Fine Mapping of Five Grain Size QTLs Which Affect Grain Yield and Quality in Rice

Grain size is a quantitative trait with a complex genetic mechanism, characterized by the combination of grain length (GL), grain width (GW), length to width ration (LWR), and grain thickness (GT). In this study, we conducted quantitative trait loci (QTL) analysis to investigate the genetic basis of grain size using BC1F2 and BC1F2:3 populations derived from two indica lines, Guangzhan 63-4S (GZ63-4S) and TGMS29 (core germplasm number W240). A total of twenty-four QTLs for grain size were identified, among which, three QTLs (qGW1, qGW7, and qGW12) controlling GL and two QTLs (qGW5 and qGL9) controlling GW were validated and subsequently fine mapped to regions ranging from 128 kb to 624 kb. Scanning electron microscopic (SEM) analysis and expression analysis revealed that qGW7 influences cell expansion, while qGL9 affects cell division. Conversely, qGW1, qGW5, and qGW12 promoted both cell division and expansion. Furthermore, negative correlations were observed between grain yield and quality for both qGW7 and qGW12. Nevertheless, qGW5 exhibited the potential to enhance quality without compromising yield. Importantly, we identified two promising QTLs, qGW1 and qGL9, which simultaneously improved both grain yield and quality. In summary, our results laid the foundation for cloning these five QTLs and provided valuable resources for breeding rice varieties with high yield and superior quality.


Introduction
Rice (Oryza sativa L.) is one of the most important food crops worldwide, providing staple food for more than half of the world's population [1].Rice yield is largely determined by three major components: grain weight, number of grains per panicle, and number of effective tillers per plant [2].Among these, grain weight exhibits a strong correlation with grain size [3].Grain size is a crucial agronomic trait that has a major impact on the market values of rice grain produce [4,5].
Previous studies have uncovered several signaling pathways controlling grain size, including the guanine nucleotide-binding protein (G protein) signal pathway, the ubiquitinproteasome pathway, the mitogen-activated protein kinase (MAPK) signaling cascade, the transcriptional regulators pathway, and the phytohormone signaling pathway [3,[6][7][8].G proteins have been demonstrated to regulate grain size in rice.GS3 is the first identified major QTL that negatively controls grain length and weight [9,10].The GS3-3 allele has a longer grain and heavier weight than the GS3-4 allele [11,12].GS3 reduces grain length by interacting with RGB1 [13].Recently, the combination of OsMADS1 lgy3 , dep1, and gs3 simultaneously increased both grain yield and quality [14].Furthermore, the ubiquitinproteasome pathway could directly or indirectly regulate grain size.GW2 encodes a RING-type E3 ubiquitin ligase [15] and negatively regulates grain width, weight, and yield [16,17]. gw2.1, a new allele of GW2, affects grain size by causing cell proliferation, which could be used to improve grain yield and appearance in hybrid breeding [18].WTG1 determines grain size by influencing cell expansion [19].The combination of npt1 and dep1-1 has the potential to increase grain yield in rice [20].In addition, the conserved module of the MAPK signaling cascade plays a key role in the regulation of grain size.The OsMKKK10-OsMKK4-OsMAPK6 cascade modulates grain size by promoting cell division [21][22][23][24][25]. OsWRKY53 can be phosphorylated by OsMAPK6, indicating a cascade of OsMKKK10, OsMKK4, OsMAPK6, and OsWRKY53 in grain size regulation [26].Moreover, the QTL gene GLW7 could increase grain length, thickness, and weight by promoting cell expansion in the spikelet hull [27].Similarly, GW8 [28,29], GS2 [30,31], GS9 [32], and GL6 [33,34] are all transcriptional regulators that regulate grain size.Additionally, many phytohormones play significant roles in determining grain size.GW5 was identified as a major QTL for grain width and weight [35].The decreased expression of GW5 causes wide and heavy grains, resulting in improved yield [36]. OsPPKL1 is encoded by a major QTL for grain length, GL3.1 [37][38][39].GL3.1 regulates grain size through interaction with OsGSK3 [40].The qgl3 allele could increase grain yield without affecting grain quality [39].The TGW6 loss-of-function allele enhances grain weight and yield without affecting grain quality [41].qTGW3 negatively regulates grain size and weight [42][43][44].A loss-of-function mutation in qTGW3 results in large and heavy grains, suggesting this locus has potential in breeding high-yield rice [44].
So far, numerous grain size QTLs have been identified, among which a few major QTLs have been cloned.However, the genetic basis of grain size is still not well elucidated, and mining novel QTLs for grain size is of great importance to gain a better understanding of regulation mechanisms and provide gene resources for breeding applications.In this study, we mapped QTLs for grain size using BC 1 F 2 and BC 1 F 2:3 populations derived from the cross between GZ63-4S and W240.A total of twenty-four QTLs were identified, among which five major QTLs were confirmed and further fine mapped.Scanning electron microscopic analysis and expression analysis revealed the cytological basis underlying the five major QTLs on grain size.What is more, an investigation of yield-and quality-related traits demonstrated the potential of these QTLs for breeding rice with high yield and superior quality.

