【作者】 李成奇；

【导师】 张天真；

【作者基本信息】 南京农业大学 ， 作物遗传育种， 2007， 博士

【摘要】 提高棉花产量依然是目前棉花育种的主目标,研究棉花产量性状的遗传规律对指导棉花产量育种具有十分重要的意义。衣分是棉花重要的产量构成因素,大量资料表明,棉花产量的改进与衣分的提高有密切关系。同时,衣分又与其它产量性状、产量构成因素以及纤维品质性状间存在不同程度的相关关系。因此,对棉花衣分性状的遗传及相关研究尤为重要。本文以不同衣分的棉花品种（系）为材料,研究了棉花衣分及相关产量性状的遗传、QTL定位,以及不同衣分棉花材料在纤维初始发育过程中的异同。结果如下:1.双列杂交遗传分析利用6个不同衣分的陆地棉品种（系）半半棉、泗棉3号、苏棉12、关农1号、石短5号和黔农465,配制6×（6-1）个杂交组合,通过亲本和F₁的随机区组试验,利用加性-显性（AD）遗传模型分析表明,陆地棉各产量性状都有较高的遗传主效应方差,产量性状受加性效应和显性效应共同控制,其中,衣分、衣指以加性效应为主;籽棉产量、铃重和籽指以显性效应为主;皮棉产量和单株铃数以加性和显性效应为主。衣分和衣指的普通广义遗传率和普通狭义遗传率均最高,加性方差分量分别为61.8%和65.6%。相关和通径分析一致表明,产量构成因素中单株铃数对皮棉产量的贡献最大,衣分次之,铃重最小。对F₁、F₂杂种优势预测结果表明,籽棉产量、皮棉产量和铃重具有极显著的正向群体平均优势;衣分和衣指的群体平均优势为负优势,衣指达极显著;单株铃数和籽指的群体平均优势不显著。群体超亲优势衣分、衣指和籽指为极显著负优势;其他产量性状群体超亲优势不显著。2.主基因-多基因遗传分析通过构建泗棉3号×石短5号（组合Ⅰ）和泗棉3号×黔农465（组合Ⅱ）的P₁、P₂、F₁、F₂、B₁、B₂六世代群体,采用主基因-多基因混合遗传模型对衣分等产量性状进行了多世代联合分析。结果发现,尽管2个组合各产量性状的最适遗传模型并不完全一致,且主基因和多基因的遗传效应、遗传率存在差异,但所有性状至少在1个组合中检测到主基因的存在,表明棉花产量性状主基因存在的普遍性。对2个组合各产量性状的主基因、多基因遗传率比较得出,各性状的主基因遗传率比多基因遗传率在不同组合间趋势变化相对较稳定。各性状的主基因、多基因遗传率分量在2个组合中也不完全相同。籽棉产量和皮棉产量在2个组合中均以主基因遗传为主;衣指在组合Ⅰ中以多基因遗传为主,在组合Ⅱ中属于典型的多基因遗传;单株铃数和铃重在2个组合中分别以主基因遗传为主和以多基因遗传为主;籽指在2个组合中分别以多基因遗传为主和以主基因遗传为主;衣分在组合Ⅰ中以主基因和多基因遗传并重,在组合Ⅱ中属于典型的多基因遗传。3.陆地棉衣分差异群体产量及产量构成因素的QTL标记和定位利用泗棉3号×石短5号（组合Ⅰ）的F₂、F_2:3和泗棉3号×黔农465（组合Ⅱ）的F₂衣分差异群体,对产量及产量构成因素进行了QTL分子标记和定位。应用6111对SSR引物对亲本进行多态性检测,组合Ⅰ和组合Ⅱ分别获得了123个和107个多态性位点。组合Ⅰ共鉴定出了18个控制产量及产量构成因素变异的QTLs,包括2个衣分QTLs、4个籽棉产量QTLs、4个皮棉产量QTLs、2个衣指QTLs、3个单株铃数QTLs、2个铃重QTLs和1个籽指QTL,解释的表型变异分别为6.9%～16.9%、5.6%～16.2%、4.8%～15.6%、7.7%～13.3%、8.2～11.6%、6.1%～7%和6.6%。单标记分析还检测到8个分子标记分别与各产量性状连锁。组合Ⅱ共鉴定出了7个产量及产量构成因素的QTLs,包括1个衣分QTL、1个籽棉产量QTL、1个皮棉产量QTL、1个衣指QTL、2个籽指QTLs和1个单株铃数QTL,解释的表型变异分别10.0%、41.9%、36.5%、12.6%、6.8%～10.9%和15.9%。单标记分析还检测到11个分子标记分别与各产量性状连锁。2个组合中均发现不同性状QTLs在染色体相同或相邻区段上成簇分布,表明与产量性状相关的基因可能紧密连锁或一因多效。对所有检测到的QTL遗传效应分析发现,控制产量性状的QTLs可能主要以显性和超显性效应遗传为主。本研究所检测到的主效QTLs可以用于棉花产量性状的标记辅助选择。通过与前人不同群体不同组合产量性状QTL的定位结果比较得出,Chr3（A3）、Chr6（A6）、Chr9（A9）、Chr11（A11）以及Chr12（A12）/Chr26（D12）很可能包含产量性状QTLs的富集区。其中,在组合ⅠF_2:3群体中Chr3（A3）上检测到的1个衣分QTL（qLP03,解释的表型变异为16.9%）可能与Shen等（2007）在同一染色体上检测到的衣分QTL（同时在3个环境中检测到）是相同的或紧密连锁的,因为他们都与桥梁标记NAU1190连锁,对其稳定性需做进一步的试验验证。4.陆地棉不同衣分材料纤维初始发育的比较研究利用扫描电镜技术,对5个不同衣分（28.53%～43.10%）的陆地棉品种进行了纤维初始发育的超微结构观察。结果表明,在开花前1 d所有的陆地棉品种均出现纤维细胞突起,但突起分布因基因型不同存在差异,且与衣分高低无直接关系。在开花当天,所有材料的胚珠表面均出现大量纤维细胞突起。在开花后1 d这些突起已明显伸长。除泗棉3号（衣分最高为43.10%）外,所有材料在开花后1 d的纤维密度均明显低于开花当天,这说明随着棉胚珠体积的膨大,纤维“稀释”程度可能与衣分高低有关。不论开花当天和开花后一天,黔农465（衣分最低为28.53%）的纤维密度均最低,而且在开花当天,该材料的胚珠表面发现有许多非正常的凹陷形纤维突起,推测这些突起将最终不能发育成成熟的棉纤维。相关分析结果表明,开花后1 d的纤维伸长密度与开花当天的纤维突起密度均与衣分的相关性最高（分别为0.8642和0.8141）,其次为衣指（分别为0.8055和0.7968）,与铃重和籽指均为负相关。灰色关联分析结果表明,开花后1 d的纤维伸长密度与衣分的关联度最高（0.7029）,其次是衣指（0.6820）,与铃重和籽指关联度最低;开花当天的纤维突起密度与衣指的关联度最高（0.7111）,其次是衣分（0.6571）,与铃重和籽指关联度依然最低。综合两种分析结果可以得出,在纤维初始发育阶段,纤维密度与衣分和衣指关系最密切,为棉花的纤维产量育种提供了重要的信息。5.棉花四个栽培种纤维初始发育的比较研究棉花4个栽培种具有明显的衣分差异,对陆地棉（泗棉3号）、海岛棉（海7124）、亚洲棉（定远小花）以及非洲棉(A_1-50)开花前后的胚珠进行了扫描电镜比较观察,并探讨了延迟授粉对各棉种纤维初始发育的影响。结果表明,除亚洲棉外其他3个棉种均在开花前1 d出现纤维细胞突起。在开花当天和开花后1 d,所有棉种的纤维突起（或伸长）密度均在珠柄顶部的脊突处最大,合点端和胚珠中部其次,珠孔附近最小。除陆地棉泗棉3号（衣分为43.10%）外,其他棉种在开花后1 d的纤维伸长密度均低于开花当天的纤维突起密度。说明随着棉胚珠体积的膨大,其他3棉种胚珠表面的纤维“稀释”程度大于高衣分的陆地棉。在棉纤维初始发育阶段,纤维密度不仅受气象因子如温度的影响,而且与授粉时柱头的生活力密切相关;延迟授粉对纤维初始发育的影响以非洲棉较大,海岛棉较小,亚洲棉和陆地棉最小。该研究为理解4个栽培棉种纤维分化和发育机理、通过人工授粉棉纤维的杂种优势利用提供了一定的理论依据;对探讨4个栽培棉种纤维发育的亲缘进化关系可能具有一定的参考价值。更多还原

