Assessment of genetic diversity within the Merodon ruficornis species group ( Diptera : Syrphidae ) by RAPD analysis

As one of the most distinct groups in the hoverfly genus Merodon, the monophyletic ruficornis species group has been the focus of several studies using different approaches. Molecular methods have shown incongruences between morphological and molecular data. In the present study, we investigated four species of the Merodon ruficornis group (i.e. M. loewi, M. armipes, M. papillus and M. hoplitis) with the aim of detecting intraand interspecific genetic diversity, and we examined the usefulness of random amplified polymorphic DNA (RAPD) in an integrative taxonomic approach to species delimitation. Analysis of Nei’s genetic variation over all loci showed that genetic diversity for the analyzed Merodon species was h=0.24. Based on UPGMA, PCoA and Bayesian clustering analyses, our results clearly differentiated four groups that correspond to the four morphologically-defined Merodon species. Among the analyzed species, M. armipes and M. hoplitis showed the lowest level of genetic divergence; M. loewi was clearly separated from both M. armipes and M. papillus. Based on our data, we propose the use of RAPD-PCR as an additional tool for resolving taxonomic problems within Merodon.


INTRODUCTION
Hoverflies (Diptera: Syrphidae) are one of the most species-rich dipteran families [1].Among their genera, Merodon Meigen, 1803, is comprised of more than 160 species, distributed throughout the Palaearctic and Afrotropical regions [2,3].This genus has been the subject of numerous studies dealing with its phenotypic diversity as well as its molecular diversity, aimed at providing additional tools for taxon identification and phylogenetic revision [2][3][4].Although both the systematics and taxonomy of this genus have recently received attention [2][3][4][5][6][7][8][9][10][11][12][13], the status of some Merodon species still requires clarification.The high diversity and presence of cryptic taxa make the inference of phylogenetic relationships within this genus challenging, especially concerning their evolutionary and conservation biology [9,14,15].Different Merodon species and species complexes have been studied primarily using allozymes and mtDNA cytochrome c oxidase subunit I (COI) sequences (M.albifrons Meigen, 1822 [11], the M. avidus complex [10,15,16] and the M. aureus and M. cinereus species complexes [9]).These markers have proven useful in the detection of intra-and interspecific genetic diversity, although results have not always been congruent.Both markers appear to be powerful in integrative taxonomic studies, with an a priori advantage of COI for distinguishing species due to its sequence variability.However, allozyme data with species-specific alleles have occasionally proven more informative for species delimitation than mtD-NA markers [15].In addition, the results of certain morphological studies have not always agreed with the results of molecular analyses [5,14].
Discordance between the results of different studies using different approaches to reveal taxonomic boundaries within this genus is especially relevant in the Merodon ruficornis species group (sensu Milankov et al. 2002 [17] in [13]).Predominantly distributed in the northern and eastern Mediterranean, with the greatest diversity on the Balkan Peninsula, in Turkey and the Caucasus region, this group is characterized by a significantly high level of endemism.Among the 18 recorded species, 12 are known as limited-range endemics, present in small, often isolated parts of the total range of the group [13].The M. ruficornis species group is comprised of closely related Palaearctic taxa that are morphologically extremely similar, with the shape of metalegs being the main diagnostic character in almost all the species [13].The distinct phylogenetic position of the ruficornis species group has been resolved, and molecular and morphological data support its monophyly [13].The taxonomic boundaries of the several species within the group have been well defined based on wing shape and size.Even interpopulation wing morphometric variability implies high structuring within morphologically clearly defined species, which is in accordance with molecular diversity [5].However, analyses of mitochondrial and nuclear regions revealed ambiguities regarding interspecies relationships within the ruficornis group (i.e.Merodon loewi van der Goot, 1964, M. armipes Rondani, 1843, M. papillus Vujić, Radenković, Pérez-Bañón, 2007) [2,8,13].
The integration of multiple data sources in order to re-evaluate taxonomic species delimitation within the Merodon genus is prudent [6,7,11,[14][15][16]. Ever since random amplified polymorphic DNA (RAPD) genetic markers were described [18,19], they have been widely used for the estimation of genetic diversity within and among various insect species and populations [20].Some of the first applications of such 10-base primers for amplification of random regions of genomic DNA were used to identify subspecies and geographic populations of mosquitoes (Diptera: Culicidae) [21] and to differentiate morphologically indistinguishable cryptic species [22].Being relatively simple, quick, technically accessible and independent of a priori sequence information, RAPD continues to be used successfully in Dipteran studies to the present day, e.g. for determining the population genetic variation of blowflies (Calliphoridae) [23], species characterization of leaf-miner flies (Agromyzidae) [24], and interspecific genetic relationships of fruit flies (Drosophilidae) [25] and houseflies (Muscidae) [26].Even though they have been shown to be a powerful tool in various molecular analyses of many fly families, to date there are no published data on the application of RAPD markers to Syrphidae.
In the present study, we investigated four Merodon species of the ruficornis group: M. loewi, M. armipes, M. papillus and M. hoplitis Hurkmans, 2012 [13].Since these species are morphologically differentiated and defined, we attempted to resolve a persisting incongruence among the results of analyses based on different molecular markers and morphological characters.The main goal of this study was to detect the genetic diversity of these four species and to clarify their relationships using RAPD markers.Furthermore, we wanted to evaluate the use of RAPD molecular markers as an additional tool for taxon delimitation within Merodon.

