Predictable allele frequency changes due to habitat fragmentation in the Glanville fritillary butterfly
Contributed by Ilkka Hanski, January 22, 2016 (sent for review June 15, 2015; reviewed by Andrew P. Hendry)
Significance
Understanding how species respond to environmental changes is essential for comprehending their ecological and evolutionary dynamics. Changes in land use have led to widespread loss and fragmentation of many habitats, resulting in the extinction of innumerable populations and species. Here, using historical DNA samples from now extinct populations that used to live in a highly fragmented landscape, we demonstrate past natural selection for better colonizing capacity. We also found highly reduced genetic diversity in the isolated populations prior to their extinction. These results demonstrate that, although species may show adaptation to rapidly changing landscapes, the evolutionary changes may not be strong enough to offset the direct adverse effect of environmental change and thereby to rescue populations from extinction.
Abstract
Describing the evolutionary dynamics of now extinct populations is challenging, as their genetic composition before extinction is generally unknown. The Glanville fritillary butterfly has a large extant metapopulation in the Åland Islands in Finland, but declined to extinction in the nearby fragmented southwestern (SW) Finnish archipelago in the 20th century. We genotyped museum samples for 222 SNPs across the genome, including SNPs from candidate genes and neutral regions. SW Finnish populations had significantly reduced genetic diversity before extinction, and their allele frequencies gradually diverged from those in contemporary Åland populations over 80 y. We identified 15 outlier loci among candidate SNPs, mostly related to flight, in which allele frequencies have changed more than the neutral expectation. At outlier loci, allele frequencies in SW Finland shifted in the same direction as newly established populations deviated from old local populations in contemporary Åland. Moreover, outlier allele frequencies in SW Finland resemble those in fragmented landscapes as opposed to continuous landscapes in the Baltic region. These results indicate selection for genotypes associated with good colonization capacity in the highly fragmented landscape before the extinction of the populations. Evolutionary response to habitat fragmentation may have enhanced the viability of the populations, but it did not save the species from regional extinction in the face of severe habitat loss and fragmentation. These results highlight a potentially common situation in changing environments: evolutionary changes are not strong enough to fully compensate for the direct adverse effects of environmental change and thereby rescue populations from extinction.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Changes in land use have greatly reduced the abundance and caused the extinction of innumerable populations and of many species (1–5). The remaining isolated remnant populations are expected (6, 7) and observed (8–10) to lose genetic variation via drift and inbreeding (with a few notable exceptions, e.g., ref. 11), which may further hasten their decline (12). Thus, gradual genetic deterioration along with demographic processes (13, 14) can contribute to the extinction debt in communities inhabiting fragmented landscapes. At the same time, increasing habitat fragmentation may strengthen selection on traits that reduce the risk of extinction of local populations and increase the rate of establishment of new populations to compensate for local extinctions (15). For example, species occupying recently fragmented habitats may evolve an increased (16) or decreased (17, 18) dispersal capacity, depending on factors such as the opportunity for successful colonization, the cost of dispersal, and the degree of temporal demographic variation (19). Although several studies have investigated the genetic composition of declining populations using neutral markers [e.g., Florida panthers (20) and greater prairie chickens (21)], there is a more limited understanding of the adaptive evolutionary dynamics of populations along the trajectory to extinction (although see ref. 22).
European butterflies have been greatly affected by land use changes, especially by changing agricultural practices and habitat fragmentation (23). The Glanville fritillary butterfly (Melitaea cinxia) has greatly declined in western and northern Europe. In Finland, the species went extinct from the mainland and southwestern (SW) archipelago in the 1970s (24), and it is now restricted to the Åland Islands (Fig. 1), where it has a large metapopulation in a network of 4,000 dry meadows (16, 25). Although the extinction risk of small local populations is increased by inbreeding (26), along with many demographic processes, the metapopulation as a whole has shown no declining trend for the past 25 y (25), although the amplitude of fluctuations in the metapopulation as a whole has increased, probably due to climate change (27). In contrast, an isolated population on a small island, Pikku Tytärsaari, in the northern Baltic, shows great loss of genetic diversity and reduction of fitness due to accumulation of genetic load (9).
Fig. 1.
There is both phenotypic and genetic evidence demonstrating that the Glanville fritillary has the capacity to adapt to fragmented habitat. Butterflies inhabiting fragmented landscapes, including the Åland Islands, have significantly higher flight metabolic rate than butterflies from continuous landscapes (28). Increased flight metabolic rate in turn enhances dispersal rate (29) and colonization rate (30) in fragmented landscapes, thereby facilitating long-term survival (31). In an RNA-seq study of gene expression, 1,841 genes were differentially expressed between fragmented and continuous landscapes. Genes that were systematically up-regulated following an experimental flight treatment had higher basal expression in fragmented than in continuous landscapes (32).
