landscape structure, clonal propagation, and ... - Semantic Scholar

American Journal of Botany 94(7): 1146–1155. 2007.

LANDSCAPE

STRUCTURE, CLONAL PROPAGATION, AND GENETIC

SCANDINAVIAN POPULATIONS ARABIDOPSIS LYRATA (BRASSICACEAE)1

DIVERSITY IN

MYRIAM GAUDEUL,2 HANS K. STENØIEN,3

AND

OF

˚ GREN4 JON A

Department of Ecology and Evolution, Evolutionary Biology Centre, Uppsala University, Villava¨gen 14, SE-752 36 Uppsala, Sweden Colonization history, landscape structure, and environmental conditions may influence patterns of neutral genetic variation because of their effects on gene flow and reproductive mode. We compared variation at microsatellite loci within and among 26 Arabidopsis lyrata populations in two disjunct areas of its distribution in northern Europe (Norway and Sweden). The two areas probably share a common colonization history but differ in size (Norwegian range markedly larger than Swedish range), landscape structure (mountains vs. coast), and habitat conditions likely to affect patterns of gene flow and opportunities for sexual reproduction. Within-population genetic diversity was not related to latitude but was higher in Sweden than in Norway. Population differentiation was stronger among Norwegian than among Swedish populations (FST ¼ 0.23 vs. FST ¼ 0.18). The frequency of clonal propagation (proportion of identical multilocus genotypes) increased with decreasing population size, was higher in Norwegian than in Swedish populations, but was not related to altitude or substrate. Differences in genetic structure are discussed in relation to population characteristics and range size in the two areas. The results demonstrate that the possibility of clonal propagation should be considered when developing strategies for sampling and analyzing data in ecological and genetic studies of this emerging model species. Key words:

Arabidopsis lyrata; Brassicaceae; clonality; genetic structure; microsatellites; Scandinavia.

Demographic history of populations, gene dispersal, and breeding system are major factors influencing patterns of neutral genetic diversity and differentiation. Plant populations of the northern hemisphere were subject to range contractions and expansions as a consequence of climate fluctuations during the Quaternary period. In northern Europe, most populations located in previously glaciated regions are the products of postglacial colonization from ancestral populations that survived in more southern refugia outside the glaciated area (Comes and Kadereit, 1998; Taberlet et al., 1998; Hewitt, 2000, 2004; Despres et al., 2002; Scho¨nswetter et al., 2003). Successive founder events along colonization routes can explain a gradual decline in within-population diversity and an increase in among-population differentiation away from refugia (Hewitt, 2000, 2004). At more restricted historical and spatial scales, landscape structure and degree of habitat fragmentation influence patterns of gene flow and thereby genetic diversity within populations and genetic differentiation among populations (Young et al., 1996). For lowland plants, alpine areas represent unsuitable habitat and may constitute physical barriers to gene flow, promoting genetic differentiation among populations (Heuertz et al., 2004; Hewitt, 2004). In contrast, other landscape characteristics, such as river systems, may promote gene flow by facilitating seed or pollen 1

Manuscript received 13 September 2006; revision accepted 29 May 2007.

The authors thank S. Sandring and P. Tora¨ ng for help with sampling and M. Heuertz for technical assistance. This work was financially supported by grants from the Swedish Research Council (to H.K.S. ˚ .) and Formas (to J.A ˚ .). and J.A 2 Current address: UMR 5202 (Origine, Structure et Evolution de la Biodiversite´ ), De´ partement Syste´ matique et Evolution, Museum National d’Histoire Naturelle, 16 rue Buffon, F-75005 Paris, France 3 Current address: Department of Biology, Norwegian University of Science and Technology, N-7491 Trondheim, Norway 4 Author for correspondence (e-mail: [email protected])

dispersal and thereby reduce among-population differentiation (Kudoh and Whigham, 1997; DeWoody et al., 2004). However, few studies have examined the extent to which geographic variation in landscape physiognomy has resulted in contrasting patterns of spatial genetic variation in plant species (Till-Bottraud and Gaudeul, 2002; Manel et al., 2003; Hirao and Kudo, 2004; Dechaine and Martin, 2005). Neutral genetic diversity and differentiation are also influenced by intrinsic factors, among which the breeding system, in particular the selfing rate and the propensity for clonal propagation, is of crucial importance (Hamrick and Godt, 1989, 1996). Strict or partial clonal propagation should result in nonrandom distribution of genotypes within populations, i.e., genetic substructuring, and generalized linkage disequilibrium within the genome (McLellan et al., 1997; Eckert, 2002; Balloux et al., 2003; Eckert et al., 2003; Halkett et al., 2005). The extent to which clonal propagation affects genetic diversity and its distribution among and within populations is less clear. Both theoretical and empirical studies indicate that low levels of sexual recruitment can maintain high levels of genotypic variation in predominantly clonal populations (Ellstrand and Roose, 1987; Watkinson and Powell, 1993; Wide´n et al., 1994; Holderegger et al., 1998; Auge et al., 2001; Bengtsson, 2003). Recent theoretical treatments indicate that the frequency of clonal propagation is critical for the effects on heterozygosity and genetic structure. They suggest that high rates of asexual reproduction should reduce genotypic diversity, but may increase heterozygosity and lead to negative FIS values and reduced population differentiation (Balloux et al., 2003; Halkett et al., 2005). In contrast, mixed clonal and sexual reproduction may be very similar to strict sexual reproduction in terms of population genetic consequences (Balloux et al., 2003; Halkett et al., 2005). Empirical studies have found evidence for both weak (Edwards and Sharitz, 2003) and substantial effects of clonal propagation on measures of genetic diversity and genetic structure in plant populations

1146

July 2007]

G AUDEUL

ET AL .—P OPULATION STRUCTURE IN

(Reusch et al., 1999; Ivey and Richards, 2001; Stoeckel et al., 2006). Additional empirical and theoretical work are thus needed (Stoeckel et al., 2006). The importance of clonal growth should vary along environmental gradients affecting conditions for sexual and vegetative propagation (Eckert and Barrett, 1993; Eckert, 2002; Eckert et al., 2003). Shifts from sexual to vegetative reproduction with increasing altitude and/or latitude have been documented in both vascular plants and mosses (Young et al., 2002; Hassel et al., 2005) and may reflect reduced sexual reproduction or less successful establishment of sexually produced progeny at low temperatures and when the growing season is short (Eckert, 2002). Poor conditions for sexual reproduction may increase the importance of vegetative propagation also where the availability of mates is low such as in small and sparse populations, and where resources are scarce (Williams, 1975; Sun et al., 2001). Finally, vegetative propagation may increase where soil conditions allow rhizomes to spread easily, for example, in sandy soils compared to rocky habitats (Hangelbroek et al., 2002, 2003). In this study, we quantified variation at microsatellite loci to examine genetic structuring in local populations of the outcrossing perennial herb Arabidopsis lyrata subsp. petraea across two of its main distribution areas in northern Europe, i.e., southwestern Norway and the coast of eastern Sweden. Both areas were probably colonized from ancestral populations that survived in unglaciated refugia south of Scandinavia and thus are likely to share a common colonization history. However, the two areas differ in size and in landscape structure and habitat conditions likely to affect patterns of gene flow and reproductive mode (sexual reproduction vs. vegetative propagation). In southwestern Norway, A. lyrata is found over a large mountainous area in riverbeds and landslides, from sea level up to 1400 m a.s.l. High alpine areas may represent dispersal barriers, and the opportunities for sexual reproduction may become reduced with increasing altitude because of increasingly harsh weather conditions. In Sweden, A. lyrata occurs in a geographically restricted area on exposed shores on the coast of the Bothnian Bay (Jonsell et al., 1995), which should facilitate gene flow among populations. Arabidopsis lyrata is capable of vegetative propagation through the production of subterranean runners, but the frequency of vegetative propagation has not been quantified in the field. We addressed the following questions: Is clonal propagation more frequent in populations on sandy soils than in populations inhabiting rock crevices, and does clonal propagation increase with increasing altitude and decreasing population size? Does genetic diversity within populations decrease towards the north and with decreasing population size, and does it differ between Norway and Sweden? Are populations more strongly differentiated in the Norwegian than in the Swedish range? Moreover, because clonal propagation has not been considered in previous population genetic studies of A. lyrata, we examined the effects of including more than one individual of repeated multilocus genotypes when estimating population genetic parameters. MATERIALS AND METHODS Study species—Arabidopsis lyrata (L.) O’Kane & Al-Shehbaz is an outcrossing perennial with a circumpolar distribution. Three subspecies are recognized (O’Kane and Al-Shehbaz, 1997), one of which occurs in Europe.

