【作者】 周学标；

【导师】 陆作楣；

【作者基本信息】 南京农业大学 ， 作物遗传育种， 2007， 博士

【摘要】 目前，我国杂交水稻年种植面积稳定在1500万hm²左右，占全国水稻种植面积的50％以上。国内推广应用的杂交籼稻不育系除了红莲型细胞质外，野败型、冈型、印尼水田谷型、矮败型等都可用其它类型的保持系保持其不育性。这些细胞核不同的品种称为异型保持系或第二保持系。不育系与异型保持系杂交产生的杂种不育系通称为二元不育系。其中双亲之间没有明显亲缘关系的称为非同胞二元不育系，例如珍汕97A/协清早B；双亲为姊妹系的二元不育系称为全同胞型二元不育系，双亲之间只有一个共同亲本的二元不育系，称为半同胞型二元不育系；双亲之一为另一个亲本的衍生系的二元不育系，称为回交型二元不育系，这三类二元不育系统称为近亲型二元不育系。非同胞型二元不育系与恢复系杂交得到的杂种F₁称为水稻三交种，近亲型二元不育系与恢复系杂交所得到的杂种F₁称为改良单交种。据研究三交种和改良单交种由于聚合了三个亲本的优良基因，常常表现出较好的综合抗性，较强的群体缓冲性和对不良环境的广泛适应性，因此在玉米、油菜等许多作物中广泛应用。我国水稻三交种的研究始于20世纪70年代末，到目前为止仅有少量报道，虽然达成了利用三交种可提高水稻制种产量的共识，并已经筛选出少量比单交种有一定增产潜力的三交组合，但一直没有大面积生产应用，其主要原因在于用来配制三交种的二元不育系双亲遗传差异较大，引起三交种F₁代生育期和株型分离较严重，群体整齐度较差，难以为群众所接受。类似的问题在玉米杂种优势利用过程中也遇到过，最终通过筛选特定的组合，特别是运用姊妹种配制改良单交种的方法，既提高了杂交种的制种产量，又解决了玉米三交种整齐度差的问题，但迄今为止未见关于水稻改良单交种的研究报道。为了解决当前杂交水稻生产上品种适应性降低，稳产性差的问题，减少因高温冷害等异常气候条件对杂交稻生产所造成的损失，增强杂交稻对环境的适应能力，实现杂交稻生产的稳产高产，提高制种产量，本研究借鉴玉米育种的经验，选取了与纯合不育系有近亲关系的品系作为第二保持系，配制了三类近亲型二元不育系，再与不同的恢复系杂交配制了三种类型的改良单交种，对二元不育系本身的杂种优势、一般配合力表现、以及改良单交种的产量表现、群体整齐度和稻米的部分品质性状进行研究，一方面探讨杂交稻改良单交种生产应用的可行性，以解决水稻三交种因整齐度较差不能在大面积生产上应用的问题，为培育高产稳产且具有广泛适应性的杂交稻品种探索一条新的路子。另一方面筛选产量较高，整齐度较好，能为大面积生产所接受的改良单交种。1、用10个具有不同亲缘关系的纯合不育系及其保持系配制了3个全同胞型二元不育系、11个回交型二元不育系和12个半同胞型二元不育系。鉴定了其育性、开花习性、异交特性及部分农艺性状的表现，分析了相应的保持系间杂交F₁代单株产量的杂种优势，结果表明：（1）所有二元不育系的自交结实率均为零，表现完全不育，花粉组成与双亲不育系相比变化不大；（2）多数二元不育系的开花习性和异交特性比相应的纯合不育系有明显改善，特别是柱头双外露率尤为明显；（3）半同胞型二元不育系株高、剑叶长、剑叶宽、穗长及部分农艺性状的杂种优势较为明显；（4）三类B×B组合F₁的平均单株产量随双亲的遗传差异的加大而增加，说明二元不育系随着双亲遗传差异增加杂种优势逐渐增强；26个保持系间杂交组合中有23个产量超过高亲，说明利用三类二元不育系的杂种优势都有可能提高制种产量。2、利用3个全同胞型二元不育系、11个回交型二元不育系和10个半同胞型二元不育系及其10个亲本纯合不育系，分别与南恢04、明恢86、辐恢838、先恢207和密阳46等5个恢复系按照NCⅡ交配设计配制了50个单交种和120个改良单交种，按随机区组的随机模型分析方法，对二元不育系主要性状进行了配合力分析。结果表明：（1）所有性状的一般配合力方差和特殊配合力方差均达到了极显著水平，说明所有性状均受加性效应和非加性效应的共同影响，其中千粒重、有效穗数、每穗总粒数、结实率、株高、播始历期等以基因的加性效应作用为主，单株产量却以非加性效应为主．在多数性状上，恢复系对F₁的影响大于不育系及互作效应，但在单株产量上不育系的影响超过了恢复系。（2）三类二元不育系单株产量和有效穗数的一般配合力都比纯合不育系极显著增加，回交型二元不育系和半同胞型二元不育系结实率的一般配合力也极显著高于纯合不育系，其他性状上二元不育系和纯合不育系没有明显差异。（3）在水稻改良单交种育种中，以双亲具有近亲关系的二元不育系作母本进行适当配组，有可能选育出有效穗数较多，结实率较好，比相应单交种产量更高的水稻改良单交种。3、分析了三类二元不育系所配的改良单交种及其相应的B×B组合F₂的抽穗历期和始齐历期以及株高、有效穗数，主穗长、剑叶长、剑叶宽和播始历期等性状的变异系数，计算了两者之间各性状上的相关系数，结果表明：（1）三类二元不育系所配的改良单交种除株高变异系数的平均值显著地高于纯合不育系的平均值外，其余株型性状上差异不显著；从纯合不育系到全同胞型二元不育系、回交型二元不育系、再到半同胞型二元不育系，播始历期变异系数、始齐历期（10％植株抽穗日期至80％植株抽穗日期）、抽穗历期（各小区第一株抽穗至最后一株抽稳的总天数）等三个性状的平均值都表现出了逐渐增大的趋势；部分全同胞型和回交型二元不育系配出改良单交种的平均群体整齐度与普通单交种接近。（2）二元不育系所配的改良单交种与其相应的B×B组合F₂在抽穗历期、始齐历期、播始历期变异系数、株高的变异系数等性状上的相关系数分别为0.499、0.372、0.343和0.251，达到了极显著水平，在剑叶宽变异系数和主穗长变异系数上的相关系数分别为0.226和0.196，达到了显著水平；其它性状上相关不显著。说明随着二元不育系双亲遗传差异的增大，其所配改良单交种的生育期和株高、穗长等性状的整齐度可能会降低，在改良单交种的育种工作中，适当减少二元不育系双亲的遗传差异是完全有必要的。在普通单交种和各类改良单交种群内，单交种或改良单交种与其相应的保持系或B×B组合F₂在生育期有关的性状上相关均不显著，而大多数株型有关性状也都不显著。说明当把二元不育系双亲的遗传差异控制在一定的范围内时，改良单交种中在大多数性状上的整齐度与其母本不育系之间的相关不显著，在此范围内对二元不育系的选择主要根据其一般配合力的高低及其自身的杂种优势表现。（3）通过综合分析B×B组合F₂以及二元不育系所配的改良单交种的整齐度，筛选得到了几个较好的二元不育系，主要包括Ⅱ协1A/Ⅱ协2B、金协黄A/金协紫B、金23A/金协紫B、金协黄A/金23B、Ⅱ-32A/Ⅱ协1B、Ⅱ-32A/Ⅱ协2B，和新协紫A/Ⅱ协2B。4、研究比较了各改良单交组合与普通单交种单株产量的差异，以及株高、剑叶长、剑叶宽、主穗长、抽穗期等与群体整齐度关系较为密切的性状的变异情况。结果表明：（1）三种类型的改良单交种平均单株产量均比普通单交种的平均产量增加；有74个组合比其两个对应单交种都增产，占所有改良单交种的61.67％，其中全同胞型改良单交种9个，回交型改良单交种37个，半同胞型改良单交种28个；（2）三类改良单交种都有部分组合在株高、剑叶长、剑叶宽、主穗长、单株播始历期等性状的变异系数，以及始齐历期、抽穗历期上与普通单交种接近。