Diversity in sea buckthorn (Hippophae rhamnoides L.) accessions with different origins based on morphological characteristics, oil traits, and microsatellite markers

Sea buckthorn (Hippophae rhamnoides) is an ecologically and economically important species. Here, we assessed the diversity of 78 accessions cultivated in northern China using 8 agronomic characteristics, oil traits (including oil content and fatty acid composition) in seeds and fruit pulp, and SSR markers at 23 loci. The 78 accessions included 52 from ssp. mongolica, 6 from ssp. sinensis, and 20 hybrids. To assess the phenotypic diversity of these accessions, 8 agronomic fruit traits were recorded and analyzed using principal component analysis (PCA). The first two PCs accounted for approximately 78% of the variation among accessions. The oil contents were higher in pulp (3.46–38.56%) than in seeds (3.88–8.82%), especially in ssp. mongolica accessions. The polyunsaturated fatty acid (PUFA) ratio was slightly lower in the seed oil of hybrids (76.06%) than that of in ssp. mongolica (77.66%) and higher than that of in ssp. sinensis (72.22%). The monounsaturated fatty acid (MUFA) ratio in the pulp oil of ssp. sinensis (57.00%) was highest, and that in ssp. mongolica (51.00%) was equal to the ratio in the hybrids (51.20%). Using canonical correspondence analysis (CCA), we examined the correlation between agronomic traits and oil characteristics in pulp and seeds. Oil traits in pulp from different origins were correlated with morphological groupings (r = 0.8725, p = 0.0000). To assess the genotypic diversity, 23 SSR markers (including 17 loci previously reported) were used among the 78 accessions with 59 polymorphic amplified fragments obtained and an average PIC value of 0.2845. All accessions were classified into two groups based on the UPGMA method. The accessions of ssp. sinensis and ssp. mongolica were genetically distant. The hybrid accessions were close to ssp. mongolica accessions. The 8 agronomic traits, oil characteristics in seed and pulp oils, and 23 SSR markers successfully distinguished the 78 accessions. These results will be valuable for cultivar identification and genetic diversity analysis in cultivated sea buckthorn.


Introduction
In the present study, 78 accessions of sea buckthorn with variation in fruit traits were selected as materials. The aim of this study is to report the phenotypic characteristics and oil traits in fruit pulp and seeds and the genetic diversity of the 78 sea buckthorn accessions in northern China, providing a foundation for MAB in sea buckthorn.
The young leaves of each plant were kept at −80˚C for use. The berries of each accession were pooled and frozen as quickly as possible at −20˚C. When all plant materials were harvested, the berries were transferred to -50˚C for storage until analysis.

Morphological characteristics of fruit
Hundred berry weight (HBW) was the weight of 100 fresh berries after they were picked from bushes. Hundred seed weight (HSW) was the weight of 100 seeds after air drying at room temperature (25˚C) for 2 weeks [20]. There were three biological replicates for each measurement. The transverse and longitudinal diameters of berries (BTD and BLD) and the length, width and thickness of seeds (SL, SW and ST) were measured over 20 times each (on average) by micrometer calipers. The berry shape indices (BSIs) were estimated by the ratio of BLD to BTD. The minimum (Min), maximum (Max), mean ± standard deviation (SD), and coefficient of variation (CV%) were reported.

Oil extraction and FA analysis in seeds and pulp
The methods of lipid extraction, transesterification (methylation) and purification of methyl esters of the lipid extracts were described by Yang and Kallio [11]. Briefly, samples (1 g) of seeds and fruit pulp were isolated from freeze-dried berries and lipids from the samples were extracted with chloroform/methanol (2:1, v/v) with mechanical homogenization of the tissues. The purified oils were filtered before the solvent was removed on a rotary evaporator. The lipids were weighed, and the oil contents (percentages) in seeds and fruit pulp were calculated. Three biological replicates were taken for analysis. Lipids were stored in chloroform at −20˚C until analysis.
The oil (10 mg) was transesterified by sodium methoxide catalysis [11,21]. It was dissolved in sodium-dried diethyl ether (1 ml) and methyl acetate (20 μl). Then, 1 M sodium methoxide in dry methanol (20 μl) was added, and the solution was agitated briefly and incubated for 5 min at room temperature. The reaction was stopped by adding a saturated solution of oxalic acid in diethyl ether (30 μl) with brief agitation. The mixture was centrifuged at 1500 g for 2 min, and the supernatant was dried in a gentle stream of nitrogen. Fresh hexane (1 ml) was added and the solution was filtered with microporous filtering films (0.22 μm) for analysis.