A RABIDOPSIS

LYRATA

1147

We studied subspecies petraea (L.) O’Kane & Al-Shehbaz [syn.: Arabis petraea L., Cardaminopsis petraea (L.) Hiit.], which has a disjunct European distribution (Jalas and Suominen, 1994). Studies of A. lyrata have documented high levels of within-population variability in allozyme, microsatellite, and DNA sequence markers (Jonsell et al., 1995; Van Treuren et al., 1997; Schierup, 1998; Savolainen et al., 2000; Clauss et al., 2002; Charlesworth et al., 2003; Ka¨rkka¨inen et al., 2004; Clauss and Mitchell-Olds, 2006). At the same time, A. lyrata populations also seem highly structured with estimated fixation indices FST ranging from 0.12 to 0.59 (Jonsell et al., 1995; Van Treuren et al., 1997; Schierup, 1998; Ka¨rkka¨inen et al., 2004). However, because most studies have included a limited number of populations, and populations sampled on large geographical scales, the regional patterns of neutral genetic structure remain largely undocumented in A. lyrata. Sampling—We sampled leaf material from 14 Norwegian and 12 Swedish populations of A. lyrata (Fig. 1, Table 1). In Sweden, all study populations were located close to sea level, whereas the altitude of the Norwegian study populations ranged from 10 to 1360 m a.s.l. In Norway, all study populations occurred on sand and gravel, whereas in Sweden some populations occurred in rock crevices (Table 1). Average population sizes did not differ between the two areas (test conducted on log [number of plants], t ¼ 1.4, P ¼ 0.16). In each population, samples were collected from 20 plants spread across the population. Sampled plants were separated by at least 50 cm except in the small populations S6, N3, N7, and N14, where some sampled plants were separated by only 10– 20 cm. The samples were immediately dried on silica gel and stored at room temperature. Microsatellite procedure—DNA was extracted using the DNeasy 96 Plant Kit (Qiagen, Crawley, UK), following the manufacturer’s protocol. We screened for genetic variability using 22 microsatellite markers described in Clauss et al. (2002). Multiplex PCR was developed following the recommendations of Henegariu et al. (1997) and yielded 22 markers amplified in four reactions (Table 2). Amplification reactions were performed in a 25-lL volume containing 1.5 lL DNA template, 1 3 Taq Buffer, 2 mM MgCl2, 0.2mM of each dNTP, 2 U AmpliTaqGold Polymerase (all from Applied Biosystems, Stockholm, Sweden), and varying concentrations of primers (determined to optimize the intensity of each microsatellite band, see Table 2). For each locus, one primer was fluorescently labeled (Applied Biosystems for NED-labeled primers; Invitrogen, Stockholm, Sweden for HEX- and 6-FAM-labeled primers; Table 2), allowing further multiplexing after amplification. The following cycling conditions were used: 10 min at 958C followed by 40 cycles composed of 30 s denaturation at 958C, 30 s annealing at 468C, and 4 min extension at 658C. PCR products from multiplex 1 and 2 and from multiplex 3 and 4, respectively, were mixed in equal proportions and loaded on 5% polyacrylamide gels; electrophoresis was run for 3 h on a 377 ABI Prism automated sequencer. Fragments were detected with the Genescan software, and their fragment sizes were determined with the Genotyper software (Applied Biosystems). Three loci were not genotyped because of ambiguous banding patterns (AthDET) or numerous missing data (ICE15 and ca72), possibly due to null alleles. One primer pair (AthGAPAb) produced two microsatellite patterns in distinct size ranges, and one locus (MDC16) was completely monomorphic, yielding 19 polymorphic loci. Statistical analyses—For each population, we calculated the expected heterozygosity (He; Nei, 1987), the allelic richness (mean number of alleles per locus based on the minimal sample size; El Mousadik and Petit, 1996), and the fixation index FIS (Weir and Cockerham, 1984) using the program FSTAT (Goudet, 1995). The statistical significance of FIS was assessed by 1000 random permutations of alleles in each population at each locus. We calculated Spearman rank correlations (rs) to examine whether heterozygosity and allelic richness were related to latitude and population size. Differences in variability measures between Norway and Sweden were tested by 1000 random permutations of populations across countries. The presence of linkage disequilibrium (LD) was tested for all pairs of markers within each population using the log-likelihood ratio G statistic implemented in FSTAT. The percentage of loci in linkage disequilibrium in relation to the maximum possible number of loci in linkage disequilibrium (Pd) was estimated for each population (Stenøien and Sa˚ stad, 1999). To detect signs of recent bottlenecks, we examined deviations in heterozygosity from mutation–drift equilibrium in each population with the software Bottleneck (Cornuet and Luikart, 1996). Under the infinite allele

1148

A MERICAN J OURNAL

OF

B OTANY

[Vol. 94

Fig. 1. Differentiation at microsatellite loci among populations of Arabidopsis lyrata in Norway and Sweden estimated by F statistics and two Bayesian methods. For each population, a pie represents the relative contribution of the five clusters identified by the program Structure. Dashed lines group populations clustered by F statistics (nonsignificant pairwise FST), while solid lines group populations clustered by the program BAPS. The gray area in the detailed map to the left indicates the distribution of A. lyrata in Norway. model (IAM) or as soon as the mutation model slightly deviates from a pure one-step stepwise mutation model (SMM), the loss of rare alleles in recently bottlenecked populations leads to an excess of heterozygosity relative to the expected heterozygosity with the same number of alleles at mutation–drift equilibrium. Microsatellite loci probably follow an intermediate mutation model between the extreme IAM and SMM (Estoup et al., 2002; Symonds and Lloyd, 2003). We ran the test by consecutively assuming the IAM, the SMM, and the two-phase mutation model (TPM) with 70% of single-step mutations (the remaining consisting of multiple-step mutations). We performed the implemented Wilcoxon test, considered the most powerful and robust among the proposed tests in the Bottleneck software. To partition the genetic variance into among-country, among-population within-country, and within-population components, we conducted an analysis of molecular variance (AMOVA; Excoffier et al., 1992) in Arlequin (Schneider et al., 2000). We estimated the fixation index FST (Weir and Cockerham, 1984) based on the overall data set, within each country, and among all pairs of populations. The statistical significance of FST estimates was tested by 1000 random permutations of individuals across populations using Genetix (Belkhir et al., 1996). Using FSTAT, we compared the level of population differentiation in the two countries by conducting 1000 random permutations of population across countries. This analysis was also performed between Swedish populations and subsets of Norwegian populations (namely N1 to N4, N5 to N10 excluding N7, and N13 to N16; cf. Fig. 1) to account for the difference in range size between Norway and Sweden. We tested for isolation by distance by performing Mantel tests, i.e., testing for correlations between FST/(1  FST) and the logarithm of geographic distances between pairs of populations. Mantel tests were performed on the overall data set and separately by country using FSTAT. The geographical distribution of genetic variability was further investigated with two Bayesian methods implemented in the programs Structure (Pritchard

