The characterisation of microsatellite markers reveals tetraploidy in the Greater Water Parsnip, Sium latifolium (Apiaceae)

The Greater Water Parsnip, Sium latifolium (Apiaceae), is a marginal aquatic perennial currently endangered in England and consequently the focus of a number of conservation translocation projects. Microsatellite markers were developed for S. latifolium to facilitate comparison of genetic diversity and composition between natural and introduced populations. We selected 65 S. latifolium microsatellite (MiSeq) sequences and designed primer pairs for these. Primer sets were tested in 32 individuals. We found 15 polymorphic loci that amplified consistently. For the selected 15 loci, the number of alleles per locus ranged from 8 to 17. For all loci, S. latifolium individuals displayed up to four alleles indicating polyploidy in this species. These are the first microsatellite loci developed for S. latifolium and each individual displayed 1–4 alleles per locus, suggesting polyploidy in this species. These markers provide a valuable resource in evaluating the population genetic composition of this endangered species and thus will be useful for guiding conservation and future translocations of the species.


Background
Plant translocation is a common occurrence, with an estimated 600 species of plants having been relocated as population introduction, re-introduction or augmentation [1,2]. Whilst a tactic for large scale habitat restoration is through the planting of multiple species, translocation is also an important conservation strategy for specific plants at risk [3]. Guidance on plant translocations recommends consideration of genetic composition [4] however projects infrequently utilise genetic techniques in planning and evaluating reintroductions ( [5]; although see [6,7] as examples).
One species that has been widely translocated in the UK is Sium latifolium L., the Greater Water Parsnip. S. latifolium is a herbaceous, marginal aquatic perennial in the plant family Apiaceae, tribe Oenantheae; one of nine species within the genus, it is found across Europe and Asia [8]. With large, conspicuous, umbel inflorescences and growing to 2 m tall [9], S. latifolium was once a noticeable dominant in wetland areas of England, where it grows in habitats of fen, pond margins and grazing marsh ditches [10]. However, the population of S. latifolium has much declined over the past 40 years, due to habitat loss and change in wetland management [11]. It is now classified as 'endangered' on the vascular plant red list for England [12]. As a response to the marked decline in populations, conservation projects involving translocations of S. latifolium have occurred independently in at least seven counties of England, re-introducing the species in regions where it has been lost or declined, however the success of these translocations has been mixed.
The goal of this study was to generate a suite of microsatellite markers specifically developed for S. latifolium in order to evaluate and compare the genetic composition of populations, both old and new, with the view to guide practitioners in the best approaches for further translocations of this species. With many independent reintroductions it can also be used as a case study for exploring broader questions relating to genetic management of plant translocations.