PLOS ONE
Diversity in sea buckthorn accessions based on morphological characteristics, oil traits, and SSR markers program: initial temperature 50˚C, hold for 1 min, heat to 175˚C at 25˚C/min, then heat to 215˚C at 3˚C/min and hold for 10 min, heat to 230˚C at 3˚C/min and hold for 5 min. The inlet was operated in split mode (1:20) at a temperature of 250˚C with helium as the carrier gas at constant flow of 1.0 ml/min. The transfer line temperature was 215˚C, and the MS ion source was set to 230˚C. MS detection was carried out in electron impact (EI) ionization mode, scanning all masses from 45-400 amu. FAME components were identified based on mass spectral comparison with an external standard (Supelco 37 Component FAME Mix, Sigma-Aldrich, St. Louis, MO, USA) and previous studies [10][11]. The main FA composition was expressed as a weight percentage of the total FAs from three replicates. The minimum, maximum, mean ± SD, and CV% were reported.

Statistical analysis
The data analysis for morphological traits and oil characteristics was performed with SPSS 1 24.0 (IBM 1 ). The following parameters were evaluated: mean, minimum value, maximum value, SD and CV%. One-way analysis of variance (ANOVA) was used in the comparison of all traits among subsp. of sinensis, subsp. of mongolica and hybrids. Pearson correlation coefficients were calculated to analyze the relationship between pairs of 8 agronomic traits. Principal component analysis (PCA) was used to determine relationships among the accessions. In

PLOS ONE
Diversity in sea buckthorn accessions based on morphological characteristics, oil traits, and SSR markers addition, a canonical correspondence analysis (CCA) was applied to the data between morphological characteristics and oil traits in different tissues (seeds and pulp).

DNA extraction and SSR analysis
Total genomic DNA was extracted from young leaves using the TaKaRa MiniBEST Plant Genomic DNA Extraction Kit (TaKaRa, Beijing, China) based on the manufacturer's protocol. The purity and quantity of extracted DNA were evaluated by gel electrophoresis and a Nano-Drop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Twenty-three polymorphic microsatellite loci (SB1-SB23) developed using RNA-Seq were evaluated. Of these, 17 (SB1-SB17) had been deployed in a previous study by the group [14] (S4 Table). The microsatellites were scored as codominant markers for genetic diversity analysis. The number of alleles (Na), effective number of alleles (Ne), observed and expected heterozygosity (Ho and He), Shannon's information index (Is) and polymorphic information content (PIC) for each of the genic SSR markers were calculated using GenAlEx 6.5 [22][23] and PowerMarker version 3.25 [24] software packages. A genetic similarity matrix based on the proportion of shared alleles was generated, and a UPGMA tree was constructed using PowerMarker. The dendrogram was displayed using MEGA 6 software [25] to reveal genetic relationships between the 78 sea buckthorn accessions.

Morphological characterization of berries and seeds
Descriptive statistical analysis of 8 agronomic fruit traits for the 78 sea buckthorn accessions is shown in Table 2, S5 and S6 Tables. Relatively high CV values were observed for the HBW, BLD, and HSW (> 20%). The highest CV% was observed for the HBW (39.12%), which varied   Table). The HBWs values in Suiling (38.33-67.59 g) were higher than those in Dengkou (32.87-63.85 g). For all the introduced cultivars, the HBW values in the two experimental fields were lower than those in their country of origin, Russia. The phenotypic characteristics of sea buckthorn berries showed differences due to their origins, different parts of fruit analyzed, climatic and growing conditions. In this study, 78 accessions were selected for their good adaptabilities to growth sites.
PCA was performed using fruit characteristics (Fig 2). The first two PCs explained 78.11% of the total morphological variance. The first PC accounted for 41.74% of the variance. It was associated with BTD, HBW, ST, HSW, and SW in descending order. Therefore, these traits were important attributes for the classification of sea buckthorn accessions. The second PC accounted for 36.37%, which were correlated with BSI, SL, and BLD in descending order. The plot shows the distribution of 78 sea buckthorn accessions on PC1 and PC2 (Fig 2). The ssp. mongolica accessions with larger berries tended to cluster together, mainly positive on PC2. Six accessions of ssp. sinensis with the smallest berries were negative on both PC1 and PC2. The hybrids were largely distributed between the above two groups. Some hybrids (including ZCY1, ZCY10, ZCY12, XZC26, SCY2, and SCY5) were close to the accessions from ssp. sinensis.