et al., 2000; Falush et al., 2003) and BAPS (Corander et al., 2003). In these analyses, Markov chain Monte Carlo (MCMC) algorithms are used to group individuals (in Structure) or populations (in BAPS) in clusters (where the numbers of clusters is a priori unknown) in such a way as to achieve Hardy– Weinberg equilibrium and linkage equilibrium within each cluster. The program BAPS was run 10 times with a burn-in period of 10 000 iterations followed by 50 000 iterations. For determining the most adequate numbers of clusters (K), the program Structure was run 10 times for each K value from one to 20, each run comprising a burn-in period of 30 000 iterations followed by 100 000 iterations. In Structure, we used the admixture model (each individual may have inherited some fraction of its genome from ancestors in several clusters) and the independent allele frequencies model (allele frequencies differ randomly across populations; Falush et al., 2003; Pritchard and Wen, 2003). After the number of clusters (K) was estimated, an AMOVA was used to estimate the proportion of genetic variation explained by the division in K clusters. We used analysis of covariance to test whether clonality (proportion of identical multilocus genotypes) increased with decreasing population size and differed between areas (Norway vs. Sweden). Because variation in altitude and substrate was observed in only one area each, we examined the relationship between clonality and altitude using data on Norwegian populations (multiple regression model that also included population size as an independent variable), and the relationship between clonality and substrate (sand and gravel vs. rock crevices) using data on Swedish populations (Mann–Whitney U test). There were too few populations sampled on sand and gravel in Sweden to evaluate simultaneously the effects of population size and substrate on clonality. To assess the probability that observed identical multilocus genotypes were the product of sexual reproduction rather than of clonal propagation, we calculated the product of genotypic probabilities across loci following the procedure for codominant loci outlined by Sydes and Peakall (1998). For pairs

July 2007]

G AUDEUL

ET AL .—P OPULATION STRUCTURE IN

A RABIDOPSIS

1149

LYRATA

TABLE 1. Characteristics of the 26 Scandinavian populations of Arabidopsis lyrata. Population size was estimated as approximate number of established vegetative and flowering plants. Population

Location

Latitude (N)

Longitude (E)

Altitude (m a.s.l.)

Size

Area (m2)

Norway N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N13 N14 N15 N16

Sunndalsøra Gjøra Grøvudalen Nedre Kamtjern Visdal Spiterstulen ˚ heim A Briksdal ˚ rdal Øvre A Lærdal Sæbo Hjølmodalen Roaldkvam Proststøldalen

62841 0 62834 0 62830 0 62839 0 61842 0 61841 0 62803 0 61842 0 61818 0 61806 0 60825 0 60824 0 59839 0 59828 0

8835 0 9807 0 9805 0 9839 0 8823 0 8825 0 5834 0 6851 0 7851 0 7830 0 7809 0 7810 0 6858 0 6849 0

10 210 900 1360 500 1100 70 100 50 60 30 80 80 820

150 150 50 500 350 3000 300 500 250 350 350 90 500 110

500 100 80 800 150 .1000 20 180 600 100 500 8 600 300

Coarse Coarse Coarse Coarse Coarse Coarse Sand Coarse Coarse Coarse Coarse Sand Coarse Coarse

Sweden S1 S2 S3 S6 S7 S8 S10 S13 S15 S16 S17 S19

Skagsudde Stubbsand 1 Stubbsand 2 Storstensudden Svartlandsudden Storsanden Norrfa¨llsviken Rotsidan Klubben Pra¨sthushamn Notsand Sjo¨viken

63811 0 63812 0 63813 0 63811 0 63811 0 62859 0 62857 0 62850 0 62844 0 62842 0 62836 0 62835 0

19800 0 18857 0 18858 0 18848 0 18846 0 18831 0 18831 0 18822 0 18809 0 18808 0 18803 0 18801 0

,5 ,5 ,5 ,5 ,5 ,5 ,5 ,5 ,5 ,5 ,5 ,5

200 2000 2000 120 800 300 500 2000 300 80 1000 400

800 .1000 .1000 100 300 .1000 400 .1000 800 400 .1000 600

Rock crevices Sand þ gravel Rock crevices Sand Sand þ gravel Sand þ gravel Rock crevices Rock crevices Rock crevices Rock crevices Rock crevices Rock crevices

Soil

gravel gravel gravel gravel gravel gravel gravel gravel gravel gravel gravel gravel

RESULTS

population size interaction, P . 0.24). Multiple regression analysis conducted on data from Norwegian populations indicated that the proportion of identical genotypes was negatively related to log of population size (P ¼ 0.01) but not significantly related to altitude (P ¼ 0.41). Among Swedish populations, the proportion of identical genotypes did not differ between populations on sand and gravel and populations in rock crevices (Mann-Whitney U test, P ¼ 0.91); evidence of clonal propagation was detected in one of four populations on sand and gravel and in two of eight populations in rock crevices (Tables 1 and 3).

Redundancy of multilocus genotypes—The 19 polymorphic microsatellite markers displayed a total of 77 alleles, with two to 13 alleles per locus (mean 6 SD: 4.1 6 3.0). Despite this considerable variation, identical multilocus genotypes were detected in 12 of 14 populations sampled in Norway and in three of 12 populations sampled in Sweden (Table 3). Genetically identical samples originated from the same populations except for one multilocus genotype, which was found in populations N4 and N14 in Norway. In all other cases, multilocus genotypes were shared by closest neighbors among sampled plants within populations. The probability that the observed number of samples with identical multilocus genotypes represented separate genets with shared genotypes was 0.0032 or lower in all cases (median [range], 2.7 3 106 [6.3 3 1019  0.0032], N ¼ 40). Analysis of covariance indicated that the proportion of samples with identical multilocus genotypes decreased with increasing population size (F1,23 ¼ 9.3, P ¼ 0.006) and was higher in Norway than in Sweden (F1,23 ¼ 9.6, P ¼ 0.005; no significant country 3

Genetic variability within populations—Within-population diversity (He and allelic richness) was higher in Swedish populations than in Norwegian populations (He, P ¼ 0.002; allelic richness, P ¼ 0.032; Table 3) but did not vary with latitude (Spearman rank correlations, P . 0.29) or population size (P . 0.07) in either the Norwegian or Swedish range. Multilocus FIS was significantly positive in N15 and S19, and significantly negative in N7 and S15. After Bonferroni correction, there were indications of linkage disequilibrium (LD) between three locus pairs (ICE14-AthGAPAb1 in N4, AthZFPG-AthGAPAb2 in N7, and F20D22-AthCDPK9 in N14) leading to Pd values, i.e., percentage of possible locus pairs in significant LD, ranging from 1.5% to 1.8% in the three populations. Observed heterozygosities suggested that relatively few populations had undergone recent bottlenecks. Assuming the infinite-allele model, the one-tail Wilcoxon test for heterozygosity excess indicated recent bottlenecks in five Swedish populations (after Bonferroni correction). However, heterozy-

of loci that showed evidence of linkage disequilibrium, only one of the two loci was considered. All measures of genetic variation within and among populations were calculated on a data set from which redundant genotypes had been excluded, i.e., each multilocus genotype was allowed to be represented by not more than one individual in each population. To assess the influence of vegetative propagation on measures of genetic diversity (paired sampled t tests on expected heterozygosity and allelic richness measures) and genetic structure, we compared these results to those obtained based on a data set that included all sampled plants.

