Introduction
Camellia oleifera is one of the most important healthy oil-bearing tree species. It has the highest total yields and acreage in China (Zhuang, Reference Zhuang2008). It is very important to increase the yield of C. oleifera to meet Chinese people's increasing requirement for edible oils. Variety improvement through molecular breeding, a promising method to breed elite varieties with precision and efficiency (Philips, Reference Philips2006; Xu, Reference Xu2010), is the key factor to increase yields (Chen et al., Reference Chen, Yang, Peng, Li, Wang and Duan2005; Zhuang, Reference Zhuang2010; Shi et al., Reference Shi, Yin and Shi2011; Tan et al., Reference Tan, Ma, Li, Yao, Dong, Su, Mao, Li and Yuan2012).
Microsatellite is a kind of codominant, multi-allelic, even covering and high genetic informational marker (Powell et al., Reference Powell, Machray and Provan1996; Varshney et al., Reference Varshney, Graner and Sorrells2005a), therefore is very useful in molecular breeding (Chen et al., Reference Chen, Dai, Hou, Guan, Wang, Li and Yin2016; Wang et al., Reference Wang, Yang, Zhang, Qin, Wang, Liao, Wang and Gao2017). Microsatellite markers of C. oleifera have been developed from the transcriptomes of several tissues, e.g. shoots, leaves, buds and flowers, through 454 GS-FLX platform (Xia et al., Reference Xia, Jiang, Huang, Zhang, Zhang and Gao2014) or Illumina HiSeq 2000 platform (Chen et al., Reference Chen, Yang, Huang, Duan, Long and Rong2017). However, these microsatellite sites were not validated by experiments. Microsatellites in the transcriptomes of leaves and seeds were also detected, then used in genetic diversity (Li et al., Reference Li, Wang, Ding, Chen, Hu, Li and Wei2017) and transferability of these microsatellites across two relatives were evaluated (Jia et al., Reference Jia, Lin, Zhang, Tan, Lei, Hu and Shao2014, Reference Jia, Lin, Feng, Hu, Tan, Shao and Zhang2015).
The formation of different tissues was determined by differential expression and interaction of large numbers of genes (Slovak et al., Reference Slovak, Ogura, Satbhai, Ristova and Busch2016; Drapek et al., Reference Drapek, Sparks and Benfey2017). The deficiency of markers from root transcriptomes may limit the molecular breeding of root-related traits. Hence, here we developed a novel set of microsatellites from root transcriptomes of C. oleifera, and then reported the transferability of these markers to its relatives in genus Camellia.
Experimental
The root samples were collected from tissue cultured plants (‘Cenruan No.3’) at the three key time points, i.e. the second day after shoot induced in rooting medium (T1), the 14th day after induced (root point appearance) (T2) and the 50th day after induced (root growth) (T3). The RNA extraction, cDNA library construction and Illumina paired-end sequencing were carried out. After removing low-quality reads, adaptors, possible contaminations and ribosomal RNA sequences, the clean reads were assembled into unigenes using Trinity (Grabherr et al., Reference Grabherr, Haas, Yassour, Levin, Thompson, Amit, Adiconis, Fan, Raychowdhury, Zeng, Chen, Mauceli, Hacohen, Gnirke, Rhind, di Palma, Birren, Nusbaum, Lindblad-Toh, Friedman and Regev2011). Microsatellites were identified in the assembled unigenes using MISA script (Thiel et al., Reference Thiel, Michalek, Varshney and Graner2003) against the criteria as follows: a minimum of 12, 6, 5, 5, 4 and 4 times of repeats for 1–6 nucleotide motifs, respectively. Primer pairs covering the microsatellites were designed by Primer 3. Forty-eight primer pairs (online Supplementary Table S1) were randomly selected and synthesized for validation and application.
A total of 18 accessions (Table 1) from C. oleifera (three accessions) and its eight relatives (15 accessions) were sampled for validation and transferability. Their genomic DNAs were extracted respectively with a modified CTAB method. SSR-PCRs were performed in a volume of 10 µL containing 20 ng of DNA, 0.25 µmol/L of each primer, 5 µL 2 × Taq PCR MasterMix and 2 µL ddH2O. The SSR-PCRs were 94°C for 4 min; 32 cycles of 94°C for 45 s, 60°C for 45 s and a 72°C extension for 45 s; followed by a final extension at 72°C for 7 min. The PCR products were separated by 6.0% non-denatured polyacrylamide gels, visualized via silver staining according to the previous reports (Wang et al., Reference Wang, Wang, Zheng, Lv, Liu and Wang2014, Reference Wang, Yang, Zhang, Qin, Wang, Liao, Wang and Gao2017).
