Production of grains with low glutelin and high eating quality by using dominant allele Lgc‐1 in three‐line japonica hybrid rice

Grain proteins constitute the second most storage substance in rice, of which glutelin accounts for 60%–80% of total protein and is easy to be absorbed by humans (Kumamaru et al., 1988). However, for patients with kidney disease and diabetes, excessive glutelin intake is not conducive to recovery. The lgc-1 mutant is the earliest discovered low-glutelin material, and Lgc-1 regulates glutelin content in rice grains (Iida et al., 1993; Kusaba et al., 2003), which makes this allele have more extensive application prospects in the cultivation of low-glutelin varieties. Rice eating quality (REQ) is influenced by protein content and composition. Studies have shown that with the increase in protein content, the REQ decreases and the palatability becomes worse (Huang et al., 2020). Exogenous glutelin and prolamin could affect REQ, whereas knockout of glutelin-related genes could significantly improve the hardness, appearance and REQ. Therefore, the effect of glutelin on REQ may be greater than prolamin and total protein (Furukawa et al., 2006; Huang et al., 1998; Yang et al., 2022). Furthermore, it was found that the expression of genes related to glutelin synthesis had an important effect on protein content and REQ. As the expression of Nhd1 increased, the expression of GluA2 was inhibited, resulting in the decrease of glutelin content and protein content, thus improving REQ (Zhang et al., 2023). These studies indicated that glutelin can significantly affect REQ, however, fewer studies have been done on japonica hybrid rice with high eating quality and low glutelin. Combining with molecular marker and phenotypic screening, three low-glutelin restorer lines, HL8005, HL8023 and HL8027, were screened by crossing two varieties L9037 and R228 (Figure 1a). L9037 is a low-glutelin variety with the genotype Lgc-1 (without restoration gene), and R228 is a wide compatibility restorer line (without genotype Lgc-1). The amplified bands of HL8005, HL8023 and HL8027 were consistent with L9037 by molecular markers (Figure 1c; Figure S1). Compared with L9037, the number of grains per panicle decreased by 4.0% for HL8005, increased by 11.8% and 32.2% for HL8023 and HL8027, respectively, and the 1000-grain weight increased by 4.1% and 9.2% for HL8005 and HL8023, respectively, while the 1000-grain weight of HL8027 decreased by 4.6%, and the seed setting rate were all above 75% (Figure 1b). The heading time of three restorer lines was significantly shorter than L9037, and the single plant yield and population yield of three restorer lines were significantly higher than L9037 (Figure S2). The glutelin content of HL8005, HL8023 and HL8027 were significantly lower than R228, but higher than L9037 (Figure 1d). SDS-PAGE of storage profiles showed that the protein compositions of HL8005, HL8023 and HL8027 were consistent with L9037, but the glutelin precursors, acidic and basic subunits were less than R228 (Figure 1e). HL8005, HL8023 and HL8027 rice grains were longer than L9037, the grain width and thickness were greater than R228, and the length/width ratio was between them (Figure S3a–e). The results of physicochemical properties analysis showed that R228, HL8005 and HL8027 starch powder began to dissolve in 4 M urea, with no difference in solubility, while L9037 and HL8023 starch powder began to dissolve in 5 M urea. The solubility of HL8023 starch powder was the highest among the three restorer lines and higher than L9037 (Figure S3f). The amylose content of HL8005, HL8023 and HL8027 were significantly higher than two parents (Figure S3g). HL8005, HL8023 and HL8027 were significantly higher than L9037 and lower than R228 in terms of onset, peak and endset gelatinization temperatures; the enthalpies of HL8005 and HL8027 were significantly higher than L9037, whereas the enthalpy of HL8023 was similar to L9037 (Figure S3h). REQ analysis showed that HL8027 had better REQ than HL8005 and HL8023 (Figure S3i–k). These results indicated that the three low-glutelin restorer lines HL8005, HL8023 and HL8027 containing Lgc-1 had better agronomic traits than L9037. Among three restorer lines, HL8027 had the best REQ, and the glutelin content decreased to 3.0%, which was 38.8% lower than the non-low-glutelin variety R228. Nine hybrid rice combinations