Heterosis And Genetic Diversity Of Japonica Rice (Oryza Sativa L.) In Taihu Lake Valley

During the past 30 years, more than one hundred hybrid varieties in japonica rice(Oryza sativa L.) were released in China. However, the planting areas of japonica hybridrice was about 2.5×10⁴ hectares each year, occupied only 3ï¼…of total areas of japonica ricein China. Compared with indica hybrid which occupied 55ï¼…of total areasof indica rice,there are a great of areas to be planted with japonica hybrid. Taihu Lake Valley is one ofareas of high yield of japonica rice in China. From the beginning of 1990's, japonica hybridhas been cultivated in this region. And now the region has become one of the most activeareas for japonica hybrid breeding and popularization. Nevertheless, with the developmentof pure line breeding, the genetic basis of parents for hybrid breeding becomes more andmore narrow and the competitive heterosis of hybrid becomes more and more weak, whichconstrained further development of japonica hybrid in this region. In order to benefitjaponica hybrid breeding, heterosis and combining ability of parents newly bred, andgenetic diversity of landrace in japonica rice in Taihu Lake Valley were studied in thisdissertation.1. In order to evaluate the heterosis of late-mature japonica rice and combining abilityof their parents used currently in Taihu Lake Valley, a diallel cross study was conducted.Seven CMS lines with BT cytoplasm and seven restorers were used as parents. CMS lineswere newly bred in Wuzhong Institute of Agricultural Sciences. Restorers were introducedfrom other provinces or newly bred in Suzhou. With NCâ…¡mating design, 49 combinations(7×7) were made by paper bag isolation in Hainan. The plots of F₁s and parents werearranged in the field in a randomized complete block design with three replications inSuzhou. Ten agronomic characters were investigated. Research results were as follows:(1) Both positive and negative heterosis existed in all the 10 characters. Amongthem, heterosis of spikelets per panicle was the highest, showing the average heterosis22.6ï¼…of the control and over 8.6ï¼…of high parents. The average value of the heterosis overhigh parent in yield per plant is 9.5ï¼…, the highest among ten characters, ranging from-38.0ï¼…to 43.9ï¼…. The competitive heterosis of yield per plant ranged from -54.8ï¼…to 10.3ï¼…. while panicles per plant had the lowest competitive heterosis and heterosis overhigh parents, average values were -17.6ï¼…and -13.3ï¼…respectively.(2) For character of yield, 852-77A, Hou 36A and 9522A were elite CMS lines;161-10, 3402, HP121 were excellent restorers; 9522A/161-10, 9522A/C166, 852-77A/3402were superior combinations.(3) Both additive and non-additive genetic effects were existed in all the 10characters. For the character of yield per plant, filled grains per panicle, panicles per plant,non-additive effects were more important than those of additive effects. The other sevencharacters, days from sowing to heading, plant height, panicle length, grain length, grainwidth, spikelets per panicle, 1000-grain weight, additive effects were more important thanthose of non-additive effects. Among the seven characters, the variance of GCA of CMSlines was larger than those of restorers for character of 1000-grain weight, and reverse forthe other six characters.From the results above, it could be seen that yield heterosis of japonica hybrid dependson SCA of parents, while heterosis of other agronomic characters was mainly due to GCAof restorers. Therefore, enhancing GCA of hybrid parents, especially in restorer lines, is thekey step in japonica hybrid breeding in this region.2. Genetic diversity of japonica rice landraces in Taihu lake valley was evaluated by19 characters and SSR markers for the purpose of genetic improvement of hybrid parents.In this section, phenotype diversity, core collection construction and DNA diversity basedon SSR marker were studied.(1) Japonica rice landraces (Oryza sativa L.) in Taihu lake valley consists of 823accessions (Total population, T), of which 72 are medium japonica rice (â… subset), 172late early-maturity japonica rice (â…¡subset), and 579 late medium and late-maturityjaponica rice (â…¢subset). Nineteen characters, include four morphological descriptors andfifteen agronomic characteristics were used to estimate its phenotypic diversity. The resultsshown that, four morphological descriptors, there were seventeen phenotypes for huskcolor, twelve for apicutulus color, five for awn type and four for grain type, respectively.The value of fifteen quantitative characters had large range, the range of coefficientvariance (CV) was 4.6％～33.7ï¼…. the value of yield per plant, filled grains per panicle,spikeletes per panicle were the hightest, and grain length, 1000-grain weight were thelowest. For the 19 traits, the phenotypic diversity index (H′) ranged from 0.757 (grain shape) to 1.930 (filled grains per panicle), averaging 1.559. Phenotypic variance ofaccessions headed after 10 September was larger than that of accessions headed before 10September (including this day). Principal component analysis showed that plant height,panicles per plant, spikelets per panicle, filled grains per panicle, spikelet density,1,000-grain weight, yield per plant, and panicle neck length were important traits inexplaining the phenotypic diversity of japonica rice landraces from Taihu Lake region.There were significant phenotypic correlations among the various characters. Four ofthese, between panicles per plant and yield per plant, spikeletes per panicle and filledgrains per panicle, spikeletes per panicle and spikeletes density, filled grains per panicleand spikeletes density, explained 49ï¼…variation in the other trait in all four groups.Among the 3 subsets,â…¢subset had the most abundant phenotypes in characters of huskcolor and apicutulus color, and largest value of eight quantitative characters' variance,nine quantitative characters' CV, and average H' of eighteen characters. The resultsindicated that genetic redundance existed in the original total population, and it is necessaryto construct core collection for easy to study further.