ZHANG Meng-di,ZHU Xiang-mei,ZHANG Hong-yan,ZHUANG He-zhen,JIN Yi-fan,HU Kai-lin,CHENG Jiao-wen

College of Horticulture,South China Agricultural University,Guangzhou 510642,PRC

Abstract In this study,the cultivated bitter gourd inbred line ‘Foli 112’ and the wild bitter gourd accession THMC170 were used as female (P1) and male (P2)parent,respectively,to obtain 4 generations (P1,P2,F1and F2) through self-crossing and hybridization.The mixed major gene plus polygene inheritance model of plant quantitative trait was used to perform genetic analysis for 5 plant type-related traits including the number of internode,leaf length,leaf width,stem width and internode length in bitter gourd.Results showed that the inheritance of internode number,leaf length and leaf width was in accordance with E-6 model which involved two pairs of equi-dominant major genes with additive-dominant polygene model;the stem width was controlled by additive-dominance-epistasis polygene inheritance model(C-0 model);the inheritance of internode length involved 2 pairs of equi-additive major genes with additive-dominant polygene model (E-4 model).Meanwhile,the heritability (h2) of stem width and internode number was 20.08% and 39.28%,respectively,indicating that they were vulnerable to environment impacts;leaf length,leaf width and internode length had a higher heritability of 51.50%,52.16% and 64.36%,respectively;the inheritance of leaf length was mainly controlled by major gene (the heritability of major gene was 40.75%),indicating that early-generation selection was better for leaf length,while the inheritance of leaf width and internode length was mainly controlled by mutiple minor genes,indicating that it was better to select these two parameters for late generation.

Key words Bitter gourd;Plant type-related trait;Segregation analysis of quantitative trait;Genetic analysis

1.Introduction

Bitter gourd (Momordicaspp.) is an annual herb ofMomordicain Cucurbitaceae.It originated in tropical Africa and is widely distributed in tropical,subtropical and temperate regions of Africa and Asia[1].According to theCompendium of Materia Medica,bitter gourd is bitter in taste,cold in property and nontoxic in nature.It has the effects of eliminating pathogenic heat,relieving fatigue and heart-relieving and eyesight-improving.The root,stem,leaf,flower,fruit and seed of bitter gourd can be used as medicine.It is a traditional medicine in developing countries such as China,India,Brazil and Malaysia.Modern medical research results show that bitter gourd has hypoglycemic,antibacterial,antiviral and anti-tumor effects[2].Meanwhile,bitter gourd fruit is rich in minerals,vitamins,polysaccharides,flavonoids,folic acid and saponins.It is a kind of melon vegetable with unique flavor and high nutritional value.In recent years,bitter gourd is more and more favored by consumers as a healthy vegetable for both medicine and food.According to the statistics of the Ministry of Agriculture,department of rural areas,in 2016,the sown area of bitter gourd was 132 300 hm2,and the yield was 3.806 million t.

The mixed genetic model of“major gene+polygene”is a general model for quantitative traits of plants.Based on this model,GAI J Yet al.[3]developed a set of separation analysis methods suitable for genetic analysis of plant quantitative traits,which can identify major genes,minor genes and their gene effects,and estimate the corresponding genetic parameters.These genetic information can provide important theoretical guidance for breeding practice.At present,this analysis method has been widely used to study the genetic characteristics of quantitative traits in various plants[4-7].

In terms of genetic research on important agronomic traits of bitter gourd,predecessors mainly carried out genetic analysis on fruit color[8-10],single fruit weight[11-12],fruit surface structure[13-15],seed number[16],seed coat color[16-17],stigma color[14],gynoecious[13,18-20],node position of first female flower[12,21]and powdery mildew resistance[22-23],etc.In production practice,plant type is an important factor affecting the planting density and yield of crops.Improving plant type can improve crop population structure and photosynthetic situation,increase receiving light area of leaf,prolong canopy light duration andetc.,so as to improve the photosynthetic efficiency and yield of the whole population.Node number,leaf length,leaf width,stem diameter and internode length are important components of bitter gourd plant type.So far,there are few reports on genetic analysis of the above plant type traits in bitter gourd.

In this study,the bitter gourd inbred line Foli 112 (M.charantiavar.charantia) and wild bitter gourd inbred line THMC170 (M.charantiassp.macroloba) was used as female parent (P1) and male parent (P2).Four genetic generations (P1,P2,F1,and F2) were constructed by crossing and selfing.The genetic analysis of five plant type traits (node number,leaf length,leaf width,stem diameter and internode length) of bitter gourd was carried out by using the method of major genes of quantitative traits in plants+genetic segregation analysis of polygenic mixtures,the purpose of this study was to provide reference for plant type breeding of bitter gourd.

2.Materials and Methods

2.1.Test materials

F1and F2populations were obtained by crossing and selfing with bitter gourd inbred line Foli 112 as female parent (P1) and wild bitter gourd inbred line THMC170 as male parent (P2).On March 15th,2018,the seeds of P1,P2,F1and F2generations were soaked,germinated and seeded at the same time.On March 28th,the seedlings were planted in the vegetable experimental base of South China Agricultural University when they grew up to two leaves.