Phenotypic Variation and Correlation in the BC 1 F 2 and BC 1 F 2:3 Populations
The BC 1 F 2 and BC 1 F 2:3 populations, as well as the two parents, exhibited great phenotypic variation in grain size.Compared with GZ63-4S, W240 exhibited greater grain size (Table S1).In two years of repeated field trials, GL ranged from 8.64 mm to 10.27 mm in the BC 1 F 2 population and from 8.58 mm to 10.23 mm in the BC 1 F 2:3 population, respectively (Figure 1A).The ranges of GW were 2.37-2.98mm and 2.43-2.92mm (Figure 1B), LWR's were 2.96-3.92and 3.11-4.04(Figure 1C), and GT's were 1.92-2.19mm and 1.94-2.22mm (Figure 1D).Notably, all these traits exhibited normal distribution in both years, indicating typical patterns of quantitative variation.The results suggested that both BC 1 F 2 and BC 1 F 2:3 populations met the requirement for QTL mapping.
The correlation coefficients among the four grain traits in two years showed that both GL and GW exhibited weak to moderate correlations with LWR and GT in 2016 or 2017.LWR displayed no and weak correlation with GT in 2016 and 2017, respectively (Figure S1A,B).Moreover, GL16 and GL17, GW16 and GW17, LWR16 and LWR17, and GT16 and GT17 all displayed moderate correlation (Figure S1C-F).

QTL Mapping and Verification
To identify QTLs associated with GL, GW, GT, and TGW, 177 polymorphic markers were used to screen 327 plants from the BC1F2 population.We generated a BC1F2:3 population by propagating single plants from the BC1F2 populations for an additional generation.The phenotypic evaluation was conducted using mixed single plants from each family in the BC1F2:3 population, with the average phenotype serving as the phenotype for each individual.Subsequent QTL analysis revealed the presence of twenty-four QTLs, including eleven for GL, five for GW, six for LWR, and two for GT (Figure 2, Table 1).
Based on the above QTL mapping, qGW1, qGW5, qGW7, qGL9, and qGW12 were identified as significant contributors to grain size (Table 1).Five BC1F6 and BC1F7 populations developed from the BC1F2:3 line that were heterozygous in the target QTL region and homozygous for most of the other grain size QTLs were chosen to precisely evaluate their effects on grain size, respectively.

QTL Mapping and Verification
To identify QTLs associated with GL, GW, GT, and TGW, 177 polymorphic markers were used to screen 327 plants from the BC 1 F 2 population.We generated a BC 1 F 2:3 population by propagating single plants from the BC 1 F 2 populations for an additional generation.The phenotypic evaluation was conducted using mixed single plants from each family in the BC 1 F 2:3 population, with the average phenotype serving as the phenotype for each individual.Subsequent QTL analysis revealed the presence of twenty-four QTLs, including eleven for GL, five for GW, six for LWR, and two for GT (Figure 2, Table 1).
Based on the above QTL mapping, qGW1, qGW5, qGW7, qGL9, and qGW12 were identified as significant contributors to grain size (Table 1).Five BC 1 F 6 and BC 1 F 7 populations developed from the BC 1 F 2:3 line that were heterozygous in the target QTL region and homozygous for most of the other grain size QTLs were chosen to precisely evaluate their effects on grain size, respectively.

The Genetic Effect of qGW1
The grain width showed significant differences between NIL with homologous W240 (NIL-W) and NIL with homologous GZ63-4S (NIL-G) for qGW1 (Figures 3A,B and S2).To fine map qGW1, we developed a BC 1 F 8 population consisting of 2994 plants and subsequently identified 478 recombinants using markers RM128 and RM319.Eventually, we localized qGW1 to a high-resolution linkage map by progeny testing 36 recombinants and narrowed the qGW1 locus to a 231 kb region between markers R1344 and R1346 (Figure 3C).Grain size is restricted by the spikelet hull, which is determined by cell division and Grain size is restricted by the spikelet hull, which is determined by cell division and expansion [45].We conducted scanning electron microscopic (SEM) analysis to uncover the cytological basis underlying the regulation of grain size in qGW1.The values of cell width and the number of transverse cells were higher in NIL-W qGW1 than in NIL-G qGW1 (Figure 3D-I).Furthermore, we examined the expression of cell cycle and cell expansion genes in young panicles of NILs using qRT-PCR.Our findings indicated that four cell cycle related-genes (E2F2, CDKA1, CYCA3;1, and CYCT1), as well as four cell expansion related-genes (EXPA10, EXPA5, EXPA6, and EXPA7), were up-regulated in NIL-W qGW1 , suggesting that qGW1 regulates grain width by promoting both cell division and cell expansion (Figures 3J and 9).

