Evaluation of additive and nonadditive genetic variation in elite maize germplasm for root, shoot, and ear characteristics

For the hybrid seed corn industry, corn breeding and research in the public sector ideally would serve the role of identifying novel genes, alleles, and breeding procedures that have promising commercial applications. However, despite increasing research on public germplasm improvement, the role is stymied due to limited access to commercial quality lines and hybrids. Utilization of lines recently released from Plant Variety Protection (PVP) may be a solution. To this end, a selected set of inbred lines representative of North American proprietary germplasm accessed following expiration of their U.S. PVP certificates were thoroughly characterized for agronomically relevant vegetative and reproductive traits. The twelve inbreds used in this study span a broad range of genetic diversity found within contemporary proprietary germplasm that we believe forms a representative subset. These inbreds contain assemblies of key alleles produced from multiple cycles of selection in the U.S. This selection was primarily focused on agronomically important traits such as grain yield, maturity, uniformity of plant architecture to accommodate mechanized farming, increased ear size, increased nitrogen use efficiency, disease resistance, stalk and root strength, as well as generally improved performance under a range of biotic and abiotic stresses. Ten proprietary and two public inbreds were evaluated with a generation means design emanating from the diallel cross of the 12 parents. Parental inbreds, F1 hybrids, and segregating F1 derived F2 populations were phenotyped for an extensive number of leaf, tassel, and ear traits, including yield (Chapter 1). Parents and hybrids were also evaluated for root system architecture and complexity (Chapter 2), as well as susceptibility to naturally occurring common rust, grey leaf spot, and northern corn leaf blight leaf diseases (Chapter 3). The Eberhart and iii Gardner (1966) general model was employed for the analysis of traits in the first three chapters. Chapter 1 of this thesis explains the rationale and selection criteria for the inbreds selected for study and development into a new public resource. The inbreds chosen to be founders of an elite mapping population are shown to be genetically and phenotypically diverse, important for construction of association mapping populations and the applicability of the population to the investigation of a wide range of traits. Analysis of genetic effects for the above ground traits is presented, along with a comparison of the F1 hybrids to current commercial hybrid checks. Significant heterosis was observed for nearly every trait, and heterotic effects were frequently significant, but additive effects were responsible for most of the variation in height, flowering time, leaf width, tassel, and yield components. There was substantial genetic variation for the primary yield components, kernels per row, row number, and kernel weight, which will be of prime interest in future work with populations derived from this germplasm. Evaluation of mature post-embryonic maize root systems is rare, yet they are responsible for water and nutrient acquisition, which are key factors in the health and performance of a plant. The study presented in Chapter 2 describes the use of a high through-put image analysis based phenotyping method to collect measurements on stem diameter and root angle, which model root structure, and estimate fractal dimensions, which estimate the amount of branching in a root sample. Using generation means analysis, significant additive effects were identified for the two root architecture and two complexity traits phenotyped. Additive heterosis was not significant for any of the traits, but significant specific heterosis was iv observed for root angle and one of the measures of complexity. The additive effect-by-environment was significant for all traits and explained a large portion of the variation compared to the heterotic genetic effects. Nevertheless, significant heterosis was detected for root complexity and angle. This study is unique in that it evaluated heterosis for mature post-embryonic root systems grown in field conditions. Chapter 3 characterizes the ex-PVP parents and hybrids for their response to naturally occurring common rust, grey leaf spot, and northern corn leaf blight leaf. Pervasive late season common rust was observed in a replicated trial, along with apparent variation for severity of infection, so the plots were rated for the amount of diseased leaf area. Symptoms of grey leaf spot and northern corn leaf blight were also visible, but only certain genotypes appeared to be susceptible. Since there did not appear to be much variation for these diseases, when present, only qualitative measures were recorded. For common rust, additive effects and additive effect-by-replication interactions were highly significant. The means of the parents and the F1 generation were not significantly different from each other, indicating a lack of an appreciable amount of total heterosis in the population. However, both additive and specific heterosis were both highly significant, indicating that although the distribution of hybrid disease levels was similar to that of the inbreds, the hybrids deviated from the expectation based on the mid-parent value. Interestingly, inbreds with additive effects for lower rust ratings and many hybrids with significant specific heterosis for reduced rust tended to be low yielding in the environment in which the experiment was grown. Rust severity was not associated with maturity, yield, or kernel weight. Hybrid checks had levels of leaf rust comparable to the F1 entries and superior performance for yield. Overall, contrary to expectation, post-flowering rust infection was not a v major factor contributing to variation in yield, and the material appeared to exhibit some level of tolerance and resistance to the foliar disease strain(s) present that year. The results of the first three chapters confirm that substantial dominance and epistatic variation, in addition to additive variation, would be present in hybrid association mapping populations produced by crossing among recombinant inbred lines (RIL) derived from the 66 F1 hybrids utilized in the generation means experiment. Experimental materials derived from ex-PVP lines will be an important component of corn breeding research on increased productivity. In this research, hybrid mapping populations will be invaluable for detection of non-additive QTL and association effects in yield related traits. This should facilitate the identification of functional diversity related to productivity and the study of the molecular basis of heterosis in maize. Chapter 4 compares the metabolic response of seedling roots of a hybrid maize genotype subjected to western corn rootworm (WCR) (Diabrotica vergifera vergifera) larvae feeding, mechanical wounding, and untreated controls. Large batches of root tips from treated seedlings grown in a replicated design within a growth chamber were collected, bulked, and processed for metabolite analysis. Shotgun gas-chromatography mass-spectrometry (GC-MS) was performed on biological replicates using six different extraction protocols. Analysis of variance was performed on relative metabolite abundances for the 95 metabolites in common across the treatments to detect significant variation for the effect of technical machine error, extraction protocol, and treatment. Significant variation due to technical error was rare, but the effect of the protocol on metabolite abundances was significant for a third of the metabolites tested. Differences between treatments were tested with three contrasts per metabolite. A full vi model containing protocol, treatment, and protocol-by-treatment interaction parameters, along with a reduced model only containing the treatment effect, identified 30 metabolites with at least one significant treatment contrast. The metabolites with significant contrasts were then classified according to patterns of contrast significance and direction of effect. Many metabolites were observed to be more abundant in the WCR-fed roots compared to mechanically wounded roots, but in many cases both of those treatments were not significantly different than the control. On the whole, abundances of many of the significant metabolites are reduced in mechanically wounded roots and enriched in WCR-fed roots. A large portion of the metabolites with significant differences between WCR and the other treatments are associated with cell wall metabolism and defense related signaling pathways. Canonical discriminate analysis was able to clearly separate samples from the different treatments using a subset of significant metabolites as predictors. Cluster analysis using the significant metabolites isolated WCR treated samples from wounded and control samples, but the quality of separation appeared to relate to the extraction protocol used. This is the first known public study identifying the metabolic response of maize roots to WCR feeding