Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Plant-Species Diversity Correlates with Genetic Variation of an Oligophagous Seed Predator

Plant-Species Diversity Correlates with Genetic Variation of an Oligophagous Seed Predator

  • Liisa Laukkanen, 
  • Pia Mutikainen, 
  • Anne Muola, 
  • Roosa Leimu
PLOS
x

Abstract

Several characteristics of habitats of herbivores and their food-plant communities, such as plant-species composition and plant quality, influence population genetics of both herbivores and their host plants. We investigated how different ecological and geographic factors affect genetic variation in and differentiation of 23 populations of the oligophagous seed predator Lygaeus equestris (Heteroptera) in southwestern Finland and in eastern Sweden. We tested whether genetic differentiation of the L. equestris populations was related to the similarity of vegetation, and whether there was more within-population genetic variation in habitats with a high number of plant species or in those with a large population of the primary food plant, Vincetoxicum hirundinaria. We also tested whether genetic differentiation of the populations was related to the geographic distance, and whether location of the populations on islands or on mainland, island size, or population size affected within-population genetic variation. Pairwise FST ranged from 0 to 0.1 indicating low to moderate genetic differentiation of populations. Differentiation increased with geographic distance between the populations, but was not related to the similarity of vegetation between the habitats. Genetic variation within the L. equestris populations did not increase with the population size of the primary food plant. However, the more diverse the plant community the higher was the level of genetic variation within the L. equestris population. Furthermore, the level of genetic variation did not vary significantly between island and mainland populations. The effect of the population size on within-population genetic variation was related to island size. Usually small populations are susceptible to loss of genetic variation, but small L. equestris populations on large islands seemed to maintain a relatively high level of within-population genetic variation. Our findings suggest that, in addition to geographic and species-specific ecological factors, the plant community affects population genetic structure of oligophagous herbivores.

Introduction

Herbivores have to adapt to geographically and temporally varying communities of their host plants as well as to genetically determined plant defenses and the nutritive content of their food-plant species. Plant-species composition and plant quality exert selection on herbivores and may affect their population genetics (e.g. [1]). On the other hand, the level of genetic variation in a population may restrict the ability of an herbivore to use certain plant individuals, populations, or species as food (e.g. [2][5]). Distinctive selective pressures have promoted some herbivore species to evolve to be dietary generalists and others specialists [6]. Generalists have been observed to harbour higher within-population genetic variation than specialists. The level of genetic differentiation between populations may also differ between specialists and generalists [7][9].

The occurrence and nutritive quality of a single food-plant species determine the survival and performance of specialist herbivores (e.g. [10][12]). Accordingly, population size, population genetic structure, and secondary chemistry of the primary food-plant species affect the genetic variation and differentiation of populations of specialist herbivores (e.g. [13]). High plant-species diversity may reduce the density of specialist herbivores due to the lower concentration of a certain preferred resource in the environment or due to more abundant and effective enemies [14][15]. In contrast to specialists, generalist herbivores with a broad diet may not be strongly affected by a single plant species. Within-population genetic variation and population differentiation of oligophagous herbivores that utilise few food-plant species may be influenced both by the occurrence, abundance, and quality of their primary food-plant species and the diversity of alternative food-plant species in a plant community. We predict that large variety of alternative food-plant species may support stable populations of oligophagous herbivores and thus maintain high level of within-population genetic variation. To our knowledge, there are no studies on the effects of plant-species diversity on the population genetic structure of insect herbivores.

In addition to food-plant quality and diversity, several other, more general, biological and geographic characteristics of the habitat and population may influence the genetic structure of herbivore populations. Small and isolated populations are in general more susceptible to the loss of within-population genetic variation that affects the viability of populations and species [16]. Small and isolated populations may also be more differentiated than large populations or populations with many migrants from other populations [8], [13]. Isolation may be due to geographic distance or the presence of unsuitable habitats for the species to survive or reproduce in [16], [17]. Populations that have persisted over longer times are expected to be genetically more differentiated than younger populations, because random genetic drift, accumulation of mutations, and differential natural selection have had more time to influence their genetic structure [18], [19]. These genetic mechanisms may modify the genetic structure of herbivore populations both independently and in combination with the effects of species interactions. For instance, it is generally expected that genetic differentiation of food-plant populations may lead to differentiation among herbivore populations, but in practice, evidence for this is contradictory (e.g. [20][23]).

We investigated the distribution of genetic variation in 23 populations of the oligophagous seed predator Lygaeus equestris (Heteroptera) in southwestern Finland and in eastern Sweden. We used amplified fragment length polymorphisms (AFLP) to assess genetic variability. We specifically tested whether genetic variation of L. equestris was higher in populations that occurred with large populations of the primary food plant, Vincetoxicum hirundinaria [5], [24]. Large V. hirundinaria populations may sustain larger and more stable L. equestris populations than smaller V. hirundinaria populations. Furthermore, while L. equestris may use other plant species than V. hirundinaria as alternative food [5], [24], we tested whether there was more within-population genetic variation in L. equestris populations occurring in habitats with a high total number of vascular plant species. Since L. equestris may feed on a large number of alternative food-plant species when the seeds of V. hirundinaria are not available ([5], [24], L. Laukkanen, A. Muola and R. Leimu, pers. obs.), we assumed that the total number of plant species correlates with the number of alternative food-plant species in a habitat. A high number of alternative food-plant species may maintain more stable L. equestris populations. We also tested whether genetic differentiation between the L. equestris populations was related to the similarity of the vegetation in the habitats. Furthermore, we examined how other geographic and ecological factors were related to the genetic structure of the populations. First, we tested whether genetic differentiation between the L. equestris populations was related to geographic distance between the populations. We also tested whether island populations had less within-population genetic variation than mainland populations, as migration among the island populations may be limited by water [7], [16]. The size of the island may indirectly affect the viability of the L. equestris populations: for instance, large islands may have more diverse habitats for feeding, oviposition, and hibernation compared to smaller islands. Therefore, we tested whether large islands had more genetically variable L. equestris populations than small islands. Population size generally influences the level of genetic variation [16], and thus we also tested the impact of population size on within-population genetic variation of L. equestris.