The above results raise questions about the populations that went extinct in SW Finland in the 20th century. Did they show reduced genetic variation that might have contributed to their extinction? How related were these populations to the large Åland metapopulation and was there increasing genetic divergence in the declining populations? Are there signs of selection in flight-related genes (or other genes) in the extinct populations that inhabited the highly fragmented landscape? Here, we genotype extinct and extant populations for 222 SNPs distributed across the genome of the Glanville fritillary (33) to answer these questions. We compare 34 putative neutral SNPs with SNPs from potential candidate genes to identify outliers and hence potential targets for selection. We use three other, independent datasets to test the hypothesis that selection for improved dispersal and colonization ability due to habitat fragmentation and high population turnover shifted outlier allele frequencies in the now extinct populations. First, we compare the magnitude and direction of allele frequency changes in the extinct (SW Finnish) populations with those that have occurred over 24 y in an introduced metapopulation, representing another example of a fragmented landscape. Second, using the difference in allele frequencies between newly established versus old local populations in the contemporary Åland metapopulation, we examine whether allele frequencies had shifted in the same direction as in the offspring of colonizing females. And third, by comparing allele frequencies in two metapopulations inhabiting highly fragmented landscapes versus two metapopulations inhabiting much more continuous landscapes, we test whether allele frequencies shifted toward frequencies characterizing individuals living in fragmented landscapes (population names and a summary of the analyses are given in Fig. 1). In brief, the first comparison tests how repeatable the observed allele frequency shifts are; the second focuses on the genetic consequences of recolonization and high population turnover; and the third examines the outcome of long-term selection.
Results
Genetic Diversity and Divergence of Museum Samples.
To investigate the divergence of museum populations from contemporary Åland populations, we aggregated museum samples from Åland and from the extinct populations in SW Finland into three temporal groups each, namely Åland1905 (1889–1921; n = 21), Åland1945 (1936–1949; n = 43), and Åland1965 (1957–1974; n = 10) and SW Finland1900 (1880–1920; n = 2), SW Finland1950 (1945–1954; n = 9), and SW Finland1960 (1962–1968; n = 8), respectively (Dataset S1). The yearly limits were selected to have temporally aggregated groups with a roughly equal sample size in each group for each population, although subsequent filtering for sample quality produced some variation in sample size. The Discriminant Analysis of Principal Components (DAPC) indicates that, although museum samples from Åland (n = 74) cluster with the contemporary Åland population (n = 39) sampled from the Saltvik region in Åland between 2007 and 2011 (Fig. 1), there is a systematic temporal change in the allele frequencies in the SW Finnish populations: the oldest samples are close to the Åland samples, whereas the more recent samples show increasing genetic divergence (Fig. 2 and SI Appendix, Table S1). These results suggest that the SW Finnish archipelago was colonized from Åland, most likely because people who traveled frequently between Åland and the archipelago close to the Finnish mainland accidentally translocated them. The reverse colonization is unlikely because the amount of suitable habitat has always been much more limited in the SW archipelago than in Åland, and one of the two host plants in Åland (Veronica spicata) has been very rare in SW Finland.
Fig. 2.
The extinct populations in SW Finland had much lower genetic diversity [standardized multilocus heterozygosity (sMLH) = 0.689] than Åland museum samples (sMLH = 0.993; Tukey test, P = 1.1e−11) and Saltvik (contemporary Åland) samples (sMLH = 1.068; Tukey test, P = 2.0e−14.). Heterozygosity of Saltvik samples was somewhat higher than heterozygosity of old museum Åland samples (Tukey test, P = 0.04), indicating that there has been no loss of genetic diversity in the Åland metapopulation as a whole in the past 100 y.
FST values calculated for the museum samples in relation to the contemporary samples were significantly greater in the SW Finnish populations than in Åland (SI Appendix, Fig. S1; linear model (LM): F = 65.9, P = 7.1e−14; analysis in SI Appendix, Table S2). The results indicate a decline in the FST values toward the present in the Åland samples, but not in the samples from SW Finland, consistent with the results in Fig. 2. In the Åland samples, the neutral markers showed significantly greater FST than the SNPs in the candidate genes (SI Appendix, Fig. S1; LM F = 15.8, P = 0.0001), suggesting that some of the latter markers have been affected by stabilizing selection.
Systematic Allele Frequency Changes in Candidate Genes.
We identified 10 outlier loci among the candidate genes (n = 164) in museum samples from Åland and 10 from SW Finnish populations (with all time periods pooled together), of which five were common to both datasets (Dataset S2). For the 34 neutral SNPs, the respective numbers were four and one (Dataset S3). Given that we tested 150 candidate SNPs, and assuming that the two datasets are independent, the expected number of outlier SNPs in both datasets is 0.6, conditional on the observed numbers in each dataset. The probability of having at least five shared outliers in the two sets is very low—4.8e−05 (from the hypergeometric distribution). However, after correcting for multiple testing, no SNP remained significant. Using BayeScan, an alternative method, Mc1:2673:141336 was identified as a marginally significant outlier after correcting for multiple testing [false discovery rate (FDR) = 0.05] (it was significant without correction in the first analysis). Although the 15 outliers were not significant after correcting for multiple testing, they nonetheless represent the most divergent loci among the set of candidate SNPs. Moreover, lack of significance does not invalidate the analyses below, where we test possible correlations in allele frequencies in the extinct SW Finnish populations and in three other independent datasets (the results of these comparisons for each outlier locus are summarized in SI Appendix, Table S3). We excluded two SNPs that are located in the sex chromosome (Mc1:1791:57299 and Mc1:3568:15483) and are hence affected by varying sex ratio in the small samples, and one marker (Mc1:4593:27912), which is known to be associated with host plant preference (34) and is therefore affected by the varying distribution of the two host plants across the study areas (no V. spicata in SW Finland).
