The fgw1 mutant produces broad grains by affecting cell proliferation and expansion
To understand the genetic and molecular mechanisms that regulate rice grain size and quality, we initiated a genetic screen for mutants with altered grain appearance. The fgw1 mutant was isolated from the M2 generation of the rice cultivar ‘Jinhui10’ subjected to mutagenesis with ethyl methanesulfonate (EMS), and it also showed semi-dwarfism, reduced seed setting rate, fewer tillers, less branches and shorten root with slight curled leaves (Supplemental Fig. S1, table S1). In addition, the fgw1 displayed deep green leaves with higher chlorophyll content, containing more number of mesophyll cells in fgw1 compared with those in wild type, which caused the significant increase in net photosynthetic rate (Pn), stomatal conductance (Gs) and transpiration rate (Tr) (Supplemental Fig. S2).
The grain width of the fgw1 mutant was increased by 28.38% compared with that of the wild type, whereas the grain length of the fgw1 mutant was slightly decreased, but not significantly, compared with the wild type. These changes contributed to an increase in 1000-grain weight in comparison with the wild type (Fig. 1A–D). Anatomical observation of the glumes revealed that the outer parenchyma cell layer of the fgw1 mutant was longer (by 10.39%), contained a substantially higher number of cells (1.83% more), and the cell length was increased (by 8.40%) compared with the wild type (Fig. 1E–H). We also measured the width and length of cells in the outer and inner epidermal tissues of the lemma of the wild type and fgw1 mutant using scanning electron microscopy (Fig. 1I). Compared with those of the wild type, in the fgw1 mutant the number of outer epidermal cells in the transverse and longitudinal planes increased by 25.59% and 8.79%, respectively (Fig. 1J, K). The length of the inner epidermal cells was significantly (P < 0.001) decreased by 12.20%, whereas the cell width increased by 23.80%, compared with the wild type (Fig. 1L, M). Similar results were obtained for the cell width of the outer epidermis (Supplemental Fig. S1). Consistent with these results, a number of genes associated with cell division and expansion were upregulated to varying degrees in the panicle of the fgw1 mutant (Supplemental Fig. S3A). All of these suggested that FGW1 regulated the development of both cell proliferation and expansion.
FGW1 regulates opaque and white endosperm through its effects on grain filling
The fgw1 mutant produced a larger hull and the grains were wrinkled and opaque at maturity, which suggests that FGW1 may be associated with incomplete grain filling. Therefore, the grain milk filling rates in the wild type and fgw1 mutant were studied in detail. Prior to 6 days after fertilization, the endosperm fresh weight and dry weight of the wild type were higher than those of the fgw1 mutant. The endosperm fresh weight and dry weight of the fgw1 mutant were significantly higher than those of the wild type from 9 days after fertilization, with the differences peaking 21 days after fertilization (Fig. 2A–C). The endosperm fresh weight and dry weight of the fgw1 mutant was increased by 8.96% and 9.81%, respectively, on day 21 after fertilization compared with the wild type. These results indicate that the larger endosperm and heavier grain of the fgw1 mutant resulted from accelerated accumulation of dry matter, which suggests that FGW1 may negatively regulate the rate of dry matter accumulation.
To further clarify the altered grain quality of the fgw1 mutant, transverse sections of mature grains from the wild type and mutant were compared to investigate differences in endosperm content. The endosperm of the fgw1 mutant was opaque and uniformly white (Fig. 2D, E). Scanning electron micrographs of the transverse sections indicated that the starch grains of the wild-type endosperm cells were polyhedral and densely packed (Fig. 2F), whereas the fgw1 mutant showed markedly more grain chalkiness as a result of abnormally developed and loosely packed starch granules (Fig. 2G). To analyse the sugar content in grains of the wild type and fgw1 mutant, we measured the starch, glucose, and sucrose contents at the early grain-filling stage. The starch content was dramatically lower, whereas the glucose and sucrose contents were higher, in grains of the fgw1 mutant, despite the presence of defective starch grains (Fig. 2H–J). qPCR showed that some genes regulating both grain size and quality had a significantly changing in transcript level, such as D11 and GL7. D11 is a BR-synthesis gene regulating sugar accumulation and seed size (Zhu et al. 2015), which expression increased by 3.7 times in fgw1 compared with those in wild type, the other gene GL7 (Wang et al. 2015a, b), regulating both the seed size and grain quality, increased the expression by 2.4 times (Supplemental Fig. S3B).
