【作者】 张轲；

【导师】 张正圣；

【作者基本信息】 西南大学 ， 作物遗传育种， 2011， 博士

【摘要】 棉花(Gossypium spp.)是世界上最重要的天然纤维作物。棉属包括约45个二倍体种和5个异源四倍体种,其中栽培种4个,分别是亚洲棉(G. arboreum),草棉(G. herbaceum),陆地棉(G. hirsutum)和海岛棉(G. barbadense).陆地棉(G. hirsutum)提供了世界棉花总产量的95%以上。但是,作为纺织工业的主要原材料,陆地棉(G. hirsutum)的纤维长度、比强度和细度等纤维品质性状,均比海岛棉(G. barbadense)差。尽管我国自主育成棉种的纤维品质能基本满足当前纺织工业生产的要求,但由于其长度过于单一、纤维比强度较低并且纤维较粗,不能达到高档纺织品的原料要求。随着人们生活水平的提高及棉纺织技术的革新,对棉花纤维品质的要求日益增高。在兼顾产量的同时,大幅、快速的改良纤维品质成为了棉花遗传改良的新焦点。传统的遗传改良已经为我们育成了一批高产、优质的陆地棉品种。但是由于其选育周期长、效率低、并且棉花纤维品质和产量性状受数量性状基因控制且呈负相关关系,所以传统的遗传改良已经不能满足我们的需要。DNA分子标记技术的出现,为棉花的遗传改良提供了一种快速、准确的选择方法。通过对棉花高品质纤维QTL的分子标记筛选,利用与纤维品质QTL紧密连锁的分子标记进行辅助选择,能够有效的提高选择效率,有助于我国棉花品种的纤维品质水平的提高。目前,陆地棉种间(interspecific)图谱虽然较为饱和,但不能直接用于陆地棉栽培种的改良。而陆地棉种内(intraspecific)图谱的覆盖度及标记密度远达不到分子标记辅助选择和图位克隆的要求。现在以SSR标记为主的陆地棉种内图谱的最大覆盖度约为70%。复合杂交群体是由三个或三个以上亲本杂交产生的群体,这类群体经常用于作物栽培品种的选育,通过表型选择累积有利的加性效应等位基因来达到对目的性状的改良。然而,目前大多数陆地棉遗传连锁分析和QTL定位都基于两亲本衍生出的群体,如：F2,BC和RIL等。利用两亲本群体进行连锁分析和QTL定位具有很大的局限性,因为早期陆地棉品种选育多通过种内品种间杂交选育品种,使得陆地棉的遗传背景变得十分狭窄,两亲本间的遗传差异较小,获得多态性标记较为困难,所以两亲本的陆地棉种内遗传连锁图谱都表现出较低的标记密度和覆盖度。并且使用两亲本群体进行QTL定位,所检测的到的QTL仅仅是该性状的部分遗传结构,大量的QTL信息都没有被检测到。同时,最多只有两个等位基因的效应能够被检测到,这也大大限制了两亲本遗传连锁图谱定位的QTL在分子标记辅助选择中的应用。而利用复合杂交群体进行连锁分析和QTL定位不仅可以增加遗传连锁图谱的密度,还可以检测到更多的QTL信息,从而定位的QTL在分子标记辅助育种有较好的应用。目前,利用多亲本构建遗传连锁图谱和QTL定位的报道还较少,本研究旨在利用分子标记技术和多亲本分离群体,构建陆地棉种内高密度遗传连锁图谱,并分析陆地棉纤维品质相关性状的QTL位点。为达到这一目的,本研究用陆地棉丰产品种中棉所35、高纤维强度品种渝棉1号、以及异常棉(G. anomalum)渐渗系7235建立[(渝棉1号×中棉所35)×(渝棉1号×7235)]复合杂交群体及F1 2/3家系,利用多态性稳定的SSR标记构建复合杂交群体遗传连锁图谱,并结合F1(2007)群体和F12/3(2008、2009)家系表型数据对纤维长度、马克隆值、比强度、整齐度、伸长度5个纤维品质性状进行QTL检测,主要研究结果如下：1.SSR标记多态性与群体基因型检测利用本实验室的16052对SSR引物,对亲本渝棉1号、中棉所35和7235进行多态性引物筛选,共获得有效多态性引物1057对,多态性比例为6.6%。用1057对多态性SSR引物对陆地棉三亲本复合杂交群体[(渝棉1号×中棉所35)×(渝棉1号×7235)]172个单株进行基因型检测,共获得1067个位点。卡方测验结果显示,1067个标记位点中有375个偏离孟德尔分离比例(P<0.05),偏分离比例35.19﹪。2.三亲本复合杂交群体遗传连锁图谱利用Joinmap4.0软件对1067个多态性位点进行遗传连锁分析,构建的遗传连锁图谱包含978个位点,标记间的平均距离为4.3 cM,遗传连锁图覆盖4184.4 cM,约占异源四倍体棉花基因组重组总长的94.1%。该图谱含69个连锁群,4个连锁群未定到染色体上,其余的65个连锁群被定位到26个染色体上,每个染色体含1-6个连锁群。A染色体亚组由32个连锁群组成,总长1954.3 cM,含366个多态性位点,标记间平均距离为5.3 cM；D染色体亚组由33个连锁群组成,总长2122.1 cM,含599个多态性位点,标记间平均距离为3.5 cM。3.亲本与F,群体和F1:2/3家系的性状表现2007、2008、2009年重庆三个环境下,亲本7235纤维长度分别为31.9、34.5、34.1 mm,高于渝棉1号的29.9、30.3、30.6 mm和中棉所35的29.3、30.2、31.4 mm。亲本7235和渝棉1号纤维比强度三年分别为38.1、34.2、41.3 cN/tex和38.6、36.2、41.3 cN/tex,高于中棉所35的33.5、30.7、34.9 cN/tex,其它纤维品质性状三亲本都比较接近。F1群体纤维长度平均为31.2 mm,分布范围为28.2-34.4 mmm；整齐度平均为85.3%,分布范围为80.9-88.1%；马克隆值平均为4.2,分布范围为3.5-5.0；伸长度平均值为6.5%,分布范围为6.2-6.7%；比强度平均值为33.9 cN/tex,分布范围为29.2-39.7 cN/tex。F1：2家系纤维长度平均为31.9 mm,分布范围为28.4-34.8 mm；整齐度平均为85.4%,分布在81.6-87.5%之间；马克隆值平均为4.1,分布范围为3.1-5.2；伸长度平均值为6.6%,分布在6.3-6.8%之间；比强度平均值为34.0 cN/tex,分布在29.8-39.3 cN/tex之间。F13家系纤维长度分布在29.4-36.2 mm,平均32.0 mm；整齐度分布在82.3-87.4%,平均85.3%；马克隆值平均值为4.1,分布范围为3.1-5.1；比强度平均值为37.3 cN/tex,分布范围为31.6-43.9 cN/tex。除纤维伸长度外,其余各纤维品质性状在三个环境中都表现出超亲分离。CP群体纤维品质性状的相关性分析结果显示,除纤维马克隆值与长度和比强度呈极显著负相关。其他性状间均成极显著正相关,相关系数从0.276-0.687。方差分析结果显示,纤维马克隆值受基因型和环境影响均达到极显著；纤维整齐度和纤维伸长度受环境和基因型影响均不显著；纤维长度受环境影响显著,受基因型影响不显著；纤维比强度受环境影响极显著,受基因型影响显著。4.