Insect material
Genetic diversity was analyzed for 52 specimens of four Merodon species belonging to the ruficornis group.Sampling was carried out in 2008 and 2010 from April to June.The study area comprised three different localities on the Balkan Peninsula: Lesvos Island (Greece) -M.papillus (15 specimens); Đerdap Gorge (Serbia) -M.loewi (15 specimens); Orjen Mountain (Montenegro) -M.armipes (12 specimens) and M. hoplitis (10 specimens).Collection was carried out by hand-net and insects were frozen at -20°C until further use.

DNA extraction and PCR amplification
Total genomic DNA was isolated from 2-3 legs of each specimen following Hondelmann et al. [27].Genome polymorphism was assessed using a RAPD-PCR marker system.Ten RAPD primers (Operon Technologies Inc., USA) gave clear banding patterns.RAPD-PCR reactions were performed in a volume of 20 μL containing 1 U Taq polymerase (Thermo Scientific), 1×PCR Buffer (Thermo Scientific), 2.5 mM MgCl 2 , 0.2 mM of each dNTP, 0.1 µM of primers and about 100 ng genomic DNA.Amplifications were carried out in an Eppendorf Personal mastercycler in 40 cycles using the following program: initial denaturation at 95°C for 5 min, followed by denaturation at 94°C for 2 min, primer annealing at 36°C for 1 min, elongation at 72°C for 2 min and a final extension at 72°C for 5 min.The amplified products of RAPD-PCR reactions were separated by horizontal electrophoresis in 1.5% agarose gels, and photographs of the gels were obtained under UV light.Product sizes were determined using a 1 kbp DNA ladder (Serva), and a gene ruler, 100 bp DNA ladder (Thermo Scientific).Clear bands were read (using Adobe Photoshop 7.0) and scored for further statistical analysis.

Data analysis
The binary matrix for further statistical analyses was made according to the DNA banding patterns that were scored for presence (1) or absence (0) of the fragment in each sample.POPGENE ver.1.32 [28] was used for estimation of genetic variation statistics for all loci [29], i.e. calculation of genetic diversity parameters: h -Nei's [30] gene diversity; I -Shannon's [31] information index, and p -the percentage of polymorphic loci.Arlequin version 3.0 [32] was employed to perform AMOVA (analysis of molecular variance) with 10000 permutations.
Free Tree [33] software was used to calculate genetic similarity among specimens by pairwise comparisons [34].The resulting coefficients were used to construct a UPGMA (unweighted pair group method with arithmetic mean) dendrogram.Support for the tree topology was assessed by bootstrap analysis using 5000 repetitions, in Free Tree [33].For graphical representation, data were processed using Tree View [35] software.Principal coordinate analysis (PCoA) was performed using GenAlEx [36] based on the Nei and Li [37] coefficients of genetic distances.
Using STRUCTURE ver.2.3.3 [38,39], an analysis of population structure based on a Bayesian clustering approach was conducted.Each of the genotype classes was treated as a haploid allele, as recommended by the software documentation, Oliveira et al. [40] and Pinheiro et al. [41].The admixture ancestry model and the correlated allele frequency model were employed.Runs with a burn-in of 20000 and a Monte Carlo Markov chain (MCMC) of 200000, with 10 repetitions for K=1-10, were set.The results were visualized by CLUMPAK [42].Using the same software package, an estimate of the most likely number of clusters (K) according to the ΔK method [43] was performed by calculating the log probability of data (lnD(P)) for each value of K [38,44].

