Genetic Architecture of Anther Extrusion in Spring and Winter Wheat

Abstract

Hybrid wheat breeding is gaining prominence worldwide because it ensures higher and more static yield than conventionally bred varieties. The cleistogamous floral architecture of wheat (Triticum aestivum L.) impedes anthers inside the floret, making it largely an inbreeder. For hybrid seed production, high anther extrusion is needed to promote cross pollination and to ensure a high level of pollen availability for the seed plant. This study, therefore, aimed at the genetic dissection of anther extrusion (AE) in panels of spring (SP), and winter wheat (WP) accessions by genome wide association studies (GWAS). We performed GWAS to identify the SNP markers potentially linked with AE in each panel separately. Phenotypic data were collected for 3 years for each panel. The average levels of Pearson's correlation (r) among all years and their best linear unbiased estimates (BLUEs) within both panels were high (r(SP) = 0.75, P < 0.0001;r(WP) = 0.72, P < 0.0001). Genotypic data (with minimum of 0.05 minor allele frequency applied) included 12,066 and 12,191 SNP markers for SP and WP, respectively. Both genotypes and environment influenced the magnitude of AE. In total, 23 significant (|log₁₀(P)| > 3.0) marker trait associations (MTAs) were detected (SP = 11; WP = 12). Anther extrusion behaved as a complex trait with significant markers having either favorable or unfavorable additive effects and imparting minor to moderate levels of phenotypic variance (R²(SP) = 9.75-14.24%; R² (WP) = 9.44-16.98%). All mapped significant markers as well as the markers within their significant linkage disequilibrium (r² ≥ 0.30) regions were blasted against wheat genome assembly (IWGSC1+popseq) to find the corresponding genes and their high confidence descriptions were retrieved. These genes and their orthologs in Hordeum vulgare, Brachypodium distachyon, Oryza sativa, and Sorghum bicolor revealed syntenic genomic regions potentially involved in flowering-related traits. Moreover, the expression data of these genes suggested potential candidates for AE. Our results suggest that the use of significant markers can help to introduce AE in high yielding varieties to increase cross fertilization rates and improve hybrid-seed production in wheat.

Keywords: QTL; Triticum aestivum L.; anther extrusion; hybrid wheat; linkage disequilibrium; marker trait associations.

Figure 1

Boxplots and Pearson's product moment correlations among growing years and their Best Linear Unbiased Estimates (BLUEs) in spring and winter wheat panels. Numbers on the Y-axis represent the number of extruded anthers. xs indicate the outlier accessions.

Figure 2

Anther extrusion (AE) performance of the top anther extruding accessions based on Best Linear Unbiased Estimates (BLUEs) ±SE. Six spring and 12 winter wheat accessions showed AE >80%. The growth habit of the accessions is indicated within the bars. Numbers on the Y-axis represent the number of extruded anthers.

Figure 3

Principal component analysis (PCA) of spring and winter wheat panels based on SNP genotypes (MAF >5%). The spring wheat accessions bred in Europe and Asia form recognizable clusters, but those bred in the Americas do not. There is no pronounced clustering among winter wheat accessions. Color code is given in the figure.

Figure 4

Summary of genome-wide association studies results of anther extrusion (AE) in spring wheat panel (SP). (A) Distribution of Best Linear Unbiased Estimates of AE (B) Manhattan plots based on mixed linear model using kinship matrix (K) to correct for population stratification. Horizontal dashed line corresponds to the threshold (|log₁₀(P)| > 3.0) for MTA estimation. (C) Quantile-quantile plots depicting expected vs. observed |log₁₀(P)| values.

Figure 5

Summary of genome-wide association studies results for anther extrusion (AE) in winter wheat panel (WP). (A) Distribution of Best Linear Unbiased Estimates of AE (B) Manhattan plots based on mixed linear model using kinship matrix (K) to correct for population stratification. Horizontal dashed line corresponds to the threshold (|log₁₀(P)| > 3.0) for MTA estimation. (C) Quantile-quantile plots depicting expected vs. observed |log₁₀(P)| values.

Figure 6

Expression of genes from spring wheat panel in above ground tissues. Expression values are given at log2 scale of transcripts per million (tpm): red, high expression; yellow, moderate expression; blue, low expression. Each individual gene (transcript_ID) is represented as horizontal row and different tissues are described in vertical columns. Tissues from pistil to spikelets show higher expression values for most genes compared to other tissues. Gene expression data were obtained from http://www.wheat-expression.com.

Figure 7

Expression of genes from winter wheat panel in above ground tissues. Expression values are given at log2 scale of transcripts per million (tpm): red, high expression; yellow, moderate expression; blue, low expression. Each individual gene (transcript_ID) is represented as horizontal row and different tissues are described in vertical columns. Tissues from pistil to spikelets show higher expression values for most genes compared to other tissues. Gene expression data were obtained from http://www.wheat-expression.com.