The first analysis asks how repeatable the allele frequency changes are in response to habitat fragmentation. For this purpose, we compared the extinct SW Finnish populations with the 24-y-old introduced Sottunga metapopulation, which also inhabits a highly fragmented landscape, a small sparse network of 51 habitat patches. Allele frequencies in SW Finland and Sottunga have shifted mostly in the same direction from Saltvik frequencies (Fig. 3 and SI Appendix, Fig. S2; LM: R2 = 0.36, df = 11, P = 0.02). There is no comparable agreement in allele frequency changes in neutral markers (SI Appendix, Figs. S3 and S4; LM: R2 = −0.004, df = 19, P = 0.35). Importantly, allele frequency changes in the 24-y-old introduced Sottunga metapopulation have been significantly smaller than in the SW Finnish populations, which have had on average 80 y to diverge (Fig. 3; paired t test: t = 2.59, df = 11, P = 0.03). These results point to systematic changes in allele frequencies in the outlier loci in populations inhabiting fragmented landscapes, and hence the results suggest that nonrandom processes have influenced the changes.
Fig. 3.
Population Turnover and Allele Frequency Changes.
The Sottunga metapopulation consists of, and the now extinct SW Finnish populations have consisted of, small local populations with frequent population turnover. Repeated recolonizations may have selected for particular genotypes. For example, individuals sampled from newly established populations have a high frequency of a genotype associated with high dispersal capacity (16, 35, 36). To test this possibility, we calculated the difference in allele frequencies of outlier loci in newly colonized versus old local populations in contemporary Åland samples and correlated this difference with changes in allele frequencies in the samples from Sottunga and SW Finland in relation to the pooled sample from old large populations in the contemporary Åland metapopulation. Fig. 4 shows that allele frequencies have generally shifted in the same direction as newly established populations differ from old local populations (SW Finland: LM, R2 = 0.36, df = 11, P = 0.02; Sottunga: LM, R2 = 0.14, df = 11, P = 0.13). These results suggest that dispersal and recolonization processes that lead to the allele frequency difference between new and old local populations have influenced allele frequency changes in the introduced Sottunga metapopulation during its 24-y history and in the museum samples from the extinct SW Finnish populations between 1880 and 1968.
Fig. 4.
Allele Frequency Changes and Habitat Fragmentation.
The above results suggest that allele frequencies of outlier loci in Sottunga and SW Finland have shifted in the direction expected to occur in highly fragmented landscapes with frequent local extinctions and recolonizations. To test this hypothesis directly, we compared the allele frequency changes of outlier loci in Sottunga and SW Finland with allele frequencies in four regional populations in the Baltic region (data from ref. 32; Fig. 1 shows their locations). Two regional populations, Åland in Finland and Uppland in Sweden, inhabit highly fragmented landscapes, whereas the two others, Saaremaa in Estonia and Öland in Sweden, occur in landscapes with much more continuous habitat (28). Three outliers had to be excluded because their allele frequencies could not be extracted from the RNA-seq dataset for all four regional populations. In a principal component analysis of outlier allele frequencies for these four regional populations, the second principal component (PC2, R2 = 0.33) was strongly positively correlated with allele frequencies in fragmented populations and strongly negatively correlated with allele frequencies in the continuous populations (SI Appendix, Table S4). Large values of PC2 thus indicate a pattern of allele frequencies in fragmented landscapes. Allele frequencies in Sottunga and SW Finland have shifted in the direction of allele frequencies in fragmented landscapes (SI Appendix, Fig. S5; LM: R2 = 0.30, df = 8, P = 0.07).
Discussion
To summarize the findings of this study, allele frequencies at 12 outlier loci, which have been under putative positive selection or linked to genes under selection, have shifted in the same direction in two independent datasets, a 24-y-old small introduced metapopulation on the island of Sottunga and museum material from now extinct populations in SW Finland during the 20th century. Both landscapes are highly fragmented, and we suggest that nonrandom processes related to rapid population turnover in highly fragmented landscapes have affected these allele frequency changes. The landscape in SW Finland is an archipelago, a naturally highly fragmented landscape, and hence the sparse and small butterfly populations in the 20th century likely exhibited much turnover before the eventual regional extinction. For Sottunga, which was introduced as an experiment 24 y ago, we have a complete record of demographic dynamics in the network of 51 small meadows. All local populations have gone extinct at least once during this period, and hence the very persistence of the metapopulation is definitely due to frequent recolonizations compensating for local extinctions; the current butterflies are the offspring of recent colonizers. The hypothesis that dispersal and recolonizations have affected the allele frequency changes is further supported by the result that these changes are consistent with allele frequency differences between newly established versus old local populations in the contemporary metapopulation in Åland, which reflects the immediate effect of recolonizations over one generation. Moreover, the allele frequencies in the introduced Sottunga metapopulation and the extinct SW Finnish populations have shifted in the direction observed in highly fragmented landscapes elsewhere in northern Europe. In brief, the likely mechanism leading to these results is high population turnover in Sottunga and SW Finland having selected genotypes that enhance dispersal and recolonization, an adaptive genetic response to habitat fragmentation.