【Abstract】 Improving cotton yield is still the main goal of present cotton breeding;it is meaningful for studying the genetics of cotton yield traits for yield breeding.Lint percentage is one of the important yield components,and plenty of data show that the raising of cotton yield has close relation with lint percentage improvement.At the same time,the different level correlation is existed between lint percentage and other yield traits,yield components and fiber quality characters.Therefore,the inheritance and related research of lint percentage is more important than other traits.In the present paper,the inheritance and QTL tagging of lint percentage and its related yield traits,as well as fiber initiation development of cottons varied in lint percentage were studied.The results are given as follows:1.Diallel crossing genetic analysis6×（6-1） crosses were made among 6 upland cotton varieties or lines varied in lint percentage including Banbanmian,Simian 3,Sumian 12,Guannong 1,Shiduan 5 and Qiannong 465.The parents and hybrids were planted in 3 replications with randomized plots.Genetic analysis for each of yield traits was conducted by Zhu’s AD （additive-dominance） with MINQUE（Ⅰ） approach.The results indicated that each yield trait of Upland cotton had always higher variance of genetic major effects,and yield traits were controlled by additive effects and dominant effects altogether.In which,lint percentage,lint index was mainly controlled by additive effects;seed yield,boll size and seed index were mainly controlled by dominant effects;lint yield and bolls per plant were mainly controlled by additive effects and dominant effects.Both common broad heredity and common narrow heredity of lint percentage and lint index were always the highest,whose additive variance percentages were 61.8%and 65.6%,respectively.Correlation analysis and path analysis of yield traits always indicated that among yield components,bolls per plant had the most contribution to lint yield,secondly by lint percentage,and the least by boll size.Heterosis prediction of F₁ and F₂ showed population mean heterosis values for seed yield,lint yield and boll size were always more positively significant;lint percentage and lint index had negative population mean heterosis,that of lint index was more significant;bolls per plant and seed index had not significant population mean heterosis.Population over-parent heterosis values of lint percentage,lint index and seed index were always more negative significant,for other traits,however,were not significant.2.Major gene-polygene genetic analysisJoint analyses of multiple generations of P₁、P₂、F₁、F₂、B₁ and B₂ was used to analyze the genetics of lint percentage and its related yield traits in two crosses Simian 3×Shiduan 5（corssⅠ） and Simian3×Qiannong 465（crossⅡ）,by using the method of major gene-polygene mixed inheritance model.It was found from the results that major genes controlling each of yield traits were always detected for at least in one cross although the optimum models for each trait were not completely consistent and the genetic effects,the heritability of major gene and polygene had also differences between two crosses, indicating that the major genes controlling yield traits were existed generally.The comparison of major gene and polygene heritability for all traits between two crosses showed,major gene heritability had comparatively stable tendency change than polygene heritability.Heritability proportion of each trait was also different between two crosses. seed yield and lint yield was mainly controlled by major gene in two crosses;lint index was mainly controlled by polygene in crossⅠand belonged to typical polygene