (46.15 ± 5.39%) and Fennoscandian populations (43.25 ± combinations of populations (Table 2) revealed three 2.22%), whereas these two latter groups did not signi- population clusters: one group includes all the Pyrenean ficantly differ (Mann–Whitney U-test; Alps vs. Pyrenees: populations, another groups the five Alpine populations P = 0.035; Alps vs. Fennoscandia: P = 0.008; Pyrenees vs. with the German population, and the third groups Fennoscandia: P = 0.6). Nei’s and Shannon diversity Fennoscandian populations with the Polish and Romanian indices were also higher in the Alps (0.229 ± 0.024 and populations in a basal position (Fig. 2). This latter group is 0.338 ± 0.033, respectively) but it was significant only when

Fig. 2 Neighbour-joining dendrogram based on Nei’s distances for the 18 studied populations, and the corresponding regions. Bootstrap values over loci (based on 2000 replicates) are indicated for each node.

2342 L. DESPRES, S. LORIOT and M . GAUDEUL ns 0.0917 ns ns 0.0715 ns ns ion 0.0785 0.1262 0.1720 0.1102 ns ns ns ‡ . 0.0513 0.0718 0.1265 0.1533 < 0.001. ‡ ‡ ‡ ‡ ‡ P 0.1646 0.1385 0.1842 0.1934 0.1939 < 0.01; ‡ P ns 0.0624 ‡ ‡ < 0.05; †0.001 0.1981 P ns ‡ 0.2717 0.09580.0923 0.2549 ns ns ns ‡ 0.0630 0.2940 0.0886 0.1157 : non significant. *0.01 < ns ns ‡ ‡ † ns 0.1211 0.3588 0.1692 0.1646 0.1131

ns ns ns ns 0.0829 0.0751 0.1022 ns ns ns ns 0.0739 0.0925 0.1059 0.1223 0.1648† 0.1414 ns ns ns ns ns ‡ 0.1501 0.1096 0.1830 0.1158 0.1147 0.1386 ns ns ns ns ns ns ns 0.1166 0.1210 0.0916 0.1401 0.1249 0.1141 0.1541†0.1406 0.1443* 0.1544‡ 0.1551† 0.1473† 0.2228‡ 0.1679‡ 0.1890 0.1570† 0.1811‡ 0.1588‡ 0.3074‡ 0.1602‡ 0.1421‡ 0.1567 0.1445 0.1344† 0.1434* 0.1580‡ ns ns ns ns ns ns ns ns ns ns Pairwise Nei’s unbiased genetic distance among populations of European globeflowers and exact tests population differentiat 0.0608 0.089 0.0964 0.0687 0.0952 13 0.1688‡ 0.1979‡ 0.2119‡ 0.1896‡ 0.1713‡ 0.1594‡ 0.2095‡ 0.1645‡ 0.3006‡ 0.2109‡ Table 2 456 789101112131415161718 123 2 3 4 5 6 78 0.0969 9 0.0809 10 0.3158‡ 0.1162 0.3176‡12 0.2601‡ 0.1589‡ 0.3464‡14 0.1850‡ 0.3243‡ 0.1529‡15 0.2200‡ 0.1672‡16 0.1761‡ 0.1997‡ 0.2098‡ 0.1622‡ 0.1874‡ 0.1816‡ 0.2206‡ 0.1931‡17 0.1672‡ 0.2230‡ 0.1932‡ 0.1196 18 0.1699‡ 0.1683‡ 0.1941‡ 0.2107‡ 0.1768‡ 0.2066‡ 0.2233‡ 0.2364‡Populations numbered as in Fig. 1. are grouped into regions. Pairwise comparisons within regions italic type 0.2178‡ 0.1994‡ 0.1670‡ 0.1696‡Significance levels of exact tests differentiation are shown by: 0.1964‡ 0.1980‡ 0.1954‡ 0.3054‡ 0.1982‡ 0.3447‡ 0.1881‡ 0.2224‡ 0.2183‡ 0.3291‡ 0.2292‡ 0.1837‡ 0.1998‡ 0.2251‡ 0.1627‡ 0.2073‡ 0.3628‡ 0.1662‡ 0.2049‡ 0.1606‡ 0.3047‡ 0.1900‡ 0.1710‡ 0.2169‡ 0.2181‡ 0.2176‡ 0.2223‡ 0.1855‡ 0.1867‡ 11 0.1269