RESULTS
RAPD analyses of the four Merodon species (52 specimens) with ten different primers resulted in 241 amplified fragments.The number of detected loci varied between primers from 18 (OPA20) to 28 (OPA02) (Table 1).The fragment size ranged from 50 bp to 2 kbp.The total number of polymorphic fragments by species was between 145 (60%) for M. hoplitis and 182 (76%) for M. armipes (Table 2).The analysis of genetic variation statistics for all loci showed that the Nei's genetic diversity for all analyzed Merodon species was h=0.24 and the Shannon index was I=0.39.
Analysis of the distribution of genetic variation among and within species revealed a significantly high variance among the analyzed species (24.53%, p<10 - 6 after Bonferroni correction [45]), despite the high variation detected within species (75.47%) (Table 3).
The UPGMA dendrogram based on Jaccard genetic similarity coefficients showed clear subclustering among different species (Fig. 1).The group consisting of all M. papillus specimens formed a cluster that was      separated from the others with high support (bootstrap value of 100).The M. loewi group was the next to branch out (86), and M. hoplitis and M. armipes were resolved as sister groups (with a bootstrap value of 72).A similar grouping of samples was revealed by PCoA (Fig. 2).The first two principal coordinates, explaining a total of 21.02% of the genetic variation, strongly separated samples of M. papillus and clearly differentiated M. loewi, whereas individuals of M. hoplitis and M. armipes showed slightly lower genetic divergence from each other.
Results of the Bayesian clustering analysis were in accordance with the previously described results.According to the ΔK method (Fig. 3A) and lnP(D) values (Fig. 3B), four was deemed the most likely number of clusters in the analyzed data.All samples were strongly assigned to one of the four different clusters, clearly differentiating the four analyzed species (Fig. 4).