1150

A MERICAN J OURNAL

OF

B OTANY

[Vol. 94

TABLE 2. Conditions for multiplex PCR and characteristics of the Arabidopsis lyrata microsatellites. Microsatellite loci that could not be genotyped reliably or that were monomorphic are in parentheses.

Microsatellite locus

Repeat

Primer concentration (lM)

Fluorophorea

No. of alleles

Size range (base pairs)

PCR multiplex 1 nga162 nga151 (AthDET1) ICE5 (ICE15) (ca72) (MDC16) ICE6 ICE12

GA CT CT GT TA CT GA CCA CT

0.15 0.10 0.20 0.20 0.45 0.77 0.15 0.40 0.40

6-FAM 6-FAM 6-FAM 6-FAM 6-FAM 6-FAM HEX HEX HEX

2 3 — 2 — — 1 3 3

79–81 98–106 — 167–171 — — 117 146–151 229–233

PCR multiplex 2 AthCDPK9 F19K23–483 SLL2 nga280 ATTS0392 ICE13

TC TTC CA AG AAG ATC

0.30 0.30 0.40 0.40 0.30 0.30

HEX HEX HEX NED NED NED

4 2 2 2 6 10

88–96 178–188 300–311 76–82 140–158 222–267

PCR multiplex 3 ICE7 AthPHYC ICE14 ICE11 AthGAPAb AthGAPAb 0

GAA TC GAT GA GTT —

0.25 0.40 0.30 0.30 0.45 —

6-FAM 6-FAM 6-FAM HEX NED NED

2 2 6 4 2 3

95–98 184–210 225–240 136–147 123–132 159–162

PCR multiplex 4 AthZFPG F20D22

CT GTTT

0.50 0.40

6-FAM HEX

13 6

137–168 173–187

a Fluorophores are chemical dyes used for fluorescent labeling of oligonucleotides; they differ by their excitation wavelength and color display (6-FAM ¼ blue; HEX ¼ green; NED ¼ yellow).

gosity excess was statistically significant only in population S2 under the two-phase mutation model 70% (corrected P ¼ 0.035) and in no population under the stepwise mutation model. Genetic variability among populations—Populations were substantially differentiated at microsatellite loci both among and within the two areas. According to the AMOVA, 9.1% of the total genetic variance was found among countries, 22.5% of the variance was within countries among populations, and 68.4% was within populations. Estimates of FST were 0.28 across all populations, 0.23 in Norway, and 0.18 in Sweden. All FST estimates were statistically significant between pairs of populations except between N1 and N3, N5 and N6, N13 and N16, S2 and S3, and S10 and S13 (Fig. 1). Over all loci, the FST estimate was significantly lower in Sweden than in Norway (P ¼ 0.012). However, if Swedish populations were compared to groups of Norwegian populations from an area of similar size (Swedish populations vs. each of N1 to N4, N5 to N10 excluding N7, and N13 to N16), there was no significant difference between differentiation indices in Norway and Sweden (P ¼ 0.90, 0.77, and 0.70, respectively). Finally, genetic and geographic distances were positively correlated, both in the overall data set (P , 0.001), in Norway (P , 0.001, Fig. 2A), and in Sweden (P ¼ 0.019, Fig. 2B). In agreement with F statistics and Mantel tests, Bayesian

Fig. 2. Relationships between the log of interpopulation distance (km) and FST/(1  FST) among (A) Norwegian and (B) Swedish populations of Arabidopsis lyrata. The Mantel test was significant in both cases (P , 0.001 and P ¼ 0.004, respectively).

methods indicated substantial genetic divergence among populations and a certain amount of geographic structuring of genetic variation. In all 10 independent runs of the program BAPS, the existence of 17 clusters was associated with a 0.90 probability, while the partition in 18 clusters had a probability of 0.09. All runs inferred a unique, identical grouping of the populations in 17 clusters (Fig. 1). Using Structure, the loglikelihood of the data did not reach a maximal value for one given K, but rather progressively increased when K was increased (Fig. 3a). In such a situation, Pritchard and Wen (2003) suggest selecting the smallest K value that captures the major structure in the data (see also Rosenberg et al., 2001; Heuertz et al., 2004). Here, the likelihood of the data considerably increased from K ¼ 1 to K ¼ 2 and from K ¼ 2 to K ¼ 3, but this increase was minor with higher values of K (Fig. 3B). Moreover, separate runs converged on different clustering solutions when K . 5. Thus, graphical representations were drawn for K ¼ 5 (Fig. 1). In all 10 runs assuming K ¼ 2, Norwegian and Swedish populations were clearly separated (Fig. 4). The third cluster always differentiated northern Norwegian populations from southern Norwegian populations and from Swedish populations. When K ¼ 4, nine of 10 runs suggested a further division of populations in southern Norway into two groups. All populations could unambiguously be assigned to one cluster, although populations N9 and N10 showed some genetic admixture. Finally, when K ¼ 5, populations in Sweden divided into two groups (this division of Sweden was also suggested by one run with K ¼ 4). An AMOVA showed that

July 2007]

G AUDEUL

ET AL .—P OPULATION STRUCTURE IN

A RABIDOPSIS

1151

LYRATA

TABLE 3. Within-population diversity estimates of the 26 Scandinavian populations of Arabidopsis lyrata. FIS values significantly different from zero are indicated in boldface. For each population, expected heterozygosity (He), allelic richness and FIS were calculated with only one individual per genotype (Genets) and with all plants sampled including repeated multilocus genotypes (Ramets). Genets

Ramets

No. samples

No. distinct genotypes

Identical genotypes (%)

Polymorphic markers (%)

He

Allelic richness

FIS

He

Allelic richness

FIS

N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N13 N14 N15 N16 Mean 6 SD

23 20 20 32 20 20 22 20 20 20 20 31 20 23

18 18 16 22 19 20 17 20 19 19 17 17 17 16

21.7 10.0 20.0 31.3 5.0 0.0 22.7 0.0 5.0 5.0 15.0 45.2 15.0 30.4 16.2 6 13.3

63.2 63.2 63.2 63.2 57.9 57.9 57.9 57.9 73.7 63.2 78.9 63.2 52.6 73.7 63.6 6 7.0

0.249 0.237 0.255 0.225 0.279 0.270 0.187 0.214 0.294 0.261 0.357 0.226 0.189 0.284 0.252 6 0.045

1.96 1.91 2.03 1.89 2.02 1.90 1.71 1.78 2.32 2.07 2.62 1.76 1.82 2.29 2.01 6 0.24

0.067 0.039 0.052 0.090 0.050 0.092 0.228 0.136 0.032 0.077 0.093 0.096 0.257 0.100

0.246 0.234 0.252 0.224 0.276 0.270 0.191 0.214 0.298 0.266 0.341 0.216 0.179 0.274 0.249 6 0.043

1.94 1.89 2.01 1.83 2.01 1.90 1.67 1.78 2.31 2.07 2.57 1.69 1.79 2.24 1.98 6 0.24

0.027 0.013 0.085 0.042 0.046 0.092 0.267 0.136 0.028 0.076 0.078 0.201 0.229 0.078

