Ozone-Tolerant Rice for Air-Polluted Environments

Abstract

Recent decades have witnessed a surge in tropospheric ozone concentrations across many Asian countries, including Bangladesh, posing a significant threat to rice yields and regional food security. This study presents a comprehensive marker-assisted breeding program designed to introgress two quantitative trait loci (QTL) conferring ozone tolerance into elite Bangladeshi rice varieties. We crossed the donor landrace Kasalath with high-yielding varieties BRRI dhan28 and Binadhan-11, generating BC₂F₄ lines with an average of 87% recipient parent genome recovery. These lines harbored diverse combinations of positive alleles at the target QTLs while retaining the agronomic traits of modern rice varieties. In controlled open-top-chamber experiments, the presence of these QTLs in elite backgrounds mitigated ozone stress symptoms, reduced physiological damage such as lipid peroxidation or loss in photosynthetic CO₂ assimilation rates, and alleviated yield losses. Field trials conducted in a high ozone stress environment in Bangladesh, with and without the ozone protectant ethylenediurea (EDU), revealed that sensitive recipient parents suffered under ambient ozone concentrations. In contrast, the ozone-tolerant lines exhibited significantly less yield loss and were less affected by pollution. Our findings demonstrate the efficacy of marker-assisted breeding in adapting rice to changing atmospheric conditions, offering a robust strategy for safeguarding food security in the face of global change.

Keywords: air pollution; cereals; food security; molecular breeding; plant stress; yield losses.

FIGURE 1

Schematic representation of the backcross breeding scheme for the introgression of ozone tolerance QTLs (OzT8 and OzT9) into rice cultivars. The cross between the recurrent parent (BRRI dhan28 or Binadhan‐11) and the donor parent (Kasalath) was followed by backcrossing and selection for desired loci using marker‐assisted selection (MAS). The selection process integrates both phenotyping and genotyping. Foreground and recombinant selection were performed from BC₁F₁ to BC₂F₄ and BC₃F₃ progeny to identify favorable allelic combinations on chromosomes 8 and 9. Background selection was conducted on BC₂F₄ and BC₃F₃ progeny to evaluate the recovery of the recurrent parent's genetic background. Self‐pollination in advanced backcross generations ensured the fixation of homozygous alleles for the target loci. Chromosome diagrams highlight the positions of KASP markers, with OzT8 (yellow star), and OzT9 (pink star) explicitly indicated.

FIGURE 2

Foliar traits of selected breeding lines carrying OzT8 and/or OzT9 derived from the cross BRRI dhan28 × Kasalath (A, C, E) and Binadhan‐11 × Kasalath (B, D, F) under ozone stress and control conditions. (A, B) Leaf bronzing score (LBS) presented as absolute values; asterisks within the bars indicate significant differences between the genotypes and the recipient parent (BRRI dhan28 or Binadhan‐11), as determined by Dunnett's test (p < 0.05, *p < 0.01, **p < 0.001). (C, D) Normalized difference vegetation index (NDVI) and (E, F) Lichtenthaler index 2, presented as relative values (ozone/control). The bar graphs display the mean and standard error (n = 4). Circles (○) and triangles (△) represent absolute values for the control and ozone‐treated groups, respectively. Asterisks within the bars indicate significant differences between the genotypes and the recipient parent as determined by Dunnett's test, while letters shown beside the circles or triangles indicate significant differences among treatments within each genotype based on Šidák‐adjusted pairwise comparisons (p < 0.05). DP, donor parent (Kasalath), and RP, recipient parent (BRRI dhan28 or Binadhan‐11).