inheritance in crossⅡ;bolls per plant and boll size were controlled by major gene and polygene in two crosses,respectively;seed index was controlled by polygene and major gene in two crosses, respectively;lint percentage was controlled by major gene and polygene altogether in crossⅠand belonged to typical polygene inheritance in crossⅡ.3.Tagging and mapping of QTLs for yield and its components in Upland cotton （Gossypium hirsutum L.） population with varied lint percentageQTLs for yield and its components were tagged and mapped in populations with varied lint percentage of Simian 3×Shiduan 5（CrossⅠ） F₂,F_2:3 and Simian 3×Qiannong 465（CrossⅡ） F₂.123 and 107 polymorphism loci from 6111 pair SSR primers were obtained in crossⅠand crossⅡ,respectively.18 QTLs for yield and its components were identified in CrossⅠincluding 2 QTLs for lint percentage,4 QTLs for seed yield,4 QTLs for lint yield,2 QTLs for lint index,3 QTLs for bolls per plant,2 QTLs for boll size and 1 QTL for seed index,which separately explained 6.9%～16.9%、5.6%～16.2%、4.8%～15.6%、7.7%～13.3%、8.2～11.6%、6.1%～7.0%and 6.6%of the phenotypic variance.In addition,8 molecular markers were detected linked with yield traits using the single-marker analysis.7 QTLs for yield and its components were identified in CrossⅡincluding 1 QTL for lint percentage,1 QTL for seed yield,1 QTL for lint yield,1 QTL for lint index,1 QTL for bolls per plant,2 QTLs for seed index.They separately explained 10.0%、41.9%、36.5%、12.6%、15.9%和6.8%～10.9%of the phenotypic variance,QTLs for boll size were not identified.In addition,11 molecular markers were detected linked with yield traits using the single-marker analysis.Different QTLs affecting yield traits were always detected within the same or closer chromosome region in two crosses,suggesting that genes controlling yield traits may be linked closely or the result of pleiotropy.The QTLs controlling yield and its components might mainly be dominant and over-dominant.The molecular markers linked closely to the major QTLs may be used in MAS（marker-assisted selection） to improve cotton yield.The conclusion was drawn from the comparison of QTL tagging among different crosses or different populations that Chr3（A3）,Chr6（A6）,Chr9（A9）,Chr11（A11） and Chr12（A12）/Chr26（D12） might respectively contain the cluster of QTL for yield traits,in which the QTL for lint percentage identified in F_2:3 population of CrossⅠ（qLP03, explained 16.9%of the phenotypic variance） might be same as or closely linked to the QTL identified in three environments simultaneity reported by Shen（2007）,because they were always linked to the bridge marker BNL1190.The stability of this QTL would be tested in next experiments.4.Fiber initiation development in Upland cotton（Gossypium hirsutum L.） varieties varied in lint percentageFiber initiation was observed in five Upland cotton（Gossypium hirsutum L.） varieties that vary in lint percentage,from 28.53%to 43.10%,using scanning electron microscopy （SEM）.The results indicated that at-1 days post-anthesis（dpa）,fiber cell protrusions were found in all varieties,but these protrusions varied among the materials and the differences did not correlate with lint percentage.At 0 dpa,a large amount of fiber cell protrusions appeared on the ovular surface of all samples,and these protrusions had been elongated significantly by +1 dpa.Interestingly,fiber density at +1 dpa of almost all samples was always