S1 S2 S3 S6 S7 S8 S10 S13 S15 S16 S17 S19 Mean 6 SD

20 15 16 21 20 20 20 20 20 20 20 21

18 15 16 21 17 20 20 20 20 17 20 21

10.0 0.0 0.0 0.0 15.0 0.0 0.0 0.0 0.0 15.0 0.0 0.0 3.3 6 6.2

84.2 78.9 78.9 84.2 63.2 73.7 84.2 78.9 84.2 68.4 84.2 84.2 78.9 6 6.8

0.332 0.362 0.354 0.269 0.274 0.315 0.346 0.357 0.375 0.219 0.319 0.370 0.324 6 0.048

2.18 2.25 2.28 2.04 1.93 2.17 2.37 2.36 2.42 1.82 2.46 2.24 2.21 6 0.19

0.032 0.019 0.102 0.064 0.107 0.063 0.026 0.107 0.034 0.285 0.082 0.223

0.329 0.362 0.354 0.269 0.273 0.315 0.346 0.357 0.375 0.213 0.319 0.370 0.323 6 0.049

2.17 2.25 2.28 2.04 1.92 2.17 2.37 2.36 2.42 1.80 2.46 2.24 2.20 6 0.19

0.001 0.019 0.102 0.064 0.128 0.063 0.026 0.107 0.034 0.325 0.082 0.223

Population

14.7% of the total genetic variance was found among the five clusters. Effects of clonality on measures of genetic variation— Compared to the original analyses, the following differences were observed when redundant genotypes were included. (1) Estimates of both expected heterozygosity and allelic richness were 0.8% lower (He, t ¼ 2.3, P ¼ 0.03; allelic richness, t ¼ 3.9, P ¼ 0.001; Table 3). (2) Multilocus FIS was reduced by 0.014 (t ¼ 2.6, P ¼ 0.02) and became significantly negative in one additional Norwegian population (N14). The reduction in FIS observed when all plants were included in the analysis increased with the proportion of identical genotypes present in the population (regression coefficient, b ¼0.00182, N ¼ 26, P , 0.0001, R2 ¼ 0.64). (3) The number of populations with locus pairs in LD after Bonferroni correction increased from three to nine, with Pd values ranging from 1% to 23%. A fivefold increase in LD was observed in population N4 (from 1.5 to 7.6%) and a 15-fold increase in population N14 (from 1.5 to 23%). (4) The estimate of FST among Norwegian populations increased considerably, from 0.23 to 0.32, while the changes in FST estimates among Swedish populations (from 0.18 to 0.19) and across all populations (from 0.28 to 0.29) were small. DISCUSSION This study documented considerable variation at microsatellite loci within and among natural populations of Arabidopsis

lyrata subsp. petraea in Norway and Sweden, two disjunct areas in the northern part of its distribution. Indices of withinpopulation genetic diversity were lower in Norway than in Sweden, indicating a greater influence of genetic drift in the former area. Both multilocus Bayesian approaches and F statistics indicated stronger population differentiation across the Norwegian compared to the Swedish range, but this could apparently simply be explained by a difference in range size. Finally, the results indicate that the frequency of clonal propagation can be high, particularly in Norwegian populations. Genetic diversity within populations—Populations were less diverse in Norway than in Sweden (He ¼ 0.252 vs. 0.324 and allelic richness ¼ 2.0 vs. 2.2 alleles per locus, respectively). This suggests that genetic drift has influenced the structuring of genetic variation in A. lyrata more strongly in the Norwegian range and may reflect a difference in effective population size. Field estimates of current population sizes did not differ between the two areas, but a higher frequency of vegetative propagation should reduce effective population sizes in the Norwegian range (Chung and Kang, 1996; Jones and Gliddon, 1999). Recent bottlenecks could strongly influence present-day genetic structure. However, in most populations observed, heterozygosities did not deviate significantly from expectations under different mutation models, providing little evidence of recent bottlenecks. Fossil and genetic data indicate that, after the last glacial maximum, most temperate plant species colonized northern Europe from refugia in southern Europe (Hewitt, 2000, 2004).

1152

A MERICAN J OURNAL

OF

B OTANY

[Vol. 94

genetic clusters by Structure (three clusters were identified in Norway before Swedish populations were split into two clusters). In Norway, A. lyrata populations are found in a large number of different watersheds, whereas in Sweden all populations are located on the coast, which may facilitate gene flow among populations. Contrary to expectations, however, FST estimates did not differ between countries when areas of similar size were compared. The results suggest that the wider distribution area can explain why population differentiation is greater across the Norwegian than across the Swedish range (mean distance between sampled populations was 163 km in Norway vs. 40 km in Sweden) and that gene flow among neighboring A. lyrata populations is not necessarily greater in Sweden than in Norway. Although mountains may represent effective barriers to gene flow in lowland plant species (e.g., Heuertz et al., 2004), they need not strongly reduce gene flow in plants reaching into alpine areas. In a recent review, TillBottraud and Gaudeul (2002) found no difference in the magnitude of genetic differentiation among populations in mountain compared to lowland environments. More studies are clearly needed to determine how spatial variation in landscape structure affects patterns of gene flow and population structure (Manel et al., 2003).

Fig. 3. Structuring of variation at microsatellite loci in Norwegian and Swedish populations of Arabidopsis lyrata analyzed with the program Structure. (A) Relationship between the log-transformed probability of the data and the hypothesized number of clusters, K. (B) Difference between consecutive log-transformed probabilities of the data when K increases.

Based on its present ecological amplitude, A. lyrata could have colonized the previously glaciated regions of northern Europe from areas south of the ice sheet, but north of the Alps (Clauss and Mitchell-Olds, 2006). Consistent with this view, the 26 populations sampled in the present study had lower genetic diversity than eight central European populations in Clauss and Mitchell-Olds (2006); He ¼ 0.27 6 0.013 vs. He ¼ 0.52 6 0.024, estimates based on 12 microsatellite loci included in both studies; M. Clauss [Max Planck Institute of Chemical Ecology], ˚ gren unpublished data). M. Gaudeul, H. K. Stenøien, and J. A Among-population differentiation—Among-population genetic differentiation was strong, as shown by high FST values (0.28 across Norway and Sweden), numerous clusters inferred by BAPS, and continued increase of data likelihood when the number of clusters was increased in Structure. This pronounced genetic structure had a geographical pattern, with the main divergence between the Norwegian and Swedish ranges, followed by the distinction of three genetic clusters in Norway—clearly separating the northern and southern parts of Norway (N1 to N4 and N5 to N16)—and two genetic clusters in the northern and southern parts of Sweden (S1 to S7 and S8 to S19). Genetic distances were correlated with geographic distances among populations, which is consistent with the model of isolation by distance and with populations at migration–drift equilibrium. Among-population differentiation was higher in Norway than in Sweden, as reflected by F statistics (FST ¼ 0.23 vs. FST ¼ 0.18, respectively) and by the sequential disjunction of

Clonal propagation and genetic diversity—Identical multilocus genotypes were detected in many populations, and several observations suggest that they are the result of clonal propagation. First, because of their high polymorphism, microsatellite markers represent a powerful tool when it comes to detecting unique genotypes (Reusch et al., 1999; Van der Strate et al., 2002; Halkett et al., 2005). In the study populations, the number of polymorphic loci ranged from 10 to 16, and the number of alleles per polymorphic locus from two to 13. The probability that samples with identical multilocus genotypes represented separate genets with shared genotypes was in no case greater than 0.0032 and in most cases markedly lower (median 2.7 3 106). It thus seems highly unlikely that insufficient resolution of markers can explain the identical multilocus genotypes observed in several populations. Second, identical genotypes were found within the same populations in all cases but one. And third, in each population they were closest neighbors among sampled plants. The present survey suggests that sexual reproduction predominates in A. lyrata, but that vegetative propagation is substantial in some populations and more common in the Norwegian than in the Swedish range. Across all populations, the proportion of identical genotypes ranged from 0 to 0.45 (median 0.05; N ¼ 26). Considerable among-population variation in the frequency of clonal propagation has also been documented in other perennial herbs and may be related to population characteristics and environmental conditions (Silander, 1985; Richards, 1986; Eckert, 2002). In the present study, the proportion of identical genotypes was negatively associated with population size. This may be because vegetative propagation is negatively correlated with the number of potential mates and magnitude of sexual reproduction. Alternatively, it may simply reflect a higher probability of detecting clonal propagation with a fixed sample size as the population size decreases. In three Norwegian populations covering limited areas (N3, N7, and N14; range 8–80 m2, see Table 1), sampled plants were separated by relatively short distances, which probably made the detection of vegetative propagation more likely. However, substantial proportions of