The Genetic Effect of qGW5
Compared with NIL-W qGW5 , the values of grain length were significantly higher in NIL-G qGW5 (Figures 4A,B and S3).A progeny test of homozygous segregates further narrowed down the qGW5 locus to a 128 kb region between R5101 and R51142 (Figure 4C).

The Genetic Effect of qGW7
The significant variations in grain width were observed among lines carrying different genotypes of qGW7 (Figures 5A,B and S4).The qGW7 locus was ultimately mapped to a region of 444 kb between R7277 and R7281 using 1680 plants from the BC1F9 population (Figure 5C).
The NIL-W qGW7 spikelet hull had a wider cell size than NIL-G qGW7 , while no difference was observed in cell number (Figure 5D-I).We further observed an up-regulation in the expression of six cell expansion related-genes (EXPA10, EXPA3, EXPA5, EXPA6, EXPA7 and EXPB7) in NIL-W qGW7 , indicating that qGW7 regulates grain width through the alteration of cell expansion (Figures 5J and 9).NIL-G qGW5 displayed significantly larger cell length and a higher number of longitudinal cells than that of NIL-W qGW5 (Figure 4D-I).Expression analysis revealed that five genes related to the cell cycle (E2F2, CDKA2, CYCA3;1, MAPK, and MCM3) and four genes related to the cell expansion (EXPA10, EXPA3, EXPA4, and EXPB7) showed higher expression levels in NIL-G qGW5 .These findings revealed that qGW5 regulates grain length by influencing both cell division and expansion in the spikelet hull (Figures 4J and 9).

The Genetic Effect of qGW7
The significant variations in grain width were observed among lines carrying different genotypes of qGW7 (Figures 5A,B and S4).The qGW7 locus was ultimately mapped to a region of 444 kb between R7277 and R7281 using 1680 plants from the BC 1 F 9 population (Figure 5C).

The Genetic Effect of qGL9
NIL-G qGL9 had significantly larger grain length than NIL-W qGL9 (Figures 6A,B and  S5).Subsequently, we mapped the qGL9 to a 335 kb region between markers R9194 and R9197 in the BC1F9 generation, which included a total of 1584 plants (Figure 6C).
The number of longitudinal cells in the spikelet hull was higher in NIL-G qGL9 than in NIL-W qGL9 .However, there was no difference in cell length between these two NILs (Figure 6D-I).Notably, NIL-G qGL9 exhibited up-regulated expression levels of fifteen genes related to cell cycle (CDC20, CDKA2, CDKB, CYCA2.1,CYCA3;2, CYCB1;1, CYCB2.2,CYCD1;1, CYCD4, CYCD6, H1, KN, MAPK, MCM2, and MCM4), resulting in increased cell division within the spikelet and ultimately leading to an increase in grain length (Figures 6J and 9).The NIL-W qGW7 spikelet hull had a wider cell size than NIL-G qGW7 , while no difference was observed in cell number (Figure 5D-I).We further observed an up-regulation in the expression of six cell expansion related-genes (EXPA10, EXPA3, EXPA5, EXPA6, EXPA7 and EXPB7) in NIL-W qGW7 , indicating that qGW7 regulates grain width through the alteration of cell expansion (Figures 5J and 9).

The Genetic Effect of qGL9
NIL-G qGL9 had significantly larger grain length than NIL-W qGL9 (Figures 6A,B and S5).Subsequently, we mapped the qGL9 to a 335 kb region between markers R9194 and R9197 in the BC 1 F 9 generation, which included a total of 1584 plants (Figure 6C).

The Genetic Effect of qGW12
We constructed NIL-G qGW12 and confirmed that this allele could significantly increase grain width (Figures 7A,B and S6).Subsequently, qGW12 was mapped within a 624 kb interval flanked by markers R12246 and R12252 (Figure 7C).
Analysis of the outer glume found that NIL-G qGW12 exhibited increased cell number and larger cell size in the grain-width direction, causing wider grain (Figure 7D-I).In addition, the expression levels of six genes related to cell cycle (CDKA1, CYCB1;1, CYCD1;1, CYCD3, CYCIaZm, and KN) and three genes related to cell expansion (EXPA6, EXPB4, and EXPB7) were higher in NIL-G qGW12 than in NIL-W qGW12 .Thus, the up-regulation of both cell division and expansion genes was responsible for the increase in the grain width of NIL-G qGW12 (Figures 7J and 9).The number of longitudinal cells in the spikelet hull was higher in NIL-G qGL9 than in NIL-W qGL9 .However, there was no difference in cell length between these two NILs (Figure 6D-I).Notably, NIL-G qGL9 exhibited up-regulated expression levels of fifteen genes related to cell cycle (CDC20, CDKA2, CDKB, CYCA2.1,CYCA3;2, CYCB1;1, CYCB2.2,CYCD1;1, CYCD4, CYCD6, H1, KN, MAPK, MCM2, and MCM4), resulting in increased cell division within the spikelet and ultimately leading to an increase in grain length (Figures 6J and 9).