Oil characterization in seeds and seedless parts
The oil characteristics of seeds and seedless parts (pulp and peel) among the 78 accessions are summarized in Table 3 and Table 4. One special feature of sea buckthorn fruit was the high oil content in the pulp and peel (20.41%), in contrast to the oil content in the seeds (8.82%). A higher CV% was observed in pulp oil (42.72%) and varied over a wide range, from 3.46 to 38.56%. The pulp fraction of berries of ssp. mongolica had the highest oil content (24.68%) based on dry weight. The lowest pulp oil content (7.10%) on average was found in the berries of ssp. sinensis. In hybrids, the berries of ZJ2 contained 27.22% pulp oil, which slightly exceeded that of ssp. mongolica on average (S6 Table). The seed oil content varied from 3.88 to 12.75% with an average of 8.82%. The seeds of ssp. mongolica had the highest oil contents, with an average of 9.46%, and those of the other two groups did not differ significantly.
For sea buckthorn, the FA composition in the seed oil differed significantly from that in the pulp oil. The proportions of FAs ranked from high to low as linoleic (18:2n6), α-linolenic (18:3n3), oleic (18:1n9), palmitic (16:0), stearic (18:0) and vaccenic (18:1n7) acids in the seed oil of most accessions (Table 4). Linoleic acid varied from 34.22 to 52.75% with an average of 42.17%. The proportion of α-linolenic acid varied from 21.37 to 47.16% with an average of 34.67%. High CV values were observed in oleic (30.50%) and vaccenic (39.17%) acids. Furthermore, the level of palmitoleic acid (16:1n7, < 0.5%) was extremely low in the seed oil. The FA composition of sea buckthorn seeds was similar among berries of the two different subspecies and hybrid accessions. Small variations were found in the proportion of linoleic acid in seed oil (40.44-42.87%). Its proportion in hybrids was slightly higher than in ssp. mongolica (42.87% vs 42.10%), and showed the highest mean value among the two different subspecies and hybrid accessions. α-Linolenic acid showed little variation, with a larger proportion in ssp. mongolica than in ssp. sinensis (35.56% vs 31.78%). A higher proportion of palmitic (7.41% vs 6.38%) and oleic (16.37% vs 13.96%) acids and a lower proportion of stearic acid (2.19% vs 2.23%) were discovered between the accessions of ssp. sinensis and hybrids. The polyunsaturated fatty acids (PUFA) ratio in hybrids (76.06%) was slightly lower than that in ssp. mongolica (77.66%) and higher than that in ssp. sinensis (72.22%). Some hybrids (including ZJ1, Za1-2, Za13-25, Za05-6, LFZ, and ZCY12) contained a high proportion of PUFAs (> 80%) in seed oil, which was more than the average level of ssp. mongolica accessions (S6 Table).

PLOS ONE
Diversity in sea buckthorn accessions based on morphological characteristics, oil traits, and SSR markers In pulp oil, the dominant FAs were palmitoleic, palmitic, linoleic, oleic, and vaccenic acids ( Table 3). Major differences were observed in the proportion of palmitoleic (17.93-57.75%), oleic (1.44-23.43%) and vaccenic (3.51-24.24%) acids. The special feature of pulp oil is high proportions (> 35%) of palmitoleic and palmitic acids. Compared to ssp. sinensis, ssp. mongolica contained a higher proportion of palmitoleic and palmitic acids in the berry pulp (p < 0.05) ( Table 4). In particular, the proportions of oleic and vaccenic acids were highest in ssp. sinensis, much higher than those in ssp. mongolica and hybrid accessions. The relative levels of α-linolenic and stearic acids in pulp of ssp. sinensis were higher than ssp. mongolica (p < 0.05) ( Table 4). For hybrids, the proportions of most FAs were between ssp. mongolica and ssp. sinensis accessions, except for linoleic acid. Similar to the results in seed oils, the hybrids had the highest proportions of linoleic acid (11.53%) and PUFA (12.60%). The monounsaturated fatty acid (MUFA) ratio in the pulp oil of ssp. sinensis (57.00%) was highest and that of ssp. mongolica (51.00%) was almost equal to that of the hybrids (51.20%). In the

PLOS ONE
Diversity in sea buckthorn accessions based on morphological characteristics, oil traits, and SSR markers hybrids, the pulp oil of SCY2 contained 39.16% palmitoleic acid, and the content of MUFAs was 60.77%, which was higher than that in ssp. sinensis (S6 Table).

