A genetic population analysis was performed for the plant bug M. pygmaeus throughout a wide range of its geographic distribution using nine microsatellite markers. Samples of M. pygmaeus were collected on S. lycopersicon in 15 localities (Fig. 2), and a sample from a commercial producer was also included in the analyses. The moderate number of alleles per locus (5.5 alleles per locus per population) showed the utility of these molecular markers for population studies in M. pygmaeus. The average number of alleles per locus in M. pygmaeus was lower than that found in other mirid species such as Lygus lineolaris (Palisot de Beauvois) (Hemiptera: Miridae) (12.25 alleles per locus) (Perera et al. 2007) and Stenotus rubrovittatus (Matsumura) (Hemiptera: Miridae) (11.4 alleles per locus) (Kobayashi et al. 2011). Lygus heperus Knight (Hemiptera: Miridae) had similar numbers (4.5 alleles per locus) than M. pygmaeus (Shrestha et al. 2007). Significant linkage disequilibrium was observed among several loci in one of the Turkish and the Greek population. Inbreeding, small population size, genetic isolation between lineages, mixing of populations during sampling, and population subdivision are some of the factors that may lead to linkage disequilibrium (Rafalski and Morgante 2004; Gupta et al. 2005). The variation in linkage disequilibrium has been also associated with bottlenecks (Tishkoff et al. 1996). In the case of M. pygmaeus, the linkage in the populations from Turkey and Greece might be due to isolation and bottlenecks. This is in agreement with the low genetic diversity and the low allelic richness in these two geographic areas, and to the presence of individuals with the same score for the nine microsatellite loci. Population bottlenecks may have resulted in the loss of alleles and allele diversity may have not been compensated with the migration of individuals from other areas. The presence of individuals with the same score for the nine microsatellites might be explained by the low genetic diversity, the founder effect, and the high reproduction rates of this mirid species. Tomatoes are temporal plants where M. pygmaus populations have to re-establish on every yearly cycle; populations may arise from very few founder adults and many individuals in a population may come from the same progenitors in the initial growing phase. Linkage disequilibrium due to inbreeding is excluded because it would produce significant departure from HWE in most of the loci (Schlötterer 1998), and in our case, heterozygosity deficit happens only in a few of them. The heterozygosity deficit detected in a few of the loci could be most likely due to null alleles. The high diversity in the Italian and Iberian Peninsula suggest that M. pygmaeus in these areas did not suffer from the bottlenecks of the Balkans and Asia Minor. The lower diversity in the Canary Islands may be a consequence of the loss of diversity in the process of colonization of the islands from the mainland. The fixation of two of the loci (Mp27 and Mp33) is also symptoms of a bottleneck. Seabra et al. (2009) argued that the lower microsatellite variability and heterozygosity of Cicada orni L. (Hemiptera: Cicadidae) in the Aegean islands compared with mainland Greek was possibly due to a reduced gene flow and bottlenecks during the process of colonization of the islands. When just a small numbers of individuals takes part in the colonization of new areas, the genetic variability of the founder population is expected to be lower than the original one (Zygouridis et al. 2009).