Material and Methods

Study species

Lygaeus equestris L. (Heteroptera: Lygaeidae) is a seed-eating true bug that is specialised to feed on Vincetoxicum hirundinaria Med. ( =  Cynanchum vincetoxicum (L.) Pers.) (Apocynaceae), a highly poisonous long-lived perennial herb [5], [25]. Although L. equestris may occasionally also feed on several other plant species than V. hirundinaria, in Scandinavia it is found merely in V. hirundinaria populations [24], [26], [27]. Lygaeus equestris is locally common in the distribution area of V. hirundinaria in Scandinavia, but its population sizes vary considerably among years and populations ([26], [28], L. Laukkanen, unpubl. data). Lygaeus equestris is usually univoltine and overwinters as adult. In the study area, the female L. equestris oviposits on the ground-layer vegetation in June and July. Adults of the new generation commonly appear from late-July onwards ([24], L. Laukkanen, pers. obs.).

Study populations

Samples for genetic analysis were collected from 23 L. equestris populations in June and July 2007. Fifteen of the populations are located in southwestern Finland and eight in eastern Sweden (Table 1, Fig. 1). One Finnish and six Swedish populations are located on the coastal mainland, and the rest of the populations are located on separate islands of the Baltic Sea. We collected on average 27.3±1.12 (mean ± se) individuals from each population (628 individuals in total). The samples were stored in 96.1% ethanol and kept at 4°C before DNA extraction.

All necessary permits were obtained for the described study. Forest administration of Finland (Metsähallitus) granted a licence to collect L. equestris from the area of Archipelago national park. Our studies did not involve endangered or protected species.

DNA extraction and genetic analysis

DNA of L. equestris was extracted with NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's standard protocol. All laboratory work and genotyping were performed by the Center of Evolutionary Applications (University of Turku, Finland). The genotyping was based on amplified fragment length polymorphism (AFLP) profiles that were generated using a protocol modified from Vos et al. [29]. This method has been used in many former population genetic studies of insects (e.g. [30][32]). AFLP is based on selective PCR amplification of restriction fragments from a total digest of genomic DNA. The amplification is done by using adapter sequences as target sites for primer annealing. AFLP products are separated according to size using gel electrophoresis. The AFLP band patterns may be used for monitoring the degree of similarity or differentiation among populations, and as the band patterns map to specific loci, the individuals can be genotyped based on the alleles they carry. The visualisation of genetic differences among individuals and populations using AFLP as molecular marker tool does not require prior knowledge about nucleotide sequence [29], [33]. For detailed description of methods see Supporting Information S1.

Marker reproducibility and genotyping-error rate were estimated using replicates in different phases of the AFLP protocol (Supporting Information S1). It is recommended that the number of running replicates is 5–10% of total number of samples [33][35]. We analysed duplicate AFLP profiles of 38 samples, which is 6.1% of the total number of samples. The replicates were run as separate reactions starting from separate DNA extractions, so that the error rates were reflective of the entire genotyping process. After DNA extraction the AFLP protocol was performed on 96 well plates, and positive controls were used to monitor for differences among the plates. We had four control samples on every plate in randomised positions (the same four samples on each plate: two samples were of ligated DNA, and two samples were of DNA). Usually the error rates fall in a 2–5% range [33], [34], [36]. In this study, the error rates were 0.1–0.5%, and thus the results were highly reproducible. Loci with a varying fragment size and potentially overlapping loci were discarded. In total, we used 86 loci in the further analyses.

Detecting non-variable and non-independent loci.

Allele frequencies were calculated with the square-root method [37] to delete loci showing less than 5% variation, i.e., the proportion of an allele either present or absent was less than 5%. In addition, all loci were checked for linkage equilibrium (independence of loci) to be able to remove loci expressing patterns similar to other loci [17], [34]. This was done according to the formula presented in Gaudeul et al. [38]. For the non-independent loci pairs (less than 5% difference among loci), the locus showing less variation was deleted from the final data set. In many cases the non-independent loci also showed lack of variation. In total, 27 loci and 21 individuals were removed from the data, and we used 59 polymorphic loci and 607 individuals in the final data set.

Estimation of genetic differentiation and within-population genetic variation.

We estimated the levels of genetic differentiation and within-population genetic variation of L. equestris using the AFLP-SURV 1.0 program [39]. We assumed that each marker had only two alleles (a dominant marker allele coding for the presence of a band, and a recessive null allele coding for the absence of the band) and assumed Hardy-Weinberg genotypic proportions. The allele frequencies (p and q) at each marker locus were calculated with the Bayesian method with non-uniform prior distribution of allele frequencies [40]. The frequency of the recessive null allele (q) at each locus was computed using the number of individuals in the sample that lacked the AFLP band and the sample size. The distribution of allele frequencies for each population was then estimated based on the variation over loci, and the statistics of genetic differentiation and within-population genetic variation were computed according to Lynch and Milligan [41]. We used the percentage of polymorphic loci (PLP) at 5% level and expected heterozygosity over all loci (Nei's gene diversity, He) as measures of within-population genetic variation for each of the 23 populations [16]. To estimate He, we first calculated the expected heterozygosity per locus, hj, by subtracting the expected frequencies of homozygotes from 1 (hj = 1 – p2 – q2, where p and q are the allele frequencies). This was repeated for all loci and the average was calculated to obtain expected heterozygosity over all loci, He. We also estimated the average genetic variation within populations by calculating the average PLP and mean expected heterozygosity, HS, and standard errors for these values.

In addition to estimating the level of within-population genetic variation, we calculated average Wright's fixation index, FST, to estimate genetic differentiation among populations in terms of allele frequencies [16]. To be able to evaluate genetic differentiation between the individual L. equestris populations, we also calculated pairwise genetic differentiation as pairwise FST for all population pairs.

