Abstract
Taste is crucial for the economic value of rice (Oryza sativa L.) and determines consumer preference. However, the mechanisms underlying taste formation have remained unclear. Here, we show that OsGATA7 contributes to desirable taste quality by affecting the swelling properties, texture, and taste value of cooked rice. OsGATA7 binds to the promoter of SMOS1, and activates its expression, thereby regulating taste quality. Furthermore, SMOS1 binds to the promoter of the protein biosynthesis gene OsGluA2, and recruits the PRC2 complex to repress its expression, leading to increased protein content. The overexpression of both OsGATA7 and SMOS1 reduces protein content and enhances taste quality. The haplotypes OsGATA7Hap1 and SMOS1Hap1 maintain low protein content and improve taste scores. Collectively, these findings reveal a regulatory mechanism for taste quality formation mediated by the OsGATA7–SMOS1 protein content module, and identify the elite haplotypes OsGATA7Hap1 and SMOS1Hap1 as a means to improve taste quality.
Similar content being viewed by others
Introduction
Rice (Oryza sativa L.) is a major source of carbohydrates, satisfying the daily need for the dietary calories and protein of more than two billion people worldwide1. Consumers prefer for rice grains with a specific and desirable taste, flavor, and visual appearance2. The establishment of rice taste quality is complex and largely evaluated on the basis of sensory attributes, textural properties, and/or taste value3. The starch and protein contents constitute ~80–85% and 4–10%, respectively, of the grain weight and determine the physicochemical and textural properties of cooked rice4. The starch in rice grains is composed of amylose and amylopectin, and its biosynthesis requires the cooperation of numerous enzymes, including GRANULE-BOUND STARCH SYNTHASE I (also known as Waxy), SOLUBLE STARCH SYNTHASE, ADP-GLUCOSE PYROPHOSPHORYLASE, STARCH DEBRANCHING ENZYMEs, and STARCH BRANCHING ENZYMEs5. The distinct composition and distribution of starch in rice grains affect the appearance quality and the eating and cooking quality (ECQ) of rice grains, reflecting their amylose content (AC), gel consistency (GC), gelatinization temperature (GT), viscosity, and milling quality4. Protein, the second most abundant storage substance in rice grains by weight, is negatively correlated with taste quality, with higher protein content altering the water absorption and textural properties of cooked rice, including hardness and stickiness6. However, how the taste quality of cooked rice is established, and the underlying molecular mechanisms remain largely unexplored.
In this study, we elucidate how a transcription regulatory module determines the formation of rice taste quality. We show that rice OsGATA77 binds to the promoter of SMOS18,9 and activates its expression. SMOS1 binds to the promoter of the protein biosynthesis gene OsGluA210 and recruits LC2 to repress its expression, resulting in increased protein content. Overexpression of both OsGATA7 and SMOS1 reduces protein content and enhances taste quality. Haplotypes OsGATA7Hap1 and SMOS1Hap1 maintain low protein content and improve taste value. In summary, we decipher a regulatory mechanism involved in the quality and taste of cooked rice and identify the elite haplotypes OsGATA7Hap1 and SMOS1Hap1.
Results
OsGATA7 regulates grain appearance and taste quality in rice
Transcription factor OsGATA7 was previously reported to regulate plant architecture and grain shape in rice7. To better understand the molecular mechanism of OsGATA7, we generated the osgata7-1 and osgata7-2 mutants via clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated nuclease 9 (Cas9)-mediated gene editing and the complementary lines osgata7-1 gOsGATA7#1 and osgata7-1 gOsGATA7#2 in the Zhonghua11 (ZH11) background (Supplementary Fig. 1a). The plant height, tiller number, panicle length, number of primary branches, grain length and width significantly differed between osgata7 and ZH11 (Supplementary Fig. 1b–h). Compared with wild-type (WT) ZH11, osgata7-1 and osgata7-2 presented an increased percentage of grains with chalkiness (PGWC) and higher chalkiness degree (CHD), likely caused by the loosely packed, round, and irregular starch granules (SGs) observed in the central regions of the endosperm in mutant grains (Fig. 1a–c and Supplementary Fig. 2a–c), together with the increased diameter of starch granules and their wider granulometric distribution (Supplementary Fig. 2d–e). Consistent with these results, the osgata7-3 and osgata7-4 mutants generated from the japonica rice accession ShenNong 265 (SN265) also presented increased PGWC and CHD (Supplementary Fig. 3a–c). ECQ indicators such as AC, GC, gelatinization properties, and starch viscosity properties were significantly different in osgata7-1 and osgata7-2 than in the corresponding WT, ZH11; we obtained similar results in osgata7-3 and osgata7-4 compared with SN265 (Fig. 1d–e and Supplementary Figs. 2f–h, 3d, e).
a Grain and cooked rice appearance of the ZH11, osgata7 and complementation lines. The grains in the upper row are non-cooked, whereas cooked rice is presented in the second row. b, c The percentage of grain with chalkiness (PGWC) (b) and chalkiness degree (CHD) (c) of mature grains of the ZH11, osgata7, and complementation lines. d–g Amylose content (d), gel consistency (e), taste value (f), and protein content (g) of the ZH11, osgata7, and complementation lines. n = 3 independent experiments in (b–e and g), n = 4 independent experiments in f. Data are presented as the means ± SD. and P-values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
The physicochemical properties of starch usually affect the texture of cooked rice11. To investigate the molecular role of OsGATA7, we measured the textural attributes of cooked rice with a texture profile analyzer. The values for chewiness and gumminess of cooked rice decreased in all four osgata7 mutants relative to their respective WTs, whereas hardness, springiness, and cohesiveness values were lower in the osgata7-1 mutants than in ZH11, and adhesion force was increased in osgata7-3 and osgata7-4 relative to SN265 (Table 1; Supplementary Table 1). Grains from the osgata7-1 and osgata7-2 mutants showed an increase in cooked rice elongation, water absorption rates, and volume expansion compared to that of ZH11 grains (Supplementary Fig. 2i–k). All four OsGATA7 loss-of-function mutants showed a drop in taste value, together with the lower appearance quality and palatability observed in the osgata7-3 and osgata7-4 grains (Fig. 1f and Supplementary Figs. 2l, 3f–h). Compared with that of the corresponding wild type, the protein content of the grains from all four mutants was higher (Fig. 1g and Supplementary Fig. 3i). Taken together, our results indicate that in addition to regulating rice agronomic traits, OsGATA7 also affects the swelling and textural properties of cooked rice, as well as its taste score.
