Genetic Diversity of Leaf Phenotypic Traits in Walnut Germplasm Resources

YU Qiu-xiang, LIU Jing, LI Yang, LIU Jin-li, BAI Zhong-kui

Changli Institute of Pomology, Hebei Academy of Agricultural and Forestry Sciences, Changli 066600, PRC

Abstract In order to reveal the genetic diversity and variation sources of leaf phenotypic traits in walnut, the coefficient of variation, Shannon-weaver index,phenotypic differentiation coefficient, and the correlation of 9 leaf phenotypic traits for 251 walnut germplasm resources from two walnut populations were analyzed.The results showed that the averages of the variation coefficients and Shannonweaver indexes of the leaf phenotype traits were 17.45% and 1.86 respectively. The F values of most inter-population traits and all intra-population traits were highly significant, and the F values of inter-population leaf shape index, leaflet number,length and leaf length were very large. The averages of inter-population and intrapopulation differentiation coefficients were 9.15% and 90.85% respectively. The leaf size, leaf thickness, and leaflet number were highly correlated with trunk perimeters.According to the results of principal component analyses, the 9 traits can be simplified into four principal components, which represent leaf size, petiole, leaflet number and leaf shape, respectively. There were extensive variations and rich genetic diversities in leaf phenotypic traits of walnut, and the intra-population variation was the leading source of phenotypic trait variation. Leaf shape index, leaflet number,internode length and leaf length can be used for population classification. The size and thickness of leaves and the number of leaflets will affect the accumulation of nutrients in walnut tree, and then affect its trunk perimeter.

Key words Walnut; Leaf; Phenotypic traits; Populations; Diversity

1. Introduction

Phenotypic diversity was the major indicator and research focus of species diversity, which can help understand the gene stability and breeding potential of walnut population, providing theoretical guidance for the exploration, usage, and innovation of walnut germplasm resources.Recent research on the genetic diversity of walnut phenotypic traits had offered reliable theoretical basis for resource introduction, parent selection, heterosis utilization, and genetic improvements. Walnut,which was one of the tree species native to China, had rich genetic diversity. Research on the phenotypic traits of walnut had attracted great attention. Some of the existing researchapplied molecular markers to reveal the genetic diversity of walnut germplasm resources in Xinjiang and Qinghai. LIU B Y

et al

.suggested that the intra-population genetic differentiation was higher than inter-population genetic differentiation for walnut. Previous research reporedthe phenotypic traits of Xinjiang thin-shelled walnut, Xinjiang wild walnut, Yunnan major walnut cultivar, and Xizang

Amygdalus mira

(

Koehne

) to reveal the genetic diversity of walnut. LUO X

et al

.and ZHAO S

et al

.investigated the genetic diversity of walnut phenotypic traits from the perspective of population structure. LONG J C

et al

.evaluated the correlation between the leaves and fruit shape of Xinjiang thinshelled walnut. Nevertheless, very few research had been reported on phenotypic traits of walnut leaves.Currently, only PEI X N

et al

.and ZHANG S M

et al

.studied the variation and diversity of leaf phenotypic traits of

Juglans mandshurica

clone and Dabie Mountain walnut, respectively, whereas the leaf phenotypic diversity and variation of ordinary walnut had not yet been studied. The variation of leaf traits was an important indicator for genetic variation and taxonomy of plants. Leaf, which was the key plant organ for photosynthesis and transpiration, can best reflect the effects of environmental factors on plants and their environmental adaptabilities. It was therefore natural to take leaf as a crucial indicator for plant growth, tree vigor, fruit yield, and fruit quality.

In this research, 251 walnut resources from 2 populations were sampled to evaluate the 9 phenotypic traits of walnut leaves and to reveal the correlations among the genetic diversity level, phenotypic differentiation coefficients, and walnut leaf phenotypic traits. The findings had demonstrated the genetic diversity level and variation of walnut leaf phenotype and their correlation with tree vigor, which can provide theoretical basis for the resource collection,early-stage evaluation, new germplasm cultivation,selective breeding, and genetic improvement of walnut germplasm resources.