Results
Samples of S. latifolium were collected in May 2012 and August 2013 (Table 1), permission for sampling was obtained from the landowner of each site. Three leaflets per plant were preserved in silica gel and stored at room temperature. Prior to extraction, 10-20 mg of leaf tissue was frozen overnight at −80 °C before being homogenised at 1000 Hz for 3 min using a GenoGrinder 2000 (Spex CertiPrep, Metuchen, NJ USA). Genomic DNA was isolated employing a cetyltrimethyl ammonium bromide (CTAB) protocol [13], with the addition of 1% polyvinyl pyrrolidone (PVP) to the isolation buffer to remove polyphenols [14]. Once washed and air-dried, DNA was re-suspended in 100 µl low TE (10 mM Tris-HCl, 0.1 mM EDTA, pH 8.4) and subsequently diluted to 100 ng/µl with low TE. The microsatellite library was prepared from one individual sampled at Wickhampton Marshes, Norfolk, UK (52°35′N 1°35′E; sample identification code = I50). The library was enriched for microsatellites, using magnetic beads in the hybridisation [15,16]. An Illumina pairedend library was created using 1 µg of the repeat-enriched genomic DNA. The SureSelect Library Prep Kit, ILM (Agilent Technologies Inc. Santa Clara, California) protocol was followed and 2 × 250 bp paired-end sequencing conducted using a MiSeq Benchtop Sequencer (Illumina Inc. San Diego, California).
Sequences with at least ten repeats were selected for primer design; primer sets were designed to amplify the microsatellite regions using PRIMER3 v 0.4.0 [17]. Specifications for primer selection were set at a primer length of 16-36 base pairs (optimum 20 bp), an optimal primer melting temperature of 60 °C, (min-max of 59-61 °C), a maximum of 0.5 °C between primers, presence of a 3′ GC clamp, a maximum poly-X of three and the default settings for all other parameters. Sixty-five primer sets were designed. The 5′ end of each forward primer was fluorescently-labelled with HEX or 6-FAM.
Microsatellites were amplified in 2-µl PCRs, including 1 µl (100 ng) genomic DNA (air dried), 2 µl primer mix (forward and reverse primer at 0.2 µM) and 1 µl Qiagen Multiplex PCR Master Mix including HotStar Taq DNA polymerase (Qiagen Inc.). Covered with a thin layer of mineral oil, products were amplified under the following profile: incubate at 95 °C for 15 min, followed by 35 cycles of 94 °C for 30 s, selected primer temperature (51, 53 or 58 °C, see Table 2) for 1 min 30 s and 72 °C for 1 min 30 s, and finally incubated at 72 °C for 10 min. The optimum annealing temperature for each primer set was initially All primer sets were initially tested in six unrelated individuals (Table 1), each from a different geographic population in the UK. Markers failing to amplify or appearing monomorphic at this stage were discarded. The remaining primer sets were then tested in a further 24 individuals from the same population as the individual sequenced to isolate the microsatellites (Wickhampton Marshes, England; I50; Table 1) to fully evaluate their characteristics and usefulness. Overall, of the 65 primer pairs tested, 15 (23%) loci were polymorphic and easily scoreable ( Table 2). The remainder were monomorphic (18%), not useable due to stutter and scoring difficulty (31%) or had poor/no amplification (28%).
To estimate genotyping error, extraction and scoring for a proportion of individuals was repeated to compare the data. The mean scoring error was found to be 0.02% (calculated as per [18]). All of the 15 markers tested displayed more than 2 alleles in multiple individuals and all individuals tested displayed more than 2 alleles in several markers, suggesting S. latifolium is polyploid (for data, see Additional file 1). A maximum of 4 alleles were observed per individual indicating tetraploidy in this species (see Additional file 2). Characteristics of each microsatellite locus were calculated for S. latifolium samples using the R package polysat [19,20]. The number of alleles per locus ranged from 8 to 17 and the mean average number of alleles was 12 (Table 3). Observed heterozygosity per locus ranged from 0.88 to 1.00, with a mean average of 0.99 (Table 3). Due to polyploidy and unknown inheritance patterns, deviation from Hardy-Weinberg equilibrium could not be calculated nor could the frequency of null alleles be estimated [21].
Initial measures of genetic diversity were calculated for the genotyped population (Wickhampton Marshes) using the programme GenoDive [22]. In this population, the mean average number of alleles per locus was 9.13 and observed heterozygosity was 0.976. Genetic distances between individuals within the library population were calculated (Bruvo distance, R package polysat [20,23]) and visualised by ordination (R package Vegan [24]). The microsatellite markers revealed variation in the genetic distance between individuals within a single population and identified clusters of individuals with similar genotypes (Fig. 1).

Table 3 Characterisation of 15 dinucleotide microsatellite loci for the Greater Water Parsnip Sium latifolium, all tested on 24 individuals sampled at Wickhampton Marshes, reveals tetraploidy in this species
Microsatellite loci, expected and observed allele sizes (with the sequenced allele underlined*; bp) of individual from which the microsatellite sequences were isolated (individual I50, sampled at Wickhampton Marshes, Norfolk), number of individuals successfully genotyped (n), number of alleles (k), allele size range (bp), observed heterozygosity (Ho). Exp. I50 (bp), Expected allele size of I50, Obs. I50 (bp), Observed amplified allele sizes of individual, I50, *Minor size differences (bp) were observed between the expected size of the allele (based on sequencing) and observed allele size (based on ABI genotyping). This error is caused by (1) the presence of the fluorescent dye label (6FAM and HEX) and/or (2)

Conclusions
We have successfully developed the first set of microsatellite markers for S. latifolium. The 15 loci amplified reliably and have been shown to be sufficiently variable for distinguishing between individuals (Fig. 1). These will be helpful in providing a genetic context for planning and managing further reintroductions of S. latifolium. Additionally, using S. latifolium as an example species, these microsatellite loci will also be helpful in interpreting the effects of genetic diversity and source population composition on plant reintroductions. We also found each S. latifolium individual genotyped displayed 1-4 alleles. We conclude that this is evidence of tetraploidy, a trait not previously reported in this species. Polyploidy occurs occasionally through the Apiaceae family, in just over 10% of species [25]. In other species of Sium intraspecific variation in ploidy levels has been recorded, with local polyploid cytotypes found within a diploid species [26]. A chromosome count of 12 or 20 has been reported in S. latifolium [27]. As these previous cytological studies used specimens from continental Europe, the chromosomal characteristics of UK S. latifolium is unknown. Differences in records suggests that there may be variation within the species and all reported counts are a multiple of 4, indicating that tetraploidy is possible. Additional cytological analyses would also consider historical polyploidy or aneuploidy as causes of the multiple alleles observed. Further work on S. latifolium is needed to determine the nature of the ploidy (i.e. the inheritance type) and the patterns of ploidy throughout the species' geographic range. Authors' contributions ND performed DNA extraction, designed primers, optimised PCR reactions, selected and validated markers, conducted data analysis, and drafted the manuscript. GH constructed the microsatellite-enriched genomic library, designed primers and helped in optimising PCR reactions and interpreting results. DD participated in primer design, marker selection and validation, discussion of results, and revised all drafts of the manuscript. All authors read and approved the final manuscript.