Genetic, genomic, and breeding approaches to further explore kernel composition traits and grain yield in maize

Maize (Zea mays L.) is a model species well suited for the dissection of complex traits which are often of commercial value. The purpose of this research was to gain a deeper understanding of the genetic control of maize kernel composition traits starch, protein, and oil concentration, and also kernel weight and grain yield. Germplasm with different genetic backgrounds and mating structures were investigated. Illinois Low Protein cycle 90 (ILP 90) was crossed and backcrossed to B73 and three populations derived, two per se and one testcross. High and low tails of distributions for starch were selected and a microarray platform was used to assess gene expression profile differential in developing kernels of selected materials sampled at 15 and 20 days after pollination. Microarray data analysis revealed a repertoire of differentially expressed genes from which one, alpha amylase inhibitor, stood out and was suggested to have an important role in starch biosynthesis in ILP/B73-derived materials. A second testcross population was developed by crossing the highest S3 line for starch concentration with the S3 line associated with highest grain yield. The new testcross population showed higher starch and comparable grain yields as compared to commercial hybrids. Further evaluation of best ILP/B73-derived hybrids alongside modern commercial hybrids under nitrogen deprivation conditions suggested better tolerance of experimental hybrids to nitrogen stress as compared to commercial hybrids. Evaluation of the intermated B73xMo17 (IBM) population revealed quantitative trait loci (QTL) for starch, protein, oil, and kernel weight. The IBM based QTL findings differ from QTL findings involving the Illinois long-term selection experiment (ILTSE) materials. Fewer QTL were detected in the IBM population than in populations derived from the ILTSE strains and only two QTL detected in the IBM population co-localized for both protein and starch. This differed from the findings from ILTSE in which it was very common to identify QTL co-localizing for starch and protein, and starch and oil. The IBM based results may be more relevant to contemporary breeding program germplasm than results involving ILTSE materials. The ILP/B73-derived materials nevertheless may be useful donor materials to contemporary breeding programs.