Genome-wide Association Analysis Of Tiller Number And Study On A Tiller Number Decreased Mutant Based On RNA-seq In Rice

Tiller number of rice diretly affects grain yield and plant architecture,which is one of the important agronomic traits.The optimum tiller number is required to form the ideal plant architecture and guarantee the high and stable yield of modem rice varieties.Traditional breeding mainly depends on the personal experience of breeders,and some traits are difficult to observe and unstable,which are easy to be lost in the selection process.With the application of genome-wide association analysis in rice research,many agronomic traits related loci have been discovered Therefore,it is of great significance to explore and utilize the genes related to tiller number,which directly affect the number of panicles per plant Although a large number of QTL analysis and association analysis results provide the location information and genetic effects of natural variation sites,natural variation can be effectively manipulated and applied to production only when the corresponding molecular mechanisms are understood.In recent years,with the development of molecular biology and genomics,more and more mutants on tiller formation have been applied to study the function of genes related to tiller fromation.However,the regulatory mechanism of rice tiller formation remains unclear.As a model plant of monocotyledons,the study of the molecular mechanism of rice tillering can not only provide a theoretical basis for breeding,but also play an important role in studying the molecular mechanism of plant branch.In this study,we compared the results of genom-wide association analysis of 219 Korean rice varieties with the RNA-seq results of a low-tiller mutant of the target gene screened in the overlapping region of QTL related to tiller number,and explored the regulation mechanism of tiller formation from both natural variation and artificial mutagenesis.The main results are as follows:(1)For 219 Korean rice accession,tiller number at early tillering stage(designated as TNE),early maximum tillering stage(designated as TNM1),and late maximum tillering stage(designated as TNM2),productive tiller number(PTN),heading date(HD),heading interval(HDI),and productive tiller ratio(PTR)were measured to perform Genome-Wide Association Analysis.There are a total of 20 association signals(p<1.0 × 10-6)represented by lead SNPs were detected for all traits,5 of them have colocalization loci with known candidate genes as Hdl,TOC1,HAM1,HAM2,RLCK57.The highest association were detected for HD and the largest number of association,six loci,were detected in GW AS for HDI.(2)In Korean rice accessions,modern cultivar groups showed the lowest TNE and TNM1 while landrace showed the highest values.For TNM2,Tong-il type cultivar group showed higher median than the other groups.PTN and PTR in all Korean rice accessions were mainly found in the range of 10-11 and 0.65-0.7,respectively The highest median values of PTN and PTR were found in Landrace group.Tong-il type cultivar group revealed the lowest phenotypic variation in all traits.However,that showed the highest variation,but also the highest median value in HDI.In HD,modern cultivar group showed higher median values than the other groups.(3)Among 5 detected locis,Hdl and heading stage,HAM1 and HAM2 were significantly correlated with early maximum tiller number,OsTOC1 and late maximum tiller number,OsRLCK57 and early tiller number,respectively.Haplotype analysis found that amnong Korean varieties,functional Hdl haplotype showed late heading and high effective tillering rate,while TOC1 was only detected in local variety materials,showing the highest maximum tiller number in late stage,while only 4 materials of RLCK57 showed high early tiller number.(4)In order to dissect the comprehensive genetic basis for determining TNM2,PTN,and PTR,linear regression analysis were performed using lead SNPs for these three traits,but also lead SNPs for the traits with correlations.Resulting linear regression model was able to explain TNM2 with coefficient of determination(R2)value of 0.49,to explain 22.3%of PTN variation and show contribution of 38.1%to PTR,respectively.Linear regression models for TNM2,PTN,and PTR could explained 52.7%,23.8%,and 36.5%of phenotypic variation,respectively,indicating that the models consisted of lead SNPs could consistently explain TNM2,PTN,and PTR similar accuracy in other independent temperate japonica accessions.(5)Using known QTLS related to tiller formation in rice,candidate genes of auxin,cytokinin and monocolactone were screened,OsPIN5a was selected,and t-DNA insertion mutants with Dongjin background were ordered from kyung Hee University mutant library in Korea.The screened homozygous lines showed plant groth retardation,significantly reduced plant height and tiller number.(6)Total RNA was extracted from the shoot materials of pin5a mutant and wild-type at 30 days after sowing,and illumine sequencing was performed.A total of 31,0183,606 raw data were obtained by sequencing.After filtering,305989414 clean data were matched on the reference genome of rice to obtain 44,540 genes.After filtering out the genes with FPKM at zero,the remaining 32,089 genes were obtained Selected with |Fold Change|≥2,p-value<0.05,a total of 927 differentially expressed genes were obtained in mutant vs wildtype,which 678 were up-regulated and 249 were down-regulated.(7)Gene ontology(GO)enrichment analysis showed that 275 down-regulated and 699 up-regulated differentially expressed genes were enriched in 19 and 61 GO terms,respectively.The significant enriched terms as follow:in biological process group,the terms are lipid transport(GO:0006869),response to hormone stimulus(GO:0009725),response to biotic stimulus(GO:0009607),and terpenoid metabolic process(GO:0006721).In Molecular Function group,the terms are lipid binding(GO:0008289),receptor activity(GO:0004872),oxidoreductase activity(GO:0016702),and amino acid transmembrane transporter activity(GO:0015171),In the cellular component group,mitochondrion(GO:0005739),integral to membrane(GO:0016021)and extracellular region(GO:0005576)etc.(8)In KEGG(Kyoto Encyclopedia of Genes and Genomes)pathway enrichment analysis,927 DEGs were enriched in 79 KEGG pathways,with 34 of it were significantly enriched(P-value<0.05).Among 34 pathways,29 of which were Metabolism.Pathways with name of plant hormone signal transduction and MAPK signaling pathway are belong to the signal transduction category.Ribosome biogenesis in eukaryotes belongs to translation of a genetic information processing classification The Peroxisome and plant-pathogen interaction belong to the classification of Transport and catabolism under cellular processes and the classification of Environmental adaptation under biological systems category,respectively.(9)By analyzing the specific differentially expressed genes,we found the metabolism of auxin(IAA),cytokinin(CK)and brassinolide(BR)in the mutant were significantly affected.The results showed that in pin5a mutants,knockout of OsPIN5a resulted in homeostasis imbalance of auxin,while cytokinin and brassinolide metabolism were significantly inhibited.At the same time,qRT-PCR validation was performed on 10 representative pin5a mutant,and the results were consistent with RNA-seq data,indicating that the RNA sequencing data were reliable.(10)In order to further verify if the IAA dynamic balance is broken or not in pin5a mutants,and IAA signaling pathways are intergrated or not,we treated wild type(WT)and mutant(MT)two different concentration IAA(1μM and 10 μM)after the seeds were soilded.The results showed that exist in low concentration IAA,wild type and mutant have been promoting plant growth,this means less rice mutants of auxin signal transduction pathways are complete,signal transduction was not blocked In the presence of high concentration of IAA,the growth of wild-type plants was inhibited,but the growth of low-tiller mutants was still promoted,indicating that the auxin dynamic balance in low-tiller mutants was broken.(11)Genome-wide association analysis showed that OsPIN5a was not associated with the above seven tiller number related traits,indicating that there was no natural variation of this gene in this population.Analysis of the RNA-seq data of the low-tiller mutants showed that the expression levels of the five known genes related to tiller number detected in the genome-wide association analysis were not significantly changed in the low-tiller mutants,which indicated that the reduction of tiller number in the low-tiller mutants was not regulated by the direct influence of the five genes.