How genetic variation is affected by geographic environments and ploidy level in Erianthus arundinaceus?

Erianthus arundinaceus is not only a candidate plant for sugarcane breeding programs, but also a potential bioenergy grass. Genetic variation that is affected by geographic environments and ploidy level is very important for the utilization of Erianthus arundinaceus. In this study, effects of geographic environments and ploidy level on genetic variation were studied through analyzing the genetic diversity, genetic similarity and cluster analysis of 46 E. arundinaceus materials from natural habitats in China by using 7 ISSRs and 15 SSRs. Results showed that: 1) Seven ISSRs generated total 66 bands, of which 77% were polymorphic bands, the Nei's genetic similarity coefficient of tested materials ranged from 0.642 to 0.904 with an average value of 0.765. Fifteen SSRs generated 138 bands, of which 81% were polymorphic bands, the Nei's genetic similarity coefficient of tested materials ranged from 0.634 to 0.963 with an average value of 0.802. The results indicated great genetic diversity existed in the tested materials. 2)The tested materials were clustered into 3 groups and 7 subgroups, which demonstrated a strong geographic effect on variation of the local E. arundinaceus, and weak relationship was found between genetic distance and geographic distance. Five tetraploid materials were not clustered together, and were clustered together with materials from similar geographical location. 3) The genetic variation and cluster results were affected by geographic landforms and environments, the gene flow was blocked by Ocean and mountains, and promoted by river. The effect of ploidy level on genetic variation was little.


Introduction
E. arundinaceus is a warm-season, and caespitose perennial tall grass in Erianthus (Poaceae) native to China which are widely distributed in south China [1]. And it is also widely distributed in other south and southeast Asian regions [2]. Tetraploid (2n = 4x = 40) and hexaploid (2n = 6x = 60) are the two major cytotypes in E. arundinaceus [3]. E. arundinaceus was onsidered to be a member of a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 the "sugarcane complex" due to its relation with taxa in Saccharum, Erianthus, Sclerostachya and Narenga [4]. Early studies of E. arundinaceus were focused on sugarcane breeding [5] to improve the traits like disease resistance and tolerance to abiotic stresses [6]. Although it is not easy to cross the two species of Saccharum and Erianthus, hybrids of Saccharum × Erianthus were successfully generated from some breeding projects [7][8][9][10]. Recently, the species has been targeted as a bioenergy perennial for its high fiber, high tillering ability and biomass yield potential on marginal lands [11].
Information of genetic diversity is important for studying the genetics of E. arundinaceus, and breeding new cultivars in the species. Morphological traits were studied as indicators in genetic diversity of E. arundinaceus, and high level genetic diversity of E. arundinaceus from China and low level genetic diversity of E. arundinaceus from Indonesia were found from previous studies [12,13]. Genetic diversity of E. arundinaceus has been well documented by using DNA markers in the last two decades. Markers like AFLP, ISSR, RAPD and RFLP used in characterizing genetic diversity in E. arundinaceus showed high genetic variation existing among these materials of China [14][15][16][17][18]. However, these reports revealed very limited information on genetic variation and laws of genetic variation affected by geographical environment and ploidy level, which could help better understanding the natural variation in the species on a large geographic scale.
In this study, 45 E. arundinaceus materials from natural habitats in China were studied, in which, 5 materials were tetraploid and other materials were hexaploid. Genetic diversity, genetic similarity and cluster analysis of 45 E. arundinaceus materials were analyzed by using ISSR and SSR markers, and the effect of geographic environments and ploidy level on genetic variation of E. arundinaceus will be revealed.

Ethics statement
This study was approved by Sichuan Academy of Grassland Science, Guizhou Normal University, Southwest University of Science and Technology, Guizhou proverty alleviation and development office. No specific permissions were required for collecting Erianthus arundinaceus samples at the locations in China, because the research was funded by the National Natural Science Foundation of China, the earmarked fund for China Agriculture Research System, the state key research and development program in the 13th five-year.