Natural variation in OsGATA7 affects the formation of taste quality
To elucidate whether the natural variation in OsGATA7 is critical for the formation of taste quality, ten haplotypes Hap1 to Hap10 were screened from 1277 indica (Ind), 750 japonica (Jap), 178 aus, 63 basmati (Bas), and 75 admix rice accessions collected from 3K rice germplasm resources, of which two major haplotypes Hap1 and Hap2 were further investigated (Fig. 2a and Supplementary Fig. 4a). One SNP (21943828 C/G) in the coding region leads to a synonymous amino acid substitution. The Hap1 haplotype of OsGATA7 presented relatively high expression within the collected germplasm resources (Supplementary Fig. 4b).
a Haplotype analysis of OsGATA7 using 2280 rice accessions. b–c The protein content (b) and taste values (c) of 51 mini core rice accessions within two major OsGATA7 haplotypes. d Grain appearance of ZH210, IL-1 and IL-2. e–f Protein content (e) and taste value (f) of ZH210, IL-1 and IL-2. g Grain and cooked rice appearance of the SN265 and OsGATA7-OE lines. The grains in the second row are non-cooked, whereas cooked rice is presented in the third row. h, i Protein content (h) and taste value (i). n = 3 independent experiments in (e and h), n = 4 independent experiments in (f and i). Data are presented as the means ± SD, and P-values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
To determine whether the OsGATA7 haplotypes are necessary for the formation of taste quality, a total of 51 mini-core germplasms were collected and used for genotypic and phenotypic analysis (Supplementary Data 1). Compared with the OsGATA7Hap2, OsGATA7Hap1 has decreased protein content and improved taste value (Fig. 2b, c). To evaluate the potential breeding value of the haplotypes of OsGATA7, introgression lines (IL) OsGATA7-IL-1 (IL-1) and OsGATA7-IL-2 (IL-2) harboring the KDML105 haplotype OsGATA7Hap2 in a restore line Zhonghui210 (ZH210) background were generated, and PGWC and CHD showed no obvious differences among those materials. However, IL-1 and IL-2 presented increased protein contents but with lower taste scores compared to WT (Fig. 2d–f). These results indicated that OsGATA7Hap1 improved the taste quality of rice. In addition, we generated the overexpression lines OsGATA7-OE#1 and OsGATA7-OE#2 in the SN265 background, Compared with SN265, the overexpression lines OsGATA7-OE#1 and OsGATA7-OE#2 presented decreased protein content and increased taste value (Fig. 2g–i). These results suggest that OsGATA7 is a positive regulator of the rice taste score.
OsGATA7 directly binds to the SMOS1 promoter and activates its transcription
OsGATA7 is constitutively expressed in all tissues, suggesting that it may regulate several agronomic traits, such as plant architecture and grain shape7 (Supplementary Fig. 5a). OsGATA7 contains a typical GATA domain and is highly conserved on the basis of multiple amino acid sequence alignment with related proteins from other plant species (Supplementary Fig. 5b, c). We examined the subcellular localization of OsGATA7 by transfecting rice protoplasts with an OsGATA7-GFP construct encoding a fusion between OsGATA7 and green fluorescent protein (GFP). We detected fluorescence signals mainly in the nucleus; moreover, the N-terminus of OsGATA7 exhibited transcriptional activation activity in yeast cells (Fig. 3a, b). To identify the downstream genes regulated by OsGATA7, we conducted RNA-seq and cleavage under targets and tagmentation (CUT&Tag) assays which led to the identification of 807 downregulated differentially expressed genes (DEGs) between ZH11 and osgata7-2 and 773 putative OsGATA7 targets. All DEGs between ZH11 and osgata7-2 enriched with OsGATA7 were identified. Notably, several candidate genes were common to the two gene lists, including SMOS1, which encodes an AP2-type transcription factor and is a multifunctional gene with roles in organ size, grain size, and rice yield8,9; the OsGATA7-binding site in the SMOS1 promoter is the conserved GATC motif. The OsGATA7-binding sites for the other six loci were identified via the Integrative Genomics Viewer (IGV) (Fig. 3c–g and Supplementary Fig. 6a–h).
a Subcellular location of OsGATA7-GFP in rice protoplasts. D53 fused to the red fluorescent protein mCherry was used as a nuclear marker. The experiments were replicated 3 times with similar results. b Transcriptional activation activity of full-length OsGATA7 and its truncated variants in yeast cells. Each OsGATA7 variant was fused to the yeast GAL4 DNA-binding domain (BD). NTR, N-terminal region; CTR, C-terminal region. c Volcano plot showing the differentially expressed genes (DEGs) between ZH11 and osgata7. Grains at 10 days after fertilization (DAF) were used for RNA-seq analysis based on log2 (FC) > 1.5 and p-value < 0.05. d CUT&Tag assay revealing the genome-wide distribution of OsGATA7 binding peaks, shown as a proportion of genomic features. UTR, untranslated region. e Venn diagram showing the extent of overlap between the DEGs identified via RNA-seq and putative OsGATA7 target genes identified via CUT&Tag. f Integrative Genome Viewer (IGV) window showing the enrichment of OsGATA7 at the SMOS1 promoter, as determined via CUT&Tag analysis. Red arrows indicate significant peaks calculated by the peak-calling prioritization pipeline PePr. LOC_Os05g32260 upstream of SMOS1 was used as a negative control. SMOS1-P1 and SMOS1-P2 represent the 1.0- and 0.35-kb SMOS1 promoter fragments, respectively. g Motifs of OsGATA7-binding peaks identified via CUT&Tag analysis. The red arrow indicates the binding motif identified in the SMOS1 promoter region. h Yeast one-hybrid assay showing that OsGATA7 can bind to the SMOS1 promoter. i Electrophoretic mobility shift assay (EMSA) showing that OsGATA7 directly binds to a probe derived from the SMOS1 promoter. The experiments were replicated 3 times with similar results. j Relative SMOS1 expression levels in 10-DAF grains of ZH11 and osgata7 mutants determined via RT−qPCR. k Dual-luciferase assay showing the transcriptional activation activity of OsGATA7 toward SMOS1 transcription in rice protoplasts. Relative luciferase activity was calculated as firefly luciferase (LUC)/Renilla luciferase (REN). n = 3 independent experiments in (j, k). Data are presented as the means ± SD, and P-values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
To test whether SMOS1 is a target of OsGATA7, we performed a yeast one-hybrid assay using two fragments of the SMOS1 promoter, P1 (1000 bp upstream of the ATG) and P2 (350 bp upstream of the ATG). Indeed, OsGATA7 bound to SMOS1-P1 and SMOS1-P2 in yeast (Fig. 3h). To test whether OsGATA7 can directly bind to the SMOS1 promoter, we performed an electrophoretic mobility shift assay (EMSA) with recombinant purified OsGATA7-histidine (His) using a 40-bp probe containing the GATC motif. The presence of OsGATA7-His in the binding reaction caused a shift in the mobility of the labeled probe, which was disrupted by the addition of the unlabeled probe (Fig. 3i). We thus wondered whether OsGATA7 might affect SMOS1 expression, which prompted us to perform RT–qPCR. The expression levels of SMOS1 were much lower in osgata7-1 and osgata7-2 than those in ZH11, reaching only approximately 20% of the wild-type level (Fig. 3j). In a transcriptional activation assay in rice protoplasts, we determined that OsGATA7 can activate transcription from the SMOS1 promoter, as evidenced by the increased relative luciferase (LUC) activity observed for the pro-SMOS1:LUC reporter consisting of the firefly luciferase (LUC) reporter gene driven by the 350-bp SMOS1 promoter (Fig. 3k). Collectively, these findings demonstrate that SMOS1 is a target gene of OsGATA7 and that OsGATA7 directly binds to the promoter of SMOS1 and activates its transcription.
