The Genetic Architecture Of Kernel Row Number In Maize And Candidate Gene Mining And Functional Analysis Of The Major QTL

Kernel row number(KRN)is an important domestication and improvement trait.It is also an important yield component in maize(Zea mays L.).Genetic improvement of KRN is an effective way to increase maize yield.In this study,genome-wide association analysis and prediction were performed on 639 maize inbred lines to analyze the genetic basis of KRN.Simultaneously,we fine mapped the qKRN5.02 by using the segregation population constructed by Nong531 and its introgression lines,IL3455(containing SLN segments)and analyzed the candidate genes.The candidate gene of qKRN5.04 was further analyzed by molecular experiments and verified by genetic transformation in maize.The main results are as follows:1.Phenotype identification was carried out on 639 maize inbred lines in three environments(Gongzhuling,Beijing and Xinxiang),and we found that the phenotypic variation of KRN was abundant in the association population and population structure had no significant effect on KRN.Genome-wide association analysis of KRN was performed using seven different methods(MLM,mrMLM,FASTmrMLM,FASTmrEMMA,pLARmEB,pKWmEB,ISIS EM-BLASSO),and it was found that the formation of KRN was regulated by multiple loci in maize genome.2.Whole-genome prediction of KRN showed that the more random tagSNPs were selected,the higher prediction accuracies of KRN were achieved,but all of them were lower than 0.5.Selecting top tagSNPs based on the results of GWAS for KRN could significantly improve the prediction accuracy of KRN.When the tagSNPs detected by seven methods were combined,the best prediction effect could be achieved,that is,only 167 tagSNPs were used to predict the BLUP values of KRN,and the prediction accuracy could reach 0.75.3.In the F2 segregation population constructed by N531 and IL3455,a major QTL for KRN,qKRN5.02,was mapped to the region of 11.8-13.7 Mb on chromosome 5.Five recombinant genotypes were screened by SNP polymorphism markers in this interval.By progeny tests over multiple years and locations,the qKRN5.02 was limited in the 151 kb region.There were 7 genes in this interval,among which only 3 genes were differentially expressed in the young ears between the parents,and only Zm00001d013502 was specifically highly expressed in the developing ears.Therefore,Zm00001d013502 was considered to be an important candidate gene for qKRN5.02.4.Zm00001d016075 was considered to be a candidate gene for qKRN5.04 based on the results of previous mapping in our laboratory.In this study,through the promoter activity analysis,we found that the promoter type of SLN had higher activity than that of N531.We identified InDe1308 in the promoter of Zm00001d016075 in the association population,and found that the KRN of the groups with insertion 308 bp or deletion 308 bp were significantly difference.Therefore,InDe1308 could be used as one of the KRN molecular markers.Based on the CRISPR-Cas9 technique,we gained three homozygous transgene-free lines,which were knocked out Zm000001d016075 in the background of C01.The KRN of the three transgenic lines were significant higher than wild type.Therefore,Zm00001d016075 negatively regulated the formation of KRN in maize.KRN-associated loci and QTLs that were identified in this study will show potential for improving the KRN by genome-wide selection and molecular marker-assisted breeding and provide theoretical basis and guidance for genetic improvement of KRN.Meanwhile,candidate gene mining and functional analysis of qKRN5.02 and qKRN5.04 also lay a foundation for further understanding the molecular mechanism of KRN in maize.