Plant materials and DNA extraction
Materials in this study were collected from 46 wild E. arundinaceus populations covering Shanxi, Sichuan, Yunnan, Guizhou, Guangxi, Guangdong and Hainan of China. Each sampled was composed of 10 individuals in each population. DNA was extracted from the mixed leaves of each sample. Two cytotypes constituted the materials, of which 5 materials were tetraploid whereas 41 materials were hexaploidy (Table 1) [19]. DNA was extracted from fresh leaves tissue of 4-week plants by using the modified CTAB method of Doyle [20]. Purity and concentration of the genomic DNA were determined by a Nanodrop spectrophotometer (NanoDrop Products, Wilmington, DE). Finally, DNA samples were diluted to 25 ng/μL, and stored at -20˚C.

ISSR reactions
Seven ISSR primers were synthesized by Shanghai Biochemical Engineering Technology (Shanghai, China) ( Table 2). PCR reactions were performed in a total volume of 20 μL

Data analysis
Clearly amplified PCR bands were visually scored for presence as "1" or absence as "0" and input into an Excel matrix for the following analysis. Total bands, polymorphic bands and percent polymorphic bands were figured out respectively. A UPGMA tree based on Nei's [21] genetic distance data was generated by FreeTree program [22] to examine genetic relationships of materials, bootstrap values were generated by using FreeTree program [22] based on resampling replacements over loci in 1000 replicates. In addition, a mantel test was conducted to calculate the correlation between pairwise geographic and Nei's genetic distances by using NTSYS software [23]. The computation of the Spearman rank correlation between SSR gene diversity (HE) values and ISSR HE values was carried out by using SAS 8.0.

Diversity of ISSR and SSR amplified markers
Seven ISSR markers which produced clear and previous bands were screened out from 100 random markers. A total of 66 bands were obtained through these 7 markers after amplification, of which 51 bands were polymorphic (S1 Table). For each marker, the bands number ranged from 6 to 14, while the number of polymorphic bands varied from 3 to 12. The average number of amplified bands was 9.4, and the average number of polymorphic bands was 7.3. The percentage of polymorphic bands (PPB) ranged from 50% (847) to 88% (813) with an average of 77% (Table 3). A total of 138 bands were obtained by 15 SSR markers from amplification, of which 112 bands were polymorphic (S2 Table). Five to sixteen bands (four to thirteen polymorphic bands) were produced by each SSR marker, of which the average number of amplified bands was 9.2, and the average for polymorphic bands was 7.5. The percentage of polymorphic bands (PPB) ranged from 57% (BML21) to 100% (BML9 and BML11) with an average of 81% (Table 3).  Table).

Genetic similarity analysis of E. arundinaceus
For  Table).