The 15 outlier loci (12 loci after filtering) were not significant following the correction for multiple testing. This is not surprising because our sample size is relatively small and hence the power of the statistical test is limited. However, we emphasize the significant and consistent patterns obtained while comparing the primary dataset from SW Finland with not just one but three other, completely independent datasets. These comparisons make the interpretation of the results much stronger than what would have been possible for significant outliers in a single outlier analysis.
Previous studies on the Glanville fritillary (16, 35, 36) and other species (37, 38) have demonstrated that colonizations select for more dispersive individuals than the average individual in the metapopulation. In the Glanville fritillary, female butterflies from newly established local populations (39, 40) and from highly fragmented landscapes (28) have a higher flight metabolic rate than those from old local populations and from continuous landscapes. Flight metabolic rate is positively correlated with dispersal rate in the field (29). The majority of the outlier loci in the present study have been implicated as being related to flight in previous studies, although the annotations are not informative (Dataset S2).
Selection for more dispersive butterflies in the declining populations in SW Finland in the 20th century may have made the metapopulation more likely to persist, although the species went extinct from this entire region in the 1970s, which has been attributed to very extensive habitat loss and fragmentation due to agricultural intensification (24). In contrast, the Åland metapopulation has survived because habitat loss and fragmentation have been limited due to topographical and cultural reasons (41). As we have seen in this study, there has been no measurable reduction in heterozygosity in Åland over the past century. In the SW Finnish archipelago, habitat loss and fragmentation were so extensive that no evolutionary response in dispersal rate could compensate for it, which is likely to be a common situation in changing environments (42). It is also possible that any beneficial effects from genes linked with the outlier loci were swamped by accumulation of deleterious mutations (6, 43, 44) and general loss of heterozygosity and fitness (45). In the Glanville fritillary, flight metabolic rate is significantly and positively correlated with genome-wide heterozygosity, a proxy of individual inbreeding (46), with a 30% fitness reduction in a small (Ne ∼100) population that has remained completely isolated for at least 75 y (9). As another example, habitat fragmentation due to climate-induced range contraction has eroded genetic diversity in an alpine mammal species (47), likely reducing the adaptive potential of the species and increasing the risk of inbreeding (48).
Many previous studies have identified likely adaptive genetic responses to environmental change by studying species along environmental gradients (e.g., refs. 49–51). A complication with such studies is possibly the dissimilar evolutionary and demographic histories of populations along the gradient. A more direct approach is to monitor changes in molecular markers (e.g., refs. 52–54) or quantitative traits (e.g., refs. 55–58) across time, although in most cases the time span that has been covered is relatively short. For example, a study of Drosophila melanogaster showed significant shifts in allele frequencies along a latitudinal cline over a 20-y period, consistent with increasing temperatures (52). Hansen et al. (42) proposed six criteria that are needed to demonstrate adaptive genetic responses to environmental change. The present study has fulfilled all these criteria: (i) Suitable genetic variation for selection to act upon, which is evidently the case here. (ii) This genetic variation is linked to a specific environmental stress; our SNP panel was selected to include especially markers associated with either flight or habitat fragmentation (32). (iii) Any identified associated alleles vary across generations, which is the case here. Moreover, we have shown that allele frequency changes have become greater with increasing length of time (Fig. 3). (iv) Allele frequency changes reflect selection, which was demonstrated by comparing allele frequency changes in candidate loci with changes in neutral markers. (v) The observed allele frequency changes are attributable to the identified environmental stress, which in our case is the process of population turnover in highly fragmented landscapes. Finally, (vi) any changes in allele frequency cannot be attributed to a new well-colonizing population simply replacing the original population, which is most directly known for the introduced metapopulation on the island of Sottunga, continuously monitored since its establishment in 1991.
In conclusion, to our knowledge, this study is the first to demonstrate predictable changes in allele frequency in response to habitat fragmentation. This study also highlights a likely common situation in changing environments, where adaptive genetic responses may not be strong enough to rescue populations from extinction due to adverse effects of rapid environmental change.
Materials and Methods
Museum Samples.
