Genetic Variability and Population Structure of the Potential Bioenergy Crop Miscanthus sinensis (Poaceae) in Southwest China Based on SRAP Markers

The genus Miscanthus has great potential as a biofuel feedstock because of its high biomass, good burning quality, environmental tolerance, and good adaptability to marginal land. In this study, the genetic diversity and the relationship of 24 different natural Miscanthus sinensis populations collected from Southwestern China were analyzed by using 33 pairs of Sequence Related Amplified Polymorphism (SRAP) primers. A total of 688 bands were detected with 646 polymorphic bands, an average of 19.58 polymorphic bands per primer pair. The average percentage of polymorphic loci (P), gene diversity (H), and Shannon’s diversity index (I) among the 24 populations are 70.59%, 0.2589, and 0.3836, respectively. The mean value of total gene diversity (HT) was 0.3373 ± 0.0221, while the allelic diversity within populations (HS) was 0.2589 ± 0.0136 and the allelic diversity among populations (DST) was 0.0784. The mean genetic differentiation coefficient (Gst = 0.2326) estimated from the detected 688 loci indicated that there was 76.74% genetic differentiation within the populations, which is consistent with the results from Analysis of Molecular Variance (AMOVA) analysis. Based upon population structure and phylogenetic analysis, five groups were formed and a special population with mixed ancestry was inferred indicating that human-mediated dispersal may have had a significant effect on population structure of M. sinensis. Evaluating the genetic structure and genetic diversity at morphological and molecular levels of the wild M. sinensis in Southwest China is critical to further utilize the wild M. sinensis germplasm in the breeding program. The results in this study will facilitate the biofuel feedstock breeding program and germplasm conservation.


Introduction
The genus Miscanthus is comprised of C 4 perennial rhizomatous grasses, originated from Eastern Asia.Owing to high biomass productivity [1], low-nutrient input [2,3], and high water-use efficiencies [4], Miscanthus have attracted considerable attention as one of the most promising non-food bioenergy crops.There are about 10-15 Miscanthus species distributed worldwide, of which seven are native to China [5][6][7][8].Since the 1970s, this genus, specifically M. × giganteus, has been intensively studied in Europe as a biomass feedstock [7,9,10].However, M. × giganteus is propagated by plant rhizomes or tissue culture and does not produce fertile flowers or seeds, and its production is heavily limited by its natural sterility and a narrow genetic base [11].As a progenitor of M. × giganteus, M. sinensis is propagated by seeds which is a favorable trait for crop adoption and provides a comparable yield in some places and could be a valuable genetic resource for biofuel crop domestication and improvement [6,12].Based on the previous tests of drought and cold tolerance of M. sinensis in Europe, a much broader range of adaptation than M. × giganteus was found in this diploid species [1,11].So with M. sinensis it is considered possible to breed varieties with similar or better yield but higher tolerance for frost and drought than M. × giganteus.Besides, it can be used in crosses to create new cultivars of M. × giganteus.
Molecular markers are essential tools for germplasm evaluation, genetic analysis, and marker-assisted breeding for crop improvement.Employing molecular markers, such as Sequence Related Amplified Polymorphism (SRAP), Amplified Fragment Length Polymorphism (AFLP), Inter-Simple Sequence Repeats (ISSR) and Simple Sequence Repeats (SSR) [13][14][15][16][17][18][19] to estimate genetic variation within species could assist the breeding program for parental and breeding line selection and desirable traits.SRAP is recognized as a new and useful molecular marker system because of its high reproducibility, low cost, and no requirement of prior knowledge of target sequences [20].Up to now, SRAP markers have been successfully used for evaluation of genetic diversity for Carthamus tinctorius, Cucurbita pepo, buchloe dactyloides, and Solanum lycopersicon [21][22][23][24] and genetic map construction for Gossypium hirsutum and Triticum aestivum [25,26].
In recent years, many reports on Miscanthus were published showing the abundant resources distributed in China [12,[27][28][29]  genetic diversity and the existence of a gene flow in M. sinensis populations in this area.However, the SSR markers used in that study were mainly derived from the non-coding regions, which may not be able to provide sufficient evidence to reveal the diversity and differentiation of M. sinensis in Southwest China; therefore, SRAP markers, derived from the coding region, were used in this study.Southwest China, as one of the 34 biodiversity hot spots around the World, has abundant wildlife resources [30].It is crucial to evaluate the genetic structure and genetic diversity of the wild M. sinensis germplasm, which is widely distributed in Southwest China and to eventually utilize this valuable germplasm for crop improvement.However, there are no thorough studies on the genetic diversity and population structure of the germplasm distributed in Southwest China.Therefore, in this study, we evaluated the genetic diversity and population structure of 24 M. sinensis natural populations collected in Southwest China using SRAP markers to facilitate the conservation of the Miscanthus germplasm and breeding in the near future.