Cluster analysis
For ISSRs, the UPGMA tree was constructed based on the values of genetic distance. Total 46 materials were divided into 3 groups (Fig 1). Group 1 included the materials from Sichuan except SAG-E17 (which was from Huili) and SAG-E1 (which was from Shanxi). Group 2 consisted of 4 materials from Yunnan, in which one material was collected from Sandu of Guizhou (SAG-E26) while one material was from Nandan of Guangxi (SAG-E29). Group 3 contained the materials from Guizhou (except SAG-E26), Guangxi (except SAG-E29), Guangdong, Hainan, and one material from Huili of Sichuan (SAG-E17). Three subgroups were consisted into group 1, The first subgroup included the materials from the middle reaches of Minjiang River basin (Dujiangyan, Wenjiang, Longquan and Xinjin, except Jiajiang), the second subgroup consisted of the materials from the Qinyi River basin (Ya 0 an, Shimian and Hanyuan), Meishan city (the middle reaches of Minjiang River basin), and Leshan (the lower reaches of Minjiang River basin), the third subgroup contained the materials from Panxi area (Panzhihua, Huili and Ningnan). In group 3, the materials from Guizhou province except SAG-E26 from Sandu were divided into one subgroup, the materials from Guangxi province except SAG-E29 of Nandan and Guangdong province except SAG-E39 of Xinyi were divided into another subgroup, the materials from Hainan province were divided into the last subgroup. For SSRs, the UPGMA tree was composed of 3 major groups from 46 materials which was consistent with the results of ISSRs (Fig 2). Group 1 included the materials from Sichuan province and Shanxi province (SAG-E1). Group 2 consisted of 4 materials from Yunnan province. Group 3 contained the materials from Guizhou, Guangxi, Guangdong, and Hainan, as well as and one material from Sichuan province. In group1, the materials from Hanzhong of Shanxi province, Guangyuan and Suining in Sichuan province were divided into one subgroup, the materials from the middle reaches of Minjiang River basin (Shuangliu, Longquan, Dujiangyan, Xinjin, Wenjiang and Meishan) were divided into one subgroup, the materials from the Qinyi River basin (Ya 0 an and Shimian) and the lower reaches of Minjiang River basin (Jiajiang and Leshan) were divided into another subgroup, the materials from Panxi area (Panzhihua, Huili and Ningnan) were divided into the last subgroup. In group 3, the materials from Guizhou province and Nandan of Guangxi province were divided into one subgroup, the materials from Guangxi province except SAG-E29 from Nandan and Guangdong province were divided into another subgroup, the materials from Hainan province were divided into the last subgroup, respectively. The Mantel tests indicated that medium relationship existed between genetic distance and geographic distance among these materials (ISSR: r = 0.520 and p< 0.05, SSR: r = 0.689, p< 0.05). The relationship between SSR gene diversity (HE) values and ISSR HE values was in a medium high level (r = 0.584, p< 0.05).
There were 4 impacts of geographic environments and ploidy level on cluster structure of E. arundinaceus. The first impact was that the low level of genetic diversity of island E. arundinaceus due to ocean isolation. Materials with similar geographical distances were gathered into different groups because of the isolation effects of Qiongzhou Strait. For example, E40 (19˚05 0 , 110˚10 0 ) was located at a distance of 128.98 km from E38 (20˚55 0 , 110˚03 0 ) and 267.34 km from E45 (18˚33 0 , 109˚37 0 ) although it was separated from E38 and gathered together with E45. The second impact showed that the low level of genetic diversity of E. arundinaceus was generated by mountains isolation, which caused materials with similar geographical distances in different groups due to the isolation effects of the mountains. For example, E23 (26˚16 0 , 107˚29 0 ) was located at a distance of 470.90 km from E18 (26˚58 0 , 102˚48 0 ) and 651.20 km from E38 (20˚55 0 , 110˚03 0 ) although it was separated from E18 and gathered together with E38. The third impact was caused by the river systems which was the reason for the materials from the same basin or from different basins gathering together at the river confluence reaches. The fourth impact that the effects of ploidy level on cluster was identified little in this study because 5 tetraploid materials from different geographical locations were not clustered together.

Effects of the ocean and the big mountains on genetic variation
The percentage of polymorphic bands (PPB) of E. arundinaceus in this study were 77% (ISSR) and 81% (SSR), lower than 99.3% (AFLP) in the study of Cai [16], higher than 64.9% (ISSR) and 70.1% (RAPD) by Zhang et al [14,15], 69.2% (AFLP) by Tsuruta [24], respectively. These results revealed a high level of genetic diversity for E. arundinaceus in China. The level of genetic diversity varied among the materials in different regions of China. The materials from Hainan province showed the lowest level of genetic diversity and the value of GS was 0.865 and 0.872 for ISSR and SSR, respectively. The genetic diversity level of E. arundinaceus in this study was similar with that detected in Cai's study in which was 0.846.
The genetic diversity level of E. arundinaceus in this study was similar with that detected in Cai's study [16]. These results also supported Zhang's speculation that the low level of genetic diversity of island E. arundinaceus was generated by ocean isolation, which may have effectively blocked gene flow from germplasm outside the islands, consequently reducing genetic diversity [25].
The "isolation effect" of ocean not only affected the level of genetic diversity, but also the cluster results of materials in different regions. From the cluster analysis of ISSRs and SSRs, the materials from Hainan, Guangxi except Nandan, and Guangdong province were divided into the same subgroup, and the materials from Hainan were separated clearly from others, which showed a large influence of the Qiongzhou Strait on the cluster structure. Similarly, mountains, especially the high mountains, could impact on the cluster results. From the cluster analysis of SSR, the materials were divided into 3 groups. The materials from Sichuan and a material from Hanzhong of Shanxi around Sichuan Basin formed the group 1. Four materials seperated by the Yunnan-Guizhou Plateau from southern tip of Yunnan were clustered in the group 2. The materials from Guizhou, Guangxi, Guangdong and Hainan province formed the group 3. The group 3 was located in the south of Sichuan and in north of the Yunnan, which were just separated by mountains of the Yunnan-Guizhou Plateau. The cluster analysis of ISSRs was consistent with SSRs. Therefore, the ocean and the big mountains could hinder the gene exchange of E. arundinaceus, and influence the diversity of E. arundinaceus as well as the genetic relationship among different groups of E. arundinaceus [26].

