British whitebeams (Sorbus aria aggr.; Rosaceae) are an emblematic case study of polyploid evolution in natural tree populations. Southwestern Britain is a “Sorbus hotspot,” with ca. 30 polyploid species (3x, 4x, and even 5x), many of them occurring at just a few localities and therefore highly valuable in regard to conservation. Recent studies have provided detailed knowledge of the morphology and ploidy levels in British populations of Sorbus L. (Rich et al., 2010; Pellicer et al., 2012). Despite this effort, some essential questions regarding the evolution of the complex remain unsolved. In addition to the sexual diploid species S. aria (L.) Crantz s. str., current evidence points at the polyploid S. porrigentiformis E. F. Warb., an endemic to the United Kingdom that shows a distribution significantly larger than the other highly endemic polyploids of the complex, as a parental species of many of these polyploid endemics. The apomictic tetraploid S. rupicola (Syme) Hedl., widely distributed in northwestern Europe, including the United Kingdom, may have also been involved. To provide diagnostic alleles for the three species, we isolated and characterized the first set of microsatellites for S. porrigentiformis and tested cross-amplification in S. aria and S. rupicola. Previous studies on Sorbus in southwestern Britain have used two nuclear microsatellites from apple (Malus Mill. sp.) and three from S. torminalis (L.) Crantz (Robertson et al., 2010; Ludwig et al., 2013).

METHODS AND RESULTS

A DNA library was generated for one sample of S. porrigentiformis and sequenced on a Roche/454 GS FLX platform (454 Life Sciences, a Roche Company, Branford, Connecticut, USA). From the 35,638 reads, 10,872 microsatellite loci were detected. Primer pairs were designed with the software QDD (Meglécz et al., 2010) using default parameters (90–320 bp PCR products, with more than five repeats of 2–6 bp motifs, 18–27 bp primer length, 57–63°C annealing temperature). We tested 20 of the primers on seven geographically separated individuals of S. porrigentiformis (Appendix 1). Fluorescent labeling was performed using three primers per locus: a reverse primer, a forward primer with a universal linker sequence (M13) at the 5′ end, and a third primer consisting of the same universal M13 sequence, labeled with 6-FAM or JOE (Schuelke, 2000). We added 7.5 µL of Multiplex Mix (10×), 0.2 µL of bovine serum albumin (BSA), 0.3 µL of each reverse primer (10 µM), 0.15 µL of dye-labeled and forward primers (10 µM), 1 µL of template DNA (ca. 10–50 ng/µL), and H₂O up to a final volume of 15 µL. Amplifications were performed as follows: 94°C (4 min); 25 or 30 cycles of 94°C (30 s), 55°C (45 s), 72°C (1 min); followed by 10 cycles each of 94°C (30 s), 53°C (45 s), 72°C (45 s); and a final extension at 60°C for 30 min. PCR products (0.7 µL) were separated on an AB I 3730 sequencer (Applied Biosystems, Lennik, The Netherlands) with 10 µL of HiDi Formamide and 0.15 µL of GeneScan 500 ROX Size Standard (Applied Biosystems). Sixteen primer combinations exhibiting robust amplification were selected (Table 1). All DNA extractions were performed with the DNeasy Plant Mini Kit (QIAGEN, Valencia, California, USA).

We set up 12 simplex reactions containing one microsatellite marker and four multiplex reactions containing up to three loci (Table 1). Markers with different amplicon sizes and similar annealing temperature were identified with Multiplex Manager (Holleley and Geerts, 2009) and combined in the same multiplex. Electropherograms were automatically scored with GeneMapper version 3.7 (Applied Biosystems) and manually corrected. Fifteen markers displayed easily interpretable electropherograms with up to two alleles per locus in the diploid individuals and up to four alleles in the tetraploid individuals. Locus SP22 exhibited up to four peaks in diploids and up to six in tetraploids. Two different size ranges with different amplification intensities and up to two peaks per individual each could be distinguished in S. aria and S. rupicola, but not in S. porrigentiformis. Therefore, locus SP22 was analyzed as a dominant marker.

Table 1.

Description of 16 newly developed microsatellite loci in Sorbus porrigentiformis in four multiplex and seven simplex reactions.

Table 2.

Genetic diversity of the 11 newly developed polymorphic microsatellites in three populations of Sorbus porrigentiformis and cross-amplification in S. aria and S. rupicola. All populations are located in southwestern Britain.

To characterize the 16 microsatellite loci, 45 individuals were genotyped (Appendix 1): 25 S. porrigentiformis from three different populations in southwestern Britain (3x and 4x), 10 S. aria (2x), and 10 S. rupicola (4x). Sorbus porrigentiformis is endemic to southwestern Britain and individuals occur scattered in the field, which explains the limited sample sizes in this study. However, given that reproduction is mostly clonal, our sampling strategy is representative of the real genetic variation of the species. Ploidy levels of all samples were known from a previous flow cytometry study (Pellicer et al., 2012). Five markers were monomorphic across all 45 samples studied (Table 1). Locus SP26 was biallelic, whereas SP21, SP24, SP25, and SP34 were monoallelic. The remaining 11 microsatellite markers were polymorphic across the three congeners (Tables 1, 2), eight of them exhibited species-specific alleles. Twenty-two private alleles were identified for S. porrigentiformis. Locus SP28, although monomorphic in terms of allele counts, exhibited species-specific differences in allele dosage between S. porrigentiformis and S. rupicola that could be clearly detected, with a ratio of peak areas of 0.45 and 1.34, respectively (Esselink et al., 2004).

For the 11 polymorphic loci, one to 11, one to six, and one to six alleles per locus were retrieved for S. porrigentiformis, S. aria, and S. rupicola, respectively (Tables 1, 2). Allele sizes, number of alleles, and number of private alleles were calculated for each polymorphic locus and species using SPAGeDi (Hardy and Vekemans, 2002). Sorbus porrigentiformis genotypes were further evaluated with GENODIVE (Meirmans and Van Tienderen, 2004) by estimating the expected and observed heterozygosity, with and without correction of allele dosages for polyploids using a maximum likelihood method. Within S. porrigentiformis, populations for most loci exhibited fixed alleles. The observed heterozygosity varied between 0.40 and 1.00. Sorbus porrigentiformis exhibited low genetic variation at the intraspecific level, but it was not completely clonal, fitting the expectations for a facultative apomict.

CONCLUSIONS

The newly developed nuclear microsatellite loci allow discrimination between the species S. porrigentiformis, S. aria, and S. rupicola. These markers will be an important tool to trace the origin of polyploid endemic species of the S. aria agg. in southwestern Britain, and to understand the relative contribution of S. aria, S. rupicola, and S. porrigentiformis as parents of these local polyploids. The resulting genetic information will be relevant for choosing the best approach for the conservation of the polyploid complex S. aria agg. in southwestern Britain either by focusing on the conservation of the local endemic taxa or by focusing on the preservation of the polyploidization process (Ennos et al., 2012) by protecting the parental species, even if they are not local endemics themselves.