纤维品质QTL定位以三亲本复合杂交群体[(渝棉1号×中棉所35)×(渝棉1号×7235)]构建的遗传连锁图谱,结合2007-2009年重庆3个环境纤维品质表型数据,利用QTL作图软件MapQTL5.0,采用多QTL模型(multiple-QTL model)进行分析,共检测到63个纤维品质性状QTL,包括27个显著性QTL(1000次排列测,全基因组5%显著水平作为显著QTL的阀值),36个可能性QTL(LOD大于3.0小于5%显著水平阀值),解释性状表型变异8.1-55.8%。其中9个纤维伸长度显著性QTL,2个纤维伸长度可能性QTL,共计11个QTL,解释性状表型变异10.5-47.1%；2个纤维长度显著性QTL,14个纤维长度可能性QTL,共计16个QTL,解释性状表型变异8.6-55.8%；6个马克隆值显著性QTL,3个马克隆值可能性QTL,共计9个QTL,解释性状表型变异9.6-34.7%；6个纤维比强度显著性QTL,4个纤维比强度可能性QTL,共计10个QTL,解释性状表型变异8.1-32.6%；4个纤维整齐度显著性QTL,13个纤维比强度可能性QTL,共计17个QTL,解释性状表型变异12.1-50.5%。在所有QTL中,有11个QTL与前人研究结果一致,6个QTL可在两个环境中检测到,这些共同QTL和环境稳定QTL可用于分子标记辅助选择。在检测到的63个纤维品质QTL中,渝棉1号等位基因在中棉所35遗传背景下增加纤维品质性状表型值的有36个,在7235遗传背景下增加纤维品质表型值的有28个。中棉所35贡献23个增加纤维品质的等位基因,7235贡献33个增加纤维品质的等位基因。更多还原

【Abstract】 Cotton (Gossypium spp.), one of the most economically important crops worldwide, is a renewable source of natural fiber and secondary products such as oil, livestock feed and cellulose. The genus Gossypium comprises approximately 45 diploid and 5 tetraploid species, including 4 cultivated species, G. arboreum, G. herbaceum, G. hirsutum and G. barbadense. Of the cultivated species, G. hirsutum supplies over 95% of the world’s total fiber production.During the past decades, cotton breeders have placed more emphasis on increasing yield. As the basic raw materials for textile industry, fiber quality is highly correlated with spinning performance and end product quality. Technological advance in the textile industry requires high fiber quality. This requirement has attracted more efforts toward improving fiber properties, especially those of upland cotton.Conventional cultivar breeding programs, primarily relying on intermating adapted upland cotton genotypes and selecting novel allele combinations based on phenotypic selection, have certainly improved fiber quality while also increasing fiber yields in upland cotton. However, the quantitative inheritance and unfavorable correlations between lint yield and fiber quality greatly limited the efficiency of conventional breeding efforts in upland cotton improvement. Therefore, it needs to develop more effective strategies in the future cotton breeding programs.Great advances in DNA maker technologies have provided new insights on genetic improvement of cotton fiber quality. Based on the detailed molecular linkage maps, quantitative trait locus (QTL) affecting fiber quality traits could be mapped, genetically evaluated and selected through linked markers (marker-assisted selection, MAS), and even cloned (map-based cloning). Combining the powerful molecular tools and conventional breeding will provide effective approaches to select and develop cotton cultivars with improved fiber quality. Since the first molecular linkage map was constructed in cotton, genetic maps derived from interspecific populations are relatively saturated. However, so far the interspecific maps have little use in upland cotton breeding programs. The most extensive coverage map of upland cotton spanned 