2.2.Character determination

40 d after planting,the number of fully flattened leaves on the main stem minus 1 was taken as the phenotypic value of node number per plant.At the same time,the leaf,stem diameter and internode length of the 18th,19thand 20thnodes of the main stem were measured.The specific methods are as follows.

Leaf length and leaf width:the maximum leaf length and maximum leaf width of three fully expanded functional leaves at the 18th,19thand 20thnodes of the main stem were measured with a ruler,and the average value was taken as the phenotypic value of leaf length and leaf width per plant.

Stem diameter:use vernier caliper to measure the maximum diameter of main stem between the 18thand 19thnodes,repeat the measurement for 6 times,and take the average value as the phenotypic value of stem diameter.

Internode length:measure the length of the main stem of the 17th～18th,18th～19thand 19th～20thinternodes on the main stem with a ruler,the average value was taken as the phenotypic value of internode length per plant.

2.3.Data analysis

Excel and SPSS software were used for statistical analysis of table data.Kolmogorov-Smirnov(D test) and Shapiro-Wilk (W test) were used to test the normal distribution of F2segregating generations.IfP＞0.05,it was consistent with normal distribution.In this paper,the major gene+polygene mixed genetic segregation analysis method[24]of plant quantitative traits was used to carry out joint analysis on four generations,and the distribution parameters of relevant components in the mixed distribution were estimated by maximum likelihood method and IECM (Iterated expectation and conditional maximization) algorithm.Then,three candidate models with lower AIC value were selected by using AIC criterion (akaike’s information criterion),and then a group of fitness tests and likelihood ratio tests were conducted.There were five statistics,namelyU12,U22,andU32,nW2andDn,among whichU12,U22,andU32,were homogeneity test,andnW2was Smirnov test,Dnwas Kolmogorov test.The optimal genetic model was selected and genetic parameters such as major gene and polygene effect value,variance and heritability were estimated.The software was provided by National Soybean Improvement Center of Nanjing Agricultural University.

3.Results and Analysis

3.1.Phenotypic analysis of 5 plant type traits in 4 generations of balsam pear

The results of statistical analysis (Table 1)showed that after Foli112 and THMC170 were planted 40 d,the number of nodes on the main stem and the leaf length,leaf width,stem diameter and internode length of the 18th,19thand 20thnodes were significantly different (P＜0.05);in addition to the positive transgressive heterosis of node number of F1hybrid plants,the other four plant type traits showed mid-parent heterosis,among which leaf length and internode length were close to high value parent Foli112 (P＞0.05),and showed complete dominant genetic characteristics.

It can be seen from Fig.1 that the five plant type traits of bitter gourd were continuous variation,which conforms to the genetic characteristics of quantitative traits;among them,except internode length which is positively skewed distribution,the other four plant type traits conform to normal distribution (P＞0.05).

3.2.Analysis of the best genetic model for five plant type traits of bitter gourd

3.2.1.Number of nodes

According to the minimum AIC value,E-4,E-5 and E-6 models were better (Table 2).The results of fitness test for the three models were shown in Table 3.The statistical significance of the three models was 1.Further,the likelihood ratio test showed that there were no significant differences between E-6 and E-4 (χ2=0.000 2,P＞0.05) and between E-6 and E-5(χ2=0.000 0,P＞0.05).Because the AIC value of E-6 model was the smallest,the E-6 model was the best genetic model for node number of bitter gourd,that was,the inheritance of node number of bitter gourd a conformed to the mixed model of 2 equal dominant major genes+additive-dominant polygenes.

Table 1 Phenotypic values of five plant type traits between bitter gourd parents and their F1generation

Fig.1 Frequency distribution histogram and normal curve of five plant type characters of bitter gourd in F2population

3.2.2.Leaf length

According to the principle of minimum AIC value,C-0,E-5 and E-6 models were better (Table 2).The results were shown in Table 4.Among the 20 statistics of four generations,the statistical significance of the three models was 2.Furthermore,there were nosignificant differences between E-6 and C-0 (χ2=0.007 5,P＞0.05) and between E-6 and E-5 (χ2=0.000 3,P＞0.05).Because the AIC value of E-6 model was the smallest,the E-6 model was the best genetic model of bitter gourd leaf length,that was,the inheritance of bitter gourd leaf length was consistent with the mixed model of 2 equal dominant major genes+additivedominant polygenes.

3.2.3.Leaf width

According to the principle of minimum AICvalue,E-3,E-4 and E-6 models were better (Table 2).The results of fitness test for the three models were shown in Table 5.The statistical significance of the three models was 2.Further,the likelihood ratio test showed that there were no significant differences between E-6 and E-3 (χ2=0.000 6,P＞0.05) and between E-6 and E-4 (χ2=0.000 3,P＞0.05).Because the AIC value of E-6 model was the smallest,E-6 model was the best genetic model of balsam pear leaf width,that was,the inheritance of leaf width of bitter gourd was consistent with the mixed model of 2 equal dominant major genes+additive-dominant polygenes.