Effect of river systems on genetic variation
The group 1 from cluster analysis of ISSRs was composed of 2 subgroups. The subgroup contained the materials from the middle reaches of Minjiang River basin. The second subgroup included the materials from the Qinyi River basin, one material from the middle reaches of Minjiang River basin, and one material from the lower reaches of Minjiang River basin. For the two subgroups in group 1 SSRs, the materials from the middle reaches of Minjiang River basin were divided into one subgroup, and the materials from the Qinyi River basin and the lower reaches of Minjiang River basin were divided into another subgroup. From the cluster analysis of ISSRs and SSRs, some interesting results were found that the materials from the same river basin were firstly grouped together, while some materials from different river basin were also grouped together, such as the materials from the Qinyi River basin and the lower reaches of Minjiang River basin, which indicated the gene exchange of E. arundinaceus may be promoted through confluence of river systems. Meanwhile, the materials from near location tended to be grouped together through gene flow, due to E. arundinaceus's characteristic of outcrossing with higher level of gene flow [27]. Gene flow of E. arundinaceus could be blocked by geographical environment (like ocean and mountains). However, the blocked effects of geographical environments on gene flow could also be removed by rivers to provide channels for gene flow, and promote the gene exchange among the materials in the same river basin. Thus, the river systems should be an important factor for the consistent genetic background among the materials from the same river basin.
The mantel tests showed a weak associated relationship existed between genetic distance and geographic distances among these materials. Because the geographical environment has a great influence on the genetic diversity of E. arundinaceus, which could hinder or promote the gene exchange between different regions. Previously, genetic variation was affected by geographical environments through limited gene exchange [16,25], which was consistently identified in this study. Because geographical environments could promote the gene exchange, change the genetic diversity and affect cluster structure among materials.
In this study, both ISSRs and SSRs revealed that the native Chinese E. arundinaceus materials had a high level of genetic diversity although the correlation coefficient between SSR gene diversity (HE) values and ISSR HE values was small (r = 0.584, p< 0.05). For the geographic factors, the results indicated that the ocean and the big mountains could hinder the gene exchange of E. arundinaceus and the gene exchange of E. arundinaceus could be promoted through confluence of river systems.

Effect of ploidy level on genetic variation
Polyploidization is a common phenomenon in plant evolution which is also useful for the formation of new species [28]. Shock of polyploidization in the genome could cause chromosome recombination and sequence elimination which would lead to genetic variation [29][30]. The effects of ploidy level on genetic variation of Chinese Dactylis glomerata was studied by Zeng, which showed the ploidy level was the primary factor of the effect on Orchardgrass cluster structure and the genetic variation. And the genetic diversity of among the materials with same ploidy level were smaller than the materials with difference ploidy level [31]. In previous study, we inferred that the altered ploidy might contribute to the genetic variation in the Chinese E. arundinaceus germplasm since gene flow between plants of altered ploidy is likely limited [25]. However, the cluster structure wasn't affected by altered ploidy in this study, so ploidy level may have little effect on genetic variation of E. arundinaceus. The effects of polyploidization on the plant genome may be little in some polyploid species, which led to chromosome recombination and sequence elimination. Moreover, polyploidization were not tested in these species [30].
Supporting information S1