Phenotypic diversity of pea ( Pisum sativum L.) varieties and the polyphenols, flavonoids, and antioxidant activity of their seeds. Phenotypic diversity of pea ( Pisum sativum L.) varieties and the polyphenols, flavonoids, and antioxidant activity of their seeds

: Pea (Pisum sativum L.) is the fourth leading legume crop in the world, and its demand is increasing. In this study, the morphological characteristics (seed shape, seed surface, seed coat color, hilum color, cotyledon color, 100-seed weight and color values), total phenolic content (TPC), total flavonoid content (TFC), 2,2’-azinobis-(3-ethylbenzthiazoline-6-sulphonate) (ABTS) free radical scavenging capacity and ferric reducing antioxidant power (FRAP) of 75 pea cultivars were investigated. Results showed rich genetic diversity and a wide range of phenolic contents and antioxidant activities. Sixteen varieties enriched with phenolic contents and high antioxidant activities were screened out. A significant correlation was reported among color values, TPC, TFC, ABTS and FRAP. Principal component analysis (PCA) extracted four principal components with a total cumulative contribution of 81.29%. Hierarchical cluster analysis based on the four extracted principal components resulted in a dendrogram dividing the peas into three groups. In addition, dark pea seeds have potential as a functional food in addition to their traditional role in providing dietary protein and fibre. This study provided a scientific basis for the breeding of pea varieties, development of new products and improvement of pea resource utilization.


INTRODUCTION
Pea (Pisum sativum L., 2n=2x=14) is an annual or perennial herb belonging to the genus Pisum in the leguminosae family, which can be classified into dry peas, green peas and snow peas (THAKUR et al., 2018). It was originated in western Asia, Mediterranean region, Ethiopia, parts of southwest Asia, and the outer Caucasus, and was traditionally cultivated in cold and wet climatic regions, due to their cold resistance, drought resistance, and barren tolerance (SAHA et al., 2018). China has a long history of pea cultivation and is the world's largest pea producer, which plays a decisive role in pea production. Currently, peas are mainly grown in Yunnan, Sichuan, Hubei, Shandong, Xinjiang and Gansu in China (LI et al., 2017). Peas are widely used as a supplement and feed, green manure, grains and Zhao et al. fresh vegetables (HAGENBLAD et al., 2014) due to their rich nutritional value and healing properties (SUN et al., 2014).
Morphological characteristics are based on the description of individual traits of plants and are the most traditional and intuitive way to detect genetic variation. These characteristics are important in establishing gene pool collections and efficient use of crop germplasm resources (SANTOS et al., 2012). Morphological characteristics are the result of the combination of environment and genetics, including regulatory and structural genes, and their variants revealed their own genetic underpinnings; thus, a change in phenotypic characteristics is a sign of genetic expression (DEAN et al., 1999). Although, some advanced molecular techniques have been used for the improvement of varieties, morphological characteristics are still the most basic tool for evaluating modern breeding. Morphological description has been widely used in plant germplasm identification and classification, breeding material selection and genetic diversity research.
Many researchers have studied the genetic diversity of pea germplasm resources. KENENI et al. (2015) used 12 agronomic traits to analyse the diversity of 148 pea cultivars from Ethiopia and reported that Ethiopian pea landraces had high heredity. With respect to morphology and quality, 271 domestic and foreign pea resources were analysed for genetic diversity, and it was reported that pea germplasm resources had rich genetic variation both within China and internationally (WAN et al., 2017). Analysis of pea genetic diversity under varied environmental conditions is conducive to pea germplasm resource mining and innovation and has played an important role in pea production and breeding.
The sophistication of people's diets and the development of science and technology had increasingly diversified, and the biological activity of plant phenolic compounds has become an important area of research. Phenolic compounds are derivatives of the pentose phosphate, phenylpropanoid and shikimate pathways in plants (RANDHIR et al., 2004), and phenolic acids, flavonoids and tannins are the most important dietary phenolic compounds (KING & YOUNG, 1999). Phenolic compounds are commonly reported in plants (legumes, vegetables and grains) and are the most abundant secondary metabolites (RAUTER et al., 2005). Legume seeds account for meaningful amounts of protein, starch, dietary fibre, carbohydrates, certain fatty acids and micronutrients (vitamins, trace minerals) (DUEñAS et al., 2015;TORRES et al., 2016). They are also a good source of many bioactive non-nutrient compounds, including phenolic antioxidants (MESSINA, 1999;SCALBERT et al., 2005;GEIL & ANDERSON, 1994).
As an essential component of the human diet, phenolic compounds play an important role in combating oxidative stress in the human body by maintaining a balance between oxidants and antioxidants (HUNG, 2016). Phenolic compounds are natural antioxidant due to their ability to scavenge free radicals, donate hydrogen atoms, electron, and chelate metal cations (AMAROWICZ et al., 2004). The DPPH, ABTS, and FRAP methods have been widely used to evaluate the antioxidant capacity of seeds, vegetables and fruits (SEGEV et al., 2010;XU et al., 2014;ZHU et al., 2010.).
In response to an increasingly complex market for high-quality and unique food legume products, it is imperative to develop improved food legume varieties. Therefore, the objective of this study was to analyse the morphological characteristics of 75 pea cultivars from 12 regions and two companies and select the varieties with rich genetic diversity to provide a reference for the study of pea germplasm resources. In addition, phenolic contents and antioxidant activity were measured, and high-quality antioxidant varieties were selected to improve the utilization of pea resources and to provide added value to the products. Based on this analysis, we described the correlation among phenotypic traits, phenolic contents, antioxidant activity and color of pea seeds and identify potential genotypes for pea breeders aiming to exploit the most desirable characteristics.

