The Interaction of Genotype and Environment Determines Variation in the Maize Kernel Ionome

Abstract

Plants obtain soil-resident elements that support growth and metabolism from the water-flow facilitated by transpiration and active transport processes. The availability of elements in the environment interacts with the genetic capacity of organisms to modulate element uptake through plastic adaptive responses, such as homeostasis. These interactions should cause the elemental contents of plants to vary such that the effects of genetic polymorphisms will be dramatically dependent on the environment in which the plant is grown. To investigate genotype by environment interactions underlying elemental accumulation, we analyzed levels of elements in maize kernels of the Intermated B73 × Mo17 (IBM) recombinant inbred population grown in 10 different environments, spanning a total of six locations and five different years. In analyses conducted separately for each environment, we identified a total of 79 quantitative trait loci (QTL) controlling seed elemental accumulation. While a set of these QTL was found in multiple environments, the majority were specific to a single environment, suggesting the presence of genetic by environment interactions. To specifically identify and quantify QTL by environment interactions (QEIs), we implemented two methods: linear modeling with environmental covariates, and QTL analysis on trait differences between growouts. With these approaches, we found several instances of QEI, indicating that elemental profiles are highly heritable, interrelated, and responsive to the environment.

Figure 1

Ionome QTL from 10 environments. QTL identified for seed weight and 20 element accumulation traits using the IBM RIL population grown in 10 environments. (A) QTL on chromosome 1 affecting variation in molybdenum accumulation. An interval of Chr1 is shown on the x-axis in centimorgans. The LOD score for the trait-genotype association is shown on the y-axis. The horizontal line is a significance threshold from 1000 random permutations (α = 0.05). The LOD profiles are plotted for all environments in which the highlighted QTL was detected. (B) Total number of QTL detected for each trait, colored by environment. (C) Significant QTL (α = 0.05) for each trait. QTL location is shown across the 10 maize chromosomes (in centimorgans) on the x-axis. Dashes indicate QTL, with environment in which QTL was found designated by color. All dashes are the same length for visibility. The two black boxes around dashes correspond to LOD profiles traces in (A) and (D). (D) Stepwise QTL mapping output for nickel on chromosome 9. LOD profiles are plotted for all environments in which the QTL is significant.

Figure 2

Significant QTL-by-location interactions reflect variation in single environment mapping. (A) Nickel QTL on chromosome 9 exhibits variation between FL, IN, and NY growouts in single environment QTL mapping. Scanone QTL mapping output for Ni on a segment of Chr9 is plotted for FL05, FL06, IN09, IN10, NY05, and NY12. LOD score is plotted on the y-axis and centimorgans position on the x-axis. Horizontal line corresponds to significance threshold (α = 0.05). (B) Scanone QTL mapping for combined Ni data from Florida (FL05 and FL06), Indiana (IN09 and IN10), and New York (NY05 and NY12) growouts. All lines within each location were included, with covariates designated based on location. QTL mapping output using model with location as an additive covariate is shown as dotted line. QTL mapping output from model with location as both an additive and interactive covariate is shown as dashed line. Subtracted LOD score profile from the two models (QTL by location interactive effect only) is shown as solid line. Horizontal line corresponds to significance threshold for QTL by location interaction effect, derived from 1000 iterations of the three step procedure using randomized data: scanone QTL mapping with the additive model, scanone QTL mapping with the additive and interactive model, and subtraction of the two models.

Figure 3

Comparison of QTL mapped on traits in single environments and trait differences between environments. Examples from stepwise QTL mapping on trait differences between two years at one location, calculated between IBM lines common to both years. Scanone QTL mapping output is also plotted for the same trait from each year separately. LOD score is shown on the y-axis and centimorgans position on the x-axis. Horizontal lines correspond to significance threshold (α = 0.05). (A) Molybdenum QTL on chromosome 1 mapped for Mo in FL05, Mo in FL06, and difference in Mo content between FL05 and FL06. (B) Iron QTL on chromosome 8 mapped for Fe in IN09, Fe in IN10, and difference in Fe content between IN09 and IN10. (C) Potassium QTL on chromosome 5 mapped for K in NY05, K in NY12, and difference in K content between NY05 and NY12.