【作者】 宋高远；

【导师】 杨代常；

【作者基本信息】 武汉大学 ， 遗传学， 2014， 博士

【摘要】 优良亲本的选择在传统育种与杂交水稻育种过程中均起到至关重要的作用。但亲本选择是一个需要耗费大量时间并且效率低下的繁重工作,更为复杂的是并非所有亲本的优良性状均能在其后代得到表现,不同亲本对杂种后代表现出不同的遗传效应,为了解释这一现象和评价亲本的优劣,Sprague与Tatum于1942年提出了一般配合力(general combining ability, GCA)的概念用于评价育种亲本的优劣,并同时提出了特殊配合力的概念(specific combining ability, SCA)用于评估杂交配组的好坏。从那时起,一般配合力便被广泛应用于传统遗传育种与杂种优势利用的亲本选择与评估。但一般配合力的遗传学与分子生物学基础至今仍不明确。为了探讨和揭示一般配合力的遗传学基础,我们构建了包含我国不同历史阶段的六个优良籼稻品种：珍汕97B、矮脚南特、广陆矮四号、93-11、特青、矮仔占,按照双列杂交设计由这些亲本杂交得到的15个F1代的双列杂交群体来评价水稻的株高、生育期与每穗粒数等三个与产量相关农艺性状的一般配合力。在此基础上,我们构建了高配合力亲本特青与低配合力亲本广陆矮4号的F2群体,对140个分离单株与五个亲本进行测交,根据测交亲本的表型计算一般配合力,以GCA值作为表型进行遗传作图,研究GCA的遗传基础；同时,对广陆矮四号、特青和93-11三个亲本和相互配组的三个F1代的叶片的转录组进行RNA测序分析,试图揭示GCA的分子基础。主要结果如下：1、一般配合力分析结果显示：亲本93-11与特青的株高、生育期和穗粒数三个与产量相关农艺性状的一般配合力均表现为正值,而其余亲本则表现为负值。2、各个亲本的一般配合力在武汉与海南省不同生态环境下基本保持稳定,受环境影响较小。3、我们构建了高配合力亲本特青与低配合力亲本广陆矮4号的140个单株的F2群体,对140个单株与五个亲本进行测交,通过F1代的生育期、株高及穗粒数表型计算出GCA,利用1,136个SSR标记对亲本特青与广陆矮4号进行全基因组扫描,并得到121个多态性标记；利用这些多态性标记,我们构建了平均遗传间隔13.1cM的遗传连锁图。随后对株高、生育期、穗粒数等三个与产量相关性状的一般配合力进行遗传作图分析,检测到控制一般配合力的两个主效位点。其中一个位点位于一号染色体SSR标记RM3148与RM462之间,控制穗粒数的一般配合力,Lod值5.7；另一个位点位于七号染色体SSR引物RM1306附近,同时控制生育期、株高、穗粒数等三个性状的一般配合力,对三个性状的Lod值分别为35.5、129、12.5。结果显示由统计分析所得出的一般配合力概念具有遗传学基础,受QTL遗传位点控制。4、对双列杂交群体中各样本的表型进一步分析发现：以93-11或特青作为亲本的杂种F1代的生育期、株高及穗粒数等三个农艺性状表型明显偏向于93-11或特青,与另一个亲本(广陆矮四号、矮脚南特或珍汕97)的表型具有较大差异。针对这一现象,我们对各亲本及其对应的杂种F1代生育期、株高、穗粒数等性状进行了相关性分析,结果显示：亲本93-11及特青与其杂种F1代的表型之间具有较高的相关性(r>0.97,P<3.80E-07),而亲本广陆矮四号、矮脚南特、珍汕97与其杂种F1代的表型之间相关性较低(r=0.81,P<1.33E-03)。这一结果显示具有高一般配合力的亲本对其杂种F1代表型的影响大于一般配合力低的亲本。5、为了揭示F1代表型偏向高一般配合力亲本的原因,对特青、93-11和广陆矮4号和相互配组的三个F1代的二次枝梗分化期叶片的转录组进行了测序分析,三个亲本和F1代的转录组测序深度为89.5到114.1兆reads,对转录组数据的分析结果显示F1代转录组偏向于高一般配合力的亲本。6、对水稻开花及GA调控相关的基因表达量的分析,证实了转录组数据与表型具有对应关系,结果揭示了F1代农艺性状表型偏向高一般配合力亲本,并与转录组的亲本偏向性相关。对开花和GA代谢的通路的相关基因的表达模式进行了分析,Ehd1、Ehd2、 Hd3a、RFT1基因在三个F1代及亲本93-11、特青中表达量显著低于亲本广陆矮四号；Hd1、OsPRR1与Ghd7基因的表达量则在亲本广陆矮四号中显著下调。在GA调控通路中,促进GA合成的基因OsCPS1与OsKAO以及受高GA含量刺激表达的基因GA2ox在三个F1代和93-11、特青中的表达量显著上调；而受高GA含量抑制的基因GA3ox与GA20ox的表达量则在上述材料中显著下调,这些结果与表型一致。7、为了探索在水稻杂种优势的产生机制及水稻F1中等位基因表达模式与杂种优势的关系,我们对上述三个F1代的杂种优势进行了分析,亲本差异较大亲本杂交产生的广陆矮4号×93-11、广陆矮4号×特青的杂种优势水平较高,反之,93-11×特青的杂种优势水平较低。随后对广陆矮4号×93-11、广陆矮4号×特青和低杂种优势水平的93-11×特青及其亲本广陆矮4号、93-11、特青进行了转录组测序及全基因组等位特异性表达分析。为了完成全基因组等位特异性表达分析,我们对三个亲本基因组进行了17.7-27.7倍于水稻基因组大小的重测序以保证亲本中至少90%的SNP位点被检出,为等位基因特异性表达分析奠定了基础。为了保证等位特异性分析的准确性,我们对三个F1代样本转录组进行了高深度的测序,得到89.5到114.1兆的reads,并使三个杂种F1代中SNP的平均reads覆盖深度分别达到8.6到10.9。8、全基因组的等位基因表达模式分析,在三个杂种F1代中,有3.4%到3.9%的基因呈现单等位表达模式,23.5%到24.2%的基因显示偏等位表达,而72.0%到73.0%的表达基因可归为共表达基因。进一步对F1代中等位基因的表达模式与其亲本之中对应基因的表达量的相关性分析,我们发现F1代中等位基因的表达模式与其亲本之中对应基因的表达量比值具有显著的正相关性；等位特异性表达基因(包括单等位表达基因与偏等位表达基因)与其亲本表达量具有更显著的正相关性；等位特异性表达在F1代种发挥重要作用。9、通过对单等位表达和偏等位表达基因的正反交的表达模式分析,发现所有分析的基因均表现为偏向单一亲本,与目前在动物中被深入研究的三种单等位基因表达模式不同,我们将这类新发现的单等位表达类型称为基因型依赖性(genotype—dependent)单等位表达基因。10、发现在杂种一代的转录组中,80.7-94.6%基因表现以顺式调控机制(cis-regulated mechanism),只有5.4-19.3%基因表现为反式调控,在水稻杂种F1代转录组以顺式调节为主；结果显示平均79.7%的基因表达具有互补效应的基因由等位特异性表达造成；对F1代等位特异性表达及亲本间基因表达量的分析显示在F1代中激活及表达量升高的基因主要由亲本之间的互补效应造成。11、对亲本与F1代的差异基因表达进行分析结果表明：等位特异性表达基因造成了F1代与亲本之间平均42.4%的差异表达基因。更重要的是等位特异性表达基因造成了79.8%的F1代与亲本之间表达量差异大于10倍的差异表达基因；这类基因的在F1代表达而在亲本之一不表达的基因占97.3%。多数F1代与亲本之间的差异表达基因的表达量偏向高一般配合力的亲本；进一步证明杂种F1代的互补效应主要由单等位表达及偏等位表达基因贡献的结果。更多还原