The Genetic Effect of qGW12
We constructed NIL-G qGW12 and confirmed that this allele could significantly increase grain width (Figures 7A,B and S6).Subsequently, qGW12 was mapped within a 624 kb interval flanked by markers R12246 and R12252 (Figure 7C).
Analysis of the outer glume found that NIL-G qGW12 exhibited increased cell number and larger cell size in the grain-width direction, causing wider grain (Figure 7D-I).In addition, the expression levels of six genes related to cell cycle (CDKA1, CYCB1;1, CYCD1;1, CYCD3, CYCIaZm, and KN) and three genes related to cell expansion (EXPA6, EXPB4, and EXPB7) were higher in NIL-G qGW12 than in NIL-W qGW12 .Thus, the up-regulation of both cell division and expansion genes was responsible for the increase in the grain width of NIL-G qGW12 (Figures 7J and 9).

Investigation of Traits Related to Rice Yield and Quality in Five NILs
Finally, we performed phenotypic comparisons of yield-and quality-related traits among five NILs.The grain length of NIL-G qGW1 , NIL-W qGW5 , and NIL-W qGL9 exhibited significant reductions than their respective NILs, while no differences were observed in other NILs.Phenotypic variations were detected among the five NILs for grain width and length to width ration.Specifically, larger variations in 1000-grain weight and the number of tillers per plant were found in NIL-W qGW1 and NIL-W qGL9 , whereas higher plant height and the number of filled grains per panicle were observed in NIL-G qGW7 and NIL-W qGW12 .Grain yield per plant was significantly higher in NIL-G qGW1 , NIL-G qGW7 , NIL-G qGL9 , and NIL-W qGW12 than that of their respective NILs.However, there were no differences in grain yield per plant between NIL-W qGW5 and NIL-G qGW5 (Figure 8A-H).These results suggested that the increased yield of qGW1 and qGL9 were primarily attributed to enhanced number of tillers per plant, while the increased yield of qGW7 and qGW12 were mainly due to enhanced numbers of filled grains per panicle.