Factors affecting genetic differentiation among and variation within L. equestris populations

The size of V. hirundinaria populations varies in our study area. Thus, the number of mature V. hirundinaria individuals was counted in all those study sites, where population size was lower than 100 individuals (Table 1). In populations with more than 100 individuals population sizes were estimated independently by two or three researchers. In most cases, each island harboured a single population. More than one V. hirundinaria population occur on the two largest islands (Mörkö, and Öland where the Resmo alvar population is located). On these two islands, the L. equestris samples were collected only from one site, and V. hirundinaria population size was estimated based on observations from this site [42], [43].

To test whether genetic differentiation between populations or within-population genetic variation of L. equestris was related to the diversity of the plant community we estimated plant-species diversity in the communities where the V. hirundinaria and L. equestris populations occurred. We conducted the vegetation survey twice during the summer 2009 (the first survey in late May – early June, the second in late August – early September) in all study sites, and all the plant species present at least once were regarded as present. Poaceae and Carex were excluded from the vegetation survey because of limited resources and the fact that these groups are not among the most important alternative food sources of L. equestris [24]. We have never observed L. equestris feeding on these plants in our study area (L. Laukkanen, A. Muola and R. Leimu, pers. obs.). Poaceae was abundant in all populations, but Carex species were present only in some of the populations. First, we investigated whether the genetic differentiation between the L. equestris populations was related to the similarity of vegetation between the study sites. For this analysis, we calculated the Sørensen's species similarity coefficients for each population pair using presence/absence data on the plant species [44]. Secondly, we tested whether within-population genetic variation of L. equestris correlated positively with the number of vascular plant species (see statistical analysis).

In August – September 2007 we estimated the population size of L. equestris in the same area where the V. hirundinaria population size was estimated. The population size of L. equestris was estimated by calculating the average number of individuals observed on 14–20 randomly chosen V. hirundinaria plants in the ten Finnish populations and multiplying the average by the estimated number of V. hirundinaria individuals in a population. The V. hirundinaria individuals from which L. equestris was counted varied in size and were chosen haphazardly from the entire area where V. hirundinaria occurred. The L. equestris individuals were counted in August – September to insure that all individuals represented the same generation. We then tested whether large L. equestris populations had more within-population genetic variation compared to smaller populations (see statistical analysis).

To investigate whether the genetic differentiation between the L. equestris populations was related to the geographic distance between populations we determined the distances for all population pairs using the geographic coordinate data registered with GPS (Garmin eTrex). The distances between the populations vary from one to 598 kilometres. We also examined whether the island populations had less within-population genetic variation than mainland populations, as water may restrict migration between populations [7], [16]. To test whether within-population genetic variation was related to island size [45] we determined the area of the islands using electronic maps (http://www.paikkatietoikkuna.fi/web/fi/kartta and http://www.gis.lst.se/lanskartor).

Statistical analysis

We tested the significance of the overall genetic differentiation among populations with permutation tests. The resampling statistics was based on 100 000 random permutations. FST values were computed after each permutation consisting of randomly permuting individuals among existing populations. The real observed FST was then tested against the distribution obtained by permutation. The permutation tests for genetic differentiation were done using AFLP-SURV 1.0 program [39].

We tested whether the genetic differentiation of L. equestris populations (pairwise FST) was related to the similarity of vegetation between the sites. We first tested whether the geographically closely located populations were more similar in vegetation by comparing the geographic distance with the Sørensen's similarity coefficients by simple Mantel test between the two matrices. Because geographically close populations were more similar in vegetation (r = −0.304, n = 253, p = 0.008), we tested whether the genetic differentiation of L. equestris was associated with the similarity of vegetation by conducting partial Mantel test correcting for the geographic distances between the populations. We also tested whether the genetic differentiation between the L. equestris populations correlated with the geographic distance between populations with simple Mantel test between the two matrices. The tests of significance were based on 100 000 random permutations [46] and were conducted using the zt statistical package [47].

We tested whether the within-population genetic variation of L. equestris was associated with geographic and ecological characteristics of the study sites. We used the percentage of polymorphic loci (PLP) and expected heterozygosity (He) as estimates of within-population genetic variation [16]. We conducted separate mixed model ANOVAs with PLP or He as the dependent variables. The location of population on an island or on mainland was included as a fixed factor and V. hirundinaria population size and the number of vascular plant species were included as continuous factors in the analyses of the 23 L. equestris populations from Finland and Sweden. Because we had some additional data collected exclusively from the ten Finnish island populations, we conducted separate analyses using the data from these populations (populations 1–10; Table 1). We conducted separate general linear models for PLP and He. The number of plant species, island size (area in ha), L. equestris population size, and the interactions between these variables were included as continuous variables in the analyses. The continuous explanatory variables were not correlated (data not shown) except for V. hirundinaria population size and L. equestris population size (Spearman correlation coefficients; r = 0.918, n = 10, p<0.001). This is due to the fact that V. hirundinaria population size was used to count L. equestris population size. The population sizes of V. hirundinaria and L. equestris were not included in same models. Because the assumptions of general linear model were not fulfilled for PLP in the analysis with the 10 L. equestris populations from Finland, these data were analysed using ranked values. We used the AIC values to select the best-fitting models in all analyses [48]. The final models consisted of factors with positive or neutral effects on model stability according to the AIC values. The complete models without any model simplification are presented in the Supporting Information S2. The analyses were conducted using the MIXED procedure of the SAS Statistical Package (version SAS 9.2) (SAS Institute Inc. 2002–2007).

Results

Population genetic differentiation

The pairwise FST values ranged from 0 to 0.100. This level of among-population differentiation can be considered low or moderate [19], [49]. In total, the FST values of 45 population pairs (17.7% of all pairs) indicated moderate differentiation (FST between 0.05 and 0.15). The average FST indicating overall genetic differentiation among all populations was 0.034±0.001 (mean ± se, 253 pairwise FST values included). This value suggests rather low genetic differentiation among the L. equestris populations. However, the permutation test indicated that populations were genetically more differentiated than random assemblages of individuals (p<0.05).