（3）通过综合考查各改良单交组合的产量优势及群体整齐度，筛选得到10个有可能在生产上应用的改良单交种，其产量比相应高产单交种增加10.43％-26.39％。5、选取3个全同胞型二元不育系、3个回交型二元不育系、3个半同胞型二元不育系以及其对应的9个纯合不育系，分别与南恢04和密阳46两个恢复系所配制成的改良单交种和普通单交种，测定稻米的垩白米率（％）、垩白大小（％）、垩白度（％）、糙米率（％）、精米率（％）和整精米率（％），以及稻谷的千粒重（g）、粒长（mm）、粒宽（mm）、长宽比，计算了粒长、粒宽和长宽比的变异系数，利用方差分析的分群比较法比较了三类改良单交种和普通单交种各性状的整体差异；同时还比较了各改良单交种各性状与其相应的普通单交种的差异，分析了改良单交种籽粒整齐度的情况，结果表明：（1）半同胞型改良单交种的千粒重极显著的高于普通单交种和回交型改良单交种；半同胞型改良单交种和全同胞型改良单交种的垩白米率均极显著地低于普通单交种，而回交型改良单交种也显著地低于普通单交种，三类改良单交种之间差异不显著；半同胞型改良单交种的垩白大小显著地低于普通单交种和其它两类改良单交种，垩白度极显著地低于普通单交种，显著地低于回交型改良单交种，与全同胞型改良单交种差异不显著；（2）三类改良单交种和普通单交种之间在糙米率、精米率、整精米率以及粒长、粒宽和长宽比等性状均未表现出明显差异；说明利用二元不育系来配制改良单交种的育种方法，不会对F₁的外观品质和碾磨品质产生不良影响，反而在垩白性状上会有所改善。（3）三类改良单交种粒长、粒宽和长/宽的变异系数平均值及变动范围均与普通单交种接近，部分改良单交种的变异系数小于普通单交种。新协黄A/新协紫B//密阳46和金23A/金协紫B//南恢04两个组合的粒长、粒宽和长/宽三个性状的变异系数均小于其两个对应单交种，另有部分组合变异系数的值也在普通单交种的变异范围之内，通过选择能够获得粒型整齐的改良单交种。更多还原

【Abstract】 The planting areas of indica hybrid rice was about 1.5×10⁷ hectares each year, occupied more than 50% of total areas of rice in China. The sterility of most indica rice CMS lines（A-lines） commercially used in China can be maintained by other maintainer lines（B-lines） except the Honglian type CMS-line. The male-sterile F₁ hybrids derived from the crosses between the A-lines and the other B-lines are named as binary CMS lines. In this paper, there are four types of binary CMS lines:（a） the binary CMS lines,in which two parents are distantly relationship are called non-sibling binary CMS lines,（b）and（c） the binary CMS lines which two parents are full sibling relationships or half sibling relationships are called full-sibling binary CMS lines and half-sibling binary CMS lines respectively,（d） if one of the two parents of the binary CMS lines was derived from the other one, these binary CMS lines are called backcross type binary CMS lines. The non-sibling binary CMS lines can be used as female parents to cross with rice restorer lines to produce three-way hybrids rice, and the other type binary CMS lines can be used to produce rice modified single-cross hybrids.The researches on three-way hybrid rice initiated in china in the late 1970s, and a consensus reached on this issue is that the three-way hybrid rice can be used to increase the yield of hybrid rice seeds production. Up to now, several combinations have been selected, but they were not put into practical use commercially. The main reason is the poor population uniformity of the three-way hybrids F₁ generation that was caused by the difference between the two parents of the binary CMS lines. The similar problem had been solved by selecting specific combinations especially using modified single-cross hybrids（MSHs） in maize breeding. The modified single-cross hybrid（A/A*//B） is a hybrid which was made by using a sister-line（A/A*） as female parent to cross with a inbred line（B）, and the sister-line was produced from a cross between two highly related inbred lines（A and A*）. It had been practical utilized in maize producing; however, there have been no any reports on the use of modified single-cross hybrids in rice breeding as yet.Just like in the maize breeding, in this study, three