GENETIC DIVERSITY IN THE EUROPEAN GLOBEFLOWER 2343

Table 3 Within-population genetic diversity indices of the 18 populations studied, and total (T) or mean (M) values for the three geographical regions defined. Standard deviations are in parentheses

Number of Percentage of Nei’s diversity

Regions Populations polymorphic loci polymorphic loci (P%) index (HE) Shannon index (I)

Alps 1 Arêche 75 64.10% 0.239 (0.205) 0.353 (0.290) 2 Menée 80 67.52% 0.270 (0.207) 0.395 (0.291) 3 Fournel 67 57.26% 0.230 (0.218) 0.335 (0.308) 4 Cherlieu 66 54.41% 0.202 (0.206) 0.301 (0.293) 5 Galibier 66 56.41% 0.213 (0.213) 0.313 (0.301) 6 Fulda 76 64.96% 0.222 (0.201) 0.333 (0.283) T: 107 M: 60.78% (5.40%) M: 0.229 (0.024) M: 0.338 (0.033) Pyrenees 7 Pradella 54 46.15% 0.177 (0.210) 0.259 (0.300) 8 Eynes 70 59.83% 0.240 (0.216) 0.349 (0.305) 9 Gazies 33 28.21% 0.220 (0.213) 0.321 (0.304) 10 Gabardères 50 42.73% 0.148 (0.193) 0.222 (0.279) 11 Puymorens 63 53.84% 0.200 (0.210) 0.295 (0.298) T: 100 M: 46.15% (12.04%) M: 0.197 (0.036) M: 0.289 (0.050) Fennoscandia 12 Kvaloeya 44 37.61% 0.138 (0.198) 0.203 (0.283) 13 Ringassoeya 50 42.73% 0.154 (0.198) 0.229 (0.285) 14 Rhindunjira 57 48.72% 0.189 (0.212) 0.277 (0.302) 15 Slattatjakka 46 39.32% 0.144 (0.201) 0.212 (0.287) 16 Oulanka 56 47.86% 0.167 (0.202) 0.249 (0.288) T: 84 M: 43.25% (4.97%) M: 0.158 (0.020) M: 0.234 (0.029) Poland 17 Biebrza 53 42.30% 0.161 (0.200) 0.241 (0.287) Romania 18 Piatra Fontanele 52 44.44% 0.173 (0.209) 0.254 (0.299)

± compared to Fennoscandian populations (0.158 0.020 nees appeared to be the more differentiated region (FST and 0.234 ± 0.029, respectively), the Pyrenean populations = 0.39, 95% CI 0.33–0.44), followed closely by Fennoscandia

being intermediate and not significantly different from (FST = 0.36, 95% CI 0.30–0.42), whereas the Alps had signi- ± the two other regions for these two indices (0.197 0.036 ficantly lower among-population differentiation (FST = 0.24, and 0.289 ± 0.050, respectively). The Polish and Romanian 95% CI 0.20–0.27). populations exhibited levels of diversity comparable to Although the major part of genetic variation was found that of Pyrenean or Fennoscandian populations. within populations (64.0%), with only 19.5% variation At the regional level, 107 markers were polymorphic in among populations within regions and 16.6% among the Alps, 100 in the Pyrenees and only 84 in Fennoscandia, regions (amova, Table 4), exact tests showed a strong indicating that although the average within population genetic differentiation among the three regions (all three percentage of polymorphic markers was not higher in comparisons P < 0.001). When each region was analysed the Pyrenees as in Fennoscandia, different markers were separately, most of the variation was again detected within polymorphic in different Pyrenean populations, resulting populations, with up to 84.8% of the variance found within in a total number of polymorphic markers in the Pyrenees populations in the Alps. Pairwise exact tests of population close to that observed in the Alps. The Pyrenees were differentiation indicate that most pairs of populations from also characterized by higher heterogeneity for all within- different regions significantly differed for their marker population diversity indices. For example, the standard frequencies (Table 2). However, the three populations deviation of the percentage of polymorphic loci was 12.04% from the oriental Pyrenees (seven, eight and 11) were not in the Pyrenees, whereas it was only 5.40% and 4.97% in the significantly differentiated from some Alpine populations Alps and in Fennoscandia, respectively. (Table 2), indicating a close relationship between these two regions. To test whether the Pyrenees and the Alps really represent two differentiated regions, we excluded Among-population and among-region differentiation Fennoscandia from the amova and partitioned total vari-