were obtained by crossing three low-glutelin restorer lines and japonica sterile lines Chunjiang23A (A1), 81A (A2) and Jiahe212A (A3), the non-low-glutelin japonica hybrid combination Jiayou#5 was used as the control (CK) (Figure 1f). The physicochemical qualities of three sterile lines were completely different, the gelatinization temperatures and the amylose content of A1 were the highest, and the glutelin content of the three sterile lines was more than 4.5% (Figure S4). The seed setting rate of C1, C2 and C3 and the 1000-grain weight of all combinations was significantly lower than CK. The single plant yield and population yield of C2, C4, C5 and C6 were significantly higher than CK (Figure 1g; Figure S5). The glutelin content of all combinations was significantly lower than CK, and the C3 was the lowest (Figure 1h). The amylose content of C2, C6 and C9 was significantly higher than CK, while the amylose content of C4 and C8 was significantly lower than CK (Figure 1i). All combinations had significantly higher onset, peak and endset gelatinization temperatures than CK, and only C2 had lower enthalpy than CK, and other combinations had significantly higher enthalpies or no difference with CK (Figure S6a). C7 started to dissolve in 4 M urea and C9 started to dissolve in 5 M urea. Among all combinations, C7 had the worst solubility and C9 had the best solubility (Figure S6b). Except for C3, C7 and C9, the number of grains per panicle in all combinations were higher than CK. Compared with CK, the fresh cooked rice of C6, C8 and C9 had brighter colour, whereas the C2, C3, C4 and C5 had darker colours and less gloss (Figure 1j). The texture analysis of fresh and retrograded cooked rice showed that there were significant differences in the texture parameters among all combinations, and the hardness and adhesiveness had a great influence on REQ, the rice with low hardness and high adhesiveness has better eating quality (Li et al., 2016). The hardness of C3, C4 and C5 fresh cooked rice were significantly greater than CK, the adhesiveness of C3 was significantly higher than CK, and the adhesiveness of C7 and C9 were significantly lower than CK. The texture characteristics of retrograded cooked rice of all combinations were inferior to fresh cooked rice. Based on a comparative analysis of the textures parameters of fresh and retrograded cooked rice, the texture properties of C6 and C8 were better (Figure 1k). Comprehensive score for taste and appearance showed that C6 had a higher score than CK, whereas the scores of C2, C3, C7 and C9 were significantly lower than CK (Figure 1l). It has been demonstrated that amylose content, gelatinization temperatures and adhesiveness are the most widely accepted indicators for evaluating REQ (Wang et al., 2024). Combining the texture analysis, taste evaluation and other related indicators, C6 (81A/HL8027) was considered to have the best REQ among nine combinations. In summary, we selected three low-glutelin restorer lines and found HL8027 had better REQ. Subsequently, three restorer lines were hybridized with three sterile lines to produce nine combinations. Among these combinations, 81A/HL8027 not only had low-glutelin characteristics, but also exhibited the best REQ, and had a clear advantage in single plant yield and population yield. The results provided new germplasm resources for breeding high-production hybrid rice with high eating quality and low glutelin. This research was supported by the National Natural Science Foundation of China (32071991, 32188102 and 32172080), Zhejiang Provincial Natural Science Foundation of China (LDQ23C130001), Key Research and Development Program of Zhejiang Province (2022C02011) and the Project of Laboratory of Advanced Agricultural Sciences, Heilongjiang Province (ZY04JD05-005). The authors declare no competing interests. S.H. (Shikai Hu) and J.C. designed the experiments and analysed the data; S.H. (Shikai Hu), L.Y., J.C., G.J., H.Y., S.H.(Suozhen Hui), L.Z., R.C., J.W., Y.C., J.F. and Z.S. performed the experiments; S.H. (Shikai Hu) and L.Y. wrote the manuscript and prepared the illustrations; P.H., S.H. (Shikai Hu) and S.T. conceived the idea and supervised the project. The data that supports the findings of this study are available in the supplementary material of this article. Figure S1–S6. Supporting Figures. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.