OsGATA7 and SMOS1 cooperatively regulate rice appearance and taste quality
SMOS1 was highly expressed in the endosperm of grains 10 days after fertilization (DAF); SMOS1-GFP was mainly localized in the nucleus of rice protoplasts transfected with a SMOS1-GFP construct. SMOS1 contains a typical AP2 domain and is highly conserved on the basis of multiple amino acid sequence alignment with related proteins from other plant species (Supplementary Fig. 7a–d). To reveal the molecular role of SMOS1, we generated smos1-1 and smos1-2 mutants in ZH11 by CRISPR/Cas9-mediated gene editing (Supplementary Fig. 8a). Compared with wild-type ZH11 and 9311, smos1-1 to smos1-3 showed increased PGWC and CHD values with loosely packed, round, and irregular SGs, reminiscent of the phenotypes observed in smos1 grains (Supplementary Figs. 8b–d, 9a–e). The total starch content, AC, and GC were lower in all three smos1 mutants than in the corresponding wild type, whereas the protein content was higher in the mutants (Supplementary Figs. 8e–h, 9f–i). The smos1-3 mutant in 9311 also presented an altered rapid visco analyzer profile, an accumulation of shorter starch chains, and a lower starch crystallinity intensity along with changes in the swelling properties of cooked rice (Supplementary Fig. 9j–p). As expected, SMOS1 also regulates the performance of several agronomic traits, such as plant height, tiller number, panicle length, number of primary branches, grain length, and grain width (Supplementary Fig. 10a–g), suggesting that SMOS1 is necessary not only for the formation of rice appearance quality and physicochemical properties but also for agronomic traits.
To decipher the genetic roles of OsGATA7 and SMOS1, we obtained the osgata7-1 smos1-1 double mutant by crossing osgata7-1 with smos1-1. Compared with ZH11, osgata7-1 smos1-1 exhibited increased values for PGWC and CHD, together with larger SGs similar to those observed in smos1-1 (Fig. 4a–c and Supplementary Fig. 11a–c). The ECQ indicators AC, GC, GT, and viscosity were all decreased, whereas the protein content was increased in osgata7-1 smos1-1 (Fig. 4d, e and Supplementary Fig. 11d–f). The textural indicators of cooked rice, such as hardness, springiness, cohesiveness, chewiness, and gumminess, were significantly lower, whereas the swelling properties of cooked rice (cooked rice elongation, water absorption, and volume expansion) were all increased in osgata7-1 smos1-1 (Fig. 4f–h and Table 1). The taste value and palatability of cooked rice were decreased in the osgata7-1 smos1-1 (Fig. 4i, j). Accordingly, the appearance and rice taste quality of osgata7-1 smos1-1 grains resembled those of smos1, suggesting that OsGATA7 and SMOS1 cooperatively regulate rice appearance and taste quality.
a Representative photographs of grains and cooked rice appearance of the ZH11 and of osgata7, smos1, and osgata7 smos1 mutants. The grains in the first and third rows are not cooked; those in the second and fourth rows are cooked. b, c PGWC (b) and CHD (c) of mature grains of ZH11, osgata7, smos1, and osgata7 smos1. d Protein contents of grains of ZH11, osgata7, smos1, and osgata7 smos1. e Pasting properties of endosperm starch from grains of ZH11, osgata7, smos1, and osgata7 smos1. f–h Elongation of cooked rice (f), water absorption rate (g), and volume expansion (h) of grains of ZH11, osgata7, smos1, and osgata7 smos1. i, j Taste value (i), and palatability ( j) of cooked rice prepared from grains of ZH11, osgata7, smos1, and osgata7 smos1. n = 3 independent experiments in (b–d, f–h), n = 4 independent experiments in (i and j). Data are presented as the means ± SD, and P- values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
Natural variation in SMOS1 influences the formation of taste quality
We asked whether natural variation at SMOS1 might produce variation in rice appearance and taste quality across rice accessions. To this end, we analyzed the SMOS1 sequences from 1325 indica (Ind), 715 japonica (Jap), 154 aus, 55 basmati (Bas), 65 admix, and 35 wild rice accessions, all of which were collected from 3K rice germplasms. We identified 12 haplotypes, Hap1 to Hap12, among which we focused on the four major haplotypes Hap1 to Hap4 (Fig. 5a and Supplementary Fig. 12a). Hap1 was found mainly in Jap accessions distributed in high-latitude regions, whereas Hap2 to Hap4 were found mainly in Ind accessions distributed in low-latitude regions (Fig. 5b–c). To understand whether SMOS1 was domesticated, we investigated the nucleotide diversity (π) and fixation index (FST) of a 2-Mb genomic region centered on SMOS1. The π and FST values across the SMOS1 region were globally lower in the Jap and Ind accessions than in wild rice, with a difference between the Jap and Ind groups (Fig. 5d, e), indicating that SMOS1 has indeed undergone artificial selection during rice domestication. To further test the promoter activity of the above four haplotypes, qRT–PCR and transient expression analyses were performed. Hap1 is responsible for the higher expression level of SMOS1 in the collected germplasm resources compared to the other three haplotypes, along with higher LUC expression of the reporter gene. The base variation at positions 18811397 (T to G mutation) and 18811480 (G to A mutation) in Hap1 might affect the expression of SMOS1, and ultimately be involved in rice taste quality (Supplementary Fig. 12b, c).
a Haplotype analysis of SMOS1 using 1986 rice accessions. b Distribution frequency of the four major SMOS1 haplotypes in various rice subpopulations. c Geographical distribution of rice accessions harboring each of the four major SMOS1 haplotypes. d, e Nucleotide diversity (π values) (d) and fixation index (FST) values (e) over a 2-Mb genomic region centered on SMOS1. The position of SMOS1 is indicated by the red arrow. f, g Protein contents (f) and taste values (g) of a mini core set of 61 rice accessions, each harboring one of the four major SMOS12 haplotypes. h, i Protein contents (h) and taste values (i) of grains from XS134, IL-3, and IL-4. j, k Protein contents (j) and taste values (k) of grains from Nip, IL-5, and IL-6. l, m Protein contents (l) and taste values (m) of grains from ZH210, IL-7, and IL-8. n = 3 independent experiments in (h, j and l), n = 4 independent experiments in (i, k and m). Data are presented as the means ± SD, and P-values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
To reveal which SMOS1 haplotypes produce desirable taste quality, we assembled a mini core germplasm of 61 accessions harboring Hap1 to Hap4, which we used for genotypic and phenotypic analysis (Supplementary Data 1). Compared with those carrying Hap2 to Hap4, accessions harborinmg Hap1 presented decreased protein content and improved taste value (Fig. 5f, g). To evaluate the breeding potential of accessions with different haplotypes at SMOS1, we generated the introgression lines IL-3 and IL-4, harboring the 9311 haplotype of SMOS1 (Hap3) in the Xiushui 134 (XS134) background; IL-5 and IL-6, containing the Kasalath haplotype of SMOS1 (Hap3) in the Nipponbare (Nip) background; IL-7 and IL-8, containing the Koshihikari haplotype of SMOS1 (Hap1) in the ZH210 background. We observed no significant differences among these materials in terms of grain appearance quality, such as PGWC and CHD (Supplementary Fig. 13a–i). However, in comparison to those of the corresponding wild types XS134 and Nipponbare (Nip), both of which are Hap1, the protein contents of IL-3 to IL-6 were all increased, and the taste values were decreased; the protein content of IL-7 and IL-8 was decreased but the taste value was higher than that of ZH210 (Fig. 5h–m). Thus, Hap1 of SMOS1 derived from the Jap accessions may help improve rice taste quality.