The global FST value among all populations was 0.39 (95% ance among the Pyrenees and the Alps (11 populations): confidence interval (CI) 0.37–0.42), indicating a strong variance among regions fell down to 7%, but this was among-population differentiation. At a regional scale still significant (Table 4). Within regions, most pairwise (Polish and Romanian populations excluded), the Pyre- comparisons showed no significant difference for marker

2344 L. DESPRES, S. LORIOT and M . GAUDEUL

Table 4 Analysis of molecular variance (amova) based on 117 AFLP loci in three regions: Alps, Pyrenees and Fennoscandia. Global analysis (three regions), the group Alps + Pyrenees, and each region analysed separately. The Polish and Romanian populations were excluded from the analysis because they cannot be assigned to one of the three regions

Sum of squares Variance % of the Source of variation d.f. (SSD) components total variance P-value

Global analysis Among regions 2 410.919 3.01004 16.59 < 0.001 Among populations within regions 13 606.200 3.52397 19.42 < 0.001 Within populations 143 1660.567 11.61235 63.99 Total 158 2677.686 18.14636 Alps + Pyrenees Among regions 1 112.495 1.21520 7.11 < 0.005 Among populations within regions 9 422.800 3.49224 20.45 < 0.001 Within populations 98 1212.467 12.37211 72.44 Total 108 1747.761 17.07954 Alps Among populations in the Alps 5 190.950 2.44956 15.17 < 0.001 Within populations in the Alps 54 739.500 13.69444 84.83 Total 59 930.450 16.14400 Pyrenees Among populations in the Pyrenees 4 231.850 4.81968 30.96 < 0.001 Within populations in the Pyrenees 45 472.967 10.74924 69.04 Total 49 704.816 15.56892 Fennoscandia Among populations in Fennoscandia 4 183.400 3.58922 26.49 < 0.001 Within populations in Fennoscandia 45 448.100 9.95778 73.51 Total 49 631.500 13.5470

frequencies, except for the German population differing Such a finding of higher genetic diversity within rather from one south-alpine population (Fournel), one Pyrenean than among populations was reported commonly in out- population (Gazies) differing from all the other populations crossing and/or perennial plants, by contrast to selfing and in the Pyrenees, and the Finnish population differing from annual species which tend to exhibit the opposite pattern the Norwegian populations. (Hamrick et al. 1991). The highest mean within-population diversity was found in the Alps, and the lowest in Fennoscandia. This Genetic vs. geographical structure finding of lowest variability in northern populations is The overall Mantel test based on the 18 populations was probably explained by the past history of these popula- significant (r = 0.4904; P = 0.001). At a regional scale, there tions, and is in accordance with the ‘tabula rasa hypothesis’: was no correlation between genetic and geographical only a small sample of the southern genetic variability distances in the Pyrenees (r = 0. 1882; P = 0.191), nor in the was represented in the migrants that recolonized the newly Alps (r = 0.3201; P = 0.225), nor in a region grouping available space in the north (founder effect). However, the Pyrenean and Alpine populations (r = 0.3050; P = 0.09). By level of variability within Fennoscandian populations was contrast, a significant positive correlation between the still quite high (ranging from 37.6 to 48.7% of polymorphic two distance matrices was found in Fennoscandia (r = loci), and not significantly different from the variability 0.7494; P = 0.03). observed within Pyrenean populations (range 28.2–59.8%). In the Pyrenees, the lowest within-population diversity was observed in Gazies, which is also one of the smallest Discussion populations sampled in this region. By contrast, another population in the Pyrenees, Eynes, exhibited similar level Within-population diversity of genetic diversity as Alpine populations. This population Whatever the region, most of the genetic diversity was is one of the largest populations sampled in the Pyrenees. found within populations of the European globeflower. This suggests that the comparatively low genetic variability