Genetic differentiation between the L. equestris populations was not related to the similarity of vegetation between the study sites (FST vs. Sørensen's similarity coefficients, partial Mantel test: r = 0.122, n = 253, p = 0.361). Genetic differentiation of the L. equestris populations increased with increasing geographic distance between the populations (FST vs. geographic distance, Mantel test: r = 0.288, n = 253, p = 0.026; Fig. 2).

thumbnail
Figure 2. Correlation of genetic differentiation and geographic distance among Lygaeus equestris populations.

Genetic differentiation was estimated as pairwise FST.

http://dx.doi.org/10.1371/journal.pone.0094105.g002

Within-population genetic variation

We found significant level of genetic variation within the L. equestris populations. The percentage of polymorphic loci (PLP) ranged from 74.6% to 94.9% in the individual populations (Table 2). The average PLP of all 23 populations was 86.0±1.32% (mean ± se). The level of expected heterozygosity within the populations (Nei's gene diversity, He) ranged from 0.265 to 0.316 (Table 2). The mean expected heterozygosity (mean Nei's gene diversity, HS) was 0.294±0.003 (mean ± se).

Effects of ecological and geographic factors on within-population genetic variation

When all 23 populations from Finland and Sweden were included in the analysis, neither the population size of V. hirundinaria nor the number of vascular plant species affected the percentage of polymorphic loci, PLP (F1,17 = 0.46, p = 0.507 and F1,17 = 1.07, p = 0.315, respectively). PLP did not vary significantly between island and mainland populations (F1,17 = 0.51, p = 0.486). Furthermore, the effect of the population size of V. hirundinaria on PLP of L. equestris did not differ between island and mainland populations (V. hirundinaria population size by island/mainland, F1,17 = 0.50, p = 0.489). Moreover, the effect of the number of plant species on PLP did not differ between island and mainland populations (island/mainland by number of plant species, F1,17 = 0.38, p = 0.546).

When the percentage of polymorphic loci, PLP, was tested separately using the ten island populations from Finland, the two-way interaction between the number of plant species and the island size was significant indicating that the effect of plant-species diversity on PLP of L. equestris was related to island size (Table 3, Fig. 3a). On large islands, PLP seemed to be higher when plant-species diversity was high. On small islands plant-species diversity was not related to within-population genetic variation of L. equestris in terms of PLP. Furthermore, the two-way interaction between island size and the population size of L. equestris was significant (Table 3, Fig. 3b). This suggests that the effect of the population size of L. equestris on PLP was related to island size. When a small L. equestris population occurred on a small island, the percentage of polymorphic loci was on the average low, whereas small populations on larger islands seemed to maintain a higher level of within-population genetic variation (Fig. 3b). In addition, the main effect of island size on PLP of L. equestris was significant: the larger the island, the higher was the percentage of polymorphic loci (Table 3).

thumbnail
Figure 3. Interactive effects of plant diversity, island size, and population size on the percentage of polymorphic loci.

3a) Interactive effect of plant-species diversity (number of vascular plant species) and island size on the percentage of polymorphic loci (PLP) of Lygaeus equestris. 3b) Interactive effect of island size and population size on PLP of L. equestris.

http://dx.doi.org/10.1371/journal.pone.0094105.g003

thumbnail
Table 3. Results of a general linear model on the effects of number of vascular plant species, island size, and population size of Lygaeus equestris on the percentage of polymorphic loci of L. equestris.

http://dx.doi.org/10.1371/journal.pone.0094105.t003

When all 23 populations from Finland and Sweden were included in the analysis, the population size of V. hirundinaria did not influence expected heterozygosity (He) of L. equestris (F1,19 = 0.38, p = 0.547). In contrast, the effect of the number of vascular plant species on He was significant (F1,19 = 6.24, p = 0.022; Fig. 4): when the diversity of the plant community increased also the level of genetic variation within L. equestris population in terms of He increased (Fig. 4). He did not vary significantly between island and mainland populations (F1,19 = 2.44, p = 0.135).

thumbnail
Figure 4. Effect of plant-species diversity on expected heterozygosity (He) of Lygaeus equestris.

Plant-species diversity was estimated as the number of vascular plant species.

http://dx.doi.org/10.1371/journal.pone.0094105.g004

When the ten island populations from Finland were analysed separately, the effect of the number of vascular plant species on expected heterozygosity, He, was significant (F1,6 = 6.17, p = 0.048). As in the model with all 23 populations, the more plant species in the plant community, the higher was the level of genetic variation within L. equestris population in terms of He. Neither island size, nor the population size of L. equestris affected He (F1,6 = 1.48, p = 0.270 and F1,6 = 0.10, p = 0.762, respectively).

Discussion

The effect of vegetation on within-population genetic variation of an oligophagous herbivore

In line with our predictions, our results demonstrate that in terms of expected heterozygosity (Nei's gene diversity, He) genetic variation within the populations increased with the diversity of the plant community. However, the effect of plant-species diversity on the percentage of polymorphic loci depended on island size. The diverse set of alternative food-plant species together with other resources (i.e. more diverse habitats for oviposition and hibernation) occurring on large islands may enable the existence of highly viable and stable L. equestris populations with high level of within-population genetic variation. On large islands with high plant-species diversity and high within-population genetic variation L. equestris may be able to feed on several alternative species simultaneously. These populations may then survive through periods with no V. hirundinaria seeds available. Agrawal et al. [50] and Züst et al. [51] recently showed that selection by the herbivore communities modify the genetic structure of their food-plant populations. Our results suggest that the plant community might also affect the population genetics of the herbivore species.