Polymorphism of SRAP Markers
Six accessions of M. sinensis, which have significant differences among the morphological characterization and geographic location, were selected to screen 100 pairs of SRAP primers.In total, 33 of them generated robust discernible bands (Table 1).These 33 SRAP primer pairs were then used to genotype the whole collection of 260 individuals.In total, 688 bands were generated and 646 (93.90%) were polymorphic.The number of bands per primer pairs ranged from 13 to 30, with an average of 20.58 bands, of which 19.58 in average were polymorphic.Primer pairs Me6 + em10 amplified the most number of polymorphic bands (30) while Me7 + em1 amplified the least number of polymorphic bands (9).The polymorphic information content (PIC) values ranged from 0.23 (Me3 + em5) to 0.41 (Me4 + em10) with a mean of 0.34, demonstrating a good discriminatory capacity (Table 1).

Genetic Diversity and AMOVA Analysis
The 24 wild distribution populations were comprised of 260 individuals which had a varied genetic diversity reflected by the three main genetic diversity parameters including percentage of polymorphic bands (P), Nei's [31] gene diversity (H), and Shannon's Information Index of Diversity (I).Among the 24 populations, the P value ranged from 21.22% (Pop12) to 84.30% (Pop15), with an average of 70.59%.The H value ranged from 0.0787 (Pop12) to 0.3052 (Pop15), with an average of 0.2589 at the population level.The variation trend of the I value was similar to the other two parameters, with an average of 0.3836 (Table 2).The total numbers of P, H and I were 93.90%, 0.3377 and 0.5032 within species, respectively.The genetic data exhibited a high level of genetic diversity within M. sinensis species from southwest China.The total gene diversity (H T ) was 0.3373 ± 0.0221, while the gene diversity within populations (H S ) was 0.2589 ± 0.0136 and the gene diversity among populations (D ST ) were 0.0784.The mean genetic differentiation coefficient (G ST = 0.2326) estimated from the 688 bands indicated that there were 76.74% genetic variation within populations.These results demonstrated that the accessions had a higher level of genetic variation within populations than among them.The AMOVA analysis (Table 3) of the M. sinensis wild populations showed similar results that both the genetic variations within (86.0%) and among (14.0)populations were significant.In addition, there was a high frequency of gene flow (Nm = 1.6493) between populations, indicating there were more than one effective immigrants from one population into another at each generation.Miscanthus is widely distributed around the world, although its main distribution area or diversity center is in China [18,32].Knowing the relationship and population structure of M. sinensis is important for their conservation and utilization [15].In this study, the high level of genetic diversity of M. sinensis from southwest China was revealed by SRAP markers, which are similar to the previous results with EST-SSR, SSR and AFLP markers [9,12,18].Meanwhile, SRAP analysis indicated higher genetic variation (76.74%) existed within populations than among populations, which is in agreement with the results in other grass species assessed by allozymes, ISSR, RAPD, SSR, and EST-SSR [12,[33][34][35][36][37][38] and in M. sinensis of China assessed by SNP and SSR makers [18,32].The main factors determining the plant population genetic structure include the mating and reproduction system, selection pressure, adaptation, and geographic locations [39].The genetic recombination promotes genetic diversity within populations [40].In plants, gene flow events can be initiated through pollen or seed.M. sinensis is an out-crossing species that can lead to a strong gene flow (Nm = 1.6493) and introgression among populations, so it is reasonable that the genetic variation within populations is greater than that among populations [41,42].