Determination of color value
The chromaticity of the different pea seeds was measured for L * , a * and b * using a KONICA MINOLTA CM-700d colorimeter (Konica Minolta Co., Ltd., Tokyo, Japan) as described by XU et al. (2007).

Morphological characteristics
The unified standards for pea seed observation and data on morphological traits were adopted from ZONG et al. (2005). In total, five qualitative characters (seed shape, seed surface, seed coat color, hilum color, cotyledon color) were analysed in accordance with the character value of the code after digital analysis. The character grading code is presented in table 2. For the 100-grain weight of classification code 10: 1<X-2δ, 10≥X+2δ, where the mean of every 0.5δ is used for the first expression, X indicates the average, and δ indicates the standard deviation. Seventy-five varieties were planted in their source regions (Table 1).

Chemical analyses Determination of total phenolic (TPC) and total flavonoid (TFC) contents
The extraction of pea phenolic compounds was modified slightly from the methods of previous authors (SHEM-TOV et al., 2012). The samples were ground into a powder, after which a 0.5 g portion of powder was extracted and added to 5 mL of acetone/water/acetic acid (70:29.5:0.5, v/v/v) solution in an ultrasonic bath for 20 min. Samples were then centrifuged for 10 min to extract the supernatant to determine the TPC and TFC and for ABTS and FRAP analysis. The extraction procedure was carried out in triplicate.
TPC was measured using the Folin-Ciocalteu assay (LIU et al., 2016). The standard curve was expressed as mg of gallic acid equivalent per g of dry weight (mg GAE/g DW). TFC was analysed using the aluminium chloride method (PAJAK et al., 2014). Results were expressed as mg of rutin equivalents per g of dry sample (mg RUT/g DW). Each sample was repeated three times.
The ABTS free radical scavenging capacity was determined according to method described by WANG et al. (2015). The results were expressed in terms of Trolox (µM TE/g DW) equivalent per g of dry sample. The ferric reducing antioxidant power (FRAP) was measured as previously described by MÜLLER et al. (2011). The standard curve was measured with FeSO 4 , and the results were expressed as mM Fe 2+ per g of dry sample (mM Fe 2+ /g DW).
Each sample was repeated three times.