July 2007]

G AUDEUL

ET AL .—P OPULATION STRUCTURE IN

A RABIDOPSIS

LYRATA

1153

Fig. 4. Population structure of Arabidopsis lyrata in Norway (populations N1–N16) and Sweden (populations S1–S19) estimated by the program Structure and graphically displayed with the program Distruct (Rosenberg, 2004). Each individual is represented by a thin, vertical line, which is partitioned into colored K segments that represent the individual’s estimated membership fractions in K clusters. Black lines separate individuals of different populations. Ten Structure runs at each K produced highly consistent solutions, and the results associated with the highest probability are shown. K, number of clusters.

identical genotypes were found also in populations occupying rather large areas (e.g., in population N4 covering 800 m2 and population N16 covering 300 m2), and no clonal propagation was detected in the Swedish population S6, in which closely located plants were sampled. Identical multilocus genotypes were more frequent in Norwegian than in Swedish populations. The proportion of identical genotypes ranged from 0 to 0.45 (median 0.15) in Norway and from 0 to 0.15 (median 0) in Sweden. This difference could not be explained by differences in population size, substrate, or altitude, and is apparently caused by some other factor influencing the frequency of clonal propagation in A. lyrata. As expected for a predominantly outcrossing species, most genetic diversity was found within populations of A. lyrata (68% of the total genetic variance), and expected heterozygosity was high. These findings agree with previous surveys of variation at allozyme and microsatellite loci in A. lyrata (Jonsell et al., 1995; Schierup, 1998; Van Treuren et al., 1997; Savolainen et al., 2000; Clauss et al., 2002; Ka¨rkka¨inen et al., 2004; Clauss and Mitchell-Olds, 2006). However, estimates of the within-population fixation index FIS were significantly positive in two of 26 populations. Positive FIS values were probably due to population substructure (Wahlund effect) and/ or to inbreeding. A detailed study of a German A. lyrata population revealed that small-scale spatial genetic structure among ca. 2 3 2 m large rock outcrops could explain a positive FIS at the population level (Clauss and Mitchell-Olds, 2006). Morevoer, biparental inbreeding and correlations of paternity have been documented in A. lyrata populations in Iceland (Schierup, 1998). Finally, partial selfing could potentially contribute to inbreeding in some populations. Controlled crosses indicated strong self-incompatibility and severe inbreeding depression in a Swedish A. lyrata population (Ka¨rkka¨inen et al., 1999). However, self-compatible populations have been identified in the North American subspecies of A. lyrata (Mable et al., 2005), and we cannot exclude the

possibility that some Scandinavian populations contain selfcompatible genotypes. Effects of clonal propagation on measures of genetic diversity—The inclusion of replicates of multilocus genotypes tended to decrease estimates of FIS, and increase estimates of linkage disequilibrium (LD) and population differentiation, and these effects were particularly strong for Norwegian populations. In one Norwegian population (N14), the proportion of locus pairs in LD was as high as 23% when redundant genotypes were included, which can be compared to 35% of polymorphic loci in LD in a genetically variable Norwegian population of the highly inbreeding A. thaliana (Stenøien et al., 2005). Clonality has also been observed to decrease FIS and increase linkage disequilibrium in the self-incompatible, partially clonal tree Prunus avium (Stoeckel et al., 2006). In Norway, the mean FST was 0.23 when redundant genotypes were excluded and 0.32 when redundant genotypes were included in the analysis. In Sweden, where the frequency of identical multilocus genotypes was markedly lower, the inclusion of redundant genotypes affected the estimate of FST only marginally. It is unknown to what extent clonality has influenced estimates of genetic differentiation on large geographical scales (Jonsell et al., 1995; Van Treuren et al., 1997; Schierup, 1998; Ka¨rkka¨inen et al., 2004), but the present results suggest that clonal propagation may contribute to high estimates of regional population differentiation in A. lyrata. Conclusions—This survey of genetic variation at microsatellite loci has revealed considerable population differentiation in the northern part of the A. lyrata distribution in Europe and provides a baseline for interpreting genetic differentiation at loci affecting characters of putative adaptive significance, such as flowering time (Rihima¨ki and Savolainen, 2004; Sandring et al., 2007) and trichome production (Ka¨rkka¨inen et al., 2004; Kivima¨ki et al., 2007; Løe et al., 2007). Furthermore, this

1154

A MERICAN J OURNAL

survey has demonstrated that vegetative propagation should be considered in ecological and genetic studies of A. lyrata. Nonrandom sampling may compromise assumptions in standard statistical tests and influence estimates of variability and structuring (McLellan et al., 1997; Montalvo et al., 1997; Reusch et al., 1999; Ivey and Richards, 2001). As assurance that sampled plants represent distinct genetic individuals, they should be either far apart or genotyped before analysis. LITERATURE CITED AUGE, H., B. NEUFFER, F. ERLINGHAGEN, R. GRUPPE, AND R. BRANDL. 2001. Demographic and random amplified polymorphic DNA analyses reveal high levels of genetic diversity in a clonal violet. Molecular Ecology 10: 1811–1819. BALLOUX, F., L. LEHMANN, AND T. DE MEEUˆS. 2003. The population genetics of clonal and partially clonal diploids. Genetics 164: 1635–1644. BELKHIR, K., P. BORSA, L. CHIKHI, N. RAUFASTE, AND F. BONHOMME. 1996. GENETIX 4.05, logiciel sous Windows TM pour la ge´ne´tique des populations. Laboratoire Ge´nome, Populations, Interactions, CNRS UMR 5171, Universite´ de Montpellier II, Montpellier, France. BENGTSSON, B. O. 2003. Genetic variation in organisms with sexual and asexual reproduction. Journal of Evolutionary Biology 16: 189–199. CHARLESWORTH, D., C. BARLOME´, M. H. SCHIERUP, AND B. MABLE. 2003. Haplotype structure of the stigmatic self-incompatibility gene in natural populations of Arabidopsis lyrata. Molecular Biology and Evolution 20: 1741–1753. CHUNG, M. G., AND S. S. KANG. 1996. Allozyme genetic and clonal diversity within population of Chimaphila japonica and Pyrola japonica (Pyrolaceae). Israel Journal of Plant Sciences 44: 259–271. CLAUSS, M., H. COBBAN, AND T. MITCHELL-OLDS. 2002. Cross-species microsatellite markers for elucidating population genetic structure in Arabidopsis and Arabis (Brassicaceae). Molecular Ecology 11: 591– 601. CLAUSS, M., AND T. MITCHELL-OLDS. 2006. Population genetic structure of Arabidopsis lyrata in Europe. Molecular Ecology 15: 2753–2766. COMES, H. P., AND J. W. KADEREIT. 1998. The effect of Quaternary climatic changes on plant distribution and evolution. Trends in Plant Science 3: 432–438. CORANDER, J., P. WALDMANN, AND M. J. SILLANPA¨A¨. 2003. Bayesian analysis of genetic differentiation between populations. Genetics 163: 367–374. CORNUET, J. M., AND G. LUIKART. 1996. Description and power analysis of genetic differentiation between populations. Genetics 163: 367–374. Software Bottleneck, version 1.2.02, available at http://www. montpellier.inra.fr/URLB/bottleneck/bottleneck.html. DECHAINE, E. G., AND A. P. MARTIN. 2005. Genetic divergence among sky island populations of Sedum lanceolatum (Crassulaceae) in the Rocky Mountains. American Journal of Botany 92: 477–486. DESPRES, L., S. LORIOT, AND M. GAUDEUL. 2002. Geographic pattern of genetic variation in the European globeflower Trollius europaeus L. (Ranunculaceae) inferred from AFLP markers. Molecular Ecology 11: 2337–2347. DEWOODY, J., J. D. NASON, AND M. SMITH. 2004. Inferring demographic processes from the genetic structure of a metapopulation of Boltonia decurrens (Asteraceae). Conservation Genetics 5: 603–617. ECKERT, C. G. 2002. The loss of sex in clonal plants. Evolutionary Ecology 15: 501–520. ECKERT, C. G., AND S. C. H. BARRETT. 1993. Clonal reproduction and patterns of genotypic diversity in Decodon verticillatus (Lythraceae). American Journal of Botany 80: 1175–1182. ECKERT, C. G., K. LUI, K. BRONSON, P. CORRADINI, AND A. BRUNEAU. 2003. Population genetic consequences of extreme variation in sexual and clonal reproduction in an aquatic plant. Molecular Ecology 12: 331–344. EDWARDS, A. L., AND R. R. SHARITZ. 2003. Clonal diversity in two rare perennial plants: Sagittaria isoetiformis and Sagittaria teres (Alismataceae). International Journal of Plant Sciences 164: 181–188.