GENETIC DIVERSITY IN THE EUROPEAN GLOBEFLOWER 2345

observed in the Pyrenees could be the result of genetic populations = 206 km). There was no correlation between drift (random fixation of alleles) in small populations, genetic and geographical distances in the Pyrenees, as especially strong at the southern edge of their distribution. expected if genetic drift is the major factor of differenti- A positive correlation between population size and genetic ation. For example, Gazies was genetically distant from variability has been found in many plant species (Fischer & all the other Pyrenean populations, although it was only Matthies 1998; Gaudeul et al. 2000), in agreement with the 10 km apart from Gabardères. The high among-population hypothesis that small populations cannot maintain diver- differentiation observed in the Pyrenees is the result of the sities as high as those found in larger populations. No such random fixation of different alleles in different populations relationship between population size and genetic diversity (up to 84 alleles were fixed in Gazies), regardless of their can be evidenced in the present study, because large geographical location: this indicates that populations populations in Fennoscandia do not exhibit the highest evolve independently from each other, and that gene flow level of genetic variability, probably due to the founder is very limited among them. By contrast, genetic distances effect when demographic expansion occurred. Genetic drift between Fennoscandian populations increased with in the Pyrenees is also supported by the fact that different geographical distances, in agreement with an isolation markers are polymorphic in different populations, indicat- by distance model (Ellstrand & Elam 1993). This pattern ing random allele fixation, and by a higher heterogeneity may reflect a stepwise recolonization process following between within-population diversity indices. glacier retreat, and/or present day short-distance gene flow Therefore, relatively low and similar levels of within- among more or less continuous large populations. The lack population variability in the Pyrenees and in Fennoscandia of correlation between genetic and geographical distances result from two very different processes: genetic drift due in the Alps indicates that genetic drift is acting in the Alps, to current small population sizes in the Pyrenees, and a but to a lesser extent than in the Pyrenees: gene flow among founder effect during postglacial recolonization in Fenno- populations is higher in the Alps than in the Pyrenees as scandian populations, followed by demographic expansion. showed by lower among-population differentiation. Thus, The high genetic diversity observed within Alpine popu- gene flow seems to counteract the effect of genetic drift, lations reflects the proximity of a glacial refugium, and and limits genetic erosion. the fact that populations are larger than in the Pyrenees. The presence of a glacial refugium in southwestern Alps Phylogenetic relationships among populations and was evidenced in Picea abies (Scotti et al. 2000). The high colonization routes genetic variability found within Alpine populations is concordant with the richness of its associated pollinators The 18 globeflower populations studied form three clusters community. corresponding to three present-day southern refugia for T. europaeus: the Alpine, Pyrenean and Carpathian refugia. The lack of differentiation between some Alpine and Among-population differentiation Pyrenean populations, together with the low bootstrap Despite the fact that most variability was observed within- values obtained for the two clades ‘Pyrenees’ and ‘Alps’ populations, overall among-population differentiation which are clustered together, suggests that the Alpine and

was high (FST = 0.39) suggesting that there is little gene the Pyrenean regions may have exchanged migrants flow among populations. This is not very surprising given during the last glacial period. Depending on the intensity the wide geographical range surveyed (distances between and regularity of these past exchanges, the Alpine and two populations range from 2 to 3252 km). Among- Pyrenean populations could even be considered as present- populations gene flow is limited by pollen and seed day relics of a single large T. europaeus population, as the dispersal. T. europaeus being an insect-pollinated plant, French Mediterranean region was not covered with ice pollen dispersal is limited by the flying capacity of its during the last ice-age. However, we evidenced significant pollinator, a small short-living fly (Chiastocheta genus). An genetic differentiation between these two regions, and non- allozyme analysis of the genetic structure of Chiastocheta significant correlation between geographical and genetic populations at a fine geographical scale (in Denmark) distances, indicating that even if some connections did evidenced strong local differentiation and low dispersal occur at some time between the Alpine and the Pyrenean (Johannesen & Loeschcke 1996). Moreover, seed dispersal ancestral populations, genetic drift is more important is not likely to be very efficient, given the weight of T. than gene flow in shaping present-day genetic diversity europaeus seed (approximately 0.5 mg, excluding wind patterns. We found no evidence of contribution of the transport) and its lack of dispersal structures. more eastern refuge (Carpathians) to the recolonization The Pyrenees was the region with the highest among- of the Alps, but our sampling included only populations