We searched museum collections for specimens collected across the entire historical range of the species in Finland. Altogether, 119 specimens were located, originating from 11 regions in the Åland Islands and 3 regions in SW Finland (see Fig. 1 for a map). Specimens were aggregated based on spatiotemporal information (Dataset S1). As a contemporary reference, we used samples from the Saltvik region in the main Åland Island (Fig. 1), sampled in 2007–2011. For ordination and calculation of FST values and heterozygosity, we used 40 randomly selected individuals from the total of 4,547 individuals from Saltvik to have a comparable dataset with the other datasets. A similar sample of 40 randomly selected individuals was obtained from the island of Sottunga (a 24-y introduced metapopulation; Fig. 1) out of the total of 319 individuals sampled in 2007–2011. For details of DNA extraction and validation of museum sample genotyping, see SI Appendix.
SNP Selection and Genotyping.
The genomic resources for the Glanville fritillary include the published genome (33), a high-density linkage map (59), and two RNA-seq datasets (32, 46), which were used to create a SNP panel using the KASP genotyping platform (LGC Genomics) for large-scale genotyping (Dataset S3). Markers were selected on the basis of previous studies (32, 34, 46, 60), in which candidate genes were associated with flight and dispersal traits or were significantly differentially expressed after an experimental flight treatment. Putatively neutral SNPs were selected from the noncoding regions of the genome (SI Appendix). The remainder of the panel was selected to cover all of the chromosomes (“gap-filling SNP”) based on linkage map information (59) (for details of SNP calling, filtering, and validation, see SI Appendix). Following the validation process, 320 SNPs were selected as the final genotyping panel implementing KASP chemistry (LGC Genomics). Museum samples were genotyped for 320 SNPs, whereas contemporary Åland samples were genotyped for a subset of 272 SNPs. Linkage disequilibrium between markers was assessed using GENEPOP v 4.3 (61) with FDR (62) used to correct for multiple testing. At FDR < 0.05, 11 pairs of SNPs showed significant linkage disequilibrium in the Åland and one pair in SW Finnish museum samples (SI Appendix, Table S5). SNP call rates were generally high, >0.97 in contemporary samples and >0.9 in the museum samples. Any SNP with an average call rate <0.7 and any individual with an average call rate <0.8 were excluded from the analysis to eliminate unreliable genotypes (SI Appendix). This filtering resulted in 93 museum specimens and 76 contemporary Åland specimens (Dataset S1; 39 from Saltvik and 37 from Sottunga), genotyped for 222 shared SNPs (Dataset S3).
Genetic Clustering and Diversity.
A DAPC (63) using the ADEGENET 2.0 (64) package in R Version 3.2.1 (65) was performed to examine possible genetic clustering in the material. To avoid overfitting, 55 principal components were retained, accounting for ∼70% of the total variance. We calculated the FST value (66) between each museum specimen and the contemporary Saltvik population, sampled in 2007 (n = 530), to estimate the evolutionary distance of each specimen to the reference population (SI Appendix). In addition, we calculated FST between pairs of aggregated samples using ARLEQUIN v 3.5.1.3 (67) and 10,000 permutations to calculate significance (SI Appendix, Table S1). We calculated sMLH as a surrogate for individual inbreeding (68) using only autosomal markers (n = 211). sMLH is calculated for each individual as the total number of heterozygous loci divided by the sum of the observed heterozygosities (across all samples) for the successfully genotyped loci in the focal individual. This takes into account that not all individuals were successfully genotyped for the same set of loci.
Outlier Analysis.
To identify outlier loci, we used two different methods using the pooled museum samples from Åland and SW Finland and the Saltvik sample. First, we estimated locus-specific FST values for SNPs in the candidate genes between the museum samples and the contemporary Saltvik 2007 reference sample (SI Appendix). These values were compared with the posterior distribution of FST values for the neutral SNPs, and we computed the posterior probability that the FST was higher for the candidate SNP than for the neutral SNPs. A locus was called an outlier if the posterior probability was higher than 0.95. We used FDR to correct for multiple testing. Second, we tested all 222 SNPs for outliers using BayeScan v 2.1 (69), a Bayesian method that calculates the posterior probability of each locus belonging to a model assuming either selection or no selection, using default parameters.
Allele Frequency Changes in the Outlier Loci.
We compared the allele frequencies in the outlier loci between old, well-connected local populations and newly established, isolated populations in the contemporary Åland metapopulation (sampled in 2007–2012 in the Saltvik region, Fig. 1). Old populations were defined as those that had existed for at least three years before sampling, whereas newly established populations were those that had been established by dispersing females in the previous year (the habitat patch had been unoccupied in the year before colonization). Populations were split into the two categories of high or low connectivity (70). The old, well-connected populations consisted of 24 local populations with 903 individuals, and the newly established, isolated populations included 55 populations with 288 individuals. To characterize allele frequency (AF) change due to dispersal and recolonization, we calculated the difference AF(new)-AF(old). The difference AF(new)-AF(old) is weakly correlated with AF(old) (SI Appendix, Fig. S6), but using corrected values gave qualitatively the same results (SI Appendix, Table S6), and we hence retained the uncorrected values for the main analyses. We calculated in the same way the difference in allele frequencies between samples from Sottunga (366 individuals sampled in 2007–2012) and the AF(old) and between the extinct populations in SW Finland and AF(old). Finally, we characterized the allele frequencies of the outlier loci in four regional populations in the northern Baltic region (Fig. 1), two of which have a highly fragmented landscape and two of which have continuous habitat (28, 32). Allele frequencies were obtained from ref. 32. To characterize the allele frequencies in the four regional populations, we ran a principal component analysis (SI Appendix). The second principal component was strongly correlated with the degree of habitat fragmentation (SI Appendix, Table S4). We correlated PC2 with allele frequency change from the contemporary Åland to the introduced Sottunga metapopulation and the extinct SW Finnish populations using PC1 for the latter two samples. PC1 characterizes the average frequencies in the two samples (see SI Appendix for separate analyses for the two populations). We used corrected values for AF(Sottunga)-AF(old) and AF(SW Finland)-AF(old) to remove the effect of (nonsignificant) correlation with AF(old) (see above and SI Appendix) because of very small sample size.