SMOS1 overexpression improves taste value
Loss of SMOS1 function affected the appearance quality, physicochemical properties, and textural properties of cooked rice (Fig. 4 and Supplementary Figs. 8–9). We wondered whether the overexpression of SMOS1 might improve taste quality. SMOS1 contains two main splice variants (http://rice.uga.edu/). Accordingly, we generated the overexpression lines SMOS1-OE#1 and SMOS1-OE#2, which express the longer transcript isoform of SMOS1 driven by the cauliflower mosaic virus (CaMV) 35S promoter, and SMOS1-OE#3 and SMOS1-OE#4, which express the shorter SMOS1 transcript isoform cloned in-frame with the sequence encoding a FLAG tag (SMOS1-FLAG) driven by the rice ACTIN promoter. Compared with those of ZH11, all four SMOS1-OE lines showed increased PGWC and CHD values, together with loosely packed, round, and irregular SGs (Fig. 6a–c and Supplementary Fig. 14a), Similar to the smos1 loss-of-function mutants, all SMOS1-OE lines also contained larger SGs with a wider granulometric distribution (Supplementary Fig. 14b, c). An investigation of physicochemical properties revealed that GC, GT, and viscosity values were comparable between the SMOS1-OE lines and ZH11, except for a slight increase in AC observed in the SMOS1-OE lines (Supplementary Fig. 14d–g). The elongation, adhesion force, taste score, appearance, and palatability of the cooked rice were all increased in the SMOS1-OE lines compared to ZH11, whereas their water absorption, volume expansion, and hardness values were decreased (Fig. 6d–i and Table 1). The protein contents were lower in the SMOS1-OE lines relative to ZH11 (Fig. 6j, k). We conclude that SMOS1 is a positive regulator of rice taste quality.
a Representative photographs of grains and cooked rice from ZH11 and SMOS1 overexpression lines. The grains in the first and third rows are not cooked; those in the second and fourth rows are cooked. SMOS1-OE#1 to SMOS1-OE#4 are different SMOS1 overexpression lines. b, c, PGWC (b) and CHD) (c) of mature grains of the ZH11 and SMOS1 overexpression lines. d–f Elongation of cooked rice (d), water absorption rate (e), and volume expansion (f) of grains from the ZH11 and SMOS1 overexpression lines. g–i Taste value (g), appearance (h), and palatability (i) of grains from the ZH11 and SMOS1 overexpression lines. j Protein contents of grains from the ZH11 and SMOS1 overexpression lines. k Relative SMOS1 expression levels in grains from ZH11 and SMOS1 overexpression lines. n = 3 independent experiments in (b–f and j, k), n = 4 independent experiments in (g–i). Data are presented as the means ± SD, and P-values are indicated by a two-tailed Student’s t test. Source data are provided as a Source Data file.
The OsGATA7–SMOS1 module represses expression of protein biosynthesis genes to improve taste quality
Expression of SMOS1 (also known as NITROGEN-MEDIATED TILLER GROWTH RESPONSE 5 [NGR5]) was previously shown to respond to the application of nitrogen (N) fertilization to improve grain yield9. We therefore wondered whether loss of SMOS1 function would affect rice grain quality upon treatment with N fertilizer at low (LN), medium (MN), or high (HN) doses. The values for PGWC, CHD, and protein contents all gradually increased in ZH11 and smos1-1 when treated with increasing amounts of N fertilizer, whereas the taste value gradually decreased under the same conditions (Supplementary Fig. 15a–e).
SMOS1 was reported to act as a negative regulator of rice tiller development by interacting with polycomb repressive complex 2 (PRC2)9. We collected 10-DAF grains from 9311 and smos1-3 for RNA-seq analysis to identify downstream SMOS1 target genes. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that SMOS1 affects various biological processes, namely, plant hormone signal transduction, starch and sucrose, and ubiquinone and other terpenoid-quinone. Notably, pathways related to nitrogen metabolism and amino acid metabolism of alanine, aspartate, glutamate, phenylalanine, tyrosine, tryptophan, cysteine, and methionine were affected in the smos1-3 mutant relative to the wild type, suggesting that loss of SMOS1 function may affect protein production (Supplementary Fig. 16 and Supplementary Data 2). In agreement with this idea, protein synthesis–related genes, such as glutelin type-A2 (OsGluA2), alpha-globulin1 (Glb1) and OsGluD, were upregulated in smos1-3 compared with 9311 (Supplementary Data 2). SMOS1 CUT&Tag assays revealed that the promoter regions of OsGluA2 and the Cys-rich 13-kDa prolamin (RM1) could be bound by SMOS1 (Fig. 7a and Supplementary Fig. 17a). A yeast one-hybrid assay revealed that SMOS1 could bind to OsGluA2-P1 and RM1-P2 (Fig. 7b and Supplementary Fig. 17b). Further qRT–PCR assays revealed the upregulation of OsGluA2, Glb1, and OsGluD expression as well as that of OsGluB1, AMINO ACID TRANSPORTER PROTEIN 6 (OsAAP6), RM1, and Prolamin 14 (Pro14)10, 12,13,14,15 in 10-DAF grains in osgata7-1 and smos1-1 compared with ZH11 (Fig. 7c–d and Supplementary 17c–d). Furthermore, a transcriptional activation assay was conducted in rice protoplasts, and we demonstrated that SMOS1 can repress the expression of the OsGluA2 and RM1 promoters (Fig. 7e and Supplementary Fig. 17e). OsGluA2, a positive regulator of protein content, harbors a SNP in its promoter that is associated with protein content diversity11. We generated the introgression lines IL-9 and IL-10 containing the Japonica Koshihikari haplotype of OsGluA2LET in the ZH210 background (Fig. 7f–h). The protein contents of the introgression lines IL-9 and IL-10 were decreased, in contrast, an increased taste score was observed in OsGluA2HET (Fig. 7i, j).