Investigation of Traits Related to Rice Yield and Quality in Five NILs
Finally, we performed phenotypic comparisons of yield-and quality-related traits among five NILs.The grain length of NIL-G qGW1 , NIL-W qGW5 , and NIL-W qGL9 exhibited significant reductions than their respective NILs, while no differences were observed in other NILs.Phenotypic variations were detected among the five NILs for grain width and length to width ration.Specifically, larger variations in 1000-grain weight and the number of tillers per plant were found in NIL-W qGW1 and NIL-W qGL9 , whereas higher plant height and the number of filled grains per panicle were observed in NIL-G qGW7 and NIL-W qGW12 .Grain yield per plant was significantly higher in NIL-G qGW1 , NIL-G qGW7 , NIL-G qGL9 , and NIL-W qGW12 than that of their respective NILs.However, there were no differences in grain yield per plant between NIL-W qGW5 and NIL-G qGW5 (Figure 8A-H).These results suggested that the increased yield of qGW1 and qGL9 were primarily attributed to enhanced number of tillers per plant, while the increased yield of qGW7 and qGW12 were mainly due to enhanced numbers of filled grains per panicle.There were no differences in albumin content among the five NILs.In contrast, three, four and five NILs exhibited remarkable variations in globulin, prolamin, and glutenin content, respectively.Compared with NIL-W qGW5 , NIL-G qGW5 displayed higher total starch and amylose content, whereas no differences were observed in the other four NILs.NIL-W qGW5 and NIL-G qGL9 exhibited significant increases in gel consistency, while NIL-G qGW1 , NIL-G qGW5 , NIL-W qGW7 , NIL-G qGL9 , and NIL-G qGW12 showed huge improvements in taste score (Figure 8I-P).These findings indicated that the improved cooking and eating quality of qGW1 and qGL9 results from reduced globulin and glutenin content, whereas the improved quality of qGW7 and qGW12 derives from decreased prolamin and glutenin content.In addition, both decreased globulin and glutenin content and increased total starch content, amylose content, and gel consistency lead to an increase in the cooking and eating quality of qGW5.Usually, rice with lower protein content but There were no differences in albumin content among the five NILs.In contrast, three, four and five NILs exhibited remarkable variations in globulin, prolamin, and glutenin content, respectively.Compared with NIL-W qGW5 , NIL-G qGW5 displayed higher total starch and amylose content, whereas no differences were observed in the other four NILs.NIL-W qGW5 and NIL-G qGL9 exhibited significant increases in gel consistency, while NIL-G qGW1 , NIL-G qGW5 , NIL-W qGW7 , NIL-G qGL9 , and NIL-G qGW12 showed huge improvements in taste score (Figure 8I-P).These findings indicated that the improved cooking and eating quality of qGW1 and qGL9 results from reduced globulin and glutenin content, whereas the improved quality of qGW7 and qGW12 derives from decreased prolamin and glutenin content.In addition, both decreased globulin and glutenin content and increased total starch content, amylose content, and gel consistency lead to an increase in the cooking and eating quality of qGW5.Usually, rice with lower protein content but higher amylose content and gel consistency exhibits superior cooking and eating quality [46].
In summary, negative correlations were observed between yield and rice quality for both qGW7 and qGW12.The impact of qGW5 on yield was negligible.Conversely, it exerted significant influences on quality-related traits, suggesting that this locus could enhance rice quality without compromising yield.Meanwhile, we provide two promising QTLs, qGW1 and qGL9, for breeding rice with a high yield and superior quality.
As a primary isolated mapping population, the BC 1 F 2 population offers the advantages of being relatively simple to construct, requiring less time, and providing rich genetic information.We generated a BC 1 F 2:3 population by propagating single plants from the BC 1 F 2 populations for an additional generation to address issues of offspring segregation and reproductive limitations caused by the BC 1 F 2 population.Until now, many QTLs or genes have been reported from F 2 and F 2:3 populations, such as Rpsan 1, qPH9, FM1, Co-1HY, YrZ15-1949, qHBV4.2,qHBV6.1,qHBV11.1, and qHBV11.2[47][48][49][50][51][52].To identify novel QTLs controlling grain size, we developed BC 1 F 2 and BC 1 F 2:3 populations derived from W240 and GZ63-4S.A total of twenty-four QTLs for grain size were detected in this study (Figure 2, Table 1).By comparing chromosome positions and molecular markers, it was found that OsCKX1 exerts a negative regulatory effect on grain size within the same region as qGL1 [53].GW2, which negatively controls grain width, was identified at the qGL2.2locus [15].Zhao et al. [54] detected two QTLs (qGL2.1 and qGW4) in regions similar to those of qGL2.1 and qGW4.Additionally, two other QTLs for grain length (qGL8.1 and qGL5) were, respectively, located in the vicinity of QTLs detected in previous studies [55,56].
Notably, we discovered five novel QTLs, qGW1, qGW5, qGW7, qGL9, and qGW12, which made significant contributions to grain size.Taken together, these results proved the reliability of using BC 1 F 2 and BC 1 F 2:3 populations for rice QTL mapping.

Validation and Fine Mapping of the Five QTLs
In previous studies, many genes controlling grain size have been cloned using the map-based cloning approach, such as GW2 [15], GS2 [30,31], and GL3.1 [37][38][39].Therefore, QTL mapping, validation, and fine mapping are essential steps in the process of gene cloning.
In this study, we successfully validated five QTLs (qGW1, qGW5, qGW7, qGL9, and qGW12) that exert significant influences on grain size and narrow down their locations to regions ranging from 128 kb to 624 kb using a map-based cloning method.These results laid the foundation for the additional fine mapping of the five QTLs, the cloning of the candidate genes, and functional research to explore the genetic mechanisms underlying grain size.
In the qGW5 interval, LOC_Os05g03020, a C2H2 zinc finger protein, could potentially be considered as a candidate gene.LOC_Os04g36650 (NSG1) and LOC_Os06g48530 (Du13), belonging to the C2H2 zinc finger family, have been found to determine grain size [73,74].
LOC_Os07g46460 (spl32) and LOC_Os07g46590 may be the candidate genes for qGW7.Compared with the wide type, the spl32 mutant exhibited a decrease in grain size [75].Moreover, LOC_Os03g51230 (OsDDM1b), a homologous member of the gene family that contains LOC_Os07g46590, has been reported to regulate grain size [76].
LOC_Os12g40190 (OsXLG4), LOC_Os12g40570 (OsWRKY83), LOC_Os12g40460, and LOC_Os12g40830 are potential candidate genes for qGW12.Firstly, RGA1, a homologous member of the gene family containing LOC_Os12g40190 (OsXLG4), positively regulates rice grain size [83].Secondly, OsWRKY36, a homologous member of the gene family containing LOC_Os12g40570 (OsWRKY83), has been reported to suppress grain size by inhibiting GA signaling [84].Thirdly, DGS1, a homologous member of the gene family containing LOC_Os12g40460, played a positive role in regulating grain size by binding to OsBZR1 [85].The mutant of FRRP1 resulted in an increase in rice grain length [86].Finally, OsFRK3, a homologous member of the pfkB family containing LOC_Os12g40830, was identified as a positive regulator of grain width and thickness through its influence on sugar metabolism [87].In the future, transgenic studies will be conducted on these five QTLs to further elucidate the molecular mechanisms underlying grain size.