Acknowledgments
We thank all those who contributed museum samples; Alison Ollikainen, Rainer Lehtonen, Annukka Ruokolainen, and Toshka Nyman for sample preparation and DNA extraction; Panu Somervuo for providing allele frequency information from the RNA-seq data; and Arild Husby, Patrik Nosil, and Craig Primmer for comments on the manuscript. This study was funded by the European Research Council (Grant AdG232826 to I.H.) and the Academy of Finland (Grants 133132, 250444, and 256453 to I.H.).
Supporting Information
Appendix (PDF)
Supporting Information
- Download
- 835.15 KB
pnas.1600951113.sd01.xls
- Download
- 53.50 KB
pnas.1600951113.sd02.xlsx
- Download
- 51.90 KB
pnas.1600951113.sd03.xlsx
- Download
- 37.91 KB
References
1
PM Vitousek, HA Mooney, J Lubchenco, JM Melillo, Human domination of Earth’s ecosystems. Science 277, 494–499 (1997).
2
PF Donald, FJ Sanderson, IJ Burfield, FPJ van Bommel, Further evidence of continent-wide impacts of agricultural intensification on European farmland birds, 1990–2000. Agric Ecosyst Environ 116, 189–196 (2006).
3
SA Cushman, Effects of habitat loss and fragmentation on amphibians: A review and prospectus. Biol Conserv 128, 231–240 (2006).
4
I Hanski The Shrinking World (International Ecology Institute, Oldendorf/Luhe, Germany, 2005).
5
JB Hughes, GC Daily, PR Ehrlich, Population diversity: Its extent and extinction. Science 278, 689–692 (1997).
6
M Lynch, J Conery, R Burger, Mutation accumulation and the extinction of small populations. Am Nat 146, 489–518 (1995).
7
MC Whitlock, Fixation probability and time in subdivided populations. Genetics 164, 767–779 (2003).
8
E Luquet, et al., Heterozygosity-fitness correlations among wild populations of European tree frog (Hyla arborea) detect fixation load. Mol Ecol 20, 1877–1887 (2011).
9
ALK Mattila, et al., High genetic load in an old isolated butterfly population. Proc Natl Acad Sci USA 109, E2496–E2505 (2012).
10
G Athrey, KR Barr, RF Lance, PL Leberg, Birds in space and time: Genetic changes accompanying anthropogenic habitat fragmentation in the endangered black-capped vireo (Vireo atricapilla). Evol Appl 5, 540–552 (2012).
11
R Kaeuffer, DW Coltman, JL Chapuis, D Pontier, D Réale, Unexpected heterozygosity in an island mouflon population founded by a single pair of individuals. Proc Biol Sci 274, 527–533 (2007).
12
R Frankham, Genetics and extinction. Biol Conserv 126, 131–140 (2005).
13
R Lande, Risks of population extinction from demographic and environmental stochasticity and random catastrophes. Am Nat 142, 911–927 (1993).
14
I Hanski, O Ovaskainen, The metapopulation capacity of a fragmented landscape. Nature 404, 755–758 (2000).
15
O Ronce, I Olivieri, Life-history evolution in metapopulation. Ecology, Genetics and Evolution of Metapopulations, eds I Hanski, OE Gaggiotti (Elsevier Academic Press, Amsterdam), pp. 227–257 (2004).
16
IA Hanski, Eco-evolutionary spatial dynamics in the Glanville fritillary butterfly. Proc Natl Acad Sci USA 108, 14397–14404 (2011).
17
J Travis, C Dytham, Habitat persistence, habitat availability and the evolution of dispersal. Proc Biol Sci 266, 723–728 (1999).
18
PO Cheptou, O Carrue, S Rouifed, A Cantarel, Rapid evolution of seed dispersal in an urban environment in the weed Crepis sancta. Proc Natl Acad Sci USA 105, 3796–3799 (2008).
19
I Hanski, T Mononen, Eco-evolutionary dynamics of dispersal in spatially heterogeneous environments. Ecol Lett 14, 1025–1034 (2011).
20
WE Johnson, et al., Genetic restoration of the Florida panther. Science 329, 1641–1645 (2010).