Bayesian analyses, PCA, and UPGMA analyses indicates the existence of three main population clusters in M. pygmaeus: one cluster including the populations from Greece and Turkey; a second cluster including the populations from Italy and France; and a third cluster with most of the populations from the Iberian Peninsula and the Canary Islands. The consistency of nodes for the three main clusters was supported by acceptable bootstrap values in the UPGMA analyses; the ad hoc statistic (ΔK), used to determine the true number of populations using the software STRUCTURE, reached its maximum value at K = 3. Structure is not considered as suitable to describe relationships among population, but it produces the appropriate number of evolutionary cluster under many circumstances (Kalinowski 2011). AMOVA analyses also denoted a significant contribution of the difference between geographic areas to the genetic variability of the species. Macrolophus pygmaeus showed a moderate to high degree of population differentiation among the three main geographic areas of distribution, with the highest degree of divergence found between eastern (Turkey and Greece) and western populations (Italy, France, England, Iberia, and the Canary Islands) (FST values= 0.334–0.194). This high degree of population differentiation in the main peninsulas could be explained by the isolation of M. pygmaeus in these areas during long periods of time. Finding the highest number of private alleles in southern Iberia and Turkey is in agreement with the hypothesis of the old colonization of these two areas. Many species of animals and plants have evolved into different genetic lineages in the three Mediterranean peninsulas during periods of isolations in the glaciations (Hewitt 1999, 2000; Taberlet et al. 1998). Habel et al. (2005) concluded that the Melanargia galathea (L.)/Melanargia lachesis (Hübner) (Satyrinae: Lepidoptera) sibling species complex represented three major genetic lineages, each originating from one of the three major refugial areas in southern Europe. Schmitt and Seitz (2001) argued that the differentiation (FST values = 0.060) in the widely distributed and abundant butterfly Polyommatus coridon (Poda) (Lepidoptera: Lycaenidae) was best explained by the long disjunction in two ice-age refugia of the Adriato-Mediterranean (peninsular Italy) and Ponto-Mediterranean (the Balkan Peninsula, Asia Minor and the east Mediterranean coast). In the same way, Maniola jurtina (L.) (Lepidoptera: Nymphalidae) shows two major genetic lineages: the Atlantico-Mediterranean and the Adriatic-Pontic-Mediterranean subpopulations (Schmitt et al. 2005; Schmitt 2007). Three subpopulations were detected in B. oleae from the Western Mediterranean (Spain and Portugal), Central Mediterranean (Greece, Italy, and, possibly, Turkey), and Eastern Mediterranean (Cyprus) (Augustinos et al. 2005; Zygouridis et al. 2009). Zygouridis et al. (2009) found that populations from Greece, Italy, and possibly Turkey form a distinct subpopulation, likely due to extensive gene flow in the area, which may be attributed to the historical trade between the areas. In contrast, this study on M. pygmaeus is one of the few cases reporting high divergences in the subpopulations of a species between the Italian and the Balkan Peninsulas. To our knowledge, differentiations between the lineages in the Italian and the Balkan Peninsula have been reported only in Fagus sylvatica L. (Fagaceae) (Demesure et al. 1996). In other species of insects, such as the grasshopper Chorthippus parallelus (Zetterstedt) (Orthoptera: Acrididae) (Cooper et al. 1995) and the bark beetle Tomicus destruens Woll. (Coleoptera: Scolytinae) (Horn et al. 2006, 2009), some population differentiation was found between the Balkans and Asia Minor. In the case of M. pygmaeus, the Balkans and Asia Minor seem to represent a distinct population. The absence of differentiation between the populations from Portugal and southern Spain, which are about 600 km apart, indicates the existence of a continuous flow of individuals between remote localities in eastern and western Iberia. These movements would be facilitated by the absence of mountain chains between eastern and western Iberia. Hamarsheh et al. (2009) argued that genetic differentiation can be low, even between populations separated by many thousands of kilometers, unless there are major hydrographic or geographic barriers to migration. FST values (0.065–0.080) indicate a moderate population differentiation in M. pygmaeus between southern Iberia and the Canary Islands; the lack of significance and high variance in FST values between some populations in southern Iberia and the Canary Islands could be due to the low number of individuals in the Canary Island sample. The connection of M. pygmaeus subpopulations between the Iberian Peninsula and the Canary Islands was probably facilitated by the reduction in the Strait of Gibraltar and by the stepping-stone corridor of islands between the Canaries, Madeira, and the Iberian Peninsula in the interglacial periods during the Pleistocene (Whittaker and Fernández-Palacios 2007). Habel et al. (2008) reported the recent connection