Cytological Analysis of the Five QTLs
To elucidate the characterization and commonality of differentially expressed genes influencing cell number and size in qGW1, qGW5, qGW7, qGL9, and qGW12, we analyzed the association among these genes involved in cell division and expansion.Specifically, we identified one specific differential gene in qGW1, two in qGW5, ten in qGL9, and three in qGW12.The regulatory modules of qGW1-qGW5, qGW1-qGW12, qGW5-qGL9, and qGL9-qGW12 exhibited differential expression in cell cycle-related genes, including E2F2 and CYCA3;1; CDKA1; MAPK and CDKA2; and CYCD1;1, CYCB1;1, and KN, respectively (Figure 9A,B).These findings suggest that the regulation of grain width by qGW1 and qGW12 is mediated through the modulation of CDKA1, while the influence of qGW5 and qGL9 on grain length involves MAPKand CDKA2-mediated mechanisms, thereby altering cell numbers.Moreover, the differential expression of cell expansion-related genes, such as EXPA5, EXPA10, EXPA6 and EXPA7; EXPA3, EXPA10, and EXPB7; EXPA10; and EXPA6 and EXPB7, were observed within the regulatory modules of qGW1-qGW7, qGW5-qGW7, qGW1-qGW5-qGW7, qGW1-qGW7-qGW12, and qGW5-qGW7-qGW12, respectively (Figure 9A,B), suggesting qGW1-qGW7-qGW12 influences cell size and grain width by altering the expression of EXPA6.Numerous QTLs and genes have been reported to regulate grain size by influencing the expression of cell division and expansion-related genes, such as qTGW2b [88], GLW7.1 [89], GL3.1 [38], and GS2 [90].In addition, negative correlations were observed between gene expression and grain yield for both qGW1 and qGW7.Despite variations in the expression of cell cycle-and expansion-related genes, there was no alteration in the yield of qGW5.Notably, positive correlations were observed between gene expression and grain yield for both qGL9 and qGW12.Variations in the expression of genes associated with cell division and expansion not only impact grain size, but also influence grain yield [38,89,90].
qGW1-qGW7, qGW5-qGW7, qGW1-qGW5-qGW7, qGW1-qGW7-qGW12, and qGW5-qGW7-qGW12, respectively (Figure 9A,B), suggesting qGW1-qGW7-qGW12 influences cell size and grain width by altering the expression of EXPA6.Numerous QTLs and genes have been reported to regulate grain size by influencing the expression of cell division and expansion-related genes, such as qTGW2b [88], GLW7.1 [89], GL3.1 [38], and GS2 [90].In addition, negative correlations were observed between gene expression and grain yield for both qGW1 and qGW7.Despite variations in the expression of cell cycleand expansion-related genes, there was no alteration in the yield of qGW5.Notably, positive correlations were observed between gene expression and grain yield for both qGL9 and qGW12.Variations in the expression of genes associated with cell division and expansion not only impact grain size, but also influence grain yield [38,89,90].
Understanding the relationship between different genes involved in cell division and expansion is convenient for studying the shared characteristics and cytological molecular mechanisms underlying the influence of these five QTLs on rice grain size.

Potential Uses of the Five QTLs in Rice Breeding
High yield and superior quality are essential goals of rice breeders.Gene cloning and molecular breeding have become important techniques to breed high-yield and superior-quality varieties.The major genes that increase grain size, such as GW2 [15], GS2 [31], and GW5 [35,36], produce higher grain yield while simultaneously reducing grain quality.The gs9 allele has the potential to improve the appearance quality of milled rice without affecting grain yield [32].Nevertheless, research advances have revealed several genes that could be utilized to help breeders develop new elite rice varieties with high yield and superior quality.The combination of the OsMADS1 lgy3 allele with dep1-1 and gs3 alleles has the potential to simultaneously improve both grain yield and quality in rice [14].GLW7.1 also represents a novel way to breed high-yield and superior-quality varieties [89].
In our study, negative correlations were observed between yield and rice quality for both qGW7 and qGW12.However, qGW5 exhibited the potential to enhance quality without compromising yield.Interestingly, qGW1 and qGL9 displayed positive correlations between grain yield and rice quality, indicating their pleiotropic effects in simultaneously improving both yield and quality.In summary, the identification of genetic resources, such as qGW5, qGW1, and qGL9, provides a theoretical foundation for breeding strategies aimed at enhancing grain yield and quality in rice.Our results laid the foundation for cloning these five genes.Additionally, such information will help breeders to improve grain yield and quality in rice.Understanding the relationship between different genes involved in cell division and expansion is convenient for studying the shared characteristics and cytological molecular mechanisms underlying the influence of these five QTLs on rice grain size.