FGW1 encodes a protein with DUF630 and DUF632 domains
For gene mapping of FGW1, the sterile line ‘Xida 1A’ and the original parent Jinhui10 was crossed with the fgw1 mutant, respectively. All F1 plants displayed the normal phenotype and the F2 plants showed a phenotype segregation ratio of 3:1 (normal:broad grains), which indicated that the fgw1 was a recessive mutation. Therefore, the mutational plants with bigger seeds and fewer tillers were selected from the F2 group of ‘Xida 1A’ X fgw1 and used to map FGW1.
The candidate genomic region of fgw1 was narrowed to a 123.7 kb interval between the markers Ind 10-2 and Ind 10-3 on the long arm of chromosome 10 (Fig. 3A). Twenty annotated genes were located within the candidate region in the Gramene database (http://www.gramene.org/). To define the molecular basis of the fgw1 mutant, genomic DNA sequences from the fgw1 mutant and wild type were amplified by PCR and sequenced. The only mutation detected was in LOC_Os10g41310 in fgw1, where a single-nucleotide deletion caused a frame shift that led to premature termination of the predicted protein (Fig. 3A). We identified LOC_Os10g41310 as the candidate gene for the fgw1 mutation. The identity of the fgw1 gene was further confirmed by genetic complementation analysis. A 6157-bp wild-type genomic fragment of LOC_Os10g41310, including a 1828-bp sequence upstream of the start codon and a 733-bp sequence downstream of the terminal codon, was transformed into the fgw1 mutant. The mutant phenotype was completely rescued in transgenic plants (Fig. 3B–E; Supplemental Fig. S4). Additionally, FGW1 has a relatively higher expression at grains developing 9 days (Fig. 3F). These results confirm that LOC_Os10g41310 corresponded to the FGW1 gene. LOC_Os10g41310 contains a DUF630 domain at the N-terminus and a DUF632 domain at the C-terminus, such type of proteins wildly existing in plant (Supplemental Fig. S5), and has been reported to regulate leaf rolling (REL2) and development of fewer tillers (DLT10), probably by monitoring auxin homeostasis (Yang et al. 2016; Wen et al. 2020). qPCR analysis demonstrated that the auxin transport (PINs) related genes also altered the transcript level in fgw1 panicle, such as PIN1a and PIN1b, which expression increased to the significant different level compared with those of wild type (Supplemental Fig. S6). These results further suggest that FGW1 is a novel allelic gene of REL2 and DLT10.
FGW1 negatively regulates grain width and positively regulates grain quality
To verify whether the expression of FGW1 is responsible for grain development in rice, we over-expressed FGW1 cDNA under the control of the ubiquitin promoter in the wild type. The FGW1-overexpressing plants (OE) produced a slightly narrower grain, and plants that showed a higher expression level of FGW1 produced a narrower grain as well as a decreased 1000-grain weight (Fig. 4A–E). Suppression of FGW1 by RNA interference (RNAi) led to clearly opposite phenotypes, including broader grains and increased 1000-grain weight (Fig. 4F-J). Interestingly, the phenotype severity was consistent with the FGW1 expression level in transgenic plants (Fig. 4; Supplemental Fig. S7), which suggested that the expression level of FGW1 was also responsible for the rice grain appearance and that FGW1 was a negative regulator of grain size formation and a positive regulator of grain quality.
Phylogenetic analysis indicated that FGW1 was highly conserved and showed a relatively close genetic relationship with homologous genes from other graminaceous species, such as Zea mays, Sorghum bicolor, and Brachypodium distachyon, and was not closely related to the homolog from Arabidopsis thaliana (Supplemental Fig. S5; Yang et al. 2016). Given the evolutionary conservation of DUF domains, we hypothesized that FGW1 may have a conserved function in plants. Therefore, over-expressed FGW1 cDNA under the control of the ubiquitin promoter was transformed into Arabidopsis thaliana. All 12 transgenic individuals produced narrower seeds compared with those of the wild type (Supplemental Fig. S8), which supports the view that FGW1 has a conserved function in plants.
FGW1 is preferentially expressed in the young panicle and involved in development of vascular tissues
To confirm the expression pattern of FGW1, the relative transcript levels were quantified by real-time quantitative PCR (qRT-PCR) in various organs during vegetative growth and reproductive development. The qRT-PCR analysis showed that FGW1 was expressed in various tissues, but the transcript level was higher in the culm and young panicle, and gradually decreased as the panicle matured (Fig. 5A). The FGW1 promoter activity in various tissues of PROFGW1:GUS transgenic plants was also checked. In agreement with the qRT-PCR results, β-glucuronidase (GUS) staining revealed that FGW1 was expressed ubiquitously in all examined rice tissues, with the highest expression level observed in the culm and developing young panicle (Fig. 4B–H). Cross-sectioning of the GUS-stained hull and culm further showed that the GUS signal was mainly restricted to vasculature regions (Fig. 4I–L). RNA in situ hybridization of developing panicles demonstrated the preferential expression of FGW1 in the primary and secondary branch meristem, floret meristem, and hull (Fig. 5M–O). The predominant expression of FGW1 in different organs implies that it may play a role in the control of grain size and development of vascular tissues.
