Development of 15 nuclear microsatellite markers in Deuterocohnia (Pitcairnioideae; Bromeliaceae) using 454 pyrosequencing

Premise of the Study Microsatellite markers were developed in Deuterocohnia longipetala (Bromeliaceae) to investigate species and subspecies boundaries within the genus and the genetic diversity of natural populations. Methods and Results We used 454 pyrosequencing to isolate 835 microsatellite loci in D. longipetala. Of 64 loci selected for primer design, 15 were polymorphic among 23 individuals of D. longipetala and 76 individuals of the heterologous subspecies D. meziana subsp. meziana and D. meziana subsp. carmineo‐viridiflora. Twelve and 13 of these loci were also polymorphic in one population each of D. brevispicata and D. seramisiana, respectively. Numbers of alleles per locus varied from two to 14 in D. longipetala, two to 12 in D. meziana, one to nine in D. brevispicata, and one to 10 in D. seramisiana. STRUCTURE analyses clearly separated the taxa from each other. Conclusions The 15 new microsatellite markers are promising tools for studying population genetics in Deuterocohnia species.


of 7
The genus Deuterocohnia Mez (Bromeliaceae) includes 17 species that are mainly distributed in the Andes of central South America (Schütz, 2013). It comprises terrestrial or saxicolous plants with thorny leaves in dense rosettes, giving rise to woody, perennial inflorescence axes that are able to bloom for several years (Smith and Downs, 1974;Benzing, 2000). All species are adapted to extremely arid environments such as steep and rocky slopes of the Andes and inter-Andean valleys, but some also grow on rocky outcrops in lowlands of eastern Bolivia and western Brazil (Schütz, 2013(Schütz, , 2014. Species delimitation within Deuterocohnia is often difficult due to hybridization among closely related species and subspecies (Schütz, 2014). Considering data of floral display, seed and floral morphology, and pollinators (Benzing, 2000), it seems that species from Deuterocohnia may present a variety of characteristics related to outcrossing. So far, this reproductive system was previously reported for D. meziana Kuntze ex Mez, which is self-incompatible and clonal (Arruda, 2016), and has winged seeds adapted for longdistance dispersal (Schütz, 2014).
To date, very little is known about the genetic diversity and population structure in any Deuterocohnia species. However, this information is important for endangered species like D. meziana (Ministério do Meio Ambiente, 2014;Schütz, 2014). It can contribute to our understanding of microevolutionary processes of natural populations, assist in the delimitation of species and subspecies (Palma-Silva et al., 2011), and help to detect hybridization (Zanella et al., 2016) and to design management and conservation strategies (Ribeiro et al., 2013). Here, we present 15 polymorphic microsatellite loci developed for the genus Deuterocohnia using 454 pyrosequencing technology.