3,140.9 cM, accounting for 70.6% of the whole tetraploid cotton genome. Because the level of intrapecific polymorphism revealed by DNA markers is low in upland cotton, the resolution available in the existing intraspecific maps is satisfactory neither for MAS nor for map-based cloning. To obtain high-density intraspecific map, which can be used in MAS and map-based cloning in upland cotton, it needs to develop new type of markers, use upland cotton cultivars/lines with high polymorphism as mapping parents, or employ new type of mapping populations.Composite cross populations (CP), developed from three/more cultivars/parental lines, are frequently applied in crop cultivar development programs to improve agronomic and economic traits through gradually accumulating the additive effects of novel allele combinations based on phenotypic selection. However, historically linkage analysis and QTL mapping in cotton routinely capitalize on the populations derived from simple crosses (e.g., F2 and backcross). Upland cotton genetic maps in simple crosses showed limited resolution, most likely as a result of low polymorphism between two mapping parents. QTL detected in a simple cross merely represent the paucity of genetic architecture in traits, whereas a large amount of QTL information could not be exploited. Furthermore, at most two allelic effects are able to be characterized in crosses involving two inbred lines, which is a limitation to feasible options for designing optimum breeding strategies grounded on MAS. Employing CP into linkage analysis and QTL mapping may increase the marker density of upland genetic maps, exploit more adequate gene resources and facilitate MAS.In the present study, molecular markers and segregating population derived from three parents were used to develop genetic linkage map and to exploit QTL concerned with fiber quality traits. Three upland cultivars/lines, Yumianl, CRI 35 and 7235 were used to obtain the segregating population, Yumianl/CRI 35//Yumianl/7235, and its F1:2/3 inbreed lines. A linkage map developed from CP was constructed by JoinMap4.0. Based on 3 years of phenotypic data, QTL for five fiber quality traits, fiber elongation (FE), fiber length (FL), fiber micronaire reading (FM), fiber strength (FS), and fiber length uniformity ratio (FU) were analysed by MapQTL5.0. The results are followed:1. Polymorphism of SSR markers and genotypingIn total,16,052 SSR primer pairs were used to screen the polymorphism among Yumian 1, CRI 35 and 7235, and 1,057, acconting for 6.6% of the tatal primer pairs, showed effective polymorphism. The effective polymorphic primer pairs (1,057) produced 1,067 loci when they were used to genotype the CP individuals. Chi-square test showed that 375 loci significantly distorted from the expected segregation ratio (P＜0.05), accounting for 35.1% of the total loci.2. Genetic mapconstructionWhen 1,067 loci were used to construct the linkage groups, a map with 