Statistical analysis
The Shannon-Wiener diversity index (H') (MARTíNEZ et al., 2017) for morphological characters was used to measure all varieties. It was calculated according to the equation 1: Where p i is the proportion of individuals in the each class of k-class trait, and k is the number of phenotypic classes for a given trait. This index is a measure of phenotypic diversity based on frequency data. (JIAN et al., 2017).
All the original data were transformed into standardized data to eliminate the difference in the variance of each character. Pearson's correlation test (XU et al., 2009) was used to determine the correlation among variables. Principal component analysis (PCA) was performed to generate a cluster diagram (AIL-SHTAYEH, 2017). The eigenvalue and contribution percentage of each principal component axis were calculated using the correlation matrix among morphological characters, chromatic value, TPC, TFC, ABTS and FRAP for all 75 varieties. Cluster analysis was performed using scores of the first four principal components to group the varieties. All computations were carried out using SigmaPlot 12.5 and SPSS 20.0.

Variation in morphological traits
The frequency distribution for the five qualitative traits is presented in figure 1. In the collection, seed coat color was characterized by five descriptors, the cream was dominant reaching 45.3% of all collection, followed by the green, brown with stripes or spots and brown, the purple-dark was rare (Figure 1 (a)). The order of hilum color was yellow hilum＞greyish＞brown＞black (Figure 1 (b)). Four descriptors were observed for cotyledon color to differentiate pea varieties. Results showed that cream cotyledons accounted for 33 varieties, fifteen varieties had yellowish green, 14 varieties had green and 13 varieties had orange cotyledon color (Figure 1 (c)). The distribution of three seed surface was relatively uniform, with mostly dimpled surface, followed by smooth and wrinkled (Figure 1 (d)). Additionally, three seed shapes (spherical, oval-elongated, and  (e)). Hundred-mature-seed weight (100-SW) of each cultivar mainly ranged from 19.1 -21.5 g, accounting for 21.3% of the tested cultivars. A 100-SW less than 11.7 g or more than 31.4 g each accounted for 1.3% of the total (Figure 2).
For each quantitative trait and 100-SW evaluated, the descriptive statistics, including mean, maximum, minimum, coefficient of variation (CV) and H' are summarized in table 3. Among the characteristics, the ranking of CV was as follows: cotyledon color>seed coat color>hilum color>seed shape>seed surface. Overall, the qualitative traits showed considerably high H', ranging from 0.94 to 1.30. The ranking of H' was as follows: cotyledon color>seed coat color>seed shape>seed surface>hilum color. The 100-SW of pea ranged from 35.28-11.52 g. The lowest was for 16WDS005, and the highest was for New Zealand 3. Besides, the 100-SW exhibited high diversity, with an H' of 2.06.
Statistical analysis of the quality traits and quantitative trait data showed that the pea germplasm resources had abundant genetic variation. The variety of phenotypic traits of the tested resources was quite different and had greater room for improvement, which provided an excellent germplasm base for the breeding of pea varieties.

Determination of color value
The L * , a * and b * values and chroma (saturation) for pea seeds are presented in attachment. L * values ranged from 19.47 to 62.98; the greatest variety of L * was reported in Caoyuan 24, and the least variety was found in 16WDS018. The range of a * values was -0.24 to 8.73; the greatest variety was reported in Zhushawan, and the least variety was found in 16WDS003. The b * values ranged from -0.55 to 23.93, with the greatest variety in Chengwan 10 and the least variety in 16WDS018. The color of the pea seed observed with the naked eye was consistent with the results of this measurement, indicating the reliability of naked-eye observation.

Determination of the total phenolic content (TPC) and total flavonoid content (TFC)
After investigating the basic morphological characteristics of pea seeds and their rich genetic diversity, we studied the content and antioxidant properties of phenolic compounds in order to extract more potential varieties and develop functional products.
Figure 3(a) shows that the TPC of 75 pea varieties ranged from 0.27 to 1.95 mg GAE/g DW. The pea varieties with the lowest and highest TPC were 16WDS017 and 16WDS018, respectively. Twenty-two varieties had a higher-than-average TPC (0.54 mg GAE/g DW).
The TFC in this same sample ranged from 0.53 to 5.08 mg RUT/g DW (Figure 3(b)). The varieties with the lowest and highest TFC were 16WDS017 and 16WDS021, respectively. Among them, 29 varieties had a higher-than-average TFC (0.98 mg RUT/g DW).