21
JA Johnson, MR Bellinger, JE Toepfer, P Dunn, Temporal changes in allele frequencies and low effective population size in greater prairie-chickens. Mol Ecol 13, 2617–2630 (2004).
22
K Petren, PR Grant, BR Grant, AA Clack, NV Lescano, Multilocus genotypes from Charles Darwin’s finches: Biodiversity lost since the voyage of the Beagle. Philos Trans R Soc Lond B Biol Sci 365, 1009–1018 (2010).
23
C van Swaay, M Warren, G Loïs, Biotope use and trends of European butterflies. J Insect Conserv 10, 189–209 (2006).
24
M Kuussaari, M Nieminen, J Pöyry, I Hanski, Täpläverkkoperhosen (Melitaea cinxia) elinkierto ja esiintyminen Suomessa [The life cycle and occurrence of the Glanville fritillary butterfly (Melitaea cinxia) in Finland]. Baptria 20, 167–180 (1995).
25
SP Ojanen, M Nieminen, E Meyke, J Pöyry, I Hanski, Long-term metapopulation study of the Glanville fritillary butterfly (Melitaea cinxia): Survey methods, data management, and long-term population trends. Ecol Evol 3, 3713–3737 (2013).
26
I Saccheri, et al., Inbreeding and extinction in a butterfly metapopulation. Nature 392, 491–494 (1998).
27
AJM Tack, T Mononen, I Hanski, Increasing frequency of low summer precipitation synchronizes dynamics and compromises metapopulation stability in the Glanville fritillary butterfly. Proc Biol Sci 282, 20150173 (2015).
28
A Duplouy, S Ikonen, I Hanski, Life history of the Glanville fritillary butterfly in fragmented versus continuous landscapes. Ecol Evol 3, 5141–5156 (2013).
29
K Niitepõld, et al., Flight metabolic rate and Pgi genotype influence butterfly dispersal rate in the field. Ecology 90, 2223–2232 (2009).
30
CW Wheat, CR Haag, JH Marden, I Hanski, MJ Frilander, Nucleotide polymorphism at a gene (Pgi) under balancing selection in a butterfly metapopulation. Mol Biol Evol 27, 267–281 (2010).
31
C Zheng, O Ovaskainen, I Hanski, Modelling single nucleotide effects in phosphoglucose isomerase on dispersal in the Glanville fritillary butterfly: Coupling of ecological and evolutionary dynamics. Philos Trans R Soc Lond B Biol Sci 364, 1519–1532 (2009).
32
P Somervuo, et al., Transcriptome analysis reveals signature of adaptation to landscape fragmentation. PLoS One 9, e101467 (2014).
33
V Ahola, et al., The Glanville fritillary genome retains an ancient karyotype and reveals selective chromosomal fusions in Lepidoptera. Nat Commun 5, 4737 (2014).
34
MA de Jong, SC Wong, R Lehtonen, I Hanski, Cytochrome P450 gene CYP337 and heritability of fitness traits in the Glanville fritillary butterfly. Mol Ecol 23, 1994–2005 (2014).
35
I Hanski, C Erälahti, M Kankare, O Ovaskainen, H Sirén, Variation in migration propensity among individuals maintained by landscape structure. Ecol Lett 7, 958–966 (2004).
36
M Saastamoinen, Mobility and lifetime fecundity in new versus old populations of the Glanville fritillary butterfly. Oecologia 153, 569–578 (2007).
37
R Shine, Invasive species as drivers of evolutionary change: Cane toads in tropical Australia. Evol Appl 5, 107–116 (2012).
38
JK Hill, HM Griffiths, CD Thomas, Climate change and evolutionary adaptations at species’ range margins. Annu Rev Entomol 56, 143–159 (2011).
39
CR Haag, M Saastamoinen, JH Marden, I Hanski, A candidate locus for variation in dispersal rate in a butterfly metapopulation. Proc Biol Sci 272, 2449–2456 (2005).
40
CW Wheat, et al., Functional genomics of life history variation in a butterfly metapopulation. Mol Ecol 20, 1813–1828 (2011).
41
M Nieminen, M Siljander, I Hanski, Structure and dynamics of Melitaea cinxia metapopulations. On the Wings of Checkerspots: A Model System for Population Biology, eds PR Ehrlich, I Hanski (Oxford University Press, Oxford), pp. 63–91 (2004).
42
MM Hansen, I Olivieri, DM Waller, EE Nielsen, Monitoring adaptive genetic responses to environmental change. Mol Ecol; GeM Working Group 21, 1311–1329 (2012).
43
R Lande, Risk of population extinction from fixation of new deleterious mutations. Evolution 48, 1460–1469 (1994).
44
JJ Falk, CE Parent, D Agashe, DI Bolnick, Drift and selection entwined: Asymmetric reproductive isolation in an experimental niche shift. Evol Ecol Res 14, 403–423 (2012).
45
D Charlesworth, JH Willis, The genetics of inbreeding depression. Nat Rev Genet 10, 783–796 (2009).