2. Materials and Methods

2.1. Testing materials

The testing materials were formed by 251 pieces of walnut resources introduced in 2011 from the walnut producing area of Shimen (Hebei) and Hetian (Xinjiang); these plants were grown in the field of Changli Pomology Institute, Hebei Academy of Agricultural and Forestry Sciences, including 150 pieces of typical Shimen (Hebei) walnutand 101 pieces of Xinjiang walnut(which had rich genetic diversity and were taken as the major parent for improving the early-fruiting and high-yield property of walnut); the spacing in and between rows was 3 m×5 m; the age of each tree was 9. All these trees were treated under the same pruning, watering, fertilization,and pest control measures.

2.2. Testing methods

The sampling was conducted in early August,2020 and the procedures were as follows: select a fruiting branch near the periphery of the crown with a southern exposure; pick the 3compound leaf that was healthy and fully mature; sample at least 15 compound leaves from each tree; count the leaflet number of each compound leaf (LFN//leaves);measure the petiole length (PL//cm) and internode length (IL//cm); take the 2leaflets of the selected compound leaves and seal the mina Ziploc bag; put these bags in an incubator with ice cubes and bring them back to the laboratory; clean these leaflets by tap water once and then by distilled water once; dry the leaves in the air and measure the leaf fresh weight(LFW//g). Scan the leaves with the CI-202 portable laser leaf area meter and record the leaf surface area(LSA//cm), leaf length (LL//cm), leaf width (LW//cm), leaf perimeter (LP//cm), leaf shape index (LSI)

etc

. Use a digital caliper to measure the leaf thickness(LT) at the middle of the leaf where there was no leaf vein. Put the leaves in an electro thermal blowingdry box for fixation and dry them to constant weight;measure the leaf dry weight (LDW//g); measure the trunk perimeter of each tree at a distance of 30 cm to the ground (TP//cm).

2.3. Data analysis

Analyze the variance, correlation, and principal components according to each traitby SPSS 22 statistical software; calculate the coefficient of trait variation (

CV

) by the method of JIANG X B

et al

.and calculate the Shannon-weaver genetic diversity index (

H

') according to the previous methods.Use hierarchical classification design variance analysis to evaluate each trait and phenotypic differentiation coefficient (

V

) to represent the percentage of total genetic variation that was inter-population variation based on the algorithm of LI Y.

3. Results and Analysis

3.1. Variation features of walnut leaf phenotypic traits

According to Table 1, the variation coefficient and genetic diversity index of the leaf phenotypic traits of 251 walnut resources were 17.45% and 1.86 respectively, which indicated extensive variation and rich genetic diversity. Both the variation coefficient and genetic diversity index varied drastically from 11.20% to 24.15% and 0.89 to 2.07, respectively.The results of variation coefficients, ranked in descending order were as follows: internode length>leaf surface area>petiole length>leaf shape index>leaf thickness>leaflet number>leaf length>leaf perimeter>leaf width, while the genetic diversity indexes followed the order of leaf width=leaf length=petiole length>leaf perimeter>leaf surface area>leaf thickness>internode length>leaf shape index>leaflet number. For both the variation coefficient and the genetic diversity index, the leaf surface area(21.77 and 2.04) and petiole length (21.59 and 2.07)were very large, which meant these two traits had high disperse on level and homogeneous distribution frequency among the testing materials. The internode length reached its peak value at 24.15 for variation coefficient and hit the bottom level at 1.99 for genetic diversity index, indicating high dispersion level and uneven distribution frequency. On the contrary, the leaf width peaked at 11.20 for genetic diversity index and reached the bottom level at 2.07 for mutation coefficient, indicating low dispersion level and even distribution frequency. The leaflet number was very low for both variation coefficient (15.07) and genetic diversity index (0.89); there were only four figures—5, 7, 9 and 11 with a distribution frequency of 4.9%, 41.9%, 52.6% and 0.7%, respectively. This meant leaflet number had low dispersion level and uneven distribution among the testing materials.

3.2. Variation of walnut leaf phenotypic traits between populations and inside a population

According to Table 2, there were 9 leaf phenotypic traits showing highly significant variation between Shimen and Xinjiang walnut populations; the inter-population

F

values changed drastically from 1.44 to 58.91. The inter-population

F

values of 6 traits had reached highly significant levels. Specifically,the leaf shape index, leaflet number, internode length,and leaf length presented very large inter-population

F

values, indicating extensive inter-population variations, and therefore can be used for populationclassification.