【Abstract】 Selecting elite parents is of paramount importance in conventional breeding and cross breeding programs. However, the selection of parents from a phalanx of inbred lines is extremely time consuming and often no better than a random process. Furthermore, parents with excellent agronomic traits do not always transmit those traits to their progeny. Thus, Sprague and Tatum proposed the concepts of general combining ability (GCA) and specific combining ability (SCA) to evaluate breeding parents and the parent combination, respectively. Since then, GCA has been widely and successfully used as criteria to evaluate elite parents in conventional breeding and in cross breeding practices. However, the genetic and molecular basis of GCA has been largely overlooked.To explore the genetic basis of GCA, we first analyzed GCA using a diallelcross population including six rice indica varieties, Zhenshan97(ZS), Aijiaonante (AJ), Guangluai#4(GL),93-11and Teqing (TQ), and15F1hybrids that mated from these parents. The diallel cross population were planted in three seasons: the summers of2008and2009in Wuhan and the winter of2009in Hainan Island. In each season, triplicates of30plants were grown. Three agronomic traits mainly contributed to rice yield, plant height (PH), heading date (HD) and grain number of per panicle (GP), were measured. The general combining ability (GCA) for each parent was calculated according to Griffing.For the further analysis of genetic basis of GCA, we constructed a F2population with the parents GL and TQ, which exhibited significantly difference on GCA for all the three traits. To evaluate the genetic base of GCA, total of700hybrids derived from the five test parents mating with140individual plants in the F2population of GLxTQ. A genetic map for GCA loci was made using the GCA score of individuals in the F2population. To further study the molecular basis of GCA, we did the transcriptome sequencing of GL, TQ and93-11, and their F1hybrids GLxTQ, GLx93-11and93-11xTQ. The main results as follows:1, We found that the GCA effects about the three agronomic traits of the parents tested appeared significantly difference. The GCA effects of93-11and TQ exhibited positive GCA effects, while ZS, AJ and GL exhibited negative GCA effects.2, We also found that the GCA effects does not be significantly affected by the environmental conditions in Wuhan and Hainan Island.3, A total of121polymorphic SSR markers were used for genetic linkage analysis of the GCA using140individuals of the F2population of GLxTQ. Two major QTLs were detected from the genetic mapping of the GCA for the three traits. One QTL controlling the GCA of GN with a Lod score5.7located between RM3148and RM462on chromosome1; the other QTL located near the RM1306on chromosome7, which has pleiotropism on controlling the GCA of three traits, HD, PH and GP, exhibited the Lod scores of35.5,12.9and12.5, respectively. These results indicated the concept of GCA derived from statistical analysis was controlled by QTLs.4, We found that the three agronomic traits from the F1hybrids were always similar to93-11and TQ in the diallel cross population, when either93-11or TQ was used as one parent in the hybrids. The correlation analysis for the three agronomic traits between Fl hybrids and the parents showed significant correlation with the parents of93-11and TQ (r>0.97, P<3.80E-07), but less significant correlation with the parents of GL, ZS and AJ (r=0.81, P<1.33E-03). These results indicated that the parents with higher GCA effects would make more contibutions to the phenotypes of F1hybrids than that of the parents with lower GCA effects.5, To explore the machanism of the agronomic traits in F1hybrids always biased to the high GCA parents, the transcriptome sequencing of the leaves at the development stage of secondary branch meristem from three F1hybrids, GLx93-11, GLxTQ and93-11XTQ, and their parents were accomplished. The results revealed that the global transcriptome profiles of F1hybrids were biased to the positive GCA effect parents.6, The transcriptome of F1hybrids biased to the positive GCA effect parents were validated by analyzing the gene