a IGV window showing the enrichment of SMOS1 at the OsGluA2 promoter, as determined by CUT&Tag analysis. Red arrows indicate significant peaks calculated by PePr. b Yeast one-hybrid assay showing that SMOS1 can bind to the OsGluA2 promoter. c, d Relative OsGluA2 expression levels in 10-DAF grains of ZH11 and osgata7-1 (c) and smos1-1 (d) mutants using RT−qPCR. e Dual-luciferase assay showing the transcriptional activation activity of SMOS1 toward OsGluA2 transcription in rice protoplasts. Relative luciferase activity was calculated as the ratio of firefly LUC-REN. f Diagram of the genotypes of Koshihikari with OsGluA2LET, ZH210 with OsGluA2HET, and two introgression lines IL-9 and IL-10 carrying OsGluA2LET in the ZH210 background. Koshihikari and ZH210 were used as donor and recurrent parents, respectively. g Functional SNP sequencing associated with nucleotide diversity in Koshihikari, ZH210 and two introgression lines. h Grain appearance of ZH210, IL-9 and IL-10. i, j Protein content (i) and taste value (j) of ZH210, IL-9, and IL-10. k A proposed model for the interaction of OsGATA7 with SMOS1 to repress protein biosynthesis gene expression and improve taste quality. PRC2, polycomb repressive complex 2. n = 3 independent experiments in (c–e and i), n = 4 independent experiments in (j). Data are presented as the means ± SD, and P-values are indicated by a two-tailed Student’s t-test. Source data are provided as a Source Data file.
Based on the above findings, we propose a model for the roles of OsGATA7 and SMOS1 in controlling the taste quality of cooked rice (Fig. 7k). OsGATA7 binds to the GATC motif in the SMOS1 promoter region and activates its transcription; SMOS1 can recruit the PRC2 complex LC2 to repress the expression of protein biosynthesis-related genes, such as OsGluA2 and RM1, and then affects the amino acid composition and rice taste quality (Fig. 7a–j and Supplementary Fig. 17a–e and Supplementary Data 3).
Discussion
The ECQ of rice grains plays a vital role in determining the price of rice in the export market and the consumer acceptability of rice grains16. Importantly, higher ECQ parameters are not always associated with better-tasting cooked rice, as the shape, color, smell, taste, palatability, and sensory characteristics of rice after cooking also contribute to the formation of rice quality5. Furthermore, the swelling or textural properties of cooked rice also affect taste quality and consumer satisfaction, guiding the breeding of rice varieties with good taste4,17. The swelling properties of rice grains include the elongation, water absorption, and volume expansion of cooked rice, while the textural properties reflect multiple sensory properties, especially the hardness and stickiness (adhesion force) of cooked rice17,18. In rice, allelic variation affects the swelling or textural properties of cooked rice, such as at the Waxy (Wx) gene encoding a GBSS required for amylose biosynthesis and at the Alkali Spreading Value (ALK) locus encoding SSIIa for amylopectin biosynthesis19. However, additional components contributing to taste quality are limited, along with an understanding of the underlying regulatory mechanism for the formation of cooked rice taste quality. In this study, we demonstrated that OsGATA7, a zinc finger protein previously linked to brassinosteroid (BR)-mediated growth regulation of plant architecture, panicle development, and grain shape7, directly regulates the expression of SMOS1, which is required for the establishment of plant architecture9 and of characteristics affecting the taste value of rice grains. We showed that OsGATA7 and SMOS1 affect the textural properties of cooked rice and its swelling properties (Table 1). We propose that OsGATA7 and SMOS1 are transcription factors that work together and define components of a regulatory cascade that affects the textural and swelling properties of cooked rice, ultimately regulating rice taste quality.
Starch is the most abundant component in rice grain endosperm by weight, but the composition and structure of starch among rice cultivars cannot fully explain the variation in rice ECQ or taste quality (Figs. 1d, 4e and Supplementary Figs. 2f–h, 3d, 8e–f, 9f–g, 18a, b)4, as the branches and ratio of amylopectin chain length, lipid content, nitrogen fertilizer, and growth conditions affect rice quality5. Importantly, protein represents the second most abundant storage substance in rice endosperm, and its levels are negatively correlated with rice ECQ and taste quality20. In this study, we determined that the protein biosynthesis genes OsGluA2, Glb1, and OsGluD are upregulated in the osgata7 and smos1 mutants, resulting in an increased protein content with lower taste quality than those of wild-type grains. SMOS1 binds to the OsGluA2 and RM1 promoters and represses their expression (Fig. 7a–j and Supplementary Fig. 17a–e). Similarly, 57-kDa proglutelins (pGlu) in mature seeds were obviously accumulated in the endosperm of osgata7 and smos1 mutants (Supplementary Fig. 19). In contrast, SMOS1 overexpression lines exhibited better taste quality with decreased protein content (Figs. 4d, i, 6g, j), further supporting the notion that protein is a vital indicator of rice taste quality. The application of nitrogen fertilizer notably increases the content of storage protein in grains and affects ECQ21. SMOS1 expression is activated when plants are treated with nitrate, leading to an increased tiller number8, suggesting that SMOS1 expression responds to nitrate by regulating both rice appearance and taste quality. We established that the PGWC, CHD, and protein contents of rice grains gradually increased in ZH11 and smos1-1 plants treated with increasing concentrations of nitrate, whereas taste value gradually decreased (Supplementary Fig. 15a–e). Thus, the OsGATA7-SMOS1 module is required for the production of protein and the establishment of rice taste quality, which is different from previous results upon nitrate treatment22. However, how protein and starch interact during the development of rice endosperm has not been studied in detail. Protein content, starch properties such as amylose content, and taste value were investigated in this study, and we found that taste values were significantly negatively correlated with both protein content and amylose content (Supplementary Table 2). Those findings revealed that amylose content and protein content are two critical components in rice grains that are highly correlated with the formation of taste quality. Notably, knocking out and overexpressing SMOS1 resulted in similar phenotypes for PGWC and CHD (Figs. 4b, c, 6b, c, and Supplementary Figs. 8b–d, 9a–d), whereas opposite effects on taste and protein were observed (Figs. 4d, i, j, 6g–j), suggesting that their underlying regulatory mechanisms are different. These results might indicate complicated regulation of grain components and taste quality.
Grains from japonica rice cultivars usually taste better than grains from indica rice varieties23, although the underlying reason is still unknown. Here, we investigated whether the evolution of SMOS1 contributed to the differences in taste quality among rice cultivars by examining nucleotide diversity across the SMOS1 genomic region. Indeed, we found that SMOS1 had undergone artificial selection during the domestication of rice (Fig. 5a–e). Among the four main haplotypes identified for SMOS1, SMOS1Hap1 is present mainly in japonica rice accessions with lower protein content and improved taste value compared to that of accessions harboring SMOS1Hap2, SMOS1Hap3, or SMOS1Hap4 (Fig. 5f, g). We confirmed that SMOS1Hap1 improves taste quality with lower protein content by generating isogenic lines in the XS134 and Nip backgrounds (both harboring SMOS1Hap1) with introgression of the SMOS1Hap3 allele and in the ZH210 background (harboring SMOS1Hap2) with introgression of the SMOS1Hap1 allele (Fig. 5h–m and Supplementary Fig. 13). However, we found a contradiction between the expression of Hap 2 to 4 and the phenotype of protein content appears. This phenomenon is likely due to the complex genetic background of the collected rice germplasm, and other genetic factors may be present in Hap3 rice germplasm that could cause a slight increase in protein content but not a difference in SMOS1 expression. IL-1 and IL-2 harboring the KDML105 haplotype OsGATA7 (Hap2) in the ZH210 background exhibited increased protein content but decreased taste score (Fig. 2d–f). Similarly, SMOS1 and OsGATA7 were respectively identified in an 850 kb region between markers RM3563 and RM4771 using ZH210 and KDML105, where ZH210 was used as a recurrent parent, and in a 1.4-Mb region between markers RM430 and PS18 using ZH210 and Koshihikari, where ZH210 was used as a recurrent parent (Supplementary Fig. 20). Thus, we identified elite alleles, OsGATA7Hap1 and SMOS1Hap1, that improve rice taste value.