Correlations among the agronomic traits and oil characteristics
Canonical analyses allow direct comparisons of two data matrices. All sea buckthorn accessions were represented in a two-dimensional space using CCA between phenotypic traits and oil characteristics (Fig 3). For berries of the two different subspecies and hybrid accessions, phenotypic characters (BLD, HBW, BSI, and BTD) of berries and oil traits in pulp showed a close correlation (r = 0.8725, p = 0.0000). Based on CCA, accessions of ssp. mongolica were clustered on the upper side (mainly positive on D1 and D2), those of ssp. sinensis on the other, and the hybrids in the middle in Fig 3A. The positioning of samples in the first dimension was mostly related to differences in their berry characteristics that were primarily provided by the phenotypic character of BLD. The second dimension indicated differences in the oil contents and FA compositions of pulp oil among sea buckthorn accessions. Differences between pulp oil traits were primarily related to percentages of oil content, 16:0 and 16:1n7, which were highest in ssp. mongolica, followed by hybrids, and lowest in ssp. sinensis. For seeds of 78 accessions, phenotypic characteristics (SL, SW, ST, and HSW) and seed oil traits were correlated (r = 0.7482, p = 0.0000). The positioning of samples was staggered (Fig 3B), which reflected that all seed samples had relatively little variation among phenotypic traits and oil characteristics. These results verified the previous analysis ( Table 2 and Table 3).

SSR diversity
Twenty-three pairs of RNA-Seq SSR primers with good amplification and band stability were used among 78 accessions of sea buckthorn. A total of 69 bands were amplified using the 23 primer pairs, of which 59 were polymorphic, accounting for 85.51% of all bands. The number of amplified bands per locus ranged from 2 to 5, averaging 3, and Ne ranged from 1.0392 to 3.1049, averaging 1.6602 (Table 5). SB2, SB3, SB5, SB6, SB8, SB13, SB16 and SB23 were informative SSR loci, each revealing more than four effective alleles distributed among all of the accessions. Compared with Na, Ne and their average values were lower, which was caused by the uneven distribution of gene frequencies in SSR loci. In the genetic diversity analysis, Ho ranged from 0.0385 to 0.7949, with an average of 0.2965; He ranged from 0.0377 to 0.6779, with an average of 0.3291; and Is ranged from 0.0950 to 1.2152, with an average of 0.5681. The PIC value, regarded as discriminating power, varied from 0.0370 to 0.6174, with an average of 0.2845. Loci SB6 (PIC = 0.6174) and SB8 (PIC = 0.5820) showed higher effectiveness because of their high informativity and could be used to construct the fingerprint map of sea buckthorn germplasm. The characteristics of 23 polymorphic SSR markers in sea buckthorn accessions are shown in Table 5.

Genetic relationships among sea buckthorn germplasm
Using 23 polymorphic SSR markers, the UPGMA dendrogram based on the proportion of shared alleles was constructed to assess the genetic relationships between the 78 accessions (Fig 4). The results showed that all the accessions could be divided into two groups (I and II).

PLOS ONE
Diversity in sea buckthorn accessions based on morphological characteristics, oil traits, and SSR markers subgroup. Overall, the relationship between ssp. mongolica and ssp. sinensis was relatively distant. The hybrids are close to ssp. mongolica, to which their female parents belonged.

Discussion
Morphological characteristics, biochemical traits, and microsatellite markers have been used for germplasm identification and genetic diversity analysis in many horticultural plants [26][27]. The diversity at the morphological, biochemical, and molecular levels of 78 sea buckthorn accessions, composed of 52 from ssp. mongolica, 6 from ssp. sinensis, and 20 hybrids, was investigated.
The morphological characterization of plant materials with desired traits is an essential step for the effective use of germplasm [28]. Here, 8 important agronomic traits were measured among 78 sea buckthorn accessions, and a considerable amount of variation in morphological traits was found. The sizes of berries from the two different subspecies and hybrid accessions were significantly different according to the HBW value (p = 0.000). Compared to ssp. sinensis berries, ssp. mongolica berries were much larger on average. The berry size of hybrid accessions was between the two subspecies. In the PCA, we plotted 2D plots with PC1 and PC2 scores of phenotypes (Fig 2). PC1 was mainly related to BTD and HBW, which explained the largest portion of the variance in 78 accessions. The distribution of 78 accessions on PC1 and PC2 was consistent with their agronomic characteristics (Fig 2). These results estimating morphological traits are valuable tools for identifying variation among plant germplasms [26].

