Water and nitrogen fertilization are the key factors limiting maize productivity. The genetic basis of interactions between maize genotype, water, and nitrogen is unclear. A recombinant inbred line (RIL) maize population was evaluated for seven yield and five agronomic traits under four water and nitrogen conditions: water stress and low nitrogen, water stress and high nitrogen, well-watered and low nitrogen, and well-watered and high nitrogen. Respectively eight, six, and six traits varied in response to genotype–water interactions, genotype–nitrogen interactions, and genotype–water–nitrogen interactions. Using a linkage map consisting of 896 single-nucleotide polymorphism markers and multiple-environmental quantitative-trait locus (QTL) mapping, we identified 31 QTL, including 12 for genotype–water–nitrogen interaction, across the four treatments. A set of 8060 genes were differentially expressed among treatments. Integrating genetic analysis, gene co-expression, and functional annotation revealed two candidate genes controlling genotype–water–nitrogen interactions, affecting both leaf width and grain yield. Genes involved in abscisic acid biosynthesis and bZIP, NAC, and WRKY transcription factors participated in maize response to water and nitrogen conditions. These results represent a step toward understanding the genetic regulatory network of maize that responds to water and nitrogen stress and provide a theoretical basis for the genetic improvement of both water- and nitrogen-use efficiency.

Although the use of heterosis in maize breeding has increased crop productivity, the genetic causes underlying heterosis for nitrogen (N) use efficiency (NUE) have been insufficiently investigated. In this study, five N-response traits and five low-N-tolerance traits were investigated using two inbred line populations (ILs) consisting of recombinant inbred lines (RIL) and advanced backcross (ABL) populations, derived from crossing Ye478 with Wu312. Both populations were crossed with P178 to construct two testcross populations. IL populations, their testcross populations, and the midparent heterosis (MPH) for NUE were investigated. Kernel weight, kernel number, and kernel number per row were sensitive to N level and ILs showed higher N response than did the testcross populations. Based on a high-density linkage map, 138 quantitative trait loci (QTL) were mapped, each explaining 5.6%–38.8% of genetic variation. There were 52, 34 and 52 QTL for IL populations, MPH, and testcross populations, respectively. The finding that 7.6% of QTL were common to the ILs and their testcross populations and that 11.7% were common to the MPH and testcross population indicated that heterosis for NUE traits was regulated by non-additive and non-dominant loci. A QTL on chromosome 5 explained 27% of genetic variation in all of the traits and Gln1-3 was identified as a candidate gene for this QTL. Genome-wide prediction of NUE traits in the testcross populations showed 14%–51% accuracy. Our results may be useful for clarifying the genetic basis of heterosis for NUE traits and the candidate gene may be used for genetic improvement of maize NUE.

Nitrogen (N) deficiency is one of the main factors limiting maize (Zea mays L.) productivity. Genetic improvement of root traits could improve nitrogen use efficiency. An association panel of 461 maize inbred lines was assayed for root growth at seedling emergence under high-nitrate (HN, 5 mmol L⁻¹) and low-nitrate (LN, 0.05 mmol L⁻¹) conditions. Twenty-one root traits and three shoot traits were measured. Under LN conditions, the root-to-shoot ratio, root dry weight, total root length, axial root length, and lateral root length on the primary root were all increased. Under LN conditions, the heritability of the plant traits ranged from 0.43 to 0.82, a range much wider than that of 0.27 to 0.55 observed under HN conditions. The panel was genotyped with 542,796 high-density single-nucleotide polymorphism (SNP) markers. Totally 328 significant SNP markers were identified using either mixed linear model (MLM) or general linear model analysis, with 34 detected by both methods. In the 100-kb intervals flanking these SNP markers, four candidate genes were identified. Under LN conditions, the protoporphyrinogen IX oxidase 2 gene was associated with total root surface area and the DELLA protein-encoding gene was associated with the length of the visible lateral root zone of the primary root. Under HN conditions, a histone deacetylase gene was associated with plant height. Under both LN and HN conditions, the gene encoding MA3 domain-containing protein was associated with the first whorl crown root number. The phenotypic and genetic information from this study may be exploited for genetic improvement of root traits aimed at increasing NUE in maize.