types’ binary-CMS lines were made and three types rice MSHs were produced by using those binary-CMS lines to cross with the rice restorer lines; The male sterility, flowering habit, out crossing characters, heterosis and general combining ability（GCA） of all binary-CMS lines and the yields and the population uniformity of all MSHs were studied. The objectives of this study, for one thing, were to make an approach to the feasibility of the application of the MSHs in rice breeding, and for another, were to select high yield rice MSHs with good population uniformity to put into practical use commercially in rice producing.1. Three full-sibling binary CMS lines, 11 backcross binary CMS lines, 12 half-sibling binary CMS lines were made from 10 rice CMS lines and their maintainer lines, which have different sib ship. The male sterility, flowering habit, out crossing characters and some agronomic characters of all binary CMS lines and CMS lines and the heterosis of the corresponding F₁（B×B） were studied. The results were as follows:（1） All binary CMS lines were completely sterile and had no fertile seed. The percentage of stained pollen of binary CMS lines was similar to that of the corresponding parent CMS lines.（2） Majority of binary CMS lines were superior to the CMS lines in flowering habit, out crossing characters, and especially in double stigma exsertion rate.（3） There were distinct heterosis in plant height, flag-leaf length, flag-leaf width and panicle length in half-sibling binary CMS lines.（4） The results indicated that the grain weight per plant of the three types of F₁ of B×B crosses promoted with the increase of genetic difference between the parents of the corresponding binary CMS lines. 23 of 26 F₁ among B-lines were of yield superiorities to the elite parents, suggesting that the above three binary CMS lines could be used in increasing yield in rice hybrid seed production.2. Three full-sibling binary CMS lines, 11 back-cross binary CMS lines, 10 half-sibling binary CMS lines, 10 homozygous CMS lines and five restorer lines were chosen to make 50 single hybrids（SHs） and 120 modified single cross hybrids（MSHs） by the way of NCⅡdesign. The plots of F₁ were arranged in the field in a randomized complete block design with 2 replications and the combining ability of 7 characters were analyzed using random model. The results showed as follows:（1） The variance of general combining ability（GCA） and specific combining ability（SCA） were significant at 1% level in all characters.（2） The GCA of three types of MSHs and SHs were studied using analysis of variance of independent comparison of single degree of freedom in the grain weight per plant and the other yield traits. The result indicated that, three types of MSHs were higher than the SHs at 1% level in the grain weight per plant and the number of effective tillers per plant respectively, the back-cross binary CMS Lines and half-sibling binary CMS lines were higher than the SHs at 1% level in seed setting rate. There were no significant differences in other yield traits.