DISCUSSION
In our RAPD analysis of four Merodon species of the ruficornis group, the total numbers as well as the percentages of polymorphic fragments per species were the highest for M. armipes and the lowest for M. hoplitis.Genetic variation per species over all loci revealed the highest values of Nei's diversity and Shannon's index for M. armipes and the lowest values for M. papillus.Although these genetic diversity parameters did not differ much among the analyzed species, the higher variability in M. armipes could be related to the very wide distribution of this species, compared to M. hoplitis and M. papillus that are both endemic species [13].It was previously suggested that historic biogeographical processes, as well as selection in different environments, are important factors shaping the taxonomic diversity and population structure of the M. ruficornis group [5], and of the Merodon genus in general [46].Analysis of the distribution of genetic variation in our study revealed a relatively high level of genetic differentiation among analyzed species (25%), yet genetic diversity was even higher within each species (75%).Environmental factors interact with microevolutionary processes throughout a species' history, influencing the diversity of fragmented populations [5].Hence, the high genetic polymorphism detected in this research could reflect changes during the separate evolutionary paths of the studied species.
This was shown to be true especially for M. papillus from Lesvos Island, which was the most genetically distinct of the four Merodon species examined in our analyses.This result is potentially a consequence of the long-term independent evolution of this endemic species.The importance of geographic isolation in the process of speciation has previously been discussed for the ruficornis species group [13].Despite relatively strong flight ability (i.e.high dispersal potential), Merodon hoverfly species are rather sedentary because of the tight connection to the habitats of their larval host plants.The outcome is clear phylogeographic structuring, which has been proven for many Merodon species through mtDNA COI barcode haplotype diversity and distribution patterns [46].
In analyzing Merodon hoverflies sampled on Lesvos Island by mitochondrial DNA COI barcodes (5'-end of the COI gene), Ståhls et al. [2] successfully distinguished all of the recognized 22 species, except for M. papillus and M. loewi from the ruficornis group.Neither a strict consensus tree (parsimony analyses) nor neighbor-joining cluster analysis succeeded in separating those two species, which were shown to share the same haplotype, despite there being clear morphological differences between them.Since interspecific hybridization is frequently present in similar taxa with geographically overlapping distributions, mitochondrial introgressions and retained ancestral polymorphisms have been suggested as possible reasons for the shared identical haplotypes between these two species [2].
However, in the case of allopatric species, identical COI haplotypes can hardly be the result of introgression.Milankov et al. [8] investigated five species of the ruficornis group on the Balkan Peninsula using mtDNA COI sequences (3'-end of the COI gene) and concluded that M. loewi and M. armipes (from the ruficornis group) shared the same COI haplotype, classifying them as a monophyletic group (as revealed by parsimony analysis) even though these taxa are also morphologically well defined.Although both these species have wide distributions [13], M. loewi and M. armipes inhabit allopatric localities and geographically overlap only to a small extent.Therefore, the shared haplotype of these two species could be the result of incomplete lineage sorting or stabilizing selection arising from the common ancestor, rather than the result of introgression [8].
In our study, M. loewi and M. armipes were clearly separated.To our knowledge, this study provides the first DNA evidence for the separation of these species and reveals RAPD as a molecular tool that can be used in integrative taxonomic study of the Merodon ruficornis group.Our results are consistent with a UPGMA dendrogram generated from allozyme markers [8].In addition, in analyzing wing shape and size of M. armipes and M. loewi, Francuski et al. [5] revealed interspecies phenotypic differentiation between them.There are numerous reasons why delimitation of species boundaries can be limited by single-locus studies, such as those of a mitochondrial gene, including recent radiations or recently and rapidly diverged species.Furthermore, a few sequences are not likely to be representative of the molecular variability of widespread species (such as M. loewi and M. armipes), and it has previously been suggested that species identifications should be verified using additional gene regions [2].It is also important to integrate molecular with morphological characters, especially in taxonomic assessments of closely related species and for taxa in the process of divergence and speciation, such as the M. ruficornis group [5].
Our RAPD analyses confirmed conclusions based on morphological data, i.e. that M. loewi is a clearly differentiated species from M. armipes and M. papillus.Considering all of the analyzed specimens, our UPGMA, PCoA and particularly the Bayesian clustering analysis clearly partitioned the examined individuals into four groups, which correspond to the four morphologically defined Merodon species.Although M. armipes and M. hoplitis, species previously reported as having clear morphological differences [13], showed the lowest genetic divergence, specimens from both species were assigned to distinct groups, as clearly indicated by our STRUCTURE analysis.
Previous studies that used different molecular markers have not been completely successful in distinguishing some of the morphologically defined species from the M. ruficornis group, a highly diverse group with a large proportion of endemic and cryptic species [2,8,13].Nuclear and mitochondrial genes provide plentiful molecular data as an important information source besides morphological characters, especially in difficult cases of taxa delimitation.However, nuclear and mitochondrial genomes do not share the same inheritance and evolutionary patterns.Different analytical approaches may also result in a lack of consensus between datasets.
RAPD markers have been widely used to characterize insect species, to infer phylogenies and the biogeography of insect populations, and to elucidate modes of evolution [20].This technique has proven to be a reliable, cost-effective and rapid way to distinguish dipteran species [24], often detecting interspecies differentiation more accurately than other markers (e.g.allozymes) [47].Applying this method to assess intra-and interspecific genetic relationships, various studies have reached conclusions on the taxonomy and phylogeny of dipteran taxa, emphasizing the importance of an integrative approach [24,25,26,47].

CONCLUSIONS
The present study has shown that RAPD-PCR is a reliable method for examining intra-and interspecific genetic diversity and evaluating taxonomic species delimitation within the Merodon ruficornis group.Our results demonstrate the potential of RAPD in population genetic studies and as an additional tool in resolving taxonomic problems within Merodon.Considering certain inconsistencies between morphological and molecular data as revealed in previous studies, and the fact that only a few DNA markers have thus far been used for this genus, the development of additional molecular markers is an important step in verifying species delimitation.Taking into account prior research regarding the interspecific relationships of Merodon, the results of our study support previous conclusions regarding the need for an integrative taxonomic approach, implying the use of different methods and characters from different sources, i.e. combining molecular, morphological and ecological data.

Fig. 2 .
Fig. 2. Principal coordinate analysis based on Nei and Li[37] chord distance among samples.The first and second principal coordinates account for 11.35% and 9.68% of the variation, respectively.

Fig. 3 .
Fig. 3.The most likely number of clusters estimated by: A -ΔK method suggesting K=4 and B -lnP(D) values supporting K=4.

Fig. 4 .
Fig. 4. Population structure of analyzed Merodon specimens inferred using the STRUCTURE version 2.3.3 software; cluster partitioning of the individuals at K=4.

Table 1 .
Attributes of the RAPD primers (length range and number of amplified fragments per primer).

Table 2 .
Genetic variation statistics per species for all loci.

Table 3 .
AMOVA results among and within Merodon species.