Nine from 20 D. distachya loci amplified robustly and proved to be polymorphic. Markers isolation involved the development of a genomic enriched library. A total of 96 positive clones were captured and 48 colonies were sequenced, of which 24 clones showed simple sequence repeats (SSR) motifs. Twenty primer pairs were designed in which the product size ranged between 100 and 300 base pairs. Nine of these primer pairs generated consistent patterns of amplification and were used for further characterization of two D. distachya populations. The primer sequences and product sizes are shown in Table 1.

In the introduced population, the number of alleles per locus ranged from 4 to 10, with an average of 6.2, and the expected and observed heterozygosities ranged from 0.136 to 0.667 and from 0.543 to 0.877, respectively. In the wild population, the number of alleles per locus varied from 2 to 8, with an average of 3.8, and the expected and observed heterozygosities ranged from 0.000 to 0.895 and from 0.050 to 0.811, respectively (Table 2). Four loci showed significant departures from the Hardy-Weinberg equilibrium (HWE) due to heterozygote deficiency in both populations (Dd04, Dd08, Dd09 and Dd20; Table 2). Such significant deviation from HWE could have been caused by the involvement of null alleles. We tested the presence of null alleles and the four loci that showed significant departures from HWE in both populations exhibited high null allele frequencies (data not shown). Other three loci showed significant departures from HWE in only one population (p < 0.001; Table 2) and did not show null alleles, being the departures probably due to inbreeding or Wahlund effects.

Cross-species amplification was successful in different degrees (Table 3). All loci amplified in species from same genus (except Dd19 for D. maritima). Locus Dd10 and Dd20 amplified in all species tested, and locus Dd07 and Dd16 were successful in 86.3% of the species evaluated. The 12 genus tested correspond to three representative Bromeliaceae subfamilies from Atlantic Rainforest, indicating high potential for use of these four loci in phylogeography and population genetic studies of many Bromeliaceae species, mainly for Dyckia genus.

In our study, introduced population showed, on average, a higher number of alleles per locus, and higher observed and expected heterozygosities compared with the wild population (Table 2). The results could indicate that the individuals rescued may be representative of the original population, retaining the genetic diversity, although further studies should be made for a more accurate conclusion. The introduced population presented 25 exclusive alleles of a total 60 alleles (data not show). The wild population showed only three exclusive alleles and low genetic diversity. The microsatellites characterized here will be used to assess more thoroughly the genetic diversity of natural and introduced populations. However, one hypothesis for lower genetic diversity found in natural population would be the edge effect [5], since this population occurs at the range edge of the original geographical distribution of the species. This phenomenon has also been observed in Vriesea gigantea, another species of bromeliad [6].

To identify and characterize microsatellites, genomic DNA was extracted from fresh leaves of one D. distachya individual following the protocol of Doyle and Doyle [7]. Microsatellite markers were developed as described by Kijas et al. [8] with modifications by Billote et al. [9], using an enriched library methodology. In brief, total DNA was cleaved with the restriction enzyme RsaI (Invitrogen, Carlsbad, CA, USA). The cleaved DNA was linked to two complementary oligo adapters. Polymerase chain reaction (PCR) was performed using one of the adapters as a primer. The PCR product was purified using a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany). Fragments that contained microsatellite repeats were selected using biotinylated probes (dCT)₈ and (dGT)₈ and isolated with Dynabeads M-280 Streptavidin magnetic particles (Invitrogen, Carlsbad, CA, USA). An additional PCR was performed using one of the adapters as a primer. The PCR products were inserted into the pGEM-T Easy vector (Promega, Madison, WI, USA) and used to transform XL1-Blue competent cells (Escherichia coli, Stratagene, La Jolla, CA, USA). Recombinant colonies were selected by blue/white screening. Ninety-six recombinant colonies were randomly selected, and 48 were double-sequenced using the BigDye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) on an ABI PRISM 377 DNA Sequencer (Applied Biosystems, Foster City, CA, USA). Every the forward and reverse sequences obtained were aligned and edited in SeqMan (DNASTAR package, Madison, WI, USA), and primers were designed using the Primer3 program [10].

To assess polymorphisms, 22 individuals from one D. distachya population introduced in Anita Garibaldi, Santa Catarina State, Brazil (27°58′81″S, 50°39′11″W) and 21 individuals from the wild population were genotyped. The wild population was collected in Salto do Yucumã, in Turvo State Park, Derrubadas, Rio Grande do Sul State (27°09′62″S, 53°53′45″W), southern edge of species original distribution and the introduced population came from the ex situ collection.

For each SSR, the forward primers were synthesized with a 19-bp M13 tail (5′-CACGACGTTGTAAAACGAC-3′) following the method of Schuelke [11], which involved three primers: a forward SSR-specific primer with the M13 tail at its 5′ end, a reverse locus-specific primer, and a universal M13 primer labeled with one of four fluorescent dyes (6-FAM, VIC, NED, and PET; Applied Biosystems, Foster City, CA, USA). All PCR amplifications were performed in a TC-412 thermal cycler (Techne, Burlington, NJ, USA) in 10 μL reactions containing 10 ng of DNA template, 1× Taq buffer, 2 mM MgCl₂, 200 μM of each dNTP, 1 pmol forward primer, 4 pmol reverse primer, 0.4 pmol universal M13 primer, and 0.5 U Taq DNA polymerase (GoTaq; Promega, Madison, WI, USA). The touchdown cycling program was used: 95 °C for 3 min, 10 cycles of 94 °C for 30 s, 58 °C decreasing to 48 °C at 1 °C per cycle for 30 s, 72 °C for 30 s, followed by 30 cycles of 94 °C for 30 s, 48 °C for 30 s, and 72 °C for 30 s, and concluding with a 10 min extension at 72 °C. The microsatellite alleles were resolved on a ABI 3100 DNA Analyzer Sequencer (Applied Biosystems, Foster City, CA, USA) and sized against the GS500 LIZ molecular size standard (Applied Biosystems, Foster City, CA, USA) using GeneMarker DEMO version 1.97 (SoftGenetics, State College, PA, USA). The number of alleles, and expected and observed heterozygosities were analyzed with MSA 4.00 software [12], and Genepop 4.1 [13] was used to test for departure from the Hardy-Weinberg equilibrium (HWE). Null allele frequencies were calculated following [14] for each population, using the software Micro-Checker version 2.2.3 [15].

The polymorphic markers were also tested for cross-transferability in 22 related species from three subfamilies of Bromeliaceae. The PCR conditions were the same as described above. The amplification products were separated on 2% agarose gels (Agarose LE; USB, Cleveland, OH, USA) stained with GelRed (Biotium, Hayward, CA, USA) and visualized under ultraviolet light. A locus was considered to have been amplified successfully when at least one band of the expected size was visualized. A 50-bp DNA ladder (Ludwig Biotec, Alvorada, RS, Brazil) was used as a molecular size marker. All voucher specimens of the species investigated are identified in Table S1.