（3） It suggested that the above three types of close-sibling binary CMS lines could be used increasing grain weight per plant, number of effective tillers per plant and seed setting rate in modified single cross hybrid rice breeding.3. The F₂ of B×B crosses and the three types of MSHs corresponding with the binary CMS lines were studied in the days from 10% to 80% plants heading in a plot（DPH）, the days from the first plant heading to the latest plant heading in a plot（DFL）, and the coefficient of variability of the plant height, flag-leaf length, flag-leaf width, number of effective tillers per plant, panicle length of main culms, days from sowing to initial heading（DIH）. The correlation coefficients between the F₂ of B×B crosses and the three types of MSHs were investigated. The results showed that:（1） The mean values of the coefficient of variability of plant height of three types of MSHs were significantly higher than that of the single cross hybrids; there were no significant differences in that of other characters. The DPH, DFL, and the coefficient of variability of DIH have a tendency to increase from single cross hybrids, to full-sibling MSHs, backcross MSHs, and half-sibling MSHs. The mean population uniformity of some crosses in the full-sibling MSHs and backcross MSHs were close to the single cross hybrids.（2） There were highly significant positive correlations between the MSHs and the F₂ of the B×B crosses on the DFL, DPH, the coefficient of variability of DIH, plant height, and the correlation coefficients were 0.499、0.372、0.343 and 0.251 respectively. There were significant positive correlations on the coefficient of variability of flag-leaf width and panicle length of main culms, and the correlation coefficients were 0.226 and 0.196. The results indicated that the population uniformity of the three types of MSHs decreased with the increase of genetic difference between the parents of the corresponding binary CMS lines. It is necessary to reduce the genetic difference between the parents of the binary CMS lines in rice modified single cross hybrid breeding.（3） According to the population uniformity of there corresponding MSHs and the F₂ of the B×B crosses, the attention should be paid to the full-sibling binary CMS lines and the backcross binary CMS lines, especiallyⅡxie1A/Ⅱxie2 B, JinxiehuangA/JinxieziB, Jin23A/JinxieziB, JinxiehuangA/Jin23B,Ⅱ-32A/Ⅱxie1B,Ⅱ-32A/Ⅱxie2B and XinxieziA/Ⅱxie2B.4. Three full-sibling binary CMS lines, 11 back-cross binary CMS lines, 10 half-sibling binary CMS lines, 10 CMS lines and 5 restorer lines Nanhui 04, Minghui 86, Fuhui 838, Xianhui 207 and Milyang 46 were chosen to make 50 single cross hybrids（SHs） and 120 MSHs by the way of NCⅡdesign. The grain yields, DPH, DFL, and the coefficient of variability of DIH, plant height, flag-leaf length, flag-leaf width, number of effective tillers per plant and panicle length of main culms were studied respectively. The results showed that:（1） The grain yields of three types of MSH rice were higher than single cross hybrids on average. Among all MSHs, 74 of which had superiority to the two corresponding single cross hybrids in grain yield, including 9 full-sibling type MSHs, 37 back-cross type MSHs and 28 half-sibling type MSHs.