Multiple Small-Effect Alleles of Indica Origin Enhance High Iron-Associated Stress Tolerance in Rice Under Field Conditions in West Africa

Abstract

Understanding the genetics of field-based tolerance to high iron-associated (HIA) stress in rice can accelerate the development of new varieties with enhanced yield performance in West African lowland ecosystems. To date, few field-based studies have been undertaken to rigorously evaluate rice yield performance under HIA stress conditions. In this study, two NERICA × O. sativa bi-parental rice populations and one O.sativa diversity panel consisting of 296 rice accessions were evaluated for grain yield and leaf bronzing symptoms over multiple years in four West African HIA stress and control sites. Mapping of these traits identified a large number of QTLs and single nucleotide polymorphisms (SNPs) associated with stress tolerance in the field. Favorable alleles associated with tolerance to high levels of iron in anaerobic rice soils were rare and almost exclusively derived from the indica subpopulation, including the most favorable alleles identified in NERICA varieties. These findings highlight the complex genetic architecture underlying rice response to HIA stress and suggest that a recurrent selection program focusing on an expanded indica genepool could be productively used in combination with genomic selection to increase the efficiency of selection in breeding programs designed to enhance tolerance to this prevalent abiotic stress in West Africa.

Keywords: NERICA; Oryza sativa; QTL mapping; genome wide association studies (GWAS); haplotype analysis; high iron associated (HIA) stress; iron toxicity.

FIGURE 1

Correlations between traits in the N-L-19 × IR64-Sub1 (Pop1) and N-L-43 × IR64-Sub1 populations (Pop2) and RDP1 panel. Correlations between traits scored in the field trials conducted in different West African sites during the 2012, 2013, 2016 and 2017 for Pop1 (A), Pop2 (B) and RDP1 (C). For RDP1, correlations are displayed for all AllPOP, INDICA and JAPONICA varietal groups. Lofa and Ibadan: control sites; Suakoko, Edozhigi and Vallee du Kou: HIA stress sites. White cells indicate a non-significant (p > 0.01) correlation value. The strength of significant (p < 0.01) correlations is indicated by values and colors (red if negative and blue if positive). FLW: days to flowering; PHT: plant height; GY: grain yield; GY loss: grain yield loss; LBS63: leaf bronzing score at 63 days after sowing; LBS84: leaf bronzing score at 84 days after sowing; n/a: not available.

FIGURE 2

Location of QTLs mapped in N-L-19 × IR64-Sub1 (Pop1) and N-L-43 × IR64-Sub1 populations (Pop2). Percentage of phenotypic variance explained (PVE) by the top marker of each QTL is reported in brackets (gray color if PVE < 7.5%). Positions of the semi-dwarf gene on chromosome 1 (sd1) and of the heading date gene on chromosome 3 (DTH3) are displayed. Chromosome regions showing high density of overlapping QTLs (≥3) related to HIA stress tolerance are indicated in red. FLW: days to flowering; PHT: plant height; GY: grain yield; GY loss: grain yield loss; LBS63: leaf bronzing score at 63 days after sowing; LBS84: leaf bronzing score at 84 days after sowing.

FIGURE 3

Location of GWA-QTLs mapped in RDP1. QTL positions indicated based on the bp coordinates of the most significant SNP (msSNP ± 200 kb). Red letters showing “Region 1-3 (bi-parental)” refer to the bi-parental QTL regions indicated in Figure 2. Purple letters showing “G1-G5” refer to the regions with clusters of GWA-QTLs in HIA stress sites. GY: grain yield; GY-loss: grain yield loss; LBS84: leaf bronzing score at 84 days after seeding. Iba: Ibadan; Edo: Edozhigi; Sua: Suakoko; VdK: Vallee du Kou; HIA-All: all HIA stress sites. QTLs detected in AllPOP shown in green; in INDICA shown in orange; in JAPONICA shown in blue.

FIGURE 4

Main haplotypes in the high-density bi-parental QTL regions related to HIA stress tolerance in AfricaRice and RDP1 germplasm. Haplotypes found in Regions 1, 2, and 3 (see Figure 2) constructed based on sets of 15 SNPs each; For Region 1, the haplotypes found in the nested GWAS region based on a set of 10 SNPs are shown; GWA-QTL-msSNPs overlapping with the bi-parental regions are marked by “ms” in red; blue rectangles = reference (cv Nipponbare) allele; yellow = alternate allele; gray = missing data. SNP positions (bp) shown next to reference (ref.) and alternative (alt.) alleles; SNPs predicted to cause non-synonymous substitutions highlighted in red and indicated with an asterisk (^∗). In Regions 1, 2, and 3, the main “extended” haplotypes found are named EH1 = N-L-19 (female parent), EH2 = IR64-Sub1 (male parent), EH3 = aus-like, EH4 = O. glaberrima-like. In the nested GWA region, the main “nested” haplotypes found are named NH1 = N-L-19 (female parent), NH2 = IR64-Sub1 (male parent), NH3 = aus-like, NH4 = O. glaberrima-like. Frequencies of each haplotype in different groups of accessions are indicated in tables where indica - AfricaRice group = 10 indica and 1 NERICA; indica - RDP1 group = 65 indica and 6 admixed-indica; O. glab – AfricaRice group = 21 O. glaberrima; aus - RDP1 group = 49 aus lines (see Supplementary Tables S1, S3). The complete absence of a haplotype in a group is indicated by (-).