Taxus cuspidata Siebold & Zucc. (Taxaceae) is an evergreen needle-leaved tree with a straight trunk (up to 20 m). It grows at low density in mixed broadleaved forests throughout the cool temperate zone of Japan, Korea, northeastern China, and the extreme southeast of Russia (Hayashi, 1954). In contrast, its dwarf variety, T. cuspidata var. nana Hort. ex Rehder, is a small to medium-sized shrub (≤2 m) that locally dominates in alpine regions with heavy snowfall along the Sea of Japan (Hayashi, 1954), thus demonstrating a clearly disjunct distribution from that of T. cuspidata. Taxus species, including T. cuspidata and T. cuspidata var. nana, are well known as sources of paclitaxel, an anticancer metabolite first found in the bark of the Pacific yew (T. brevifolia Nutt.; Wani et al., 1971). Because paclitaxel has been difficult to synthesize on an industrial scale (Glowniak et al., 1996; Sottani et al., 2000), its production has continued to rely heavily on natural resources until relatively recently. Consequently, some Taxus species have been overexploited.

Although the overexploitation of T. cuspidata and T. cuspidata var. nana has not been reported, the assessment of the distribution of genetic resources formed through speciation and of species' range formation is essential to the long-term management of economically valuable natural resources. Being derived from transcripts, expressed sequence tag (EST)–simple sequence repeat (SSR) markers are useful for assaying functional diversity in natural populations (Varshney et al., 2005). Ueno et al. (2015) developed 80 EST-SSR markers for T. cuspidata; however, levels of diversity are often lower in EST-SSRs than in genomic SSRs. Although genomic SSR markers were developed in some relatives of T. cuspidata (e.g., Dubreuil et al., 2008; Cheng et al., 2015a), their utility for T. cuspidata was limited and the level of polymorphism in T. cuspidata populations was low (Cheng et al., 2015b). Therefore, we have developed highly polymorphic, genomic microsatellite markers to investigate the spatial genetic structure of T. cuspidata. This paper reports 15 genomic microsatellite markers developed for T. cuspidata by using next-generation sequencing technology and their utility for T. cuspidata var. nana.

METHODS AND RESULTS

Three T. cuspidata populations were sampled throughout the species range in Japan: Mt. Rausu, Hokkaido (the north-easternmost habitat in Japan; 44°04′54″N, 145°07′33″E); Mt. Kurai, Gifu Prefecture (36°02′30″N, 137°11′80″E); and Mt. Ohnogara, Kagoshima Prefecture (the southernmost habitat; 31°29′17″N, 130°49′10″E). One T. cuspidata var. nana population was also sampled on Mt. Hyono, Hyogo Prefecture (35°21′23″N, 134°30′81″E; Appendix 1). A single leaf from each of 10 adult trees in each population was collected. Total genomic DNA was isolated from ∼50 mg of leaf tissue from each tree by using the hexadecyltrimethylammonium bromide mini-prep procedure (Stewart and Via, 1993).

Approximately 200 ng of DNA extracted from one individual of T. cuspidata collected in Mt. Kurai was used for library preparation with a TruSeq Nano DNA Library Prep Kit (Illumina, San Diego, California, USA). Sequencing was performed on a MiSeq Benchtop Sequencer (Illumina) in 2 × 300-bp read mode. The data were assembled into contigs in fastq-join software (Aronesty, 2011). Microsatellite regions were mined among contigs of >400 bp in MSATCOMMANDER version 1.0.8 software (Faircloth, 2008). The search parameter was restricted to dinucleotide motifs with a minimum of 16 repeats. Primer pairs for microsatellite amplification were designed in Primer3 version 2.2.3 software (Rozen and Skaletsky, 1999) with default parameter settings.

Table 1.

Characteristics of 15 microsatellite loci for Taxus cuspidata and T. cuspidata var. nana.

PCR amplifications followed the standard protocol of the QIAGEN Multiplex PCR Kit (QIAGEN, Valencia, California, USA) in a final volume of 10 µL, which contained 5 ng of extracted DNA, 5 µL of 2× Multiplex PCR Master Mix, and 0.2 µM of each multiplexed primer. Forward primers were labeled with fluorochromes 6-FAM or VIC (Life Technologies, Carlsbad, California, USA). Amplification was performed in a Veriti Thermal Cycler (Life Technologies) under the following conditions: initial denaturation at 95°C for 15 min; 30 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 1 min 30 s, and extension at 72°C for 1 min; and a final extension at 60°C for 30 min. The size of the PCR products was measured in an ABI PRISM 3130XL Genetic Analyzer (Life Technologies) by GeneMapper software (Life Technologies). Because microsatellite markers with high numbers of repeat units may cause unstable results due to PCR error (Shinde et al., 2003), we conducted PCR amplifications twice in four randomly selected individuals.

Table 2.

Genetic variation of the 13 polymorphic microsatellite loci for three populations of Taxus cuspidata and one population of T. cuspidata var. nana.

The genetic polymorphism at each locus was assessed by calculating the observed number of alleles, observed heterozygosity (H_o) and expected heterozygosity (H_e). Genotypic linkage disequilibrium was tested for all combinations of locus pairs within a population by using the Markov chain method provided by Web version 4.2 of GENEPOP software (Raymond and Rousset, 1995). Significance values were computed for each population by using Fisher's method for combining independent test results. The significance of deviations from Hardy– Weinberg equilibrium (HWE) in each population, represented by the deviation of the fixation index from 0, was tested by 1000 random permutations of alleles in each population at each locus in FSTAT version 2.9.3 (Goudet, 1995). Bonferroni's correction was applied to all pairwise test results to adjust for multiple comparisons. The potential presence of null alleles was assessed in MICRO-CHECKER version 2.2.3 software (van Oosterhout et al., 2004), using the second method of Brookfield (1996) to calculate the expected frequency of null alleles.

One individual of T. cuspidata provided a total of 2046 Mbp and 6,796,562 reads. A total of 1,797,717 contigs were assembled, and dinucleotide motifs with a minimum of 16 repeats were identified in 3981 contigs. Screening of 24 randomly selected loci with a minimum of 16 repeats identified 15 loci with a clear, strong, single band for each allele (Table 1). Of the 15 loci, 13 were polymorphic and two were monomorphic. PCR error was not observed. The evaluation of polymorphism in the 30 adult trees showed that the 13 polymorphic loci were hypervariable (Table 2), with 11 (TC43389) to 31 (TC48340) alleles per locus (average: 18.5), and H_e from 0.78 (TC99217) to 0.95 (TC48340) (average: 0.89). At the population level, number of alleles ranged from four to 16 (average: 8.8), H_o from 0.00 to 1.00 (average: 0.70), and H_e from 0.64 to 0.93 (average: 0.80). Null alleles were significant at two loci (TC43389 and TC63749), at a frequency of 0.21 and 0.20, respectively. Among all loci in the three full-size populations, eight of 39 combinations deviated significantly from HWE (P < 0.01, Table 2). This lack of equilibrium could be explained by the small number of samples in each population. There was no evidence of significant linkage disequilibrium (P < 0.05) in any pair of loci. All 15 loci were successfully amplified in the T. cuspidata var. nana population and showed high polymorphism (Table 2).

CONCLUSIONS

In this study, 13 novel polymorphic microsatellite markers for T. cuspidata and T. cuspidata var. nana were developed. These microsatellite markers will be useful for investigating speciation and range formation of T. cuspidata and T. cuspidata var. nana in Japan, and the results will provide crucial information for conservation of Taxus species as sources of antitumor agents.