populations differentiation (FST = 0.39), despite being the from the French Alps, i.e. western Alps. A recent genetic smallest region surveyed (maximum distance between study of the Alpine populations of Picea abies evidenced the

2346 L. DESPRES, S. LORIOT and M . GAUDEUL

presence of two lineages in the Alps, with the western from one single southeastern lineage, supporting the populations being well differentiated from the eastern ‘tabula rasa’ hypothesis. Moreover, the analysis and populations (Scotti et al. 2000). interpretation of our data at several spatial and temporal The population from Germany, Fulda, is grouped with scales suggests that distinct evolutionary processes are the Alpine populations. This suggests that there is a route responsible for the patterns of within and among popul- northward from the Alpine refugium. All the Fennoscan- ation genetic diversity observed in the Alps, Pyrenees and dian populations sampled nest inside a cluster with Polish Fennoscandia. Whereas genetic drift seems to be the major and Romanian population in a basal position, indicating factor acting in the small Pyrenean populations, leading that north Arctic was probably recolonized from an eastern to decreased within-population variability and increased Carpathian refugia, with no admixture of populations among-population differentiation with no correlation from the Pyrenean or Alpine refugia. Several recent between genetic and geographical distances (random molecular phylogeographies involving animals and plants fixation of alleles), the low diversity in Fennoscandia would evidenced the postglacial recolonization of northern rather be due to a past founder effect during northward Fennoscandia from the Balkanic refugia, together with the recolonization. This past recolonization wave and/or recolonization of southern Scandinavia from the Iberian the present gene exchange between adjacent large popu- or Italian refuges (Taberlet et al. 1998; Hewitt 2001). These lations are concordant with the isolation-by-distance expansion routes meet in central Scandinavia where the model. Last, the Alpine populations retain most of the ice-cap melted some 9000 years ago and form an hybrid ancestral genetic variability, in a moderately fragmented zone. Our sampling did not include globeflowers from habitat, and therefore exhibit the maximum within- southern Scandinavia, but such an east–west colonization population diversity. The geographical pattern of genetic pattern for T. europaeus is suggested by the German variation observed in T. europaeus supports sympatric population expanding northward from the Alps. Although speciation of associated flies, rather than allopatric a number of studies demonstrated recolonization by cold- speciations in well-isolated refugial populations. Local tolerant plant species of boreal regions from high arctic species extinction could have occurred latter, during refugia (Abbott et al. 1995; Lagercrantz & Ryman 1990; host-plant range fragmentation in the south, and during Tremblay & Schoen 1999), our study shows no evidence northern expansion. of a northern refugium for T. europaeus. This cold-tolerant plant is very sensitive to dryness, and is totally dependent Acknowledgements on Chiastocheta fly activity for its reproduction, so that its survival during dry glacial periods in high arctic ice-free We thank all the T. europaeus collectors: Jes Johannesen (Germany), refugia was probably precluded. The lack of northern Frank d’Amico (France), Ada Wrodlewska (Bierbza National refugia for the host-plant precludes the possibility for Park, Poland), Ioan Chintauan (Carpathians, Romania), Gilles Yoccoz (Norway), Asa Hemborg (Abisko Biological station, Chiastocheta allopatric speciation in differentiated southern/ Sweden) and Pirkko Siikamaki (Oulanka Biological Station, northern refugia. Furthermore, the lack of strong genetic Finland). We aknowledge Ludovic Gielly for his help with the differentiation between the Alpine and Pyrenean globe- automated sequencer, Jean-Marie Cornuet and Pierre Berthier for flower populations suggests that these two regions were providing the computer programs NJBS and DDM, respectively, connected during Pleistocene, presumably preventing and Pierre Taberlet for helpful comments on the manuscript. allopatric fly speciation. No Chiastocheta species is specific to a geographical region. This suggests that sympatric References speciation may have occurred in large, ancestral, southern host-plant populations throughout disruptive selection on Abbott RJ, Chapman HM, Crawford RMM, Forbes DG (1995) the date of oviposition. In this perspective, the absence of Molecular diversity and derivations of populations of Silene some Chiastocheta species in the Pyrenees and Fennoscandia acaulis and saxifraga oppositifolia from the high Arctic and southerly latitudes. Molecular Ecology, 4, 199–207. could result from species extinction during fragmentation Birks HJ (1996) Statistical approaches to interpreting diversity of the host-plant populations (in the Pyrenees), and species patterns in the Norwegian mountain flora. Ecography, 19, 332–340. loss during northward expansion (in Fennoscandia). Comes HP, Kadereit JW (1998) The effects of Quaternary climatic changes on plant distribution and evolution. Trends in Plant Science, 3, 432–438. Conclusion Després L, Jaeger N (1999) Evolution of oviposition strategies and Our study shows that T. europaeus probably survived speciation in the globeflower flies Chiastocheta spp. (Anthomyiidae). Journal of Evolutionary Biology, 12, 822–831. the last glaciation in more or less interconnected south- Després L, Pettex E, Plaisance V, Pompanon F (2002) Speciation in ern populations, represented nowadays by the relictual the globeflower flies Chiastocheta ssp. (Diptera: Anthomyiidae) in populations found at high elevation in the Pyrenees, Alps relation to host plant species, biogeography and morphology. and Carpathians. The high arctic populations originated Molecular Phylogenetics and Evolution, 22, 258–268.

