Protorhus deflexa H. Perrier (Anacardiaceae) is a dioecious tree that can reach 10–20 m in height (Schatz, 2001) and is found in tropical dry forests in western Madagascar (Sato, 2012). Based on the phylogenetic analysis of Randrianasolo (2003), Malagasy species of Protorhus Engl. will be formally validated as a new endemic genus, Abrahamia Randrian. & Lowry. From the end of the dry season to the beginning of the rainy season, P. deflexa blossoms and insects including bees visit the flowers (H. Sato, personal observation). Due to its breeding system without self-fertilization, this plant depends on insects for cross-pollination. During the middle of the rainy season, P. deflexa bears reddish fruits containing a large seed (Sato, 2012). Given this large seed size, large-bodied frugivorous lemurs of the genus Eulemur are the only effective seed dispersers of P. deflexa (Sato, 2012).
The vulnerability of animal-mediated gene flow in plants to human-induced disturbance has been pointed out because habitat destruction and hunting can decrease the densities of pollinators and seed dispersers (Corlett, 2007). Given the extinction crisis for Malagasy primates, seed dispersal of large-seeded plants including P. deflexa seems to be one of the most vulnerable systems. Although most of the forested areas and fauna in the Malagasy forest are threatened by human activity, we have a poor understanding of the negative impacts on gene flow via the failed services of animals for each plant. In recent years, genetic analyses using microsatellite markers have successfully demonstrated the critical roles of pollinators and seed dispersers in gene flow in plant populations (Ashley, 2010). However, because such efficient markers have not been available in P. deflexa and even in congeneric species, it is necessary to isolate a large number of microsatellite loci in this plant species. In this study, we applied shotgun 454 pyrosequencing, which is a more efficient approach for isolation of microsatellites at a fraction of the cost and effort compared to traditional Sanger methods (Zalapa et al., 2012), to develop 19 nuclear microsatellite markers for P. deflexa.
METHODS AND RESULTS
Leaf samples were collected from 20 adult trees (>5 cm dbh) of P. deflexa (identified using specimen accession numbers 3505 and 14728, Parc de Tsimbazaza [TAN] herbarium, Madagascar) in the primary dry deciduous forest of Ampijoroa Forestry Station (16°31′S, 46°82′E) in Ankarafantsika National Park. Genomic DNA was extracted from the dried leaf tissues of each individual using the modified cetyltrimethylammonium bromide (CTAB) extraction protocol of Murray and Thompson (1980). A DNA library was prepared with one individual sample of P. deflexa using a GS Junior Titanium Series Kit (Roche Diagnostics, Mannheim, Germany). A 500-ng aliquot of genomic DNA was nebulized at 0.24 MPa for 1 min. The DNA fragments were end-repaired, A-tailed, ligated to the Rapid Library Adapter (Roche Diagnostics), and suitably sized by removing short fragments (<350 bp) using an SPRIworks Fragment Library System II Kit (Beckman Coulter, Brea, California, USA). The quality and quantity of the DNA fragments were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany, USA). The fragments were then mixed with capture beads and amplified through emulsion polymerase chain reaction (emPCR) using a GS Junior Titanium emPCR Kit (Roche Diagnostics). After emPCR, the beads were collected, and those capturing the DNA library fragments were enriched before annealing with sequencing primers. The amplified fragments were sequenced using the GS Junior benchtop system (Roche Diagnostics). Because the first round of sequencing failed to yield sufficient reads, sequencing was conducted twice. A total of 112,363 DNA sequence reads were obtained (first: 22,814 reads; second: 89,549 reads).
Characteristics of 19 microsatellite markers developed in Protorhus deflexa.
To search for potential microsatellite loci, including dinucleotide and trinucleotide loci of at least seven and four repeats, respectively, the sequences were screened using MSATCOMMANDER (Faircloth, 2008). A total of 972 primer pairs, including 211 dinucleotide repeats and 761 trinucleotide repeats, were designed by the default setting of the Primer3 program embedded in MSATCOMMANDER, using the following settings: primers designed to amplify regions of 100–500 bp, an optimal oligo melting temperature range of 57–62°C, GC content range of 20–80% with an optimum rate of 50%, low levels of self- or pair-complementarity, and a maximum end-stability (AG) of 8.0 (Faircloth, 2008). Based on the structure of the repeat, 67 primer pairs (49 dinucleotide and 18 trinucleotide loci) were selected for the initial screening of microsatellites using four individuals of the sampled 20 trees of P. deflexa. To avoid labeling individual primers, an M13 tail (5′-GTTGTAAAACGACGGCCAGT-3′) was added to the 5′ end of each forward primer and labeled (Schuelke, 2000). The reaction mixture had a final volume of 5 µL, which included 10 ng of template DNA, 0.05 U of LA-Taq DNA polymerase (TaKaRa Bio Inc., Otsu, Shiga, Japan), 2.0 µM of GC Buffer I (TaKaRa Bio Inc.), 400 µM of each dNTP, 0.25 µM M13-tailed forward primer, 0.5 µM reverse primer, and 0.5 µM FAM-labeled M13 primer. The amplification profiles included an initial denaturation for 5 min at 94°C; followed by 30 cycles of 30 s at 94°C, annealing for 45 s at 55°C, 57°C, or 60°C (Table 1), and extension for 45 s at 72°C; incorporation of the fluorescent dye into the PCR product followed by eight cycles of 30 s at 94°C, 45 s at 53°C, and 45 s at 72°C; and a final extension for 15 min at 72°C (Schuelke, 2000). The product sizes were measured using an ABI PRISM 3130xl Genetic Analyzer and Peak Scanner software (Applied Biosystems, Foster City, California, USA). Primers with a monomorphic locus (13 pairs) and primers that could not amplify over half of the four samples (35 pairs) were removed from the marker set. A final set of 19 successful polymorphic markers was used to genotype 20 unrelated adult trees of P. deflexa (Table 1).
The number of alleles, observed and expected heterozygosities (Ho and He), and probability of exclusion (PE) were calculated using GenAlEx version 6.5 (Peakall and Smouse, 2012). The number of alleles per locus ranged from two to nine (mean: 4.6); the ranges of Ho and He were 0.200–0.800 (mean: 0.484) and 0.303–0.821 (mean: 0.565), respectively (Table 2). PE over all loci reached 0.98583 for the first parent and 0.99971 for the second parent, whereas PE for excluding a putative parent pair was greater than 0.99999. These values of PE reached a level high enough to detect mating system and gene flow of P. deflexa. The null allele frequency was determined for all loci using FreeNA (Chapuis and Estoup, 2007). Because the null allele frequency was <0.2 for all loci except Adf11 and Adf19 (Table 2), the results of analyses using those loci may not be changed significantly by null alleles (Latinne et al., 2011). Linkage disequilibrium (LD) between pairs of loci was tested using GENEPOP version 4.0 (Rousset, 2008). There were no pairs with significant LD after Bonferroni correction (P > 0.00029).
Genetic properties of the newly developed 19 microsatellites of Protorhus deflexa.
 The authors thank S. Ichino, M. Rasolofomanana, F. Rakotondraparany, L. L. Raharivony, M. Nakamura, A. Mori, and the staff at Ankarafantsika National Park for their support. This work was partially supported by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) Grants-in-Aid for Scientific Research (No. 21405015 to Y.T., 21310150 and 25290082 to M.I.-M., and 25870344 to H.S.) from the Japan Society for the Promotion of Science, and by the Cooperation Program of Wildlife Research Center, Kyoto University.