Potential Uses of the Five QTLs in Rice Breeding
High yield and superior quality are essential goals of rice breeders.Gene cloning and molecular breeding have become important techniques to breed high-yield and superiorquality varieties.The major genes that increase grain size, such as GW2 [15], GS2 [31], and GW5 [35,36], produce higher grain yield while simultaneously reducing grain quality.
The gs9 allele has the potential to improve the appearance quality of milled rice without affecting grain yield [32].Nevertheless, research advances have revealed several genes that could be utilized to help breeders develop new elite rice varieties with high yield and superior quality.The combination of the OsMADS1 lgy3 allele with dep1-1 and gs3 alleles has the potential to simultaneously improve both grain yield and quality in rice [14].GLW7.1 also represents a novel way to breed high-yield and superior-quality varieties [89].
In our study, negative correlations were observed between yield and rice quality for both qGW7 and qGW12.However, qGW5 exhibited the potential to enhance quality without compromising yield.Interestingly, qGW1 and qGL9 displayed positive correlations between grain yield and rice quality, indicating their pleiotropic effects in simultaneously improving both yield and quality.In summary, the identification of genetic resources, such as qGW5, qGW1, and qGL9, provides a theoretical foundation for breeding strategies aimed at enhancing grain yield and quality in rice.Our results laid the foundation for cloning these five genes.Additionally, such information will help breeders to improve grain yield and quality in rice.

Population Development and Field Experiment
The BC 1 F 2 and BC 1 F 2:3 populations were derived from two indica lines, GZ63-4S (the recurrent parent) and W240 (the donor parent).GZ63-4S is a photoperiod-thermo-sensitive genic male sterile indica rice line, carrying the infertility gene of TMS5 [91].W240 is an indica variety with larger grain size.The BC 1 F 2 and its derived BC 1 F 2:3 populations were used for QTL mapping.To validate the genetic effects of qGW1, qGW5, qGW7, qGL9, and qGW12, five BC 1 F 6 and BC 1 F 7 populations were planted at the experimental station of Huazhong Agricultural University at Wuhan, Hubei province and Lingshui, Hainan province in 2020, respectively.The progeny tests were conducted in the BC 1 F 8 and BC 1 F 9 generations.The BC 1 F 2 , BC 1 F 2:3 , BC 1 F 6 , BC 1 F 7 , BC 1 F 8 , and BC 1 F 9 populations were planted in 2016, 2017, 2020 (twice), and 2021 (twice).The detailed process of population development is shown in Figure S2.All rice plants with a density of 16 cm × 26 cm were grown under normal field management.Field management followed local practices.Ten plants were harvested from the middle of each row for trait measurement.

Trait Measurement
Harvested rice grains from each plant were air-dried and stored at room temperature for three months before testing.Grain length, grain width, grain number, grain yield, and 1000-grain weight were measured using the yield traits scorer (YTS) platform [92], whereas grain thickness was measured using vernier calipers.The plant height was measured from the main culm.The number of tillers per plant was counted as all fertile panicles in one plant.Additionally, flour ground from milled grain was used to determine the albumin content, globulin content, prolamin content, glutenin content, total starch content, amylose content, and gel consistency according to the NY/T 593-2013 standard published by the Ministry of Agriculture, China (http://www.zbgb.org/27/StandardDetail1476335.htm,accessed on 3 October 2019).Taste scores for milled rice were evaluated using a taste analyzer kit (Satake, RLTA10B-KC, Hiroshima, Japan) [93].

Genetic Map Construction and QTL Mapping
The parent varieties GZ63-4S and W240 were sequenced using the illumine HiSeq2000 (Illumina, San Diego, CA, USA), and the sequencing data were compared and assembled according to the rice reference genome (Rice Genome Annotation Project, http://rice.uga.edu/, accessed on 6 February 2022) [94].All mapping primers were designed in reference to the sequencing data of two parents.A total of 67 polymorphic simple sequence repeat (SSR) markers, 103 insert and deletion (InDel) markers, and 7 Kompetitive allele-specific PCR (KASP) markers were evenly distributed across 12 chromosomes to genotype the 327 BC 1 F 2 lines.According to the cetyltrimethylammonium bromide (CTAB) method, genomic DNA was extracted from leaves [95].The genotyping was carried out using 4% Polyacrylamide gels (PAGE) migration, as previously reported by Panaud et al. [96].DNA bands on PAGE gel were displayed by silver nitrate staining and NaOH-formaldehyde solution.
Combining the genotype data from the BC 1 F 6 lines and the phenotype data from both BC 1 F 6 and BC 1 F 7 lines, we employed the Kosambi mapping function of MapMaker/Exp3.0 program to construct a genetic linkage map [97].QTL analysis was performed using the composite interval mapping method with Windows QTL cartographer 2.5 software (WinQTLCart 2.5) [98].
Refraining from considering the QTL of cloned genes, we selected five major QTLs that have higher LOD, Add, and PVE to study.The genotypes of the five BC 1 F 6 and BC 1 F 7 lines of qGW1, qGW5, qGW7, qGL9, and qGW12 were determined using two flanking markers within the mapping interval of QTLs.To fine map these QTLs, we developed five BC 1 F 9 populations consisting of 1728, 2046, 1680, 1584, and 1440 individuals, respectively.Another 6, 5, 14, 5, and 6 specific markers were developed to genotype the recombinants of these QTLs.Relevant primer sequences are shown in Table S2.