lower than that at 0 dpa except Simian 3,the variety with the highest lint percentage 43.10%.This observation suggested that with the expanding of the cotton ovular volume, fiber "diluting" degree on the ovular surfaces might be related to higher or lower lint percentage.The cultivar with the lowest lint percentage（28.53%）,Quiannong 465,was found to exhibit the fewest fiber protrusions.Rather,many sunken,morphologically abnormal protrusions were observed on the ovular surface at 0 dpa,it was speculated that these protrusions might not yield mature cotton fibers ultimately.Correlation analysis suggested that either fiber protrusion density at 0 dpa or fiber elongation density at+1 dpa had a highest positive correlation with lint percentage（0.8642 and 0.8141,respectively）, secondly with fiber index（0.8055 and 0.7968,respectively）.The grey relational analysis suggested that fiber elongation density at+1 dpa correlated most with lint percentage （0.7029）,secondly with fiber index（0.6820）,and the least with seed index and boll weight;fiber protrusion density at 0 dpa correlated most with lint index was the highest （0.7111）,followed by lint percentage（0.6571）,and again the lowest were seed index and boll weight.A conclusion was drawn from two analyses that during fiber initiation stage, fiber density had closest relations with lint percentage and lint index,which could provide invaluable predictive information for cotton fiber yield breeding.5.Comparative studies on fiber initiation development of four cultivated cotton speciesThere are 4 cultivated species involved in cotton varied significantly in lint percentage. Here we observed the ovules before and after anthesis of 4 cultivated cotton species including Gossypium hirsutum cv.Simian 3,Gossypium barbadense cv.Hai 7124, Gossypium arboreum cv.Dingyuan xiaohua,and Gossypium herbaceum cv.A1-50 by utilizing SEM（scanning electron microscope）,and explored the influence of delay pollination on fiber initiation development of each cultivar.The results showed that under the climate conditions in the field at Nanjing,fiber cells protruded on the ovular surface in all cultivars except G.arboreum at -1 dpa.At 0 and +1 day post-anthesis（dpa）,fiber protrusion（or elongation） density of all materials was the highest at the funicular crest, higher at the chalazal cap and the middle part of ovule,and the least near the micropyle; moreover except Upland cotton cultivar Simian 3（lint percentage 43.10%）,fiber density of the other three cultivars was always lower at+1 dpa than that at 0 dpa,suggesting that with the expanding of the cotton ovular volume,the fiber "diluting" degree on the ovular surfaces was greater in G.barbadense,G.arboreum and G.herbaceum than in G.hirsutum with high lint percentage.During cotton fiber initial stage,fiber density not only was affected by weather factors such as temperature but also related to the viability of stigma. Delay pollination exerted relatively significant influence on fiber initiation development of G.herbaceum,relatively less influence on G.barbadense,and the least influence on G. hirsutum and G.arboreum.Our studies might provide some theoretical references on understanding the mechanism of fiber differentiation and growth of 4 cultivated cotton species,and exploring their phylogenetic relationships of fiber development,as well as utilizing the heterosis of cotton fiber through artificial pollination.更多还原