Comparative Mapping of Seed Dormancy Loci Between Tropical and Temperate Ecotypes of Weedy Rice (Oryza sativa L.)

Abstract

Genotypic variation at multiple loci for seed dormancy (SD) contributes to plant adaptation to diverse ecosystems. Weedy rice (Oryza sativa) was used as a model to address the similarity of SD genes between distinct ecotypes. A total of 12 quantitative trait loci (QTL) for SD were identified in one primary and two advanced backcross (BC) populations derived from a temperate ecotype of weedy rice (34.3°N Lat.). Nine (75%) of the 12 loci were mapped to the same positions as those identified from a tropical ecotype of weedy rice (7.1°N Lat.). The high similarity suggested that the majority of SD genes were conserved during the ecotype differentiation. These common loci are largely those collocated/linked with the awn, hull color, pericarp color, or plant height loci. Phenotypic correlations observed in the populations support the notion that indirect selections for the wild-type morphological characteristics, together with direct selections for germination time, were major factors influencing allelic distributions of SD genes across ecotypes. Indirect selections for crop-mimic traits (e.g., plant height and flowering time) could also alter allelic frequencies for some SD genes in agroecosystems. In addition, 3 of the 12 loci were collocated with segregation distortion loci, indicating that some gametophyte development genes could also influence the genetic equilibria of SD loci in hybrid populations. The SD genes with a major effect on germination across ecotypes could be used as silencing targets to develop transgene mitigation (TM) strategies to reduce the risk of gene flow from genetically modified crops into weed/wild relatives.

Keywords: comparative genomics; quantitative trait locus; seed dormancy; segregation distortion; weed.

Figure 1

A framework linkage map and positions of seed dormancy QTL identified in this and previous research. The map was marked with RM loci segregating in the BC₁F₁ EM93-1//EM93-1/LD population. Asterisks indicate segregation distortion loci in favor of the allele from EM93-1 (*) or LD (**). Black bars indicate C.I. (equivalent to one-LOD support lengths) of qSD, qAn, or qHC detected in this BC₁F₁ population (Figure 2B). Five-point stars indicate qSDs detected in the BC₂F₁ (9) (Figure 3) and/or BC₂F₁ (139) (Figure 4), but not in this BC₁F₁ population. Ovals indicate approximate positions of qSDs previously detected in the BC₁F₁ EM93-1//EM93-1/SS18-2 population or its advanced generations (Gu et al. 2004; Ye et al. 2010). LD and SS18-2 are temperate and tropical ecotypes of weedy rice, respectively. Ch, chromosome; LD, pure line LüDao; LOD, logarithm of the odds; qAn, QTL for awn; qHC, QTL for hull color; qSD, QTL for seed dormancy; QTL, quantitative trait loci; RM, rice microsatellite.

Figure 2

Genome-wide scan for seed dormancy QTL in the BC₁F₁ EM93-1//EM93-1/LD population. (A) Frequency distributions of percent germination for seeds or caryopses. N was the number of plants evaluated for germination at DAR. (B) Distributions of LR along the 12 Ch. Seed dormancy QTL (qSD) were inferred by peaks of the LR distributions above the threshold. Ch, chromosome; DAR, days after ripening; LR, likelihood ratio; qSD, QTL for seed dormancy; QTL, quantitative trait loci.

Figure 3

Mapping of seed dormancy QTL in the BC₂F₁ (9) population. (A) A partial linkage map. The map was constructed with markers on 10 Ch or Ch segments segregating in the population. Black bars indicate one-LOD support lengths for qSD, qAn, or qHC. Ovals indicate positions of qSDs previously detected in the BC₁F₁ EM93-1//EM93-1/SS18-2 population (Gu et al. 2004). (B) Frequency distributions of percent germination. N was the number of BC₂F₁ plants evaluated at 7, 14, or 21 DAR. (C) Distributions of LRs along the map. qSDs were inferred by peaks of the LR distributions above the threshold. Ch, chromosome; DAR, days of after-ripening; LOD, logarithm of the odds; LR, likelihood ratio; qAn, QTL for awn; qHC, QTL for hull color; qSD, QTL for seed dormancy; QTL, quantitative trait loci.

Figure 4

Mapping of qSD in the BC₂F₁ (139) population. (A) A partial linkage map. The map was constructed with markers on 10 Ch or Ch segments segregating in the population. Black bars indicate one-LOD support intervals for qSD, qAn, or qHC. Ovals indicate positions of qSDs previously detected in the BC₁F₁ EM93-1//EM93-1/SS18-2 population (Gu et al. 2004, 2005b; Ye et al. 2010). (B) Frequency distributions of percent germination. N was the number of BC₂F₁ plants evaluated at 7, 14, or 21 DAR. (C) Distributions of LRs along the map. qSDs were inferred by peaks of the LR distributions above the threshold. Ch, chromosome; DAR, days after ripening; LOD, logarithm of the odds; LR, likelihood ratio; qAn, QTL for awn; qHC, QTL for hull color; qSD, QTL for seed dormancy; QTL, quantitative trait loci.