Table 1. The accessions of Camellia species used in this study
The genetic diversity parameters including the number of alleles (Na), the effective number of alleles (Ne), observed heterozygosity (Ho), expected heterozygosity (He) and Shannon's index (I) were calculated via GenAlEx version 6.5 (Peakall and Smouse, Reference Peakall and Smouse2012).
Discussion
A total of 58,656,354 raw reads were generated from the root transcriptomes, and 55,072,968 clean reads were assembled into 57,121 unigenes with a total length of 42,633,606 nt, and the N50 of 1174 nt.
A total of 12787 sequences with 15,902 microsatellites were identified from 57,121 unigenes. Di-nucleotide repeats were the most abundant group (56.45%), then followed tri- (25.20%), mono- (12.12%), hexa- (3.21%), penta- (2.18%) and quad-nucleotide (0.84%) (Fig. 1). The microsatellite-containing sequence percentage in this study was higher than that (9–10%) of transcriptomes from many other tissues reported previously (Xia et al., Reference Xia, Jiang, Huang, Zhang, Zhang and Gao2014; Jia et al., Reference Jia, Lin, Feng, Hu, Tan, Shao and Zhang2015; Chen et al., Reference Chen, Yang, Huang, Duan, Long and Rong2017) and slightly lower than that (26.75%) of SSR developed from leaf transcriptomes (Li et al., Reference Li, Wang, Ding, Chen, Hu, Li and Wei2017). Although the tissue and identification criteria were the same, the proportions of microsatellite-containing sequences in transcriptomes were different. As to repeat motif frequencies, our results agreed to previous reports that the most frequent motif was di-nucleotides, followed by tri-nucleotides, and other motif frequencies were lower than 5% (Xia et al., Reference Xia, Jiang, Huang, Zhang, Zhang and Gao2014; Jia et al., Reference Jia, Lin, Feng, Hu, Tan, Shao and Zhang2015; Li et al., Reference Li, Wang, Ding, Chen, Hu, Li and Wei2017). However, percentages of motif frequencies and microsatellite-containing sequences were different, perhaps because of the factors such as different materials, dataset size and detection criteria (Varshney et al., Reference Varshney, Sigmund, Börner, Korzun, Stein, Sorrells, Langridge and Graner2005b).
Fig. 1. Frequencies of microsatellites in Camellia oleifera root transcriptomes according to unit size.
We then designed 6738 primer pairs. Forty-eight primer pairs were randomly selected and synthesized. Of the synthesized primers, 12 (25%) could not produce any amplicons in all species, and six (12.5%) were monomorphic. The remaining (62.5%) were codominant and polymorphic.
Each primer had different transferring ability across Camellia relatives (online Supplementary Table S1). There were 15 primers with transferring abilities of more than 75%, and only two pairs were <50%. We found that only one pair (CoSSR10) amplified across all relative species.
The relationship among the species in this study was also suggested by the transferability (online Supplementary Table S1). These primers had the highest proportion amplified in Camellia vietnamensis (96.67%), and the lowest one in Camellia nanyongensis and Camellia polyodonta (10%). This suggested that C. vietnamensis was the closest relative to C. oleifera, and C. nanyongensis and C. polyodonta seemed to be the most distant species from C. oleifera.
Genetic diversity of all samples was revealed (Table 2). The average of Na and Ne was 4.10 and 2.93. The average of Ho and He was 0.53 and 0.60. I ranged from 0.24 to 1.83, with an average of 1.12.
Table 2. Genetic diversity parameters of Camellia samples based on microsatellites
N, number of samples; Na, number of alleles; Ne, number of effective alleles; Ho, observed heterozygosity; He, expected heterozygosity; I, Shannon's index.
The novel set of microsatellites here provided a useful tool for molecular genetic improvement in C. oleifera and its relatives.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1479262119000066
Acknowledgements
This work was funded by the National Natural Science Foundation of China (31460208), Project of Science and Technology of Guangxi (GuiKeAB16380143), Guangxi Innovation-driven Project (GuiKeAA17204058) and the Project of Science and Technology of Qinzhou (201616804, 20137003).