(2) Based on phenotypic values, Methods of constructing core collection wereevaluated with combination of two genetic distance (Euclidean distances and Mahalanobisdistances), six hierarchical clustering methods (the single linkage, complete linkage,median linkage, centroid, un-weighted pair-group average and Ward's method), threesampling method (stepwise random, stepwise preferred and stepwise deviation) and ninesampling ratio (range from 10％～50ï¼…). Six testing parameters, including averageShanon-Weaver index percentage (H′％), mean difference percentage (MDï¼…), variancedifference percentage (VDï¼…), coincidence rate (CRï¼…), variable rate (VRï¼…) andphenotype retained percentage (PRï¼…), were employed to evaluated validity of samplingscheme. It was found that the best strategy for constructing core collection in this study wasthe combination of Euclidean distances, un-weighted pair-group average hierarchicalclustering methods, preferred sampling and 10～20ï¼…sampling ratio. By using the strategy,a core collection of Japonica rice landrace in Taihu Lake Valley was established. The corecollection was composed of 129 accessions and accounted for 15.7ï¼…of initial totalpopulation. Compared with initial total collection, core collection had no difference in themean, maximum and minimum value and variance of the traits., and the genotypes of 4quality characters were remained, CV(123.4ï¼…) and H' (106.91ï¼…) of traits were higher.(3) One hundred twenty-nine varieties of core collection were evaluated by means of sixty SSR markers distributed over the whole rice genome for purpose of estimatinglandraces' genetic diversity on DNA level. A total of 216 alleles were detected at fifty-threeSSR loci. The rate of polymorphic loci detected was 88.3ï¼…, and the number of alleles permarker ranged from two to seven, with an average of 4.08. The PIC value of SSR primerranged from 0.031 (RM125) to 0.773 (RM349), with an average of 0.413. Those SSR lociwith more than three alleles accounted for 71.7 percent of total SSR loci studied, anddistribution frequency of 47.7 percent alleles was less than 0.1. These results showed therewere not only abundant genetic variances but also plenty of rare alleles retained in corecollection. Comparing diversity of 12 chromosomes according to average values of allelesnumber and PIC, chromosome 5 showed the greatest genetic diversity among the others.The genetic diversity of japonica landrace rice core collection in Taihu lake Valley washigher than modern breeded japonica rice in Jiangsu & Zhejiang province according topolymorphic loci number, total alleles number, average PIC value, He, genetic distance andsimilarity coefficient among varieties. The result of clustering analysis with UPGMAmethod based on SSR markers reflected the degree of genetic similarity and their pedigree.All Taihu Lake Valley landraces could be uniquely distinguished, and at similaritycoefficient of 0.63, bred varieties and landraces could be obvious distinguished. There werebig genetic difference between landraces subset and bred varieties subset. These resultscould be used for japonica hybrid breeding. The phenomenon that varieties shared the samenames did not always show a close genetic relationship was found in this article.3. One hundred and fifteen varieties in japonica rice (Oryza sativa L.) were analyzedfor endosperm storage proteins by sodium dodecyl sulfate polyacrylamide gelelectrophoresis (SDS-PAGE) to estimate their polymorphism for the purpose of geneticimproving and variety identification. Research results were as follows:(1) Nineteen types of profile were identified according to presence/absence of 65×10³ bands, staining intensity of 70×10³, 60×10³, 57×10³, 37×10³ï½ž39×10³, 22×10³ï½ž23×10³, 13×10³ and 10×10³ bands, migration velocity of 35×10³(a-4) bandand band number at 57×10³ location.(2) Staining intensity of 60×10³band and amylose contents (AC) assayed in ricevarieties had obvious correlation, the high staining intensity of 60×10³ band representedhigh AC in rice varieties, and vice versa. This kind of correlation was also found between37×10³ï½ž39×10³,22×103³ï½ž23×10³,13×10³ band and protein contents (PC) assayed in rice varieties. This result can be used for selection of this two characters in earlygeneration of hybrid breeding.(3) The similarity coefficients ranged from 0.75 to 1.00 among the rice varietiesstudied. Cluster analysis was carried out by an unweighted-pair group average method witharithmetic mean. Three distinct groups were identified at the similarity coefficient level of0.894. The first group included eight varieties with high amylose content, the secondgroup contained fifteen varieties with high protein content, and the third group had theremaining ninety-two varieties, which accounted for 80ï¼…of total varieties. Clearrelationship between ecotypes distinguished by maturity and groups revealed by clusteranalysis was not found in this study. Only the group of high amylose linked withmedium-maturity medium japonica ecotype. This indicated low genetic diversity amongdifferent japonica rice varieties on protein level, and it was difficult to identify japonicahybrid seed purity by using seed storage protein.