Table 1 Variations of leaf phenotypic traits of walnut

Table 2 Mean value and variance analysis of leaf phenotypic traits for 2 walnut populations of walnut

The intra-population

F

values ranged mildly from 7.54 to 11.37, whereas all the values had reached highly significant levels. This meant rich variations also existed within the same population.

3.3. Phenotypic differentiation of leaf phenotypic traits between populations

The variations of all phenotypic traits were divided into inter-population variation, intrapopulation variation and individual variation (random errors) based on nest design. The percentage of variance components was an important indicator of variance resource. As shown in Table 3, the means of variance component percentage for inter-population,intra-population and inter-individual random error were 3.92%, 36.96% and 59.12%, respectively. The phenotypic differentiation coefficients of 9 phenotypic traits varied from 0.86% to 26.32%, which peaked for leaf shape index and hit rock bottom for petiole length. The mean of the inter-population phenotypic differentiation coefficients for 9 phenotypic traits was 9.15%, while the average intra-population phenotypic variation accounted for 90.85%. It meant the intrapopulation variation was much higher than interpopulation variation.

Table 3 Variance components and differentiation coefficients of leaf phenotypic traits between and within walnut populations

3.4. Correlation and principal component analysis of walnut leaf phenotypic traits

Table 4 showed the results of correlation analysis of 9 walnut leaf phenotypic traits, leaf fresh weight,leaf dry weight, and trunk perimeter. The correlations among different traits were quite complicated. Totally,there were 66 pairs of correlated traits, among which 43 pairs reached highly significant levels (

P

<0.01) and 6 pairs at significant levels (

P

<0.05). The leaf fresh weigh was highly correlated with leaf dry weight,just like its correlation with leaf surface area, leaf length, leaf width, leaf perimeter, leaf thickness, and trunk perimeter. Besides, the correlation coefficients of the leaf fresh weight with the latter 6 phenotypic traits were larger than that with leaf dry weight.Totally, there were 8 traits highly correlated with trunkperimeter, which were leaf surface area, leaf length,leaf perimeter, leaf shape index, leaf thickness, leaflet number, leaf fresh weight, and leaf dry weight. The size and thickness of the leaf would affect its dry weight and fresh weight, and then the trunk perimeter and tree growth.

Table 4 Correlation analysis of walnut leaf phenotypic traits

The analytical results of principal components were listed in Table 5. The major information of walnut leaf phenotypic traits was revealed by the first 4 principal components, which presented a contributive percentage of 38.73%, 17.95%, 15.89%and 12.93%, respectively; the total contributive percentage was 85.50%. The 1principal component showed large vector values in respect to leaf surface area, leaf length, leaf width, and leaf perimeter, which were related to leaf size. The largest vector valuesof the 2principal component appeared inpetiole length and internode length, representing petiolerelated traits. The 3principal component showed large vector values in terms of leaflet number and leaf width, indicating leaflet-related traits. The 4principal component exhibited large vector value in leaf shape index, revealing leaf shape-related traits.

Table 5 Principal component analysis of walnut leaf phenotypic traits

4. Discussion

Genetic diversity was a major part of biodiversity.There were rich genetic variations in the natural population of most species, serving as raw materials for evolution. Researches on genetic diversity were conducted to figure out the origin of species, to predict provenance adaptability, to estimate the distribution of genetic resources, to develop and utilize germplasm resources, and to investigate cross breeding and parent selection. Phenotype was the result of genotype and the environment. Phenotype research was economic and visual in comparison with molecular marker technology, and therefore had been accepted as a basic method and approach to germplasm resources.Walnut, which was a perennial woody plant, had complex genetic background and long breeding cycle. Research on the genetic principle of walnut phenotypic traits can help improve the breeding efficiency and accelerate the breeding process.

