A stepwise route to domesticate rice by controlling seed shattering and panicle shape

Abstract

Asian rice (Oryza sativa L.) is consumed by more than half of the world's population. Despite its global importance, the process of early rice domestication remains unclear. During domestication, wild rice (Oryza rufipogon Griff.) acquired non-seed-shattering behavior, allowing humans to increase grain yield. Previous studies argued that a reduction in seed shattering triggered by the sh4 mutation led to increased yield during rice domestication, but our experiments using wild introgression lines show that the domesticated sh4 allele alone is insufficient for shattering loss in O. rufipogon. The interruption of abscission layer formation requires both sh4 and qSH3 mutations, demonstrating that the selection of shattering loss in wild rice was not as simple as previously suggested. Here we identified a causal single-nucleotide polymorphism at qSH3 within the seed-shattering gene OsSh1, which is conserved in indica and japonica subspecies but absent in the circum-aus group of rice. Through harvest experiments, we further demonstrated that seed shattering alone did not significantly impact yield; rather, yield increases were observed with closed panicle formation controlled by SPR3 and further augmented by nonshattering, conferred by integration of sh4 and qSH3 alleles. Complementary manipulation of panicle shape and seed shattering results in a mechanically stable panicle structure. We propose a stepwise route for the earliest phase of rice domestication, wherein selection of visible SPR3-controlled closed panicle morphology was instrumental in the sequential recruitment of sh4 and qSH3, which together led to the loss of shattering.

Keywords: Oryza rufipogon; Oryza sativa; closed panicle; domestication; seed shattering.

Fig. 1.

Identification of a causal SNP of qSH3 associated with the degree of seed shattering. (A) Graphical genotypes of O. sativa Nipponbare and ILs for qSH3 and qSH1 in the genetic background of cultivated rice O. sativa Nipponbare. (B) Comparison of seed-shattering degree by BTS values in three ILs, IL(qSH3-W), IL(qSH1-W), and IL(qSH1-W, qSH3-W), in the Nipponbare genetic background. Data are mean ± SD of four plants. n.s. and double asterisk (**) indicate not significant and significant at the 1% level based on unpaired Student’s t test, respectively. (C) Two types of constructs carrying qSH3 complementary DNA sequences of W630 and Nipponbare (qSH3^W630 and qSH3^Npb) driven by a 3-kb region of the promoter used for transgenic analysis. (D) The BTS values observed for the transformants with qSH3^W630 and qSH3^Npb. Black triangles represent the average BTS value of IL(qSH1-W).

Fig. 2.

Lineage-specific selection at the qSH3 locus in rice. (A) Genotyping at sh4, qSH1, and qSH3 for O. rufipogon W630, O. sativa japonica Nipponbare, indica IR36, and circum-aus Kasalath based on the causal SNPs identified by derived cleaved amplified polymorphic sequence markers. (B) Allele frequency of qSH3 causal SNP (%) in cultivated rice based on the Rice 3K project data. Nipponbare type (T) and W630 type (C) are shown in yellow and green, respectively. (C) Nucleotide diversity (π) observed for domesticated rice in the physical position of 25.0 to 25.3 Mb on chromosome 3. In a flanking region of the qSH3 locus (around 25.2 Mb), π was substantially decreased in japonica and indica cultivars. π was calculated in 10-kb windows using the SNP data of the Rice 3K project (14).

Fig. 3.

Abscission layer formation partially inhibited by sh4 and qSH3 in wild rice. Graphical genotypes of three ILs in the genetic background of O. rufipogon W630 are shown on the left. Abscission layer formations in O. rufipogon W630, IL(sh4-N), IL(qSH3-N), and IL(sh4-N, qSH3-N) are shown. Each enlarged section indicated by the dotted square is shown in the right-hand panel. A black arrowhead indicates an inhibited area of the abscission layer observed for IL(sh4-N, qSH3-N). (Scale bars, 100 μm.).

Fig. 4.

Role of sh4, qSH3, and SPR3 mutations in the initial loss of seed shattering in rice domestication. (A and B) Panicle shape of the seven ILs at flowering (A) and seed-maturation (B) stages. (Scale bars, 5 cm.) (C) BTS values for the seven ILs. BTS values for O. sativa japonica Nipponbare and indica IR36 are shown as controls. n.s. indicates not significant by unpaired Student’s t test. (D) Seed-gathering rates (mean ± SD of four plot replicates, with nine plants in each plot) for O. rufipogon W630 and the seven ILs. Mean values labeled with different letters are significantly different, whereas those with the same letters are not (Tukey’s test with arcsine-transformed values, P < 0.05).

Fig. 5.

Structural mechanics analysis of panicle shape and abscission layer inhibition related to the initial loss of seed shattering. (A) Panicle shape of O. rufipogon W630. (Scale bar, 5 cm.) (B) Schematic representation of awn and grain with panicle angle. θ represents the panicle angle. l₁, l₂, m₁g, and m₂g indicate the length of grain, length of awn, weight of grain, and weight of awn, respectively. (C) Detachment of grain and pedicel (Left), and scanning electron microscopy analysis of the pedicel abscission layer of O. rufipogon W630 (Right). AL and VB indicate abscission layer and vascular bundle, respectively. (Scale bars, 1 mm [Left] and 100 μm [Right].) (D) Schematic representation of the abscission layer. D and d indicate the diameters of the abscission layer and vascular bundle, respectively. Dotted circle indicates the area of disrupted abscission layer. (E) Simulation of the bending stress exerted on the spikelet base depending on panicle shape and abscission layer inhibition. A higher stress was experienced in the open panicle with less inhibited abscission.