OF

B OTANY

[Vol. 94

EL MOUSADIK, A., AND R. J. PETIT. 1996. Chloroplast DNA phylogeography of the argan tree of Morocco. Molecular Ecology 5: 547–555. ELLSTRAND, N. C., AND M. L. ROOSE. 1987. Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74: 123–131. ESTOUP, A., P. JARNE, AND J. M. CORNUET. 2002. Homoplasy and mutation model at microsatellite loci and their consequences for population genetics analysis. Molecular Ecology 11: 1591–604. EXCOFFIER, L., P. E. SMOUSE, AND J. M. QUATTRO. 1992. Analysis of molecular variance inferred from metric distance among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131: 479–491. FALUSH, D., M. STEPHENS, AND J. K. PRITCHARD. 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164: 1567–1587. GOUDET, J. 1995. FSTAT, version 1.2. A computer program to calculate Fstatistics. Journal of Heredity 86: 485–486. HALKETT, F., J. C. SIMON, AND F. BALLOUX. 2005. Tackling the population genetics of clonal and partially clonal organisms. Trends in Ecology and Evolution 20: 194–201. HAMRICK, J. L., AND M. J. W. GODT. 1989. Allozyme diversity in plant species. In A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding and genetic resources, 43–63. Sinauer, Sunderland, Massachusetts, USA. HAMRICK, J. L., AND M. J. W GODT. 1996. Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society in London, B, Biological Sciences 351: 1291–1298. HANGELBROEK, H. H., N. J. OUBORG, L. SANTAMARI´A, AND K. SCHWENK. 2002. Clonal diversity and structure within a population of the pondweed Potamogeton pectinatus foraged by Bewick’s swans. Molecular Ecology 11: 2137–2150. HANGELBROEK, H. H., L. SANTAMARI´A, AND T. DE BOER. 2003. Local adaptation of the pondweed Potamogeton pectinatus to contrasting substrate types mediated by changes in propagule provisioning. Journal of Ecology 91: 1081–1092. HASSEL, K., S. M. SA˚STAD, U. GUNNARSON, AND L. SO¨DERSTRO¨M. 2005.

Genetic variation and structure in the expanding moss Pogonatum dentatum (Polytrichaceae) in its area of origin and in a recently colonized area. American Journal of Botany 92: 1684–1690. HENEGARIU, O., N. A. HEERMA, S. R. DLOUGHY, G. H. VANCE, AND P. H. VOGT. 1997. Multiplex PCR: critical parameters and step-by-step protocol. BioTechniques 23: 504–511. HEUERTZ, M., J.-F. HAUSMAN, O. J. HARDY, G. G. VENDRAMIN, N. FRASCARIA-LACOSTE, AND X. VEKEMANS. 2004. Nuclear microsatellites reveal contrasting patterns of genetic structure between western and southeastern European populations of the common ash (Fraxinus excelsior L). Evolution 58: 976–988. HEWITT, G. M. 2000. The genetic legacy of the Quaternary ice ages. Nature 405: 907–913. HEWITT, G. M. 2004. Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 359: 183–195. HIRAO, A. S., AND G. KUDO. 2004. Landscape genetics of alpine-snowbed plants: comparisons along geographic and snowmelt gradients. Heredity 93: 290–298. HOLDEREGGER, R., I. STEHLIK, AND J. J. SCHNELLER. 1998. Estimation of the relative importance of sexual and vegetative reproduction in the clonal woodland herb Anemone nemorosa. Oecologia 117: 105–107. IVEY, C. T., AND J. H. RICHARDS. 2001. Genetic diversity of everglades sawgrass, Cladium jamaicense (Cyperaceae). International Journal of Plant Sciences 162: 817–825. JALAS, J., AND J. SUOMINEN. 1994. Atlas flora Europaea. Distribution of vascular plants in Europe, vol. 10. Cruciferae (Sisymbrium to Aubrieta). Helsinki University Printing House, Helsinki, Finland. JONES, B., AND C. GLIDDON. 1999. Reproductive biology and genetic structure of Lloydia serotina. Plant Ecology 141: 151–161. JONSELL, B., K. KUSTA˚S, AND I. NORDAL. 1995. Genetic variation in

Arabis petraea, a disjunct species in northern Europe. Ecography 18: 321–332.

July 2007]

G AUDEUL

ET AL .—P OPULATION STRUCTURE IN

KA¨RKKA¨INEN, K., H. KUITTINEN, R. VAN TREUREN, C. VOGL, S. OIKARINEN, AND O. SAVOLAINEN. 1999. Genetic basis of inbreeding depression in Arabis petraea. Evolution 53: 1354–1365. ˚ GREN. 2004. Population structure in KA¨RKKA¨INEN, K., G. LØE, AND J. A