978 loci and 69 groups was obtained, remaining 89 loci unlinked.The map spanned 4,184.4 cM with an average distance of 4.3 cM between two markers, accounting for about 94.1% of the whole tetraploid cotton genome. Sixty-five of sixty-nine linkage groups were assigned to 26 chromosomes of tetraploid cotton. The remaining 4 linkage groups could not be associated with any chromosome and were tentatively named as "Un" following a number. Thirty-two linkage groups were assigned to A-subgenome, containing 366 loci, and spanning 1,954.3 cM with an average distance of 5.3 cM between two markers. Thirty-three linkage groups were assigned to D-subgenome, containing 599 loci, and spanning 2,122.1 cM with an average distance of 3.5 cM between two markers.3. Fiber quality phenotypes of mapping parents, F1 and F1:2/3familiesIn year 2007,2008 and 2009, the fiber length of 7235 were 31.9,34.5 and 34.1 mm, respectively, which was higher than that of Yumian 1 (29.9,30.3 and 30.6 mm) and CRI 35 (29.3, 30.2 and 31.4 mm). In year 2007,2008 and 2009, the parental lines, Yumian 1 and 7235, exhibited relatively higher fiber strength (38.1,34.2 and 41.3 cN/tex and 38.6,36.2 and 41.3 cN/tex) than CRI 35 (33.5,30.7 and 34.9 cN/tex). The other fiber quality traits of three parental lines were close to each other.In year 2007, fiber length ranged from 28.2 to 34.4 mm, and averaged 31.2 mm; Fiber length uniformity ranged from 80.9 to 88.1% and averaged 85.3%; Fiber strength ranged between 29.2 and 39.7 cN/tex, averaged 34.0 cN/tex; Fiber elongation was 6.5% on average and ranged from 6.2 to 6.7%; Fiber micronare reading ranged between 3.5-5.0, and averaged 4.2. In year 2008, fiber length ranged from 28.4 to 34.8 mm, and averaged 31.9 mm; Fiber length uniformity ranged from 81.6 to 87.5% and averaged 85.4%; Fiber strength ranged between 29.8 and 39.3 cN/tex, averaged 34.0 cN/tex; Fiber elongation was 6.6% on average and ranged from 6.3 to 6.8%; Fiber micronare reading ranged between 3.1-5.2 and averaged 4.1. In year 2009, fiber length ranged from 29.4 to 36.2 mm, and averaged 32.0 mm; Fiber length uniformity ranged from 82.3 to 87.4% and averaged 85.3%; Fiber strength ranged between 31.6 and 43.9 cN/tex, averaged 37.3 cN/tex; Fiber elongation was 6.6% on average and ranged from 6.3 and 6.9%; Fiber micronare reading ranged between 3.1-5.1, and averaged 4.1. Among all five fiber traits, transgressive segregation was oberserved except for fiber elongation.Singnificantly positive correlations were observered between different fiber traits except for fiber micronare reading and length, fiber micronare reading and strength, with correlation coefficients ranged from 0.276-0.687. Singnificantly negative correlations were observered between micronare reading and length, fiber micronare reading and strength, with correlation coefficients-0.472 and-0.321, respectively.The variance analysis indicated that fiber micronare reading was significantly influenced by both genotype (P＜0.01) and environment (P＜0.01); fiber length uniformity and elongation were influenced by neither genotype nor environment; fiber length was significantly influenced by environment (P＜0.05), but not by genotype; fiber strength was significantly influenced by both genotype (P＜0.05) and environment (P＜0.01). 