Identification Of Rice Grain Shape Genes Using Chromosome Segment Substitution Lines

As one of the important agronomic traits,seed shape not only affects grain appearance quality,but also confers grain weight,a yield component of rice.In this study,two sets of advanced backcross population were used as the basic stock to develop single-segment substitution lines and to construct the chromosome segment substitution line（CSSL）population with particular NIP and ACC10 segments introduced into the common genetic background of ZS97.Two sets of CSSL population were newly constructed and used for quantitative trait loci（QTLs）analysis of grain shape.Two novel QTLs for grain length were identified and cloned.The main results are as follows:1.Two sets of CSSL population,named DNZ and DAZ,were developed by using advanced backcross lines that were constructed from the cross of NIP/ZS97 and ACC10/ZS97 to backcross the recipient parent ZS97 with molecular marker-assisted selection strategy.The DNZ and DAZ population comprise 221 and 147 lines,respectively,with the common genetic background of ZS97.Genotyping of the CSSL populations using the 8K SNP chip showed that the DNZ and DAZ population consist of 142（64%）and 72（49%）single-segment substitution lines,respectively.Based on the recombination breakpoints of the substituted segments in the DNZ and DAZ populations,the bin maps were constructed,containing 418 and 325 bins（the smallest recombination block）with a median average length of 470 kb and 764 kb,respectively.2.QTL analyses of grain shape traits in the DNZ and DAZ populations combined with the bin maps were performed using a ridge regression method.A total of 50 QTLs for grain shape were detected in the DNZ population,of them 16 QTLs for grain length,19 QTLs for grain width and 15 QTLs for thousand-grain weight;a total of 25 loci for grain shape were detected in the DAZ population,of them 9 QTLs for grain length,12 QTLs for grain width and 4 QTLs for thousand-grain weight.There are 11 overlapping QTL regions that affect two or more grain shape traits.Among them,3 QTLs can be repeatedly detected in both populations.3.A novel QTL for grain length（q GL11-2）detected in the DNZ population was fine-mapped in the an approximately 25-kb region using a CSSL-derived segregation population,where contains 5 predicted genes.Os11g0528700（Os GH3.13）of them was the most likely candidate gene for q GL11-2.The crispr-mutated Os GH3.13 lines significantly increased grain length and thousand-grain weight compared with the wild type,indicating that Os GH3.13 is a candidate gene for q GL11-2.The near-isogenic lines（NIL）with ZS97 or 9311 background were developed.NIL1-q GL11-2^NIPwith ZS97 background significantly increased grain length and thousand-grain weight compared with NIL-q GL11-2^ZS97.Consistently,NIL2-q GL11-2^NIP with background9311 also showed a significantly increased grain length and thousand-grain weight relative to 9311.Thus q GL11-2^NIP exhibited the genetic effect of increasing grain length and thousand-grain weight in both 9311 and ZS97 genetic backgrounds.Cytological observation revealed a difference in the distance of the papillae of the longitudinal cells of the lemmas between NIL1-q GL11-2^NIP and NIL-q GL11-2^ZS97,indicating that q GL11-2 regulates grain shape possibly by affecting the size of the longitudinal cells of the lemmas in rice.4.q GL6-2 is a novel major QTL for grain length detected in the DAZ population.q GL6-2 was finely mapped into an 11-kb interval where contains the annotation gene LOC＿Os06g44620（Os4CL4）,which encodes a 4-coumarate coenzyme A ligase.Sequence analysis of Os4CL4 revealed that two nonsynonymous variations that might cause amino acid changes in the coding region,but the promoter region has greater variation.Among them,the Os4CL4^ZS97 promoter region has a 13-kb transposon at-543 bp upstream of the start codon（ATG）of the gene,while the transposon being of absence in Os4CL4^ACC10.Gene expression analysis showed that Os4CL4 was constitutively expressed in all assayed tissues,Os4CL4^ACC10 was expressed in a significantly higher level than Os4CL4^ZS97 during the panicle development.The bisulfite sequencing of the promoter region revealed that the methylation level in the Os4CL4^ZS97 promoter region near the transposon insertion was significantly higher than Os4CL4^ACC10.This result suggests that the 13 kb transposon may be a cause of the increased methylation in the promoter,resulting in a gene expression difference between Os4CL4^ZS97and Os4CL4^ACC10.Promoter activity analysis support that the transposon insertion in the promoter region is a key factor leading to a decrease in gene expression.5.Transgenic complementation experiment of Os4CL4^ACC10 showed that Os4CL4 is responsible for q GL6-2.Using a Pro^Os4CL4-ACC10 plus Os4CL4^ZS97vector to transform ZS97,we found that the transgenic positive plants had a significantly increased grain length and thousand grain weight relative to the negative plants,while there was no difference in grain width between the positive and negative.These results indicate that the variation of the gene promoter is the main cause for the difference in grain shape between NIL-q GL6-2^ACC10 and NIL-q GL6-2^ZS97.6.Cytological observation of the q GL6-2 NILs revealed there was a difference in the papillae distance within the lemmas between the NILs,indicating that q GL6-2 regulates grain shape mainly by affecting the size of the longitudinal cells of the lemmas.7.We constructed prokaryotic expression vectors to express Os4CL4^ACC10 and Os4CL4^ZS97 protein in vitro.The enzyme activity experimental results showed that the variation in the amino acid resides encoded by Os4CL4^ACC10 and Os4CL4^ZS97 does not affect their enzyme activities.However,the contents of coumaric acid and caffeic acid measured in the panicles revealed a significant difference between the NIL-q GL6-2^ACC10 and NIL-q GL6-2^ZS97.These results indicate that Os4CL4 may affect coumaric acid and caffeic acid content in the phenylpropane metabolic pathway at a transcription level.