GENETIC DIVERSITY IN THE EUROPEAN GLOBEFLOWER 2347

Ellstrand NC, Elam DR (1993) Population genetic consequences of Nordal I (1987) Tabula rasa after all? Botanical evidence for ice-free small population size: implications for plant conservation. refugia in Scandinavia reviewed. Journal of Biogeography, 14, Annual Review of Ecology and Systematics, 24, 217–242. 377–388. Ferdy JB, Després L, Godelle B (2002) Evolution of mutualism Pellmyr O (1989) The cost of mutualism between Trollius europaeus between globeflowers and their pollinating flies. Journal of and its pollinating parasites. Oecologia, 78, 53–59. Theoretical Biology, 217, 219–234. Pellmyr O (1992) The phylogeny of a mutualism: evolution and Fischer M, Matthies D (1998) RAPD variation in relation to popu- coadaptation between Trollius and its seed-parasitic pollinators. lation size and plant fitness in the rare Gentianella germanica Biological Journal of the Linnean Society, 47, 337–365. (Gentianaceae). American Journal of Botany, 85, 811–819. Raymond M, Rousset F (1995) An exact test for population differ- Gaudeul M, Taberlet P, Till-Bottraud I (2000) Genetic diversity in entiation. Evolution, 49, 1280–1283. an endangered Alpine plant, Eryngium alpinum L. (Apiaceae), Schneider S, Küffer JM, Rössli D, Excoffier L (2000) Arlequin, inferred from amplified fragment length polymorphism Version 2.0: a Software for Population Genetic Data Analysis. markers. Molecular Ecology, 9, 1625–1637. Genetics and Biometry Laboratory, University of Geneva, Hamrick JL, Godt MJW, Murawski DA, Loveless MD (1991) Geneva. Correlations betwwen species traits and allozyme diversity: Scotti I, Vendramin GG, Matteotti LS, Scarponi C, Sari-Gorla M, implications for conservation biology. In: Genetics and Conserva- Binelli G (2000) Postglacial recolonization routes for Picea tion of Rare Plants (eds Falk DA, Holsinger KE), pp. 75–86. abies in Italy as suggested by the analysis of sequence- Oxford University Press, Oxford. characterized amplified region (SCAR) markers. Molecular Hemborg A, Després L (1999) Oviposition by mutualistic seed Ecology, 9, 699–708. parasitic pollinators and its effects on annual fitness of single- Taberlet P, Fumagalli L, Wust-Saucy A-G, Cosson JF (1998) and multi-flowered host plants. Oecologia, 120, 427–436. Comparative phylogeography and post-glacial colonization Hewitt GM (1996) Some genetic consequences of ice ages, and routes in Europe. Molecular Ecology, 7, 453–464. their role in divergence and speciation. Biological Journal of the Tremblay NO, Schoen DJ (1999) Molecular phylogeography of Linnaean Society, 58, 247–276. Dryas integrifolia: glacial refugia and post-glacial recolonization. Hewitt GM (2001) Speciation, hybrid zones and phylogeography- Molecular Ecology, 8, 1187–1198. seeing genes in space and time. Molecular Ecology, 10, 537– Vendramin GG, Anzidei M, Madaghiele A, Sperisen C, Bucci G 549. (2000) Chloroplast microsatellite analysis reveals the presence Jaeger N, Després L (1998) Obligate mutualism between Trollius of population subdivision