（2） Some crosses of every types MSHs were similar to single hybrids in DBF, DFL, and the coefficient of variability of the plant height, flag-leaf length, flag-leaf width, number of effective tillers per plant, panicle length of main culms, days from sowing to initial heading.（3） A total of 10 good modified single cross hybrids were selected according to the grain yield and the population uniformity, and its grain yield increased from 10.43% to 26.39% compared with the corresponding higher single cross hybrids.5. 18 MSHs and 18 single cross hybrid, which were made from 9 binary CMS lines, 9 homozygous CMS lines and two restorer lines Nanhui 04, Milyang 46 were chosen. The characters investigated in this study include: chalky rice percentage（%）, chalky area percentage（%）, chalkiness degree（%）, brown rice rate（%）,milled rice rate（%）,head rice rate（%）, 1000-grain weight（g）, grain length（mm）, grain width（mm）, length/width ratio, and the coefficient of variability of grain length, grain width, length/width ratio. The mean value of three types of MSHs and SHs were studied using analysis of variance of single degree of freedom in the appearance quality traits and milling quality traits.The results showed as follow:（1） The mean value of 1000-grain weight of the half-sibling type MSH was significantly higher than that of single cross and the backcross MSH at 0.01 levels. The mean value of the half-sibling type MSH and the full-sibling type MSH were significantly lower than that of single cross hybrid on the chalky rice percentage at 0.01 level, and the backcross type MSH was significantly lower than that of single cross hybrid at 0.01 level. The mean value of the half-sibling type MSH were significantly lower than that of the single cross hybrid and the other two types of MSHs at 0.05 level on the chalky area percentage,and significantly lower than the single cross hybrid and the backcross type MSH at 0.01 level on the chalkiness degree.（2）There were no significant difference between the three types of MSHs and single cross hybrids on the brown rice rate, milled rice rate, head rice rate, 1000-grain weight, grain length, grain width, length/width ratio. It suggested that there are no bad influence on the appearance quality and milling quality by using the modified single cross hybrid breeding method.（3） The mean and the range of coefficient of variability of the three types of MSHs were close to that of the single cross hybrid on grain length, grain width and length/width ratio.（4）The coefficient of variability of Xinxiehuang A/Xinxiezi B//Milyang 46 and Jin23A/JinxieziB//Nanhui04 were lower than there two corresponding single cross hybrids on the grain length, grain width and length/width ratio, and the coefficient of variability of another several MSHs were in the range of single cross hybrids on the three traits. The result indicated that the MSH with uniformity rice grain could be getting by selection.更多还原