Determination of antioxidant activity
The ABTS free radical scavenging capacity ranged from 3.04 to 22.27 μM TE/g ( Figure   Figure 2 -Hundred-mature-seed weight of pea. CV, Coefficient of variation；H', Shannon-Wiener diversity index；100-SW, Hundred-mature-seed weight. Phenotypic diversity of pea (Pisum sativum L.) varieties and the polyphenols, flavonoids, and antioxidant activity of their seeds.
In general, polyphenol contents and antioxidant activity of pea seeds had a high linear correlation, which indicated that the phenolic compounds deactivate the free radicals or convert them into stable compounds (AGOSTINIC-COSTA et al., 2015). Furthermore, significant correlations were reported between seed color parameters and TPC, TFC, and antioxidant activity. In addition, we concluded that dark (brown, brown with stripes or spots, purple-dark) seeds are a good source of bioactive materials. It is important to legume breeders to have easy-to-use markers for functional foodrelated qualities.

Principal component analysis (PCA)
Because traits were interrelated, PCA was conducted to investigate the impact of individual traits. The significance level as determined by a chi-square test was 0.000 (less than 0.01), which indicated that there was a high correlation among the indicators, and that the data were thus suitable for PCA analysis.The first four principal components contributed 81.29% of the variability among the varieties. Percentages of variation explained by the first four components were 39.93%, 21.02%, 10.85% and 9.49% (Table 5).
Distribution of 75 pea varieties and the scores of the first and second principal components are shown in Figure 5. The pea varieties could be divided into three clusters. According to these components, 30 varieties with yellow seed coat color were grouped into Cluster I. Cluster II included 10 varieties, which had a dark seed coat color and the highest phenolic contents and antioxidant activity. Lastly, 31 varieties with lower phenolic contents and antioxidant activity and a green seed coat color were clustered into Cluster III.

Clustering analysis
Based on the first four principal components, 75 pea varieties were clustered by Euclidean genetic distances. The taxonomic distance ranged from 0-25. The dendrogram divided the collected varieties into three major clusters (Clusters I, II, and III) ( Figure  6). Cluster I contained 31 varieties with yellow seed coat color. Thirty varieties with green seed coat color and lower phenolic content and antioxidant activity comprised Cluster II. Fourteen varieties with dark seed coat color (brown, brown with stripes or spots, purple-dark) and with the highest phenolic content and antioxidant activity clustered into Cluster III.

