Lindera obtusiloba Blume (Lauraceae) is a deciduous plant distributed in both northern and southern floral regions of the Tertiary relict flora in East Asia (Donoghue et al., 2001; Milne and Abbott, 2002). These two regions harbor two distinct L. obtusiloba genealogies that were probably triggered by the intermediate arid belt (Ye et al., 2017), providing a perfect system to investigate the floral subdivision of the East Asian Tertiary relict flora and the effect of the west-east–oriented arid belt. Only four chloroplast fragments and six nuclear microsatellites were used in Ye et al. (2017), limiting a detailed evolutionary history inference within each floral region. The nuclear microsatellites used in Ye et al. (2017) were designed for L. melissifolia (Walter) Blume (Echt et al., 2006) or L. benzoin (L.) Blume (Edwards and Niesenbaum, 2007); therefore, in this study, we aimed to design species-specific low-copy nuclear primers for L. obtusiloba.
Transcriptome sequences are widely used in studies of plant evolutionary history (e.g., Ai et al., 2015) and can be used for development of low-copy nuclear primers (Bai and Zhang, 2014). For example, Higashi et al. (2015) developed eight primers using 100 expressed sequence tag (EST) markers of Ericaceae, and the phytogeny of Shortia Raf. was inferred through these primers. In this study, the transcriptome data of L. obtusiloba were used to develop low-copy nuclear primers, and these primers were cross-amplified in other Lindera Thunb. species.
METHODS AND RESULTS
Two L. obtusiloba leaves were collected in the populations XRD and TMSH (Appendix 1) and used for transcriptome sequencing. Total RNA was extracted using the RNeasy Plant Mini Kit (QIAGEN, Hilden, Germany), and the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, Massachusetts, USA) was used to generate sequencing libraries. An index code was added to each sample. TruSeq PE Cluster Kit v3-cBot-HS (Illumina, San Diego, California, USA) on a cBot Cluster Generation System was used to cluster the index-coded samples. The Illumina HiSeq 2500 platform was used to sequence the libraries and generate paired-end reads. The raw reads were cleaned by removing reads containing adapters, reads including more than 10% unknown base information, and reads with low quality. All clean reads were assembled by Trinity (v2012-10-05) (Grabherr et al., 2011). The transcriptome data can be accessed in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) (NCBI Resource Coordinators, 2017) under accession numbers SRR5888830 and SRR5892454. In total, 191,545 unigenes were obtained, and unigenes greater than 800 bp in length were randomly chosen for initial design of 168 primers. We BLASTed these unigenes in nucleotide collection (nr/nt) database using MEGABLAST (optimized for highly similar sequences) in the NCBI database. The exon position, intron length, and putative function were justified by the gene information of the closest gene in the NCBI database. Primer pairs were designed in separate exon regions using Primer Premier 5 (PREMIER Biosoft International, Palo Alto, California, USA) under the following criteria: (i) size of primers 17–23 bp, (ii) annealing temperature (Ta) 45–64°C, (iii) Ta difference between primer pairs less than 4°C, (iv) primer pair score greater than 90, and (v) putative amplified product length less than 1200 bp.
PCRs were performed following the procedure in Ye et al. (2017) with adjusted annealing temperatures (Table 1). Agarose gel electrophoresis was used to select primers that generated only one clear band, and these primers were amplified in eight individuals. The amplicons were sequenced and then read in CodonCode Aligner 3.6.1 (CodonCode Corporation, Centerville, Massachusetts, USA; http://www.codoncode.com/aligner/). The loci with all nucleotide sites that exhibit fewer than two types of nucleotide variants were treated as low-copy nuclear loci. Low-copy nuclear loci were tested in 90 individuals sampled from 24 populations of L. obtusiloba (Appendix 2). After reading in CodonCode Aligner 3.6.1, PHASE function in DnaSP 5.10.01 (Rozas et al., 2003) was used to determine heterozygous and polymorphic sites, determine haplotypes, and to calculate genetic diversities, including nucleotide diversity (π) and haplotype diversity (Hd), of each locus. SPADS 1.0 (Dellicour and Mardulyn, 2014) was used to calculate haplotypes, π, and allelic richness in 24 populations. Genotypic disequilibrium was assessed using all locus pairs in all populations by randomization using FSTAT 2.9.3 with Bonferroni correction (Goudet, 2001). Local BLAST function in BioEdit 7.1.9 (Hall, 1999) was used to determine the intron and exon positions of all low-copy nuclear loci, and the unigenes for primer design were used as database ( Appendix S1 (apps.1700120_S1.txt)). Low-copy nuclear genes were cross-amplified in two individuals of four other Lindera species, including L. aggregata (Sims) Kosterm., L. chunii Merr., L. erythrocarpa Makino, and L. glauca (Siebold & Zucc.) Blume (Appendix 1).
Ninety-six of the 168 tested primers did not amplify or generated multiple bands, 45 produced messy sequences, and 27 produced clear sequences (Table 1). The product length of the 27 loci ranged from 154 to 944 bp. The number of polymorphic sites and haplotypes ranged from six to 71 and five to 49, respectively, with a mean of 27 and 19, respectively. In addition, π ranged from 2.11 × 10-3 to 8.99 × 10-3 with a mean of 6.06 × 10-3, and Hd ranged from 0.57 to 0.97, with a mean of 0.77 (Table 2). In the 24 populations, the number of haplotypes ranged from 39 to 76, π ranged from 0.76 × 10-3 to 1.80 × 10-3, and allelic richness ranged from 1.43 to 1.94 (Appendix 2). No significant genotypic disequilibrium was observed among 351 locus pairs. Fifteen primers were successfully amplified in L. aggregata and L. erythrocarpa, and 14 primers were successfully amplified in L. chunii and L. glauca (Table 2).
Given that information regarding exon position and intron sequence are not included in transcriptome sequencing, the success rate of primer development using transcriptome data would be expected to be low (Bai and Zhang, 2014). In this study, we developed 27 polymorphic primers out of a set of 168 primers, with a ratio of approximately 16%. The success rate is increased twofold compared with that of Higashi et al. (2015). This methodology provides an effective approach for the development of new low-copy nuclear primers.
Twenty-seven novel polymorphic low-copy nuclear primers were developed using transcriptome data from L. obtusiloba. These primers can be used to investigate the evolutionary history of L. obtusiloba and other Lindera species.
Genetic diversity and cross-amplification of the 27 Lindera obtusiloba low-copy nuclear loci.
The authors thank for S-H. Wang for sampling assistance. This work was supported by the National Natural Science Foundation of China (grant no. 31600301, 31210103911, and 31570381).