Table 4 Suitability test of genetic model of bitter gourd leaf length

3.2.4.Stem diameter

According to the minimum AIC value,C-0,D-0 and E-1 models were better (Table 2).The results of fitness test for the three models were shown in Table 6.The statistical significance of the three models was 2.Further,the likelihood ratio test showed that there were no significant differences between C-0 and D-0 (χ2=0.001 2,P＞0.05) and between C-0 and E-1(χ2=0.003 5,P＞0.05).Because the AIC value of C-0 model was the smallest,it was determined that C-0 model was the best genetic model for stem diameter of bitter gourd,that was,the inheritance of stem diameter of bitter gourd conformed to the polygene genetic model of additive+dominance+epistasis.

Table 5 Suitability test of genetic model for leaf width of bitter gourd

Table 6 Suitability test of genetic model for stem diameter of bitter gourd

3.2.5.Internode length

According to the minimum AIC value,E-4,E-5 and E-6 models were better (Table 2).The results of fitness test for the three models were shown in Table 7,and the statistical significance number of the three models was 0.Further,the likelihood ratio test showed that there were no significant differences between E-4 and E-5 (χ2=0.001 5,P＞0.05) and between E-4 and E-6 (χ2=0.000 0,P＞0.05).Because the AIC value of E-4 model was the smallest,E-4 model was the best genetic model for the internode length of bitter gourd,that was,the inheritance of leaf internode length of bitter gourd accorded with the mixed model of 2 equal additive major genes+additive-dominant polygenes.

3.3.Estimation of genetic parameters of 5 plant type traits in bitter gourd

The estimated results of genetic parameters of the best genetic model for five plant type traits of bitter gourd were shown in Table 8.Node number,leaf length and leaf width of bitter gourd was controlled by 2 equal dominant major genes+additive-dominant polygenes (E-6 model).The additive effect of major genes was equal to the dominant effect,and the effect values were -1.75,1.16 and 0.86,respectively;the polygene genetic effect of node number and leaf width was dominant,and the polygene dominance of node number was 29.21,which showed a very high over dominance,which was consistent with the statistical results of node number in F1generation (Table 1).The internode length of bitter gourd was controlled by 2 equal additive major genes+additive-dominant polygenes (E-4 model).The additive effects of major genes and polygenes were 0.41 and 0.72 respectively,and the polygene effects were mainly additive effects.

Among the five plant type traits of bitter gourd,the major gene+polygene heritability of internode length was the highest (64.36%),and polygene was the main control (polygene heritability was 59.46%).Stem diameter was only controlled by additivedominant-epistatic polygenes without major gene effect,and the polygenic heritability was only 20.08%,indicating that it was greatly affected by environment.The heritability of major gene+polygene of node number was low (39.28%),indicating that it was easily affected by environment.The major gene+polygene heritability of leaf length and leaf width were 51.50%and 52.16%,respectively.The inheritance of leaf length was mainly controlled by major gene (40.75%of major gene inheritance),while the inheritance of leaf width was mainly controlled by polygene(polygenic heritability was 38.26%).

Table 7 Fitness test of genetic model for internode length of bitter gourd

Table 8 Estimated values of genetic parameters of the best genetic model for five plant type traits of bitter gourd

4.Discussion

Segregation analysis provides the general concept of major gene+polygene system and estimates of additive,dominant and epistatic effects of genes.Previous studies showed that segregation analysis and molecular markers had relatively consistent results in detecting major genes (QTL) of quantitative traits.Meanwhile,the application conditions of segregation analysis method were simple and the cost was low,hence it can be used as QTL mapping pre analysis.In this study,the segregation population was constructed by using the bitter gourd inbred line Foli 112 and wild bitter gourd inbred line THMC170 with significant differences in genetic background and plant type performance.The four generations of P1,P2,F1and F2were analyzed by segregation analysis of quantitative traits,and the genetic analysis of five main plant type characters of bitter gourd was carried out for the first time.The purpose of this study was to provide reference for QTL mapping and plant type breeding in bitter gourd.

The results showed that the inheritance of node number,leaf length and leaf width of bitter gourd conformed to the mixed model of 2 equal dominant major gene+additive-dominant polygene (E-6 model),the inheritance of stem diameter conformed to the additive-dominant-epistatic polygene genetic model(C-0 model),and the internode length conformed to the mixed genetic model of 2 equal additive major genes+additive-dominant polygenes (E-4 model).In addition,it was found that the genetic variance of stem diameter and node number of bitter gourd was lower than that of environment,which indicated that environment had great influence on them.The heritability of leaf length,leaf width and internode length were all high,among which the inheritance of leaf length was mainly controlled by major genes,which indicated that early generation selection should be made for leaf length.The inheritance of leaf width and internode length was mainly controlled by polygene additive effect,which indicated that the selection of leaf width and internode length should be carried out in the higher generation.