in Norway spruce (Picea abies K.). europaeus and its seed-parasite pollinators Chiastocheta flies in Genome, 43, 68–78. the Alps. Comptes Rendus de l’Academie des Sciences, Paris, Séries Vos P, Hogers R, Bleeker M et al. (1995) AFLP: a new technique for III, 321, 789–796. DNA fingerprinting. Nucleic Acids Research, 23, 4407–4414. Jaeger N, Pompanon F, Després L (2001) Variation in predation Walter R, Epperson BK (2001) Geographic pattern of genetic costs with Chiastocheta egg number on Trollius europaeus: how variation in Pinus resinosa: area of greatest diversity is not the many seeds to pay for pollination? Ecological Entomology, 26, origin of postglacial populations. Molecular Ecology, 10, 103–111. 56–62. Webb T, Bartlein PJ (1992) Global changes during the last 3 million Jaeger N, Till-Bottraud I, Després L (2000) Evolutionnary conflict years: climatic controls and biotic response. Annual Review of between Trollius europaeus and its seed-parasite pollinator Ecological Systematic, 23, 141–173. Chiastocheta flies. Evolutionary Ecology Research, 2, 885–896. Weir BS, Cockerham CC (1984) Estimating F-statistics for the ana- Johannesen J, Loeschcke V (1996) A hierarchical analysis of genetic lysis of population structure. Evolution, 38, 1358–1370. structure and variability in patchily distributed coexisting Weising K, Nybom H, Wolff K, Meyer W (1995) DNA-fingerprinting Chiastocheta species (Diptera: Anthomyiidae). Heredity, 76, in Plants and Fungi. CRC Press, Boca Raton. 437–448. Yeh F, Yang RC, Boyle T (1997) POPGENE. The User-Friendly Lagercrantz U, Ryman N (1990) Genetic structure of Norway spruce Shareware for Population Genetic Analysis. Molecular and Biotech- (Picea abies): concordance of morphological and allozymic nology Centre, University of Alberta, Edmonton. variation. Evolution, 44, 38–53. Lewontin RC (1972) The apportionment of human diversity. Evolutionary Biology, 6, 381–394. Laurence Després teaches evolutionary ecology, population Mantel N (1967) The detection of disease clustering and a general- genetics and phylogeny at the University of Grenoble. Her main ized regression approach. Cancer Research, 27, 209–220. research interests are in speciation patterns and underlying Miller MP (1997) Tools for Population Genetic Analysis (TFPGA), evolutionary processes in host-parasites and plant–insects inter- Version 1.3. Department of Biological Sciences, Northern actions. The present work is a contribution to a broad ongoing study Arizona University, AZ, USA. on the evolutionary ecology of the Trollius/Chiastocheta inter- Mueller UG, Wolfenbarger LL (1999) AFLP genotyping and action, and is part of Sandrine Loriot’s DEA (Master’s degree) on fingerprinting. Trends in Ecology and Evolution, 14, 389–395. the population genetics of Trollius europaeus. Myriam Gaudeul Nei M (1978) Estimation of average heterozygosity and genetic is a PhD student interested in the evolutionary biology and distance from a small number of individuals. Genetics, 89, conservation genetics and ecology of plant species. 583–590.