In addition to the availability of a diverse set of alternative food-plant species, the availability of the primary food, V. hirundinaria seeds, is assumed to be essential for survival and reproduction of L. equestris. In general, seed production is assumed to be higher in large plant populations compared to smaller ones, and large plant populations may support large herbivore populations that have higher level of within-population genetic variation [11], [12]. The fact that the level of genetic variation within the L. equestris populations was not influenced by the population size of V. hirundinaria could be due to the seed-production strategy of V. hirundinaria. The seed production of V. hirundinaria varies considerably among populations and among years mainly due to variation in sun exposure ([25], [52], L. Laukkanen, unpubl. data). Lygaeus equestris populations in Sweden have been observed to decrease when V. hirundinaria seeds are abundant and increase when seeds are scarce [53], [54]. The reason for this paradox is that weather conditions affect both L. equestris and seed resources directly, and these effects are negatively correlated: L. equestris thrives in dry and hot summers, whereas V. hirundinaria suffers from drought and high temperature [28]. The amount of seeds produced in the previous year is important if the summer is warm and dry [53], [54]. Thus, in years of high seed production even a small number of V. hirundinaria individuals may be able to produce enough seeds to maintain an entire L. equestris population. In these years, the population size of L. equestris may be limited by cold weather [26], [28]. On the other hand, when the seed production of V. hirundinaria is low in general, all individual V. hirundinaria plants within a population are likely to produce a low number of seeds or no seeds at all regardless of population size, and the L. equestris population suffers from a shortage of the primary food possibly leading to a population bottleneck and consequently to reduced within-population genetic variation. In addition to weather conditions, variation in the abundance of the other specialist seed predator of V. hirundinaria, the tephritid fly Euphranta connexa (L. Laukkanen, unpubl. data), affects the abundance of V. hirundinaria seeds. Consequently, the potential competitive effect of E. connexa on L. equestris is likely to vary among populations ([11], [25], L. Laukkanen, unpubl. data). Another possible explanation for the fact that the level of genetic variation within the L. equestris populations was not influenced by the population size of V. hirundinaria is the impact of contemporary migration (see below).

We used plant-species diversity and the abundance of the primary food plant as features of the plant community to explain genetic variation of an oligophagous herbivore. In nature, in addition to the number of the plant species, their spatial distribution, densities, and relative abundances contribute to the ability of herbivores to find and use the different food plants [55]. Although being a robust measure, plant-species diversity is probably a suitable indicator of the food plants available in a given habitat for oligophagous and generalist herbivores, whereas other, more specific characteristics of the plant community are important for specialist herbivores. In general, herbivores search their food plants using visual and/or olfactory cues (i.e. plant volatiles) [56]. We do not currently know how L. equestris finds its food plants. However, another Lygaeid species, Oncopeltus fasciatus, finds its food plant (Asclepias spp., Apocynaceae) mainly by olfactory cues and prefers dense patches of food plants [57], [58]. Therefore, it seems likely that olfactory cues are also used for locating suitable food plants by our study species, L. equestris.

The effect of geographic factors and population size on within-population genetic variation

Compared to other sexually reproducing insects, the L. equestris populations studied here have a high level of within-population genetic variation (e.g. [30], [31]). Migration seems to be high enough to sustain the high levels of within-population genetic variation even in the island populations. Lygaeus equestris is well able to fly distances of few kilometers [26], and we have occasionally encountered individuals even on islands where the primary food plant V. hirundinaria is not present (A. Muola, pers. obs.).

In general, small populations often have less within-population genetic variation compared to large populations due to loss of alleles caused by inbreeding and genetic drift [16], [18], [19], [59]. We found that small L. equestris populations on small islands had low percentage of polymorphic loci on average, but small populations on larger islands harboured a higher level of within-population genetic variation. In our study area, L. equestris populations on large islands may be less susceptible to population bottlenecks and extinction-recolonisation processes that both can alter allele frequencies in populations and reduce within-population genetic variation [19]. Firstly, populations in large islands may be less prone to extinctions and recolonisations compared to smaller islands due to more stable microclimates or more diverse habitats for feeding, oviposition, and hibernation. Secondly, in our study area, island size correlates positively with island age (L. Laukkanen, unpubl. data) suggesting that L. equestris populations on large islands are likely to be older than those on smaller islands. The small but older L. equestris populations on larger islands are likely to have had a longer history of inbreeding suggesting that deleterious recessive alleles have been purged in these populations. However, if the small and young populations on small islands still harbour deleterious alleles that are expressed due to inbreeding, these populations may be more prone to extinction [19], [59]. Thus, extinction-recolonisation processes may lead to reduced level of within-population genetic variation especially in the small and young populations.

Population genetic differentiation

Geographical isolation is an important determinant of genetic differentiation (e.g. [60]). Consistent with this, we found that genetic differentiation of L. equestris increased with increased geographic distance among the populations (Fig. 2). The reason for the four horizontal clusters evident in Fig. 2 is that there are no population pairs with certain distances in our study area because of the large open sea area (see Fig. 1). However, the data points are scattered evenly relative to the y-axis (no gaps in distribution of genetic differentiation) providing evidence for isolation by distance (Fig. 2) (see [61]). Peterson and Denno [8] found that isolation by distance was strongest in moderately mobile insect species. This is logical, because highly mobile species may have extensive gene flow over both small and large distances, and in sedentary species gene flow is weak across the entire spatial scale [8]. Lygaeus equestris may infrequently fly long distances, but usually the geographic scale of migration is few kilometres maximum [26], [62], [63]. Thus, L. equestris could be classified as a “moderately mobile” species that is likely to show isolation by distance, which was also observed here. However, the only low or moderate level of differentiation observed among the L. equestris populations may be explained not only by high migration, but also by the relatively short history of the L. equestris populations in the area of the Baltic Sea. In general, populations that have persisted over long time are expected to be more differentiated than younger populations [19]. The effects of genetic drift, accumulation of mutations, and/or differential selection may have acted for such a short period of time that their effects on genetic differentiation among the L. equestris populations are likely to have been minor.