FGW1 interacted with the 14-3-3 protein GF14f
To examine the subcellular localization of FGW1, the full-length coding sequence of FGW1 was fused to the N-terminus of the green fluorescent protein (GFP) and transformed into rice protoplasts. The FGW1-GFP fusion protein was exclusively co-localized with the Tracker Red plasma membrane marker (Fig. 6A). Similarly, the FGW1-GFP fusion protein was exclusively localized to the plasma membrane in root cells of transgenic rice plants (Supplemental Fig. S9). These results demonstrate that FGW1 was localized to the plasma membrane.
To understand the molecular mechanisms by which FGW1 affected rice grain development, a yeast two-hybrid assay, based on the DUAL membrane system, was utilized to screen the interacted proteins. We produced DNA constructs of FGW1 and pBT-SUC, and then co-transformed the NYM51 strain with pOst1-Nubl (positive control) and pPR3-N (negative control). The pBT-SUC plasmid showed no self-activation and was suitable for screening the membrane-based yeast two-hybrid system. The 14-3-3 protein GF14f was screened from the library of the yeast DUAL membrane system, which was previously reported to affect grain filling negatively in rice (Zhang et al. 2019). Therefore, a point-to-point verification of pBT-SUC-FGW1 and pPR3-N-GF14f was performed. The yeast freckle exhibited normal growth on SD/−Leu/−Trp/−His/−Ade medium for FGW1/GF14f, but no growth for DUF630/GF14f, DUF632/GF14f, and Bzip/GF14f. These results suggest that the unabridged protein of FGW1 is essential for the direct interaction of FGW1 and GF14f in vitro (Fig. 6B). Results of a biomolecular fluorescence complementation (BiFC) assay further confirmed this interaction in planta and showed that the location of the interaction was specific to the membrane but not in the nucleus in vivo (Fig. 6C, D), although the GF14f was localized at both cytoplasm and nucleus (Fig. 6E). Therefore, we hypothesized that GF14f may be precluded from interacting with the mutated FGW1 protein on the membrane.
Transcriptome analysis indicates that FGW1 regulates cellular processes and carbohydrate metabolism
To further explore the function of FGW1, we performed RNA sequencing (RNA-seq) analysis with panicles from wild-type and fgw1 plants. A total of 989 differentially expressed genes were detected, of which 72.8% (720 genes) were upregulated and 27.2% (269 genes) were downregulated in fgw1 plants compared with their expression in wild-type plants (Supplemental Table S6). Gene ontology enrichment analysis and Clusters of Orthologous Groups function classification showed that the expression of genes involved in carbohydrate transport and metabolism, cell division, and cell expansion were changed dramatically (Fig. 7A, B). We compared the relative transcript levels of six cell division- and expansion-related genes by qPCR: GW5 (BGIOSGA019303), GS5 (BGIOSGA018771), CycA3;1 (BGIOSGA013106), CycB1;5 (BGIOSGA020095), CYC U4;1 (BGIOSGA033453), and FSM (BGIOSGA004976). The transcript level of the six genes increased to varying degrees in fgw1 panicles, which was consistent with the results of RNA-seq analysis (Fig. S10A; Supplemental Table S6). Furthermore, the expression of five carbohydrate metabolism-related genes – OsRBCS1 (BGIOSGA007558), OsRBCS2 (BGIOSGA018498), OsRBCS3 (BGIOSGA037257), OsRBCS4 (BGIOSGA037260), and SUS1 (BGIOSGA010570) – also increased to varying degrees in fgw1 panicles compared with panicles from wild-type plants (Fig. S10B; Supplemental Table S6).
Inhibited the expression of GF14f could increase the activity of AGPase, StSase and SuSase, and produced a bigger seed (Zhang et al. 2019). Accordingly, the expression of GF14f decreased by 55%, and the activity of AGPase, StSase and SuSase increased by 37.33%, 17.8% and 44.84%, respectively, in fgw1 compared with those in wild type, all of which reached extremely significant difference in statistics (Fig. 7C-F). Collectively, these results support the conclusion that FGW1 functions as a regulator of cell division, cell expansion, and carbohydrate metabolism, at least partially, by controlling the 14-3-3 protein GF14f, which has been reported to regulate grain development and filling by interacting with enzymes involved in sucrose breakdown, starch synthesis, and glycolysis (Zhang et al. 2019).