Methods
Plant materials and growth conditions
Zhonghua11 (ZH11, Oryza sativa L. ssp. japonica ‘Zhonghua11’) and ShenNong 265 (SN265, Oryza sativa L. ssp. japonica rice accession) were used for rice transformation. All the plants were grown under natural field conditions at the China National Rice Research Institute, Hangzhou, China (30°15′N). Urea was applied at 125 or 250 kg/ha to establish medium-nitrogen (MN) and high-nitrogen (HN) treatments, respectively, with no urea applied as a low-nitrogen (LN) treatment. Seedlings were first prepared in rice seedling beds, and one day before transplanting to the field, 40% of the urea was applied, another 30% was applied at the tillering stage, and the remaining 30% was applied at panicle initiation.
Vector construction
For the generation of mutants of OsGATA7 (LOC_Os10g40810) and SMOS1 (LOC_Os05g32270), vectors were constructed using CRISPR/Cas9 Kit BGK03 (Biogle, Hangzhou, China). For SMOS1 overexpression, a cDNA template of ZH11 seedlings was used, and the full-length coding sequence of the longer SMOS1 transcript isoform amplified and driven by the cauliflower mosaic virus (CaMV) 35S promoter was cloned and inserted into the modified vector pCAMBIA139019 to obtain the SMOS1-OE#1 and #2 lines. The full-length coding sequence of the short SMOS1 transcript isoform amplified and cloned in-frame with the sequence for a FLAG tag (SMOS1-FLAG) driven by an ACTIN promoter was inserted into the modified vector pCAMBIA230019 vector to produce the SMOS1-OE#3 and #4 lines. To construct the complementation vector, the OsGATA7 and SMOS1 genomic fragments, including 2-Kb native promoter, genomic DNA, and 1.5-Kb 3′-UTR were cloned into the binary expression vector pCAMBIA1300. For subcellular localization, the full-length coding sequences of OsGATA7 and SMOS1 were individually inserted into the pAN580 vector. The fluorescence signal from GFP was detected by confocal laser scanning microscopy with 488 nm excitation and 510 nm emission (Karl Zeiss, LSM700, Germany). Primers used for plasmid construction are listed in Supplementary Data 4.
RNA extraction, RNA-seq, and RT‒qPCR analysis
For transcriptome analysis, total RNA was extracted from ZH11 and osgata7-1 grains at 10 DAF and used for RNA-seq analysis at Shanghai Majorbio Biopharm Biotechnology Co., Ltd. (Shanghai, China) according to the manufacturer’s instructions for the generation and sequencing of the sequencing libraries24 (Illumina, San Diego, CA). Differential gene expression was analyzed with DESeq2 based on log2(FC) > 1.5 and p-value < 0.05.
Total RNA was extracted from the roots, stems, leaves, sheaths, panicles, and grains at different developmental stages from ZH11 plants using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Total RNA from developing grains was isolated using a modified SDS-TRIzol method. Total RNA was reverse transcribed into first-strand cDNA with oligo (dT20) primers provided with the ReverTra Ace qPCR RT kit (Toyobo, Osaka, Japan). qPCR was then performed with SYBR Green Real-Time PCR Master Mix (Toyobo, Japan) on a Quant Studio 7 Flex Real-Time PCR System (Applied Biosystem, USA). The rice ACTIN (LOC_Os10g36650) gene was used as the reference transcript for normalization. Relative gene expression levels were calculated by the 2−ΔΔCt method25. Each assay was carried out with three biological replicates. Primers used for RT‒qPCR are listed in Supplementary Data 4.
CUT&Tag
Rice protoplasts were isolated from green leaf sheaths of the rice indica variety 9311 before being transfected with the pAN580-OsGATA7-eGFP and pAN580-SMOS1-eGFP constructs. More than 10^5 successfully transfected cells were collected, immediately fixed onto concanavalin A magnetic beads (P2156, Beyotime Institute of Biotechnology, Shanghai, China), and pretreated to bind primary and secondary antibodies and Tn5 transposase for DNA library construction26. All genes bound by OsGATA7-GFP and SMOS1-GFP were detected using PePr, a peak-calling prioritization pipeline used for ChIP-seq data.
Yeast one-hybrid assay
For the yeast one-hybrid assay, the full-length coding sequences of OsGATA7, SMOS1, and LC2 and the 1000-bp (SMOS1-P1) and 350-bp (SMOS1-P2), OsGluA2-P1, RM1-P1, and RM1-P2 promoter sequences were cloned and inserted into the pb42AD and pLacZi vectors, respectively. The plasmids were co-transformed into the EGY48 yeast strain, plated onto synthetic defined medium lacking tryptophan and uracil (SD/−Trp/−Ura), and spotted onto SD/−Trp/−Ura medium supplemented with 1 × BU salts, 2% (w/v) galactose, 1% (w/v) raffinose, and 80 mg/L X-gal for interaction tests.
Transcriptional activity assay in rice protoplasts
The full-length coding sequences of OsGATA7, SMOS1, and LC2 were subsequently cloned and inserted into an empty vector (NONE) and used as an effector construct. The 350-bp promoter sequences of SMOS1, OsGluA2, and RM1 were inserted into the p190LUC vector and used as a reporter construct. Similarly, reporters containing each of the four haplotype promoters and the SNP mutation promoter were transformed into rice protoplasts to evaluate the effect of the mutation on SMOS1 promoter activity. The Renilla LUC vector was used as an internal control. The plasmids for the effector, reporter, and internal control were co-transfected into rice protoplasts, which were subsequently incubated overnight19. The relative luciferase activity was calculated as the ratio of firefly luciferase activity to Renilla luciferase activity via a dual-luciferase reporter assay (E1910, Promega (Beijing) Biotech Co., Ltd, Beijing, China) with a TD-20/20 Luminometer (GloMax 20/20, Promega (Beijing) Biotech Co., Ltd, Beijing, China) according to the manufacturer’s instructions. Each transfection was conducted as three biological replicates. Primers used for plasmid construction are listed in Supplementary Data 4.