METHODS AND RESULTS
Total DNA was extracted from fresh leaves following the protocol of Tel-Zur et al. (1999). The source DNA for 454 sequencing was derived from one individual plant of D. longipetala (Baker) Mez that was collected along the road from Bermejo to Limal (Bolivia) and that is now cultivated in the greenhouse of the University of Kassel (accession NiSch_06-068; Appendix 1). We chose this species for microsatellite isolation and primer design because it has the widest distribution range of any Deuterocohnia species (Schütz, 2013). Library preparation and pyrosequencing of a 5-μg DNA aliquot were performed as described by Wöhrmann et al. (2012). Using default settings, 25,827 raw reads with an average length of 337 bp were obtained and imported into the pipeline iQDD (version 1.3; Meglécz et al., 2010); these sequences were also submitted to the National Center for Biotechnology Information's Sequence Read Archive (accession no. SRP126618). From those sequences, we identified 835 perfect repeats with a minimum of seven units for di-, six for tri-, five for tetra-, and four for penta-and hexanucleotide repeats, respectively. Sixty-four microsatellite loci with sufficient flanking sequence and high repeat numbers were selected for PCR primer construction (Appendix 2), following previously described criteria (Wöhrmann et al., 2012).
All primer pairs were initially tested for successful amplification in two individuals each of D. meziana subsp. carmineo-viridiflora Rauh (NiSch_06-007J, NiSch_06-007M) and D. brevispicata Rauh & L. Hrom. (NiSch_06-040F, NiSch_06-040M), as well as in one individual each of D. seramisiana R. Vásquez, Ibisch & E. Gross (NiSch_06-045K) and D. longipetala (NiSch_06-068 as a positive control). PCRs were conducted in 12.5-μL volumes in a T-Gradient thermocycler (Biometra, Göttingen, Germany) following a touchdown protocol (Wöhrmann et al., 2012). As evidenced by electrophoresis on 1.5% agarose gels, 52 of the 64 primer pairs generated single, distinct PCR products within the expected size range in the positive control (Appendix 2). Forty-seven primer pairs also performed well in one or more accessions from other Deuterocohnia species, and only 12 loci failed in all samples (Appendix 2). Of 22 primer pairs that amplified in all individuals of the test set, 15 were validated by genotyping the full set of 129 samples listed in Appendix 1 (for locus characteristics see Table 1). Fluorescencelabeled primers were used for PCR, and amplicons were electrophoresed on denaturing 6% polyacrylamide gels in 1× TBE buffer, using an automated sequencer (Li-Cor 4300 IR 2 ; Li-Cor Biosciences, Lincoln, Nebraska, USA). Fragment sizes were scored with the help of an external size standard as described by Wöhrmann et al. (2012).
Population genetic parameters are compiled in Table 2. Allele numbers as well as observed (H o ) and expected (H e ) heterozygosity values were determined with ARLEQUIN version 3.11 (Excoffier et al., 2005). Wright's inbreeding coefficients (F IS ) and deviations from Hardy-Weinberg equilibrium (HWE) were calculated with GENEPOP (Raymond and Rousset, 1995). All 15 loci proved to be polymorphic in D. longipetala and in D. meziana, whereas three and two loci, respectively, were monomorphic in D. brevispicata and D. seramisiana. Altogether 80 alleles were detected in 23 individuals of D. longipetala from various localities, showing mean heterozygosity  (Table 2).
To evaluate the potential of microsatellite markers for distinguishing between closely related taxa, a Bayesian cluster analysis was performed on a set of 129 plants comprising all samples from the two subspecies of D. meziana, D. brevispicata, D. seramisiana, and D. longipetala, using the program STRUCTURE version 2.3.4 (Pritchard et al., 2000). For the determination of the most appropriate number of genetic clusters (K value), the analysis was run for 1,000,000 generations in the burn-in period and for 100,000 generations in the Markov chain Monte Carlo simulation analyses after burn-in. Ten repetitions for each K (1 ≤ K ≤ 10) were performed, and the admixture level for each individual (Q) was also inferred. By calculating the ΔK statistic using STRUCTURE HARVESTER version 0.6.94 (Earl and von Holdt, 2012), the most likely number of clusters was identified to be four, closely followed by two and five (Fig. 1). Final plots were visualized in STRUCTURE PLOT version 2.0 (Ramasamy et al., 2014). For the three estimates of K (2, 4, and 5), there is a clear division among one cluster composed of all D. meziana subsp. meziana samples (Fig. 2). For K = 4, there is a second cluster containing all D. meziana subsp. carmineo-viridiflora plants, a third cluster combining all samples from D. brevispicata and D. seramisiana, and a fourth  containing all samples from D. longipetala (Fig. 2, middle panel). Assuming K = 5, D. brevispicata and D. seramisiana also become clearly separated from each other (Fig. 2, lower panel).

CONCLUSIONS
The 15 microsatellite markers developed from 454 sequences of D. longipetala revealed moderate levels of genetic diversity in the source species as well as in three heterologous Deuterocohnia taxa investigated. Whereas the two subspecies of D. meziana were surprisingly well separated from each other, the distinction between D. brevispicata and D. seramisiana was less pronounced, suggesting some ongoing gene flow among populations of these two species. The microsatellite markers developed here are promising tools for the study of population genetics, phylogeography, and the cohesion and delimitation of species and subspecies in Deuterocohnia. Genetic data generated by these markers will also provide important guidelines for designing management and conservation strategies in endangered species like D. meziana. Note: + = one distinct PCR product observed on 1.5% agarose gels; (+) = weak bands; -= no PCR product observed; # = number of accessions for which a successful PCR amplification could be detected. a Success of PCR amplification in a test set consisting of six Deuterocohnia individuals: 1 = NiSch_06-007J, 2 = NiSch_06-007M (both individuals of D. meziana subsp. carmineo-viridiflora); 3 = NiSch_06-040F, 4 = NiSch_06-040M (both individuals of D. brevispicata); 5 = NiSch_06-045K (D. seramisiana); 6 = NiSch_06-068 (D. longipetala as positive control).
*Primer pairs used in the present study (see also Tables 1 and 2).