Wang et al. [60] recently stated that geographical isolation explains genetic differentiation more than ecological isolation. Our results support this notion since we found that in contrast to the geographic distance, the similarity of vegetation was not related to the population differentiation of L. equestris. This, together with the result that the number of plant species present in a habitat was related to the level of within-population genetic variation (He), seems to suggest that the number of potential alternative food-plant species present in a habitat might be more important for survival of L. equestris than the identity of species per se. Indeed, most plant species have been shown to be quite poor quality food for L. equestris, and lead to reduced growth and lower reproduction [5], [64]. Population differentiation of L. equestris has previously been studied in Sweden using allozyme electrophoresis [65]. In line with our results, Sillén-Tullberg [65] found genetic differentiation of L. equestris between her two study areas (225 km distance), but not among the seven populations within one area (average distance between the populations was 3.5 km; FST = 0.010). In our study, the pairwise FST values of L. equestris population pairs with less than 10 km distance indicated low level of differentiation (n = 13, all pairwise FST<0.05).

Our findings provide valuable information about the role of ecological and geographic factors as modifiers of the population genetic structure of an oligophagous seed-eating herbivore. Several studies have shown that herbivore-species diversity increases with plant-species diversity and plant genetic diversity (e.g. [66], [67]). To our knowledge, our study is the first to suggest that plant-species diversity is related to the population genetics of an herbivore species: the more diverse was the plant community, the higher was the level of within-population genetic variation of the seed predator. Whether the higher genetic variation of the seed predator is related to the damage the predators induce on their food plants would be worth further studies. The diversity of food organisms is likely to affect predator genetic variation in other plant-herbivore and predator-prey systems as well. Therefore we believe that our results have important, and previously not extensively studied implications for studies of trophic interactions, habitat fragmentation, and landscape genetics.

Supporting Information

Supporting Information S1.

Description of DNA extraction and genetic analysis.

doi:10.1371/journal.pone.0094105.s001

(DOC)

Supporting information S2.

Results from mixed model ANOVAs and general linear models. Complete models without model simplification.

doi:10.1371/journal.pone.0094105.s002

(DOCX)

Acknowledgments

We thank Johan Ehrlén, Benjamin Jackson, Aino Kalske, Tuija Koivisto, Kukka Kyrö, Kati Pihlaja, Timo Pitkänen, Christer Solbreck, Anneli Syrjänen, Anu Veijalainen, and Eero J. Vesterinen for assistance in collecting the data. The Archipelago Research Institute of University of Turku provided facilities during the field work. The genetic analyses were conducted in the Center of Evolutionary Applications (University of Turku, Finland) with support from Meri Lindqvist. We thank Jouko Katajisto for statistical advice and Benedicte Albrectsen, Juan Fornoni, Sonja Gilbert, Aino Kalske, Kari Lehtilä, Tomas Roslin, J. F. Scheepens, and Christer Solbreck for valuable comments on the manuscript.

Author Contributions

Conceived and designed the experiments: LL PM AM RL. Analyzed the data: LL PM AM RL. Wrote the paper: LL PM AM RL. Collected the data: LL AM.