Arabidopsis lyrata: evidence for divergent selection on trichome production. Evolution 58: 2831–2836. ˚ GREN. 2007. KIVIMA¨KI, M., K. KA¨RKKA¨INEN, M. GAUDEUL, G. LøE, AND J. A Gene, phenotype and function: GLABROUS1 and resistance to herbivory in natural populations of Arabidopsis lyrata. Molecular Ecology 16: 453–462. KUDOH, H., AND D. F. WHIGHAM. 1997. Microgeographic genetic structure and gene flow in Hibiscus moscheutos (Malvaceae) populations. American Journal of Botany 84: 1285–1293. ˚ GREN. 2007. Trichome LØE, G., P. TORA¨NG, M. GAUDEUL, AND J. A

production and spatiotemporal variation in herbivory in the perennial herb Arabidopsis lyrata. Oikos 116: 134–142. MABLE, B. K., A. V. ROBERTSON, S. DART, C. DI BERARDO, AND L. WITHAM. 2005. Breakdown of self-incompatibility in the perennial Arabidopsis lyrata (Brassicaceae) and its genetic consequences. Evolution 59: 1437–1448. MANEL, S., M. K. SCHWARTZ, G. LUIKART, AND P. TABERLET. 2003. Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology and Evolution 18: 189–197. MCLELLAN, A. J., D. PRATI, O. KALTZ, AND B. SCHMID. 1997. Structure and analysis of phenotypic and genetic variation in clonal plants. In H. de Kroon and J. M. Van Gronendal [eds.], The ecology and evolution of clonal plants, 185–210. Backhuys, Leiden, Netherlands. MONTALVO, A. M., S. G. CONARD, M. THOMPSON, AND P. D. HODGSKISS. 1997. Population structure, genetic diversity and clone formation in Quercus chrysolepis (Fagaceae). American Journal of Botany 84: 1553–1564. NEI, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York, New York, USA. O’KANE, S. L., AND I. A. AL-SHEHBAZ. 1997. A synopsis of Arabidopsis (Brassicaceae). Novon 7: 323–327. PRITCHARD, J. K., M. STEPHENS, AND P. DONNELLY. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945–959. PRITCHARD, J. K., AND W. WEN. 2003. Documentation for Structure software, version 2. Website http://pritch.bsd.uchicago.edu. REUSCH, T. B. H., W. HUKRIEDE, W. T. STAM, AND J. L. OLSEN. 1999. Differentiating between clonal growth and limited gene flow using spatial autocorrelation of microsatellites. Heredity 83: 120–126. RICHARDS, A. J. 1986. Plant breeding systems. George Allen & Unwin, London, UK. RIHIMA¨KI, M., AND O. SAVOLAINEN. 2004. Environmental and genetic effects on flowering differences between northern and southern populations of Arabidopsis lyrata (Brassicaceae). American Journal of Botany 91: 1036–1045. ROSENBERG, N. A. 2004. DISTRUCT: a program for the graphical display of population structure. Molecular Ecology Notes 4: 137–138. ROSENBERG, N. A., T. BURKE, K. ELO, M. W. FELDMAN, P. J. FREIDLIN, M. A. M. GROENEN, J. HILLEL, A. MA¨KI-TANILA, M. TIXIER-BOICHARD, A. VIGNAL, K. WIMMERS, AND S. WEIGEND. 2001. Empirical evaluation of genetic clustering methods using multilocus genotypes from 20 chicken breeds. Genetics 159: 699–713. ˚ GREN. 2007. SANDRING, S., M.-A. RIIHIMA¨KI, O. SAVOLAINEN, AND J. A Selection on flowering time and floral display in an alpine and a lowland population of Arabidopsis lyrata. Journal of Evolutionary Biology 20: 558–567. SAVOLAINEN, O., C. H. LANGLEY, B. P. LAZZARO, AND H. FRE´VILLE. 2000. Contrasting patterns of nucleotide polymorphism at the alcohol dehydrogenase locus in the outcrossing Arabidopsis lyrata and the selfing Arabidopsis thaliana. Molecular Biology and Evolution 17: 645–655. SCHIERUP, M. 1998. The effect of enzyme heterozygosity on growth in a

A RABIDOPSIS

LYRATA

1155

strictly outcrossing species, the self-incompatible Arabis petraea (Brassicaceae). Hereditas 128: 21–31. SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER. 2000. Arlequin: a software for population genetics data analysis, version 2.000. Genetics and Biometry Laboratory, Department of Anthropology, University of Geneva, Geneva, Switzerland. Computer program distributed by the authors at website http://cmpg.unibe.ch/software/arlequin/. SCHO¨NSWETTER, P., O. PAUN, A. TRIBSCH, AND H. NIKLFELD. 2003. Out of the Alps: colonization of northern Europe by east alpine populations of the glacier buttercup Ranunculus glacialis L. (Ranunculaceae). Molecular Ecology 12: 3373–3381. SILANDER, J. A. JR. 1985. Microevolution in clonal plants. In J. B. C. Jackson, L. W. Buss, and R. E. Cook [eds.], Population biology and evolution of clonal organisms, 107–152. Yale University Press, London, UK. STENØIEN, H. K., C. B. FENSTER, A. TONTERI, AND O. SAVOLAINEN. 2005 Genetic variability in natural populations of Arabidopsis thaliana from northern Europe. Molecular Ecology 14: 137–148. STENØIEN, H. K., AND S. M. SA˚STAD. 1999. Genetic structure in three

haploid peat mosses (Sphagnum). Heredity 82: 391–400. STOECKEL, S., J. GRANGE, J. F. FERNA´NDEZ-MANJARRES, I. BILGER, N. FRASCARIA-LACOSTE, AND S. MARIETTE. 2006. Heterozygote excess in a self-incompatible and partially clonal forest tree species—Prunus avium L. Molecular Ecology 15: 2109–2118. SUN, S., X. GAO, AND Y. CAI. 2001. Variations in sexual and asexual reproduction of Scirpus mariqueter along an elevational gradient. Ecological Research 16: 263–274. SYDES, M. A., AND R. PEAKALL. 1998. Extensive clonality in the endangered shrub Haloragodendron lucasii (Haloragaceae) revealed by allozymes and RAPDs. Molecular Ecology 7: 87–93. SYMONDS, V. V., AND A. M. LLOYD. 2003. An analysis of microsatellite loci in Arabidopsis thaliana: mutational dynamics and application. Genetics 165: 1475–1488. TABERLET, P., L. FUMAGALLI, A.-G. WUST-SAUCY, AND J.-F. COSSON. 1998. Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 7: 453–464. TILL-BOTTRAUD, I., AND M. GAUDEUL. 2002. Intraspecific genetic diversity in alpine plants. In C. Ko¨rner and E. Spehn [eds.], Mountain biodiversity: a global assessment, 23–34. Parthenon Publishing, London, UK. VAN DER STRATE, H. J., L. VAN DE ZANDE, W. T. STAM, AND J. L. OLSEN. 2002. The contribution of haploids, diploids and clones to fine-scale population structure in the seaweed Cladophoropsis membranaceae (Chorophyta). Molecular Ecology 11: 329–345. VAN TREUREN, R., H. KUITTINEN, K. KA¨RKKA¨INEN, E. BAENA-GONZALEZ, AND O. SAVOLAINEN. 1997. Evolution of microsatellites in Arabis petraea and Arabis lyrata, outcrossing relatives of Arabidopsis thaliana. Molecular Biology and Evolution 14: 220–229. WATKINSON, A. R., AND J. C. POWELL. 1993. Seedling recruitment and the maintenance of clonal diversity in plant populations: a computer simulation of Ranunculus repens. Journal of Ecology 81: 707–717. WEIR, B. S., AND C. C. COCKERHAM. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358–1370. WIDE´N, B., N. CRONBERG, AND M. WIDE´N. 1994. Genotypic diversity, molecular markers and spatial distribution of genets in clonal plants. In L. Soukupova´, C. Marshall, T. Hara, and T. Herben [eds.], Plant clonality: biology and diversity, 139–157. Opulus Press, Uppsala, Sweden. WILLIAMS, G. C. 1975. Sex and evolution, vol. 8. Princeton University Press, Princeton, USA. YOUNG, A., T. BOYLE, AND T. BROWN. 1996. The population genetic consequences of habitat fragmentation for plants. Trends in Ecology and Evolution 11: 413–418. YOUNG, A. G., J. H. HILL, B. G. MURRAY, AND R. PEAKALL. 2002. Breeding system, genetic diversity and clonal structure in the sub-alpine forb Rutidosis leiolepis F. Muell. (Asteraceae). Biological Conservation 106: 71–78.