DISCUSSION
The evaluation and description of varieties with one or more desirable traits are the pre-requisites for breeding and processing strategies. The present analysis of the morphological traits of pea varieties revealed significant variation. Results showed that 5 quality traits had CVs ranging from 0.39 to 0.65, and the genetic diversity was from 0.94 to 1.30. Among them, the variation and the genetic diversity of cotyledon color and seed coat color were more abundant, which was in agreement with previous results (UPADHYAYA et al., 2013). Similarly, 100-SW exhibited a high H' and rich genetic variation. The 75 studied pea varieties have extensive genetic diversity related to ecological conditions, climatic conditions, geographical distribution, management measures, and heredity of resources (HOECK et al., 2000). Making full use of pea germplasm resources can lay the foundation for the improvement and breeding of improved varieties.
Many researchers have reported that flavonoids have several important biological activities, such as anti-photo aging, anti-allergenic, anti-inflammatory and cardioprotective effects (LEONARDO & DORÉ 2011;VAJDY, 2011;WOOD et al., 2004). Furthermore, the flavonoids, which is considered the main source of antioxidant activity in plants (DECKER, 1997;AMAROWICZ et al., 2000;TROSZYńSKA & CISKA 2002), was mainly located in these varieties. This result was in accordance with other studies (DUENÃS et al., 2006;KIM et al., 2012;BAI et al., 2017). Therefore, peas could be used as a natural source of antioxidants to replace the use of synthetic antioxidants in foods. Results showed that the TFC of 16WDS021 was significantly higher than in the other pea varieties; high-performance liquid chromatography (HPLC) could be used to assess the qualitative and quantitative characteristics of the flavonoids in this variety in the future (MEENU et al., 2016;FERRERES et al., 2017).
The color value, TPC, TFC, ABTS and FRAP showed wide variability within the 75 pea varieties, which depended on many factors, such as degree of maturity at harvest and environmental conditions (ZADERNOWSKI et al., 2005). The lowest and highest phenolic contents and antioxidant activity of the 75 pea varieties were all reported in the varieties from Qingdao Academy of Agricultural Sciences. The phenolic contents and antioxidant activity of the pea varieties in the same area were different, which may be related to environmental and genetic factors, such as cultivar, cultivation year, and harvest time (PÉREZ-BALIBREA et al., 2011).
Results obtained using Pearson correlation showed a significant linear correlation between TPC and TFC and between ABTS and FRAP, which were in agreement with previous results (ZHAO et al., 2014;YAO et al., 2012). The TPC and TFC of the pea seeds were strongly correlated with antioxidant activity, which could have possible beneficial effects for human health. Many studies (OU et al., 2002) reported that not only did the total content of polyphenolic compounds play an important role in antioxidant activity but also the type of phenolic compounds was essential. RICE-EVANS et al. (1996) showed the relationship between the antioxidant activity of flavones and their chemical  structure and claimed that B ring hydroxylation and glycosylation strongly influenced the antioxidant potential. In our study, sixteen varieties enriched with phenolic contents and high antioxidant activities were screened out. We can concentrate more on identifying the monomer phenols and measuring their contents in potential varieties in the future. Moreover, we can also explore the chemical structure and synthesis of phenols to better understand the mechanism. In addition, XU & CHANG (2007) reported that dark-colored beans, specifically black soybeans, had markedly higher phenolic contents and antioxidant activities than did pale-colored legumes. SHEM-TOV et al. (2012) reported that dark, yellow and high-color-intensity chickpea seeds contained more TPC, TFC and FRAP than did light-colored seeds. Moreover, a strong correlation was reported among peanut skin color parameters, TPC and antioxidant activity (CHUKWUMAH et al., 2009). Red and black sword beans had a higher phenolic content and antioxidant capacity than did white sword beans (GAN et al., 2016). Similar results were reported in our study. We found that TPC, TFC, ABTS and FRAP had significant negative correlations with L * and b * , which indicated that dark (brown, brown with stripes or spots, purple-dark) pea seeds may have potential health benefits and contain more TPC, TFC, ABTS and FRAP than light-colored seeds, which may be due to the dark seeds being associated with higher levels of tannins (GRELA et al., 2012). We can initially evaluate the color of a seed to judge its phenolic contents antioxidant properties. It is important to legume breeders to have an easy-to-use marker for functional food-related qualities.
The PCA results agreed with those of the cluster analysis. Varieties with high and low TPC, TFC, ABTS and FRAP were classified into separate groups. Some of the germplasm resources that derived from the same region were distributed in different groups, indicating that the clustering results were not significantly related to the geographical origin of the germplasm, which was in agreement with findings obtained in peas by KENENI et al. (2005) and ANNICCHIARICO et al. (2017). These differences may be caused by accelerated communication between germplasm resources. In the future, the genetic diversity of pea cultivars can be studied at the molecular level by means of molecular morphological markers, which can play an important role in improving pea germplasm resources. At the same time, analyses can be conducted to develop functional products that benefit human health in terms of the phenolic composition of pea seeds and the molecular mechanisms related to antioxidant activity. By improving the utilization of pea seeds as a functional food and new dietary raw material, as well as their antioxidant activity and other beneficial traits, the present findings can provide a theoretical basis for improving pea breeding.

CONCLUSION
The pea collection had rich genetic diversity and wide phenols contents. Sixteen varieties enriched with phenolic contents and high antioxidant activities were screened out. 16WDS021 had the highest TFC. TPC, TFC, antioxidant activity and chroma values, which were significantly related: the higher the phenolic contents were, the stronger the antioxidant activity. Dark pea seeds may have potential health benefits and contain more TPC, TFC, ABTS and FRAP than light-colored seeds.