Development of Microsatellite Markers

The microsatellite markers were developed using the magnetic bead protocol described byCullings (1992) and (Li, Huang, & Brown, 1997) and modified byGlenn and Schable (2005). Genomic DNA from a sample of five G. arborea trees was digested using HaeIII/PshA1 restriction enzymes (Invitrogen; Carlsbad, CA). Two linkers were added to the digested genomic DNA (M28 5’CTCTTGCTTGAATTCGGACTA 3’ and M29 5’pTAGTCCGAATTCAAGCAAGAGCACA 3’) and M28 was used as a primer for subsequent polymerase chain reactions (PCR). Finally, the digested genomic DNA was amplified in multiple PCR reactions and their product concentrated to gain enough DNA for the following bead hybridization process.

Two arbitrary repeat motifs (CA20 and AG17) were selected as probes for the bead hybridization reactions based uponCardle et al. (2000). The short tandem repeat (STR) probes from Integrated DNA Technologies (Coralville, IA, USA) had a biotin label on the 5’ end. The STR probes were added to a bead hybridization reaction to select for DNA fragments that contained the repeat motif of the probe. This bead hybridization process aimed to allow the fragments containing repeats to anneal to the biotin-labeled probes. After the hybridization, the selected fragments were isolated from the rest of the genomic DNA using streptavidin-coated magnetic beads, which bind to the biotin-labeled probes. These fragments were then eluted and re-amplified using the M28 primer in additional PCR reactions. The bead hybridization and PCR pre-amplification processes were repeated one more time to enrich for genomic DNA containing the selected repeats. 

After completing the bead hybridization and selection process, the repeat sequences enriched DNA was ligated into a pGEM-T vector from Promega (Madison, WI, USA) to begin the sequencing phase of this protocol. We cloned the vectors into electrocompetent Escherichia coli cells. We later plated transformed E. coli cells onto selective media containing 0.1 mg/mL ampicillin, 0.05 mg/mL X-Gal, and 1mM IPTG. All positive clones were sequenced on an ABI PRISM 377 DNA Sequencer using universal M13 forward (F) and reverse (R) primers (Schuelke, 2000). The sequencing reactions were standard 20 ml reactions using the ABI PRISM BigDye Terminator sequencing kits (Applied Biosystems, Foster City, CA, USA) and 3.2 pmol of PCR product for the template. Primers for each of the fifteen microsatellite loci were designed from sequences containing multiple copies of the repeated motif and with sufficiently long flanking regions on the 5’ and the 3’ end of the repeated region pairs using Primer 3.0 software (Rozen & Skaletsky, 2000).

Microsatellite Loci Characterization

All primer pairs were tested for amplification and polymorphism using DNA obtained from 23 promising genotypes (clones) of G. arborea belonging to five different privately operated clonal breeding programs. Two ramets from each clone were gathered from a clonal collection maintained in a greenhouse at the Instituto Tecnológico de Costa Rica to validate all alleles by genotyping them separately. As described above (Doyle & Doyle, 1987; Lodhi, Ye, Weeden, & Reisch, 1994), total genomic DNA was extracted at the Forest Molecular Genetic Laboratory, in the Forest Innovation Research Center (CIF) at the Instituto Tecnológico de Costa Rica, Cartago, Costa Rica. Copies of these clones are maintained in the mini clonal garden facility and could be made accessible upon request. 

PCR Amplification and Fragment Analysis

Polymerase chain reactions were performed in a final volume of 15 μl, containing approximately 50 ng of genomic DNA, 10 mM Tris buffer, pH 8.0, 10 mM MgCl₂, 0.2 mM dNTPs, 0.4 μM of each primer, and 1 U of Taq polymerase (Fermentas®) using an Eppendorf® Mastercycler EP thermal cycler. The PCR program used included an initial step of 2 min of denaturation at 94 °C, 30 cycles of 15 s at 94 °C, 15 s at 55 °C and 30 s at 72 °C, and a final extension cycle of 1 min at 72 °C. To genotype each individual, we conducted electrophoresis for fragment separation using a QIAxcel Advanced fragment analyzer from QIAGEN® at Centro de Investigación en Biología Celular y Molecular (CIBCM) at Universidad de Costa Rica. Once all of the data scorings were complete, random samples were re-amplified and re-ran to assess reproducibility and confirm scoring and allele sizes.

Genetic analysis

GenAlex 6.3 (Peakall & Smouse, 2006) was used to calculate common indicators of genetic diversity, including the number of alleles (N_a) per locus and the expected (H_e) and observed heterozygosity (H_o). GenAlex was also used to calculate deviations from Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium. Genotype errors due to stutter bands, allele dropout, and null alleles were estimated using the MICRO-CHECKER software (Oosterhout, Hutchinson, Wills, & Shipley, 2004).

To examine the potential of these loci for discrimination among the 23 clones, the multilocus genotype of each clone was determined using the presence and absence of alleles to estimate genetic similarities for all pairwise comparisons among clones. Genetic similarity among each pair of clones was calculated based on the number of alleles common among the clones according to the following equation proposed by Dice (1945), where $GSxy = 2a/(2a + b + c)$ , where a is the number of alleles common to clones x and y, b the number of alleles present only in clone x, and c the number of alleles present only in clone y. A cluster analysis based on sequential, agglomerative, hierarchical, and nested clustering methods (SAHN, UPGMA; NTSYS-pc-p package; (Rohlf, 1993) was conducted to describe the relationship between the clones.

Microsatellite loci

Table 1 lists the loci names, corresponding accession numbers in Genbank, repeated motifs, forward and reverse primer sequences, the size range of PCR products, and annealing temperatures for each of the fifteen microsatellite loci isolated for Gmelina arborea. All loci were polymorphic, with the number of alleles per locus ranging from 2 to 7 (Table 2). We found 75 different alleles across all loci (Supplemental Table 1), with an average of 5.00 ± 0.41 alleles per locus. Average observed and expected heterozygosities were also high (H_o = 0.504 and H_e =0.645, Table 2). Moreover, our findings did not show evidence of scoring error due to stuttering or significant allele dropout for any of the fifteen polymorphic loci.

Our analyses revealed significant deviations from Hardy-Weinberg proportion in most loci (Table 2). We observed heterozygote deficiencies in eleven loci and an excess of heterozygotes in one locus (Meldi-12; Table 2). However, given the small sample size used to validate these loci and the high number of alleles found in most of them, it is reasonable to expect that they will not be in Hardy-Weinberg equilibrium (HWE). Moreover, the clones used to validate these microsatellite loci do not represent a sample of a natural population of G. arborea, but a collection of promising genotypes selected by the timber industry. We also caution that two loci; namely, Meldi11 and Meldi11.2, which include different tandem repeats, were derived from the same sequence.

Our analysis using the software MICRO-CHECKER did not reveal evidence for genotype errors due to stutter bands or allele dropout. Our analysis suggested the presence of null alleles in nine loci (Table 2), but such findings might result from a deviation from Hardy-Weinberg proportions. MICRO-CHECKER uses deviations of Hardy-Weinberg proportions to identify loci likely to have null alleles. We need to reiterate that our sample did not represent a natural population of G. arborea. For that reason, deviations from Hardy-Weinberg proportions are likely to occur in multiple loci.

Our results also showed that all 23 clones exhibited a unique combination of alleles (Supplemental Table 1), resulting in genetic similarities (Dice) ranging from 0.36 to 0.83 (Figure 1). Overall, most of the clones clustered according to their origin or breeding program. All clones from programs PC and CA clustered together while some clones from programs MC, N, and T grouped with clones from the other programs.