46
J Kvist, et al., Flight-induced changes in gene expression in the Glanville fritillary butterfly. Mol Ecol 24, 4886–4900 (2015).
47
EM Rubidge, et al., Climate-induced range contraction drives genetic erosion in an alpine mammal. Nat Clim Chang 2, 285–288 (2012).
48
R Bijlsma, V Loeschcke, Genetic erosion impedes adaptive responses to stressful environments. Evol Appl 5, 117–129 (2012).
49
SG Byars, W Papst, AA Hoffmann, Local adaptation and cogradient selection in the alpine plant, Poa hiemata, along a narrow altitudinal gradient. Evolution 61, 2925–2941 (2007).
50
LF Jensen, et al., Local adaptation in brown trout early life-history traits: Implications for climate change adaptability. Proc Biol Sci 275, 2859–2868 (2008).
51
AJM Tack, F Horns, AL Laine, The impact of spatial scale and habitat configuration on patterns of trait variation and local adaptation in a wild plant parasite. Evolution 68, 176–189 (2014).
52
PA Umina, AR Weeks, MR Kearney, SW McKechnie, AA Hoffmann, A rapid shift in a classic clinal pattern in Drosophila reflecting climate change. Science 308, 691–693 (2005).
53
N Charbonnel, J Pemberton, A long-term genetic survey of an ungulate population reveals balancing selection acting on MHC through spatial and temporal fluctuations in selection. Heredity (Edinb) 95, 377–388 (2005).
54
J Balanyá, JM Oller, RB Huey, GW Gilchrist, L Serra, Global genetic change tracks global climate warming in Drosophila subobscura. Science 313, 1773–1775 (2006).
55
PR Grant, BR Grant, Unpredictable evolution in a 30-year study of Darwin’s finches. Science 296, 707–711 (2002).
56
JE Brommer, J Merilä, BC Sheldon, L Gustafsson, Natural selection and genetic variation for reproductive reaction norms in a wild bird population. Evolution 59, 1362–1371 (2005).
57
AP Hendry, et al., Possible human impacts on adaptive radiation: Beak size bimodality in Darwin’s finches. Proc Biol Sci 273, 1887–1894 (2006).
58
A LE Rouzic, et al., Strong and consistent natural selection associated with armour reduction in sticklebacks. Mol Ecol 20, 2483–2493 (2011).
59
P Rastas, L Paulin, I Hanski, R Lehtonen, P Auvinen, Lep-MAP: Fast and accurate linkage map construction for large SNP datasets. Bioinformatics 29, 3128–3134 (2013).
60
M Saastamoinen, S Ikonen, SC Wong, R Lehtonen, I Hanski, Plastic larval development in a butterfly has complex environmental and genetic causes and consequences for population dynamics. J Anim Ecol 82, 529–539 (2013).
61
F Rousset, genepop’007: A complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour 8, 103–106 (2008).
62
Y Benjamini, Y Hochberg, Controlling the false discovery rate: A practical and powerful approach to multiple testing. J Roy Stat Soc B Met 57, 289–300 (1995).
63
T Jombart, S Devillard, F Balloux, Discriminant analysis of principal components: A new method for the analysis of genetically structured populations. BMC Genet 11, 94 (2010).
64
T Jombart, adegenet: A R package for the multivariate analysis of genetic markers. Bioinformatics 24, 1403–1405 (2008).
65
; R Core Team, R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. Available at www.R-project.org. (2014).
66
OE Gaggiotti, M Foll, Quantifying population structure using the F-model. Mol Ecol Resour 10, 821–830 (2010).
67
L Excoffier, HEL Lischer, Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10, 564–567 (2010).
68
JI Hoffman, et al., High-throughput sequencing reveals inbreeding depression in a natural population. Proc Natl Acad Sci USA 111, 3775–3780 (2014).
69
M Foll, O Gaggiotti, A genome-scan method to identify selected loci appropriate for both dominant and codominant markers: A Bayesian perspective. Genetics 180, 977–993 (2008).
70
I Hanski, M Saastamoinen, O Ovaskainen, Dispersal-related life-history trade-offs in a butterfly metapopulation. J Anim Ecol 75, 91–100 (2006).
Information & Authors
Information
Published in
Classifications
Copyright
Freely available online through the PNAS open access option.
Submission history
Published online: February 22, 2016
Published in issue: March 8, 2016
Keywords
Acknowledgments
We thank all those who contributed museum samples; Alison Ollikainen, Rainer Lehtonen, Annukka Ruokolainen, and Toshka Nyman for sample preparation and DNA extraction; Panu Somervuo for providing allele frequency information from the RNA-seq data; and Arild Husby, Patrik Nosil, and Craig Primmer for comments on the manuscript. This study was funded by the European Research Council (Grant AdG232826 to I.H.) and the Academy of Finland (Grants 133132, 250444, and 256453 to I.H.).
Authors
Competing Interests
The authors declare no conflict of interest.
Metrics & Citations
Metrics
Citation statements
Altmetrics
Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFGet Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to get full access to it.