PLOS ONE
For biochemical traits, oil content and FA composition in the seeds and seedless parts were selected for their important roles in human health. The oil of sea buckthorn seems to be a good source of unsaturated FAs. Seed oil is rich in PUFAs, including linoleic and α-linolenic acids. The proportion of PUFAs did not differ significantly among berries from three origins, despite the differences in some morphological characteristics and in growth conditions. These results were consistent with previous studies [10]. The results of the present study and previous investigations also suggested that the berries of ssp. mongolica were a good source of palmitic and palmitoleic acids in pulp oil and that those of ssp. sinensis were a good source of oleic acid in both seeds and fruit pulp [29]. Although carefully selected for intersubspecies crosses, some hybrids displayed elite oil traits. For example, the proportion of MUFAs in the pulp of SCY2 and of PUFAs in the seeds of 6 accessions (including ZJ1, Za1-2, Za13-25, Za05-6, LFZ, and  Table 1). ▲ = ssp. mongolica; • = ssp. sinensis; � = hybrid.
https://doi.org/10.1371/journal.pone.0230356.g004 ZCY12) exceeded the average level of ssp. mongolica accessions, the subspecies that one of their parents belonged to. These results demonstrate the effectiveness of traditional cross breeding in the improvement of native accessions (ssp. sinensis), even though it is time-consuming and has low efficiency.
Previous studies found that berry size is a useful indicator of Vc, sugars and acids in population identification [19,30]. The nutrients in the seedless fraction were more concentrated in the small berries of ssp. sinensis than in the large berries of ssp. mongolica [29]. In the present study, we analyzed the correlation between agronomic characteristics and oil traits at different levels (seed and pulp) by CCA. The results showed that the phenotypic characteristics (BLD, HBW, BSI, and BTD) of berries and the oil traits in pulp were positively correlated (r = 0.8725, p = 0.0000). The BLD, as a promising marker, provided the primary difference in CCA. Our results illustrated that berry size had different correlations with various biochemical characteristics in sea buckthorn.
Variation in phenotypic traits among germplasms may be attributed to differences in genetic backgrounds, geographical location, climate, harvest period and berry maturity, while molecular markers are independent of environmental condition and growth stage [31]. Twentythree polymorphic SSR markers were used to investigate the genetic relationships among 78 sea buckthorn accessions. The 23 selected SSR markers detected 2-5 alleles, and their PIC values ranged from 0.1166 to 0.6155 and had an average of 0.3249. The PIC mean value was significantly lower than that of RAPD, ISSR and SRAP markers previously reported [15][16]32], suggesting that the gene sequences of these SSR markers were conserved in sea buckthorn germplasm.
Based on UPGMA, the 78 accessions were classified into two groups. There is a large genetic distance between accessions of ssp. sinensis and ssp. mongolica. The hybrids were in between and rather close to ssp. mongolica accessions. Coincidentally, these hybrids were also between accessions of ssp. sinensis and ssp. mongolica on the PCA plot based on 8 agronomic characteristics. This uniformity indicated that the diversity of morphological characteristics could reflect genetic diversity and be used as markers in agronomy. Ruan et al. [15] assessed 14 Chinese, Russian and Mongolian sea buckthorn accessions using RAPD markers and obtained similar results. In a previous publication, the genetic relationship of 31 sea buckthorn accessions (also contained in this study) was analyzed based on 17 RNA-Seq SSRs [14]. However, the accessions of ssp. mongolica clustered in one group and those of ssp. sinensis and hybrids were divided in the other. This revealed that genetic relationships mainly relied on the diversity of genotypes and genetic backgrounds.
With the continuous development of high-throughput sequencing technology, transcriptome databases have become a powerful resource for SSR mining. An increasing number of RNA-Seq SSRs have been developed and applied to the study of species genetic diversity and population genetic structure [33][34]. The SSRs obtained by transcriptomes are associated with many important quantitative traits [35].
The results in the present study yielded useful knowledge regarding the diversity and genetic relationships of sea buckthorn germplasm in northern China, and could therefore facilitate further studies, including the selection of mapping populations and promising candidates, marker-trait association analysis based on establishing the consistency of the traits, and characterization of parents used in future breeding programs.

Conclusion
In the present study, 8 phenotypic characteristics, oil traits in seeds and seedless parts, and 23 SSR markers successfully distinguished all 78 sea buckthorn accessions. In PCA, BTD and HBW in the first PC were the most important characteristics for distinguishing the accessions.
The agronomic traits of berries were closely correlated with the oil content and FA composition in the pulp by CCA. This information will be valuable for germplasm identification and genotypic diversity analysis in H. rhamnoides.
Supporting information S1 Fig. 78 berry samples used in this study. Numbers are the variety codes listed in Table 1