Electrophoretic mobility shift assay
The full-length coding sequence of OsGATA7 was cloned and inserted into the pET-28a vector. The fusion protein was expressed in Rosetta Competent cells (EC1002, Shanghai Weidi Biotechnology Co., Ltd, Shanghai, China) induced with 1 mmol isopropyl-1-thio-D-galactopyranoside at 16 °C for 12 h. Recombinant His-OsGATA7 was purified using amylose resin (P2229S, Beyotime Institute of Biotechnology, Shanghai, China). The probes containing the candidate motif identified by CUT&Tag in the SMOS1 promoter were synthesized and labeled with biotin at their 3′ end (Sunya Biological Technology). EMSAs were carried out using a chemiluminescent EMSA kit (GS009, Beyotime Institute of Biotechnology) according to the manufacturer’s instructions. The protein–oligonucleotide complex was separated on 6% (w/v) polyacrylamide gels and visualized on a ChemiDocTMMP Imaging System (Bio-Rad Laboratories, Inc, Segrate, Italy). The sequences of the synthesized probes are listed in Supplementary Data 4.
I-KI staining and characterization of grain appearance and starch granules
Mature rice grains were dehulled (Satake, Tokyo, Japan) and polished (ZM100 grain polisher, Xinfeng Company, Taizhou, China). The milled grains of the wild-type ZH11, the osgata7 and smos1 mutants, OsGATA7 and SMOS1 overexpression lines, and complementary lines were used for analysis of the PGWC and CHD according to the methods GB/T 17891-1999 and NY/T 83-2017 (MRS-9600TFU2L, MICROTEK, Changzhou, China).
Starch samples were prepared and fixed onto a circular aluminum sample stub using double-sided sticky tape and coated with gold19. A scanning electron microscope (SEM Hitachi 300) was used at 15 kV to observe the structure of the starch granules, and micrographs were taken at magnifications of 1500 × and 2500 ×. The size of the starch granules was measured using ImageJ software.
Grains of the rice cultivar 9311 and the smos1-3 mutant were cut through the center to expose a transverse section of endosperm, which was subsequently stained with 0.2% (w/v) iodine reagent (0.1 g I and 1 g KI in 50 mL sterile water) and examined under a fluorescence stereoscopic microscope (Leica, German) after 30 s of staining19.
Evaluation of physicochemical properties
The ECQ traits of AC and total starch content were measured using the Total Starch Assay Kit and Amylose Assay Kit (Megazyme, Ireland). 100 milligrams of milled rice flour powder was measured for GC. Protein content was determined using Auto Analyzer (Bran-Luebbe, Germany)27. 5 mg rice flour powder was digested with Pseudomonas amyloderamosa isoamylase (Megazyme, Ireland), and CLD was quantitatively analyzed using capillary electrophoresis (PA800 plus pharmaceutical analysis system, Beckman Coulter)19. 3 g rice flour and 25 mL distilled water were mixed in a box for rapid viscosity analysis (RVA, Newport Scientific, Sydney, NSW, Australia). 5 milligrams of flour and 10 µL of distilled water were mixed in an aluminum pan for differential scanning calorimetry (DSC, Mettler Toledo, Zurich, Switzerland). 2 milligrams of brown rice powder and 1 mL of 0 − 9 M urea solution were mixed in a 1.5 mL centrifuge tube to determine the gelatinization and swelling modes28.
Determination of taste quality
Approximately 10 g of polished rice grains was washed in water before being soaked in 12 mL of distilled water for 30 min, followed by cooking for 30 min and holding for 20 min in a steamer pot, with cooling for 20 min thereafter. About 8 g of cooked rice was prepared into rice balls and analyzed using a taste analyzer kit according to the manufacturer’s instructions (Satake, STA1B-RHS1A-RFDM1A, Japan). Cooked rice was mixed after the top layer, and the rice adhering to the walls of the pot were removed. Three cooked rice grains were placed centrally under a 25 mm cylinder probe on the platform and analyzed with a texture analyzer (Food Technology Corporation, UK). Each measurement consisted of three biological replicates. The settings for texture profile analysis were set as follows: pre-test speed, 60 mm/min; test speed, 30 mm/min; post-test speed, 60 mm/min; strain, 60%; time, 2.00 s; and trigger force, 0.15 N.
X-ray powder diffraction
The X-ray diffraction spectrum of the rice flour was detected using an X-ray powder diffraction instrument (XRD, X-ray diffractometer, Tokyo, Japan). Diffractogram profiles were collected at 40 kV and 40 mA with nickel-filtered Cu-Kα radiation (wavelength of 1.5406 nm). The scanning range of the rice starch was between 4 ° and 60 °, and the scanning rate was 4 °/θ. The XRD spectra of the rice starch were drawn according to the data collected from Jade 5.0 software (Materials Data Inc., Livermore, CA, USA).
Haplotype and evolutionary analyses
The genomic sequences of 3000 cultivated and 35 wild rice accessions were collected from the Rice SNP-Seek Database (https://snp-seek.irri.org/) and the Oryza Genome database (http://viewer.shigen.info/oryzagenome), respectively. Geographical information on the rice accession groups was acquired from the International Rice Informatics Consortium (http://iric.irri.org/resources/3000-genomes-project) and was marked on a map using R software.
Genetic analysis of OsGATA7 and SMOS1
Two segregating BC3F4 population were used to map OsGATA7 and SMOS1. Primer sequences used for genetic analysis are provided in Supplementary Data 4.
SDS − PAGE analysis
Total seed proteins were extracted with SDS-urea buffer by shaking overnight (50 °C). The proteins were separated via SDS-PAGE on a 4%-20% (w/v) gradient gel, followed by Coomassie brilliant blue staining and photographic imaging.
Statistical analysis
We used GraphPad Prism 8 and Microsoft Excel 2019 for statistical analysis. Data are shown as the means ± standard deviation (SD) from independent experiments, and P-values are indicated by two-tailed unpaired Student’s t-tests.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The sequence data used in this study can be found at the Rice Genome Annotation Project website under the following accessions: OsGATA7 (LOC_Os10g40810 [https://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os10g40810]), SMOS1 (LOC_Os05g32270 [https://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os05g32270]), LC2 (LOC_Os02g05840 [https://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os02g05840]), OsGluA2 (LOC_Os10g26060 [https://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os10g26060]), RM1 (LOC_Os07g10580 [https://rice.uga.edu/cgi-bin/ORF_infopage.cgi?orf=LOC_Os07g10580]). The RNA-Seq and CUT&Tag data have been deposited in the Genome Sequence Archive: OsGATA7 CUT&Tag [https://ngdc.cncb.ac.cn/gsa/browse/CRA023086], SMOS1 CUT&Tag [https://ngdc.cncb.ac.cn/gsa/browse/CRA023089], OsGATA7 RNA-Seq [https://ngdc.cncb.ac.cn/gsa/browse/CRA018645], and SMOS1 RNA-Seq [https://ngdc.cncb.ac.cn/gsa/browse/CRA018652]. Source data are provided within this paper. Source data are provided in this paper.
