Genetic Architecture Dissection Of Maize Multi-tissue Primary Metabolites

Maize(Zea mays L.)is one of the most important crops,and improving the yield and quality has always been the major target of maize genetic improvement.Plenty of metabolites are enriched in maize tissues,such as leaf and kernel,the content and species of them are closely associated with maize growth,development,quality and even the yield.Dissecting the genetic architecture and mining the key genes affecting the maize would accelerate the understanding of the complex quantitative traits,as well as the genetic improvement of maize.In this study,we conducted the metabolic profiling of different maize tissues in two diverse genetic populations,and performed genome-wide association study(GWAS)using the high density markers,and finally we dissected the genetic basis of maize primary metabolites,and obtained a batch of key candidate genes for amino acid and sugar metabolism,which can provide new perspective for a comprehensive and in-depth understanding of the genetic structure of maize metabolome.The main results are summarized as follows:1.Metabolite profiling was performed on GC-MS for samples from 4 tissues of a widely used maize association panel,and 61 annotated metabolites were identified.A total of 153 significant loci were identified for these metabolites.The QTL number for each metabolite varied from 1 to 5,with an average of 1.5 loci.The percentage of phenotypic variations explained by these QTLs ranged from 5%~25.2%,with an average of 7.8%.Combining with the transcriptome data and bioinformatics analysis,36 loci and the corresponding candidate genes were prioritized.4 candidate genes(Zm GAD1,Zmlkrts,Zm ADT and Zm Tre1)were validated through transgenic plants or re-sequencing method.This helped elucidate their function in amino acid or trehalose metabolism and provided useful gene resources for improving maize quality traits.2.The newly developed multi-parent population CUBIC(Complete-diallel design plus Unbalanced Breeding-like Inter-Cross)with 1404 progenies and 24 parents,were profiling 86 primary metabolites in leaf and mature kernel using GC-MS.These metabolites contents varied widely: a)A larger variation existed among progenies than that in parents;b)Links between metabolites showed tissue-specific pattern and a small part of metabolites,such as galactinol and raffinose,showed a constitutive pattern in different tissues with relative high correlations.3.A total of 1237 mQTL(including s QTL and h QTL)were detected with single variantgenome-wide association study(s GWAS,492 s QTL)and haplotype genome-wideassociation study(h GWAS,880 h QTL)using 11.8 million SNP and 27005 unique bingenotype in CUBIC population.It was identified 5.6 s QTL and 5.9 h QTL on averagefor each leaf metabolite respectively,the phenotypic variation jointly explained bywhich ranged from 7.3% to 62.3% with an average of 28.3%;and for kernelmetabolites,4.0 s QTL and 7.1 h QTL were detected,the phenotypic varation jointlyexplained by which varied from 9.5% to 63%,around 33.5% on average.A total of 16 mQTL hot-spots were detected,and metabolites in 12 hot-spots were enriched inspecific pathways,modules or metabolites roles,which may suggest some keyregulatory genes.4.25 loci were co-localized between leaf and kernel,which comprised 16 different metabolites.Most mQTLs showed tissue-specific patterns,which was consistent with the metabolic variations existed in the different tissues.Taking advantage of the leaf transcriptome dataset,information of genome annotation,and the priori knowledge of the plant metabolic networks,some candidate genes from 42 loci were nominated for analyzing in-depth and it would be successful to reconstruct,the proposed primary metabolic pathway focusing on amino acid,sugar and organic acid metabolism.5.Analyzing the genetic pattern of mQTL in CUBIC,we revealed that the genetic factors controlling the primary metabolites inherited in multiple different modes,including pleiotropic effects,additive effects,and epistasis.For instance,through pleiotropic effects,Zm AVT could influence the accumulation of several amino acids in leaf;two homologous genes of Zm GADs had additive effects on the accumulation of GABA by distinct subcellular locations;multiple genes regulated oligosaccharide pathway synergistically through additive effects and epistasis.This provided a lot of valuable information for an in-depth understanding of the pathways of different metabolites and the corresponding genetic regulatory networks.Furtherly,analysis of metabolites and agronomic traits in the CUBIC population revealed that metabolites can explain the phenotypic variation of agronomic traits ranging from 7.6 to 17.6%,and 680 mQTLs were co-located with agronomic QTLs,which provided insights for elucidating the genetic basis of agronomic traits using metabolites as intermediate phenotypes.