between subpopulations of the butterfly M. galathea in Iberia and northern Africa due to their genetic similarity. FST values (0.023–0.027) between the populations of M. pygmaeus in northern Spain (Argentona) and the Liguria coast in Italy were low and not significantly different from zero, which indicates null or low population differentiation between these two geographic areas. These two populations were also included in the same cluster in the UPGMA analysis. In contrast, FST values (0.099) between northern Spain and southern France indicated a moderate differentiation in M. pygmaeus populations. These results are intriguing because a higher similarity would be expected between population in northern Spain and southern France due to their geographic proximity. If these were really what happened naturally, we would have to admit that the species is able to disperse through vast open seas, through a hypothetical bridge between northern Spain, Corsica or Sardinia, and the Liguria coast (Fig. 2), which seems quite unlikely. Another striking result is that the differentiation between the M. pygmaeus population in the Liguria (Italy) and northern Spain (FST values= 0.023–0.027) was lower than that between southern and northern populations in Spain (FST values= 0.035–0.042). If these results were to be explained by physical barriers interfering with the movement of the species, we would have to admit that the mountains in central Spain are a stronger physical barrier for the dispersal of the species than the Alps and the Pyrenees put together. The high degree of similarity between population of M. pygmaeus in northern Spain and Italy could be explained by: (1) selective environmental factors (e.g., similarities in climatic conditions between northeast coast in Spain and Liguria); (2) the ad-mixture of wild population and commercial strains native to the Italian peninsula in northern Spain. Zygouridis et al. (2009) argued that trade may be in part responsible for the current population structure of B. oleae in the Mediterranean basin. Besides, the possibility of human and environmental factors interfering with the distribution of populations in M. pygmaeus the high correlation between genetic differences among populations, and geographic distances also indicates a contribution of geographic distance to the genetic structure of M. pygmaeus populations. Augustinos et al. (2005) also found a strong correlation between genetic distances and geographic distances in B. oleae.
Based on the results of this study, we predict the recent history of M. pygmaeus as follows: (1) the reduction in the geographic distribution of the species to the Iberian, Italian, and Balkan peninsulas during the Pleistocene glaciations; the survival of the species in southern France is not excluded, and M. pygmaeus quite likely also survived in northern Africa. Konnert and Bergmann (1995) concluded that southeastern France was a glacial refuge for Abies alba Mill. (Pinaceae). (2) The maintenance of high populations and diversity of M. pygmaeus in Iberia, Italy, and possibly southern France, in the periods of contraction, with a reduction in population size and bottlenecks in the Balkans. (3) The introgression of the Italian–French lineage in northern Spain (naturally or through the trade of the species), and the likely expansion of the species from the Balkans to Asia Minor, given the similarities of subpopulations and the lower diversity in the latter area. Based on the genetic similarities of the British, the Iberian, and the Italian–French populations to northern European populations, it is possible that M. pygmaeus spread to northern Europe from its southern refuges in Iberia, Italy, and likely from the south of France.
This study proves that microsatellite could be used as powerful markers to determine the genetic structure of M. pygmaeus populations. The use of microsatellite markers may have a very practical approach in biological control, especially in conservation programs, as it would allow the identification of the source of the individuals colonizing crops. Microsatellite marker may also help to understand the introgression of commercial strains into wild populations. This study gives for the first time insight into the population structure of a mirid species in the Palaearctic region, which may also help to explain the current geographic distribution and speciation of other plant bugs in the area. The findings of this study provide a better understanding of the evolution and ecology of M. pygmaeus and could be used to study other mirid species with a similar biology. Divergences in ecological traits (e.g., host plant preference, degree of phytophagy) in M. pygmaeus subpopulations in the different geographic areas may also produce diverse outcomes in population dynamics and when using the species for pest control. More populations have to be analyzed to determine the population structure of M. pygmaeus in middle and northern Europe, and to understand the role of northern Africa as a refuge for the species during the glaciations and its connections with other geographic areas in the Middle East, Asia Minor, and the Balkans.