References
Ge, J., Chen, X., Zhang, X., Dai, Q. & Wei, H. Comparisons of rice taste and starch physicochemical properties in superior and inferior grains of rice with different taste value. Food Res. Int. 169, https://doi.org/10.1016/j.foodres.2023.112886 (2023).
Fitzgerald, M. A., McCouch, S. R. & Hall, R. D. Not just a grain of rice: the quest for quality. Trends Plant Sci. 14, 133–139 (2009).
Tao, K., Yu, W., Prakash, S. & Gilbert, R. G. Investigating cooked rice textural properties by instrumental measurements. Food Sci. Hum. Wellness 9, 130–135 (2020).
Balindong, J. L. et al. Rice grain protein composition influences instrumental measures of rice cooking and eating quality. J. Cereal Sci. 79, 35–42 (2018).
Guo, X. et al. Impacts of inherent components and nitrogen fertilizer on eating and cooking quality ofrice: A Review. Foods 12, https://doi.org/10.3390/foods12132495 (2023).
Liang, H. et al. Relationship between protein structure and eating quality of rice under different nitrogen application rate. Cereal Chem. 99, 692–703 (2022).
Zhang, Y. J. et al. OsGATA7 modulates brassinosteroids‐mediated growth regulation and influences architecture and grain shape. Plant Biotechnol. J. 16, 1261–1264 (2018).
Aya, K. et al. A novel AP2-type transcription factor, SMALL ORGAN SIZE1, controls organ size downstream of an auxin signaling pathway. Plant Cell Physiol. 55, 897–912 (2014).
Wu, K. et al. Enhanced sustainable green revolution yield via nitrogen-responsive chromatin modulation in rice. Science 367, https://doi.org/10.1126/science.aaz2046 (2020).
Yang, Y. et al. Natural variation of OsGluA2 is involved in grain protein content regulation in rice. Nat. Commun. 10, 1949 (2019).
Xiong, R. et al. Relation of cooked rice texture to starch structure and physicochemical properties under different nitrogen managements. Carbohydr. Polym. 295, 119882 (2022).
Peng, B. et al. OsAAP6 functions as an important regulator of grain protein content and nutritional quality in rice. Nat. Commun. 5, 4847 (2014).
Morita, R., Kusaba, M., Iida, S., Nishio, T. & Nishimura, M. Development of PCR markers to detect the glb1 and Lgc1 mutations for the production of low easy-to-digest protein rice varieties. Theor. Appl. Genet. 119, 125–130 (2009).
Kawakatsu, T., Yamamoto, M. P., Touno, S. M., Yasuda, H. & Takaiwa, F. Compensation and interaction between RISBZ1 and RPBF during grain filling in rice. Plant J. 59, 908–920 (2009).
Wu, C. Y., Washida, H., Onodera, Y., Harada, K. & Takaiwa, F. Quantitative nature of the Prolamin‐box, ACGT and AACA motifs in a rice glutelin gene promoter: minimal cis‐element requirements for endosperm‐specific gene expression. Plant J. 23, 415–421 (2001).
Sultana, S., Faruque, M. & Islam, M. R. Rice grain quality parameters and determination tools: a review on the current developments and future prospects. Int. J. Food Prop. 25, 1063–1078 (2022).
Li, H., Fitzgerald, M. A., Prakash, S., Nicholson, T. M. & Gilbert, R. G. The molecular structural features controlling stickiness in cooked rice, a major palatability determinant. Sci. Rep. 7, https://doi.org/10.1038/srep43713 (2017).
Lu, S. et al. The textural properties of cooked convenience rice upon repeated freeze–thaw treatments are largely affected by water mobility at grain level. Food Res. Int. 163, https://doi.org/10.1016/j.foodres.2022.112254 (2023).
Zhu, M. et al. The elite eating quality alleles Wxb and ALKb are regulated by OsDOF18 and coordinately improve head rice yield. Plant Biotechnol. J. 22, 1582–1595 (2024).
Lou, G., Bhat, M. A., Tan, X., Wang, Y. & He, Y. Research progress on the relationship between rice protein content and cooking and eating quality and its influencing factors. Seed Biol. 2, 16 (2023).
Chen, Y., Wang, M. & Ouwerkerk, P. B. F. Molecular and environmental factors determining grain quality in rice. Food Energy Secur. 1, 111–132 (2012).
Zhang, Y. et al. Improving rice eating and cooking quality by enhancing endogenous expression of a nitrogen‐dependent floral regulator. Plant Biotechnol. J. 21, 2654–2670 (2023).
Hori, K. et al. Genomic regions involved in differences in eating and cooking quality other than Wx and Alk genes between indica and japonica rice cultivars. Rice 14, 8 (2021).
Guo, H.-B. et al. Transcriptome analysis of genes associated with autolysis of Coprinus comatus. Sci. Rep. 12, https://doi.org/10.1038/s41598-022-06103-z (2022).
Schmittgen, T. D. & Livak, K. J. Analyzing real-time PCR data by the comparative CT method. Nat. Protoc. 3, 1101–1108 (2008).
Wu, L. et al. A cost‐effective tsCUT&Tag method for profiling transcription factor binding landscape. J. Integr. Plant Biol. 64, 2033–2038 (2022).
Kang, H. G., Park, S., Matsuoka, M. & An, G. White‐core endosperm floury endosperm‐4 in rice is generated by knockout mutations in the C4‐type pyruvate orthophosphate dikinase gene (OsPPDKB). Plant J. 42, 901–911 (2005).
Wang, W. et al. GBSS‐BINDING PROTEIN, encoding a CBM48 domain‐containing protein, affects rice quality and yield. J. Integr. Plant Biol. 62, 948–966 (2019).
Acknowledgements
We thank Prof. Xiangdong Fu for providing the smos1-3 mutant in the 9311 background. This study was supported by the Innovation Program of the Chinese Academy of Agricultural Sciences (No. CAAS-CSCB-202402) and the Central Public-interest Scientific Institution Basal Research Fund (No. CPSIBRF-CNRRI-202403), Zhejiang Provincial Natural Science Foundation of China (LDQ24C130001), National Natural Science Foundation of China (32188102), the Agricultural Science and Technology Innovation Program (CAAS-ZDRW202401), the Project of Laboratory of Advanced Agricultural Sciences, Heilongjiang Province (ZY04JD05-005), and Yunnan Provincial Science and Technology Project (202402AE090036, 202405AF140064).
Author information
Authors and Affiliations
Contributions
S.T., G.S., and P.H. conceived this project and designed all experiments. N.C. and W.Z. performed experiments. N.C., W.Z., F.Z., A. L., M.Z., R.Z., J.W., X. Y., Y. Lv., S.H., Z.S., and X.W. analyzed the data. G.J., L.X., Y.L., and J.Y. interpreted the data. S.T., G.S., and P.H. wrote the paper.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Kiyosumi Hori, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Cao, N., Zhou, W., Zhao, F. et al. OsGATA7 and SMOS1 cooperatively determine rice taste quality by repressing OsGluA2 expression and protein biosynthesis. Nat Commun 16, 3513 (2025). https://doi.org/10.1038/s41467-025-58823-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-58823-1