Introgression of a functional epigenetic OsSPL14WFP allele into elite indica rice genomes greatly improved panicle traits and grain yield

Abstract

Rice yield potential has been stagnant since the Green Revolution in the late 1960s, especially in tropical rice cultivars. We evaluated the effect of two major genes that regulate grain number, Gn1a/OsCKX2 and IPA1/WFP/OsSPL14, in elite indica cultivar backgrounds. The yield-positive Gn1a-type 3 and OsSPL14^WFP alleles were introgressed respectively through marker-assisted selection (MAS). The grain numbers per panicle (GNPP) were compared between the recipient allele and the donor allele groups using segregating plants in BC₃F₂ and BC₃F₃ generations. There was no significant difference in GNPP between the two Gn1a alleles, suggesting that the Gn1a-type 3 allele was not effective in indica cultivars. However, the OsSPL14^WFP allele dramatically increased GNPP by 10.6-59.3% in all four different backgrounds across cropping seasons and generations, indicating that this allele provides strong genetic gain to elite indica cultivars. Eventually, five high-yielding breeding lines were bred using the OsSPL14^WFP allele by MAS with a conventional breeding approach that showed increased grain yield by 28.4-83.5% (7.87-12.89 t/ha) vis-à-vis the recipient cultivars and exhibited higher yield (~64.7%) than the top-yielding check cultivar, IRRI 156 (7.82 t/ha). We demonstrated a strong possibility to increase the genetic yield potential of indica rice varieties through allele mining and its application.

Figure 1

Schematic representation of population development for evaluation of the effect of Gn1a-type 3 and OsSPL14^WFP alleles. (A) The recipient (Re) and the donor (Do) were crossed and the true F₁ plants were selected by SSR markers. Then, the plants having the target allele were selected by marker Gn1a-17 SNP for the Gn1a-type 3 allele and by marker SPL14–04 SNP for the OsSPL14^WFP allele in BC₁F₁, BC₂F₁, and BC₃F₁ generations. A total of 32–50 BC₃F₂ plants derived from a single BC₃F₁ plant were genotyped by the above markers in each population, and then the mean values of agronomic traits were calculated from each homozygous group in 2015DS. In the following cropping season (2015WS), the phenotype was tested again using 15 BC₃F₃ plants derived from three BC₃F₂ plants. The year with cropping seasons (WS: wet season, DS: dry season) is depicted in each plant generation. (B) Line information for the Gn1a-type 3 evaluation. (C) Line information for the OsSPL14^WFP evaluation.

Figure 2

Graphical genotype maps of the BC₃F₂ plants. From the eight BC₃F₂ populations, one plant per population that had the yield-positive Gn1a-type 3 allele (A–D) or OsSPL14^WFP allele (E–H) was genotyped by using Infinium 6 K SNP markers. Polymorphic SNPs between bi-parents were used for the construction of the genotype maps.

Figure 3

The effect of Gn1a-type 3 allele (A and B) and OsSPL14^WFP allele (C and D) in different indica backgrounds. The Gn1a-type 3 allele from the different donors (Habataki, ST12, and ST6) was introgressed into three different recipients having the Gn1a-type 2 allele. Mean values of GNPP were compared between the Gn1a-recipient allele (gray bar) and the Gn1a-donor allele (black bar) in the segregating BC₃F₂ progenies in each population in 2015DS (A) and in BC₃F₃ plants in 2015WS (B). Similarly, GNPP phenotype was compared between the recipient OsSPL14 allele and the OsSPL14^WFP allele in the BC₃F₂ generation in 2015DS (C) and in the BC₃F₃ generation in 2015WS (D). The data on lines YP16–775 and YP15–786 were omitted in 2015WS because of rice Tungro virus damage. Asterisks represent a significant difference between two alleles based on Student’s t-test (* α = 0.05 and ** α = 0.01). The effect of the OsSPL14^WFP allele was shown on the top of the bar as increase rate (%). The error bar means standard deviation.

Figure 4

Morphological analyses of the selected five high-yielding lines with their recipients. (A–E) Plant and panicle images from the representative plant of the recurrent line (top) and the improved line (bottom). Scale bar = 10 cm. (A) IRRI 123 (top) and YP16–22 (bottom). (B) PR37951 and YP16–32. (C) CT5803 and YP16–37. (D) CT5805 and YP16–40. (E) IRGA427 and YP16–44. (F) Section modulus (SM) values of the third internode. Student’s t-test (* α = 0.05 and ** α = 0.01) was used. The error bar means standard deviation. (G) Cross-section images of the fourth internodes. Sample order is consistent with the above plant/panicle images.