References

  1. 1. Agrawal AA, Lau JA, Hambäck PA (2006) Community heterogeneity and the evolution of interactions between plants and insect herbivores. Q Rev Biol 81(4): 349–376. doi: 10.1086/511529
  2. 2. Futuyma DJ, Keese MC, Funk DJ (1995) Genetic constraints on macroevolution: the evolution of host affiliation in the leaf beetle genus Ophraella. Evolution 49(5): 797–809. doi: 10.2307/2410403
  3. 3. Keese MC (1998) Performance of two monophagous leaf feeding beetles (Coleoptera: Chrysomelidae) on each other's host plant: Do intrinsic factors determine host plant specialization? J Evol Biol 11(4): 403–419. doi: 10.1007/s000360050096
  4. 4. Mopper S, Stiling P, Landau K, Simberloff D, Van Zandt P (2000) Spatiotemporal variation in leafminer population structure and adaptation to individual oak trees. Ecology 81(6): 1577–1587. doi: 10.1890/0012-9658(2000)081[1577:svilps]2.0.co;2
  5. 5. Laukkanen L, Leimu R, Muola A, Lilley M, Mutikainen P (2013) Genetic factors affecting food-plant specialization of an oligophagous seed predator. J Evol Biol 26: 141–149. doi: 10.1111/jeb.12037
  6. 6. Singer MS (2008) Evolutionary ecology of polyphagy. In: Tilmon KJ, editor. Specialization, speciation and radiation: The evolutionary biology of herbivorous insects. Berkeley, Los Angeles and London: University of California Ltd.pp. 29–42.
  7. 7. Nevo E (1978) Genetic variation in natural populations: patterns and theory. Theor Popul Biol 13: 121–177. doi: 10.1016/0040-5809(78)90039-4
  8. 8. Peterson MA, Denno RF (1998) The influence of dispersal and diet breadth on patterns of genetic isolation by distance in phytophagous insects. Am Nat 152(3): 428–446. doi: 10.1086/286180
  9. 9. Kelley ST, Farrell BD, Mitton JB (2000) Effects of specialization on genetic differentiation in sister species of bark beetles. Heredity 84: 218–227. doi: 10.1046/j.1365-2540.2000.00662.x
  10. 10. Awmack CS, Leather SR (2002) Host plant quality and fecundity in herbivorous insects. Annu Rev Entomol 47: 817–844. doi: 10.1146/annurev.ento.47.091201.145300
  11. 11. Leimu R, Syrjänen K (2002) Effects of population size, seed predation and plant size on male and female reproductive success in Vincetoxicum hirundinaria (Asclepiadaceae). Oikos 98(2): 229–238. doi: 10.1034/j.1600-0706.2002.980205.x
  12. 12. Colling G, Matthies D (2004) The effects of plant population size on the interactions between the endangered plant Scorzonera humilis, a specialised herbivore, and a phytopathogenic fungus. Oikos 105: 71–78. doi: 10.1111/j.0030-1299.2004.12783.x
  13. 13. McCauley DE (1991) The effect of host plant patch size variation on the population structure of a specialist herbivore insect, Tetraopes tetraophthalmus. Evolution 45(7): 1675–1684. doi: 10.2307/2409788
  14. 14. Tahvanainen JO, Root RB (1972) The influence of vegetational diversity on the population ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera: Chrysomelidae). Oecologia 10(4): 321–346. doi: 10.1007/bf00345736
  15. 15. Bach CE (1980) Effects of plant density and diversity on the population dynamics of a specialist herbivore, the striped cucumber beetle, Acalymma vittata (Fab.). Ecology 61(6): 1515–1530. doi: 10.2307/1939058
  16. 16. Lowe A, Harris S, Ashton P (2004) Ecological genetics. Design, analysis and application. Padstow: Blackwell Science Ltd. 344 p.
  17. 17. Storfer A, Murphy MA, Spear SF, Holderegger R, Waits LP (2010) Landscape genetics: where are we now? Mol Ecol 19: 3496–3514. doi: 10.1111/j.1365-294x.2010.04691.x
  18. 18. Lande R (1976) Natural selection and random genetic drift in phenotypic evolution. Evolution 30(2): 314–334. doi: 10.2307/2407703
  19. 19. Hartl DL, Clark AG (2007) Principles of population genetics. Sunderland: Sinauer Associates, Inc. 545 p.
  20. 20. Michalakis Y, Sheppard AW, Noël V, Olivieri I (1993) Population structure of a herbivorous insect and its host plant on a microgeographical scale. Evolution 47: 1611–1616. doi: 10.2307/2410172
  21. 21. Ren Z, Zhu B, Wang D, Ma E, Su D, et al. (2008) Comparative population genetic structure of Chinese sumac aphid Schlechtendalia chinensis and its primary host-plant Rhus chinensis. Genetica 132: 103–112. doi: 10.1007/s10709-007-9153-6
  22. 22. Magalhaes IS, Gleiser G, Labouche AM, Bernasconi G (2011) Comparative population genetic structure in a plant–pollinator/seed predator system. Mol Ecol 20: 4618–4630. doi: 10.1111/j.1365-294x.2011.05296.x
  23. 23. Evans LM, Allan GJ, Meneses N, Max TL, Whitham TG (2013) Herbivore host-associated genetic differentiation depends on the scale of plant genetic variation examined. Evol Ecol 27: 65–81. doi: 10.1007/s10682-012-9582-2
  24. 24. Solbreck C, Kugelberg O (1972) Field observations on the seasonal occurrence of Lygaeus equestris (L.) (Het., Lygaeidae) with special reference to food plant phenology. Entomol Scand 3(3), 189–210.
  25. 25. Muola A, Mutikainen P, Lilley M, Laukkanen L, Salminen JP, et al. (2010) Associations of plant fitness, leaf chemistry, and damage suggest selection mosaic in plant-herbivore interactions. Ecology 91(9): 2650–2659. doi: 10.1890/09-0589.1
  26. 26. Solbreck C, Sillén-Tullberg B (1990) Population dynamics of a seed feeding bug, Lygaeus equestris. 1. Habitat patch structure and spatial dynamics. Oikos 58: 199–209. doi: 10.2307/3545427
  27. 27. Rintala T, Rinne V (2010) Suomen luteet. Helsinki: Hyönteistarvike TIBIALE Oy. 352 p.
  28. 28. Solbreck C, Sillén-Tullberg B (1990) Population dynamics of a seed feeding bug, Lygaeus equestris. 2. Temporal dynamics. Oikos 58: 210–218. doi: 10.2307/3545428
  29. 29. Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, et al. (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 23: 4407–4414. doi: 10.1093/nar/23.21.4407
  30. 30. Clark PL, Molina-Ochoa J, Martinelli S, Skoda SR, Isenhour DJ, et al.. (2007) Population variation of the fall armyworm, Spodoptera frugiperda, in the Western Hemisphere. J Insect Sci 7.05.
  31. 31. Krumm JT, Hunt TE, Skoda SR, Hein GL, Lee DJ, et al.. (2008) Genetic variability of the European corn borer, Ostrinia nubilalis, suggests gene flow between populations in the Midwestern United States. J Insect Sci 8.72.
  32. 32. Kneeland K, Coudron TA, Lindroth E, Stanley D, Foster JE (2012) Genetic variation in field and laboratory populations of the spined soldier bug, Podisus maculiventris. Entomol Exp Appl 143: 120–126. doi: 10.1111/j.1570-7458.2012.01239.x