FIGURE 3

Photosynthetic parameters of selected breeding lines carrying OzT8 and/or OzT9 derived from the cross BRRI dhan28 × Kasalath (A, C, E) and Binadhan‐11 × Kasalath (B, D, F) under ozone stress and control conditions after 65 days of ozone fumigation. (A, B) Stomatal conductance, (C, D) quantum efficiency of photosystem II, and (E, F) net CO₂ assimilation rate. The bar graphs display the mean and standard error of relative values (ozone/control, n = 4). Circles (○) and triangles (△) represent absolute values for the control and ozone‐treated groups, respectively. Asterisks within the bars indicate significant differences between the genotypes and the recipient parent (BRRI dhan28 or Binadhan‐11) as determined by Dunnett's test (p < 0.05, *p < 0.01, **p < 0.001). Letters shown beside the circles or triangles indicate significant differences among treatments within each genotype, based on Šidák‐adjusted pairwise comparisons (p < 0.05). DP, donor parent (Kasalath), and RP, recipient parent (BRRI dhan28 or Binadhan‐11).

FIGURE 4

Grain yield components of selected breeding lines carrying OzT8 and/or OzT9 derived from the cross BRRI dhan28 × Kasalath (A, C, E) and Binadhan‐11 × Kasalath (B, D, F) under ozone stress and control conditions. (A, B) Grain yield, (C, D) panicle number per plant, and (E, F) filled grain number per plant. The bar graphs display the mean and standard error of relative values (ozone/control, n = 4). Circles (○) and triangles (△) represent absolute values for the control and ozone‐treated groups, respectively. Asterisks within the bars indicate significant differences between the genotypes and the recipient parent (BRRI dhan28 or Binadhan‐11) as determined by Dunnett's test (p < 0.05, *p < 0.01, **p < 0.001). Letters shown beside the circles or triangles indicate significant differences among treatments within each genotype, based on Šidák‐adjusted pairwise comparisons (p < 0.05). DP, donor parent (Kasalath), and RP, recipient parent (BRRI dhan28 or Binadhan‐11).

FIGURE 5

Principal component analysis (PCA) biplot showing the distribution of rice breeding lines carrying OzT8 and/or OzT9 derived from the cross BRRI dhan28 × Kasalath (A) and Binadhan‐11 × Kasalath (B) under ozone stress and control conditions. Cultivar distribution is presented along PC1 and PC2, with genotypes distinguished by shape/color and trait vectors color‐scaled by their contribution to the axes. The relative positions of genotypes reflect their ranking along the principal components (see Table S18), while trait loadings and contributions to PC1 and PC2 are detailed in Table S19, highlighting the phenotypic drivers of ozone sensitivity and resistance. Relative trait values (stress/control ratio) were used, except for LBS. Trait abbreviations: LBS, leaf bronzing score; NDVI, normalized difference vegetation index; Lic2, Lichtenthaler index 2; gsw, stomatal conductance; PhiPS2, quantum efficiency of photosystem II; A, CO₂ assimilation rate; DTF, days to flowering; DTM, days to maturity; PH, plant height; PN, panicle number; FGN, filled grain number; GY, grain yield; SB, straw biomass; HI, harvest index. Sample size: n = 12 (A) and n = 11 (B).

FIGURE 6

(A) Frequency distribution of ozone concentrations measured at the field site during the experimental growth period. Measurements were recorded at 3‐min intervals from 9:00 AM to 5:00 PM daily. (B) Seasonal dynamics of ozone concentrations (minimum, maximum, and average) during the field experiment.

FIGURE 7

Grain (A, B) and straw (C, D) yield (t ha⁻¹) of selected breeding lines carrying OzT8 and/or OzT9 derived from the cross BRRI dhan28 × Kasalath (A, C) and Binadhan‐11 × Kasalath (B, D) in field conditions under ambient ozone and with the application of the antiozonant ethylenediurea (EDU). The bar graphs display the mean and standard error of relative values (EDU/ambient ozone, n = 3). Circles (○) and triangles (△) represent absolute values for EDU‐treated and ambient ozone groups, respectively. Asterisks within the bars indicate significant differences between the genotypes and the recipient parent (BRRI dhan28 or Binadhan‐11) as determined by Dunnett's test (p < 0.05, *p < 0.01, **p < 0.001). Letters shown beside the circles or triangles indicate significant differences among treatments within each genotype, based on Šidák‐adjusted pairwise comparisons (p < 0.05). DP, donor parent (Kasalath), and RP, recipient parent (BRRI dhan28 or Binadhan‐11).