In this research, 251 pieces of walnut resources from 2 populations were sampled to analyze the variation and phenotypic diversity of walnut leaf phenotypic traits. The arithmetic means of trait variation coefficients and genetic diversity indexes were 17.45% and 1.86 respectively, which had proved extensive variations and rich genetic diversities among testing materials. These results were in accord with earlier research. For most of the inter-population traits and all of the intra-population traits, the

F

value differences were at highly significant levels,indicating extensive variations between and inside walnut populations. The inter-population

F

values of leaf shape index, leaflet number, internode length and leaf length were very large and their differences had all reached highly significant levels. Therefore, these 4 traits can be used for inter-population classification.According to the data analysis and field observation,Xinjiang walnut population had large and long leaves,big leaf shape index, small leaflet number, and long internode length. In contrast, Shimen walnut had small and short leaves, small leaf shape index, a lot of leaflets and short internode length. Further analysis of variation source indicated that the average interpopulation phenotypic differentiation coefficient of 9 traits was only 9.15%, while the average intrapopulation phenotypic variation accounted for 90.85%,which meant that the intra-population variation was much higher than inter-population variation. Superior intra-population variety, in other words, was more important in the breeding process. The findings of LIU B Y

et al

.and ZENG B

et al

.about the phenotypic traits of walnut fruit also suggested that intra-population variation was the major source of phenotypic variation, which was in accord with the current research. According to ZHANG S M

et al

.and LI Y

et al

., the leaves from Dabieshan wild walnut population and the fruiting maternal branch showed rich phenotypic diversity, and the variations were mainly found between different populations.ZHANG Y

et al

.collected 548 branch and leaf samples and 625 flower samples from 13 species of

Pyrus

to analyze their phenotypic traits; rich variations were found in these phenotypic traits, whereas the variations of branches and leaves were mainly inter-population variations, while the variations of flowers were mostly intra-population variations. The phenotypic diversity of plants was a comprehensive reflection of genetic and environmental diversity.Complex intermediate links such as genetic expression, individual development and regulation do exist between genotype and phenotype; stable genetic qualitative traits and the quantitative traits of large size samples could effectively reveal the genetic and variant patterns of species. Inter-population variation was larger than intra-population variation in the case of large sample size, a broad span of species, or distinct eco-environmental difference. The phenotypic trait variation of different organs from materials with the same genetic backgrounds might be either intra-population or inter-population variations.These findings provided useful information for the follow-up research of our team. In order to reveal the phenotypic variation origin of walnut germplasm resources, it was necessary to increase the sample size of the testing materials and to expand the interpopulation span and genetic background.

There were 4 principal components in the phenotypic traits of walnut leaves, which presented a total percentage of 85.5% and represented the leaf size, leaf petiole, leaflet number and leaf shape respectively. According to the results of trait correlation analysis, the leaf surface area, leaf length,leaf width, leaf perimeter, leaf thickness and trunk perimeter were highly correlated with leaf fresh weight and leaf dry weight; the leaf surface area, leaf length, leaf perimeter, leaf shape index, leaf thickness,leaflet number, leaf fresh weight and dry weight were highly correlated with trunk perimeter. To conclude,the size and thickness of walnut leaf would influence its fresh weight and dry weight, and then affect the accumulation of nutrients in plant body and stimulate the growth of trunk perimeter. This was in accord with the sensorial understanding of our field sampling and investigation—the trunk of a walnut tree covered with dense, green, large and thick leaves was much stronger than the one with small and thin leaves. Besides, the leaflet number of the former was larger than that of the latter. This conclusion provided theoretical basis for elite breeding.

5. Conclusion

The analysis of 251 walnut resources from 2 populations was conducted to evaluate the diversity,Shannon-weaver genetic diversity index, interpopulation and intra-population genetic variation,and trait correlation of walnut leaf phenotypic traits.The analytical results had proved extensive variations and rich genetic diversities of walnut leaf phenotypic traits in and between walnut populations, and the intra-population variation was much higher than interpopulation variation. However, the inter-population differences of leaf shape index, leaflet number,internode length and leaf length were very large and have all reached highly significant levels. For this reason, these 4 traits were selected for population classification. The leaflet number, size and thickness of walnut leaf would influence the accumulation of nutrients in tree body and further affect the growth of tree trunk. The 9 leaf phenotypic traits could be simplified into 4 principal components, representing leafsize, petiole, leaflet number and leaf shape respectively.