4. QTL mapping of fiber quality traitsAll the five fiber quality traits segregated continuously and transgressive segregations were observed for fiber length, length uniformity, micronaire reading and strength. Totally,63 QTL were identified for five fiber quality traits, and mapped on 18 chromosomes and 2 Un linkage groups, including 27 significant QTL with the LOD thresholds more than the permutation-based LOD thresholds and 36 putative QTL with the LOD thresholds between 3.0 and the permutation-based LOD thresholds.For fiber elongation, nine significant QTL and two putative QTL were identified, and explained phenotypic variation between 10.5 and 47.1%. Only significant QTL qFE24.2 was detected in two environments (in year 2007 and 2008).For fiber length, sixteen QTL were identified, and explained phenotypic variation between 8.6 and 55.8%. Two significant QTL and fourteen putative QTL were detected in one environment except for putative QTL qFLunO2.1, which was detected in year 2007 and 2008.For fiber micronaire reading, six significant QTL and three putative QTL were identified, and explained phenotypic variation between 9.6 and 34.7%. Four significant QTL were detectedin two environments, and one of the four QTL. qFM03.1, was detected in year 2007 and 2008, and three others, qFM24.2, qFM24.3 and qFM24.4, were detected in year 2007 and 2009.For fiber strength, six significant QTL and four putative QTL were identified, and explained phenotypic variation between 8.1 and 32.6%. All the QTL for fiber strength were detected only in one environment.For fiber length uniformity, four significant QTL and thirteen putative QTL were identified, and explained phenotypic variation between 12.1 and 50.5%. All the QTL for fiber length uniformity were detected only in one environment.Among the QTL detected, six QTL were detected in two environments, which showed stablity in different environments. According to the sharing common markers, eleven QTL identified in the present study were also found in the same chromosome region in other populations, and these QTL included qFE24.2, qFL23.2, qFL24.2, qFM23.1, qFM24.2, qFM24.3, qFS07.1, qFS23.1, qFE24.2, qFS24.2 and qFS24.3. The QTL identified in the two environments and across different populations are of great value for MAS programs in cotton.Among the 63 QTL detected,36 and 28 alleles leading to an increase in all 5 fiber quality traits were conferred by Yumian 1 in CRI 35 and 7235 genetic background, respectively. CRI 35 contributed 23 alleles to increase five fiber quality traits.7235 contributed 33 alleles to increase fiber quality traits更多还原