Population Structure and Cluster Analysis
The population structure of the 260 individuals was estimated under the Hardy-Weinberg Equilibrium by using STRUCTURE V2.3.3 software.Based on maximum likelihood and delta K (∆K) values, the number of optimum subgroups was five (Figure 1).By using a membership probability threshold (Q) of 0.60, majority of the individuals were clearly assigned to specific groups.Among them,  S1).The genetic similarities (GS) of 260 individuals ranged from 0.565 to 0.972 with an average of 0.659 which showed a high level of genetic variation range among the accessions.The Un-weighted Pair-group Method with Arithmetic mean (UPGMA) dendrogram based on GS data obviously revealed that when at the genetic similarity coefficient value of 0.659, five major clusters were formed and group 1 accessions were mainly collected from the Yuxi area of Yunnan.The genotypes of group 2 were primarily collected from Zigong and Jian'ge in Sichuan.Group 3 contained mostly accessions from Jiangyou and Guangyuan.Group 4 accessions were mainly collected from Yaan, Daying, Banan and Zunyi (Figure S1).The rest of the accessions assigned to group 5.The results from the cluster analysis were similar with those from the structure analysis.
The genetic distances (GD) among the 24 populations were estimated by Nei's [43] unbiased measure, which could obviously reveal the genetic relationship.The GD between Pop2 from Bifengxia and Pop3 from Baoxing was the lowest (0.028), and the distance between Pop12 from Zigong and Pop24 from Yuxi was the highest (0.292) with the mean of 0.097 (Table S2).The UPGMA dendrogram based on GD data clearly showed the relationships among the 24 populations (Figure 2), which was nearly congruent with the previous cluster analysis of 260 individuals.However, in this result, we found that a new group (Group 5) including Pop22 and Pop23 collected from Guizhou.Throughout the results of the two methods at different levels, we found that combining the analysis is the best strategy to reveal the genetic structure of M. sinensis in Southwestern China.Apparently, Pop12 and Pop24 were differentiated from the other populations in both Structure and UPGMA analysis.The main reason could be the distinct geographic isolation between this two and the rest populations.In addition, Pop12 has the lowest genetic diversity parameters which are P (21.22%),H (0.0787), and I (0.1162), and a low gene flow exists between Pop12 and other populations.Furthermore, Pop12 has a narrow distribution range in this area and almost no other M. sinensis plants were found within a range of 5 km around it.Therefore, a habitat fragmentation was formed as influenced by founder effect.
Through the structure analysis, 14 out of 260 individuals with mixed ancestry were all from Pop16.In principle, all of the genetic material of the sampled individuals comes from one or more of K unobserved populations with each population characterized by a set of allele frequencies at each locus.When individuals have mixed ancestry, this means that each genotyped allele comes from one or more populations.We synthesized geographic information to analysis the accessions from Pop16 and found that they all collected along with the G42 highway in China which is the only entrance to the Dead Sea of China located in Da Ying County.The Dead Sea of China is a famous scenic spot where the total number of tourists is approximately 3 million per year.The huge traffic flow and the complex environment could help the seed spread widely.Hence a high gene flow occurred in this area and the plants there had a mixed ancestry.Although some researchers think these man-made factors contribute to the long term survival of populations, this is controversial [44], as several studies [45][46][47][48] indicate that they should not be neglected because those factors accelerate the loss of genetic variability through random genetic drift [49].
The previous clustering result of M. sinensis from southwest China assessed using SSR makers [12] was different from the dendrogram that resulted from SRAP makers.These differences could be due to the different DNA segment targets of SSR and SRAP makers.The SSR have a random distribution within the genome, while the target locus of SRAP is mainly in open reading frame regions [20,50].SSRs mostly exists in non-genic regions, could be in genic regions as well, but with low frequency [50,51].In different plant individuals, the number of repeat units varies, but the flanking sequence is conserved around the SSR.The numbers of loci studied and their coverage of the genome wide are important in obtaining reliable estimates of genetic relationships between populations and within population [52].Although, both SRAP and SSR distinguished intraspecific taxa with similar great discriminating power, the average numbers of bands generated by each primer pair of SSR (14.80) [12] were much lower than that of SRAP (20.85).Therefore, we considered that SRAP was more efficient than SSR for assessing the genetic diversity of large numbers of M. sinensis accessions.In total, as widely used PCR-based markers, SRAP has advantages over SSR markers, since no prior knowledge of target sequences is required which make it to be widely utilized.

Plant Material Collection
The experimental materials consisted of 260 individuals of M. sinensis collected from 24 natural populations in Sichuan, Chongqing, Guizhou, and Yunnan provinces in 2010 (Table 4).
The sampling locations were selected according to M. sinensis habitats based on geographic location and topography.All of approaches used in collecting samples are based on Xu's method [12] (Figure 3).Within each population, the numbers of appropriate representative individuals were selected based on the size of each population.performed on a Peltier Thermal Cycler (DNA Engine ® , Bio-Rad, Hercules, CA, USA) under the following conditions: 5 min at 94 °C, followed by 5 cycles at 94 °C for 1 min, 35 °C for 1 min, and 72 °C for 1 min, and then 35 cycles at 94 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min, extended at 72 °C for 10 min, then stored at 4 °C.The SRAP fragments were separated on 6% denatured polyacrylamide gels (acrylamide: bis-acrylamide 19:1, 1× TBE) and electrophoresis, later the gel were stained by AgNO 3 solution and photographed by the Gel Doc XR system (Bio-Rad).

Data Analysis
For the statistical analysis, the SRAP banding patterns which could be unambiguously scored across all the sampled populations were recorded manually for band presence (1) or absence (0), each of them was treated as an independent character regardless of its intensity.The discriminatory power of different SRAP primers was evaluated by means of polymorphic information content (PIC) [53].

Figure 1 .
Figure 1.Five subgroups of 260 M. sinensis accessions inferred from STRUCTURE analysis.The vertical coordinate of each subgroup indicates the membership coefficients for each individual.Red zone: SG1; Green zone: SG2; Blue zone: SG3; Yellow zone: SG4; Pink zone: SG5.

Figure 2 .
Figure 2. Dendrogram of 24 M. sinensis populations based on GD data by UPGMA cluster analysis.

Figure 3 .
Figure 3.The geographical distribution of 24 populations of M. sinensis used in this study.The accessions were mainly sampled from four provinces,Sichuan, Chongqing, Guizhou and Yunnan in China.The different colors pentagram represents the five subgroups generated by STRUCTURE V2.3.3 software.

Table 1 .
Primer sequences amplification information of the SRAP markers.

Table 2 .
Genetic diversity of M. sinensis wild populations.
Note: H, Nei's gene diversity; P, Percentage of Polymorphic Bands; I, Shannon's Information Index of Diversity.

Table 3 .
AMOVA analysis of variance distribution with and amoung M. sinensis wild populations.