  33. 33. Meudt HM, Clarke AC (2007) Almost forgotten or latest practice? AFLP applications, analyses and advances. Trends Plant Sci 12(3): 106–117. doi: 10.1016/j.tplants.2007.02.001
  34. 34. Bonin A, Ehrich D, Manel S (2007) Statistical analysis of amplified fragment length polymorphism data: a toolbox for molecular ecologists and evolutionists. Mol Ecol 16: 3737–3758. doi: 10.1111/j.1365-294x.2007.03435.x
  35. 35. Crawford LA, Koscinski D, Keyghobadi N (2012) A call for more transparent reporting of error rates: the quality of AFLP data in ecological and evolutionary research. Mol Ecol 21: 5911–5917. doi: 10.1111/mec.12069
  36. 36. Bonin A, Bellemain E, Bronken Eidesen P, Pompanon F, Brochmann C, et al. (2004) How to track and assess genotyping errors in population genetics studies. Mol Ecol 13: 3261–3273. doi: 10.1111/j.1365-294x.2004.02346.x
  37. 37. Nei M (1987) Molecular evolutionary genetics. New York: Columbia University Press. 512 p.
  38. 38. Gaudeul M, Taberlet P, Till-Bottraud I (2000) Genetic diversity in an endangered alpine plant, Eryngium alpinum L. (Apiaceae), inferred from amplified fragment length polymorphism markers. Mol Ecol 9: 1625–1637. doi: 10.1046/j.1365-294x.2000.01063.x
  39. 39. Vekemans X (2002) AFLP-SURV version 1.0. Distributed by the author. Bruxelles: Laboratoire de Génétique et Ecologie Végétale, Université Libre de Bruxelles.
  40. 40. Zhivotovsky LA (1999) Estimating population structure in diploids with multilocus dominant DNA markers. Mol Ecol 8: 907–913. doi: 10.1046/j.1365-294x.1999.00620.x
  41. 41. Lynch M, Milligan BG (1994) Analysis of population genetic structure with RAPD markers. Mol Ecol 3: 91–99. doi: 10.1111/j.1365-294x.1994.tb00109.x
  42. 42. Rydberg H, Wanntorp HE (2001) Sörmlands flora. Uppsala: SBF-forlaget. 776 p.
  43. 43. Karlsson T, Edqvist M (2007) Smålands flora. Uppsala: SBF-forlaget. 1083 p.
  44. 44. Wolda H (1981) Similarity indices, sample size and diversity. OecoIogia 50: 296–302. doi: 10.1007/bf00344966
  45. 45. Frankham R (1996) Relationship of genetic variation to population size in wildlife. Conserv Biol 10(6): 1500–1508. doi: 10.1046/j.1523-1739.1996.10061500.x
  46. 46. Legendre P (2000) Comparison of permutation methods for the partial correlation and partial mantel tests. J Statist Comput Simulation 67: 37–73. doi: 10.1080/00949650008812035
  47. 47. Bonnet E, Van de Peer Y (2002) zt: a software tool for simple and partial Mantel tests. J Stat Softw 7(10): 1–12.
  48. 48. Littell RC, Milliken GA, Stroup WW, Wolfinger RD, Schabenberger O (2006) SAS for mixed models. Cary: SAS Institute Inc.
  49. 49. Wright S (1978) Evolution and the genetics of populations. Vol. 4: Variability within and among natural populations. Chicago: University of Chicago Press. 590 p.
  50. 50. Agrawal AA, Hastings AP, Johnson MTJ, Maron JL, Salminen J-P (2012) Insect herbivores drive real-time ecological and evolutionary change in plant populations. Science 338: 113–116. doi: 10.1126/science.1225977
  51. 51. Züst T, Heichinger C, Grossniklaus U, Harrington R, Kliebenstein DJ, et al. (2012) Natural enemies drive geographic variation in plant defenses. Science 338: 116–119. doi: 10.1126/science.1226397
  52. 52. Solbreck C, Sillén-Tullberg B (1986) Seed production and seed predation in a patchy and time-varying environment. Dynamics of a milkweed – tephritid fly system. Oecologia 71: 51–58. doi: 10.1007/bf00377320
  53. 53. Solbreck C (1995) Long-term population dynamics of a seed-feeding insect in a landscape perspective. In: Cappuccino N and Price PW, eds. Population dynamics: New approaches and synthesis. San Diego and London: Academic Press.pp. 279–301.
  54. 54. Solbreck C (1995) Variable fortunes in a patchy landscape – the habitat templet of an insect migrant. Res Popul Ecol 37: 129–134. doi: 10.1007/bf02515770
  55. 55. Stanton ML (1983) Spatial patterns in the plant community and their effects upon insect search. In: Ahmad S, ed. Herbivorous Insects: Host-seeking behavior and mechanisms. New York and London: Academic press.pp. 125–160.
  56. 56. Schoonhoven LM, van Loon JJA, Dicke M (2005) Insect-plant biology. 2nd edn. Cambridge: University Press. 421 p.
  57. 57. Ralph CP (1977) Effect of host plant density on populations of a specialized, seed-sucking bug, Oncopeltus fasciatus. Ecology 58(4): 799–809. doi: 10.2307/1936215
  58. 58. Ralph CP (1977) Search behavior of the large milkweed bug, Oncopeltus fasciatus (Hemiptera: Lygaeidae). Ann Entomol Soc Am 70(3): 337–342 (6).. doi: 10.1007/springerreference_87129
  59. 59. Pekkala N, Knott KE, Kotiaho JS, Puurtinen M (2012) Inbreeding rate modifies the dynamics of genetic load in small populations. Ecol Evol 2(8): 1791–1804. doi: 10.1002/ece3.293
  60. 60. Wang IJ, Glor RE, Losos JB (2013) Quantifying the roles of ecology and geography in spatial genetic divergence. Ecol Lett 16: 175–182. doi: 10.1111/ele.12025
  61. 61. Guillot G, Leblois R, Coulon A, Frantz AC (2009) Statistical methods in spatial genetics. Mol Ecol 18: 4734–4756. doi: 10.1111/j.1365-294x.2009.04410.x
  62. 62. Southwood TRE, Leston D (1959) Land and water bugs of the British Isles. London & New York: Frederick Warne & Co. Ltd. 450 p.
  63. 63. Solbreck C (1971) Displacement of marked Lygaeus equestris (L.) (Het., Lygaeidae) during pre- and posthibernation migrations. Acta Entomol Fenn 28: 74–83.
  64. 64. Kugelberg O (1973) Laboratory studies on the effects of different natural foods on the reproductive biology of Lygaeus equestris (L.) (Het. Lygaeidae). Entomol Scand 4(3): 181–190. doi: 10.1111/j.1570-7458.1973.tb00262.x
  65. 65. Sillén-Tullberg B (1983) An electrophoretic study of allozyme frequencies in local populations of Lygaeus equestris (Heteroptera: Lygaeidae). Hereditas 99: 153–156. doi: 10.1111/j.1601-5223.1983.tb00741.x
  66. 66. Crutsinger GM, Collins MD, Fordyce JA, Gompert Z, Nice CC, et al. (2006) Plant genotypic diversity predicts community structure and governs an ecosystem process. Science 313: 966–968. doi: 10.1126/science.1128326
  67. 67. Cook-Patton SC, McArt SH, Parachnowitsch AL, Thaler JS, Agrawal AA (2011) A direct comparison of the consequences of plant genotypic and species diversity on communities and ecosystem function. Ecology 92: 915–923. doi: 10.1890/10-0999.1