expression patterns of the GA and flowering pathways that are corresponding for plant height and flowering time. In long day condition, the up-regulation of Ehdl, Ehd2, Hd3a, RFT1will induce rice flowering, these gene expression levels were obviously lower in GLx93-11, GLxTQ,93-11and TQ than that of in GL; while these gene expression levels of Hdl, OsPRR1and Ghd7that depressed the flowering in rice were significantly down-regulated in GL. In the GA regulation pathway, the gene expression levels of OsCPSl and OsKAO that enhance the GA synthesis were up-regulated in GLx93-11, GLxTQ,93-11and TQ. The similar results were obtained from expression anyalysis of GA2ox. While the expression level of GA3ox and GA20ox that feedbackly regulated by high level of GA were down-regulated.7, To explore the molecular basis of heterosis and the relationship between allelic expression and heterosis, we chose three F1hybrids, GLxTQ and GLx93-11, exhibited high heterosis, and the third,93-11XTQ, low heterosis and their parents GL, TQ and93-11for the transcriptome sequencing and allele-specific expression analysis.17.7-27.7of rice genome coverage were obtained from three parents to satisfy the minimum requirement for obtaining more than90%SNPs and a total of89.5to114.1million reads (90bp per read) were obtained from analogous tissue from the three F1hybrids used for the further allele-specific study. For the three F1hybrids, the sequencing depths were rached to8.6-11.9reads per SNP, repectively.8, We found that3.4to3.9%of the genes were classified as monoallelic expression (MAE),23.5to24.2%as preferential allelic expression (PAE) and72.0to73.0%as biallelic expression (BAE) from the three F1hybrids by the allele-specific expression analysis. The correlation analysis indcicated that correlation coefficients ranging from0.70to0.76(P<2.2E-16) in three F1hybrids between gene expression in the parents and the ASE in the F1hybrids, respectively. A higher correlation coefficient was obtained from the same analysis with ASE genes in the F1hybrids (r>0.80, P<2.2E-16). These results imply a cis-regulatory mechanism contributed to the allele-specific expression in rice hybrids.9, By comparing the allelic expression pattern of monoallelic and preferential allelic expression genes in reciprocal hybrids, we found the allelic expression pattern of all the genes tested biased to the high GCA effect parents. This allelic expression pattern was different from the three types of monoallelic expression genes that have well studied in mammals, so we designate this type of monoallelic expression gene as genotype-dependent monolallelic expression gene.10, We found80.7-94.6%expressed genes were regulated as cis-regulated mechanism in F1hybrids, while5.4-19.3%expressed genes regulated as trans-regulated effects through the gene expression level analysis. The results showed79.7%genes exhibited complementary effects, which were resulted from allele-specific expression. Our results revealed the categories of the expression activated and enhanced in the F1hybrids mainly resulted in complementary effects of superior alleles from both parents.11, The further analysis between allele-specific expression and differentially expressed genes occurred in F1hybrids indicated that these genes exhibiting allele-specific expression comprised42.4%of the genes differentially expressed between F1hybrids and their parents in average. Allele-specific expression accounted for79.8%of the genes displaying more than a10-fold expression level difference between an F1and its parents, and almost all (97.3%) of the genes expressed in F1, but non-expressed in one parent.更多还原