Scanning Election Microscopy
Lemmas of spikelets at the heading stage were collected for scanning electron microscopy, fixed in FAA solution (50% ethanol, 5% glacial acetic acid, and 3.7% formaldehyde) at 4 • C for 24 h.The young panicles were sampled at the length of about 3 cm.Then the samples were coated with gold under vacuum conditions, and observed using a scanning electron microscope (JEOL, JSM-6390LV, Tokyo, Japan) under 10 kV acceleration voltage and a 30 nm spot size.Cell number and cell size were calculated at 50 × and 100 × magnification, respectively.The spikelet epidermal cell size was measured using Image J software (NIH), and cell number was counted manually.Scanning electron microscopy analysis involved at least three biological replications of mounted specimens.

RNA Extraction, Reverse Transcription, and qRT-PCR
Total RNA was extracted from young panicles using the TRIzol method (Invitrogen, 15596026, Shanghai, China), and then treated with RNase-free DNase I (Invitrogen, 15596026, Shanghai, China).First, strand cDNA was reverse-transcribed using the M-MLV Reverse Transcriptase kit (Promega, M170A, Madison, WI, USA).All procedures were carried out according to the manufacturer's protocol.qRT-PCR was performed using ABI Real-Time PCR system with the SYBR Green I mix (TaKaRa, Shiga, Japan) according to the manufacturer's instructions.OsActin gene was used as an internal control to normalize gene expression.The gene expression levels in three biological replicates and three technical replicates were calculated to evaluate the significance of differences between samples using the student's t-test.Relevant primer sequences are given in Table S2.

Statistical Analysis
Differences between two sets of data were presented as the mean ± standard deviation and performed using the student's t-test.We conducted differential expression analysis of differentially expressed genes using Cytoscape software (3.9.1) [99] and the Metware Cloud, a free online platform for data analysis (https://cloud.metware.cn,accessed on 1 January 2024).

Figure 1 .
Figure 1.Frequency distributions of GL (A), GW (B), LWR (C), and GT (D) in the BC1F2 and BC1F2:3 populations.The vertical axis represents the number of BC1F2 and BC1F2:3 plants, with black and gray bars, respectively.

Figure 1 .
Figure 1.Frequency distributions of GL (A), GW (B), LWR (C), and GT (D) in the BC 1 F 2 and BC 1 F 2:3 populations.The vertical axis represents the number of BC 1 F 2 and BC 1 F 2:3 plants, with black and gray bars, respectively.

Figure 2 .
Figure 2. Genetic linkage map of grain-size-related QTLs detected in the BC1F2 and BC1F2:3 populations.Figure 2. Genetic linkage map of grain-size-related QTLs detected in the BC 1 F 2 and BC 1 F 2:3 populations.

Figure 2 .
Figure 2. Genetic linkage map of grain-size-related QTLs detected in the BC1F2 and BC1F2:3 populations.Figure 2. Genetic linkage map of grain-size-related QTLs detected in the BC 1 F 2 and BC 1 F 2:3 populations.

Figure 3 .
Figure 3. Analysis of qGW1 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW1.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW1 and NIL-G qGW1 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of four cell cycle related-genes and four cell expansion related-genes between NILs of qGW1.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 3 .
Figure 3. Analysis of qGW1 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW1.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW1 and NIL-G qGW1 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal Int. J. Mol.Sci.2024, 25, x FOR PEER REVIEW 7 of 21

Figure 4 .
Figure 4. Analysis of qGW5 influence grain length.(A) Grain morphology.Scale bar: 5 mm.(B) Grain length difference among three haplotypes in 2020.(C) Fine mapping of qGW5.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW5 and NIL-G qGW5 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of five cell cycle related-genes and four cell expansion related-genes between NILs of qGW5.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 4 .
Figure 4. Analysis of qGW5 influence grain length.(A) Grain morphology.Scale bar: 5 mm.(B) Grain length difference among three haplotypes in 2020.(C) Fine mapping of qGW5.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW5 and NIL-G qGW5 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of five cell cycle related-genes and four cell expansion related-genes between NILs of qGW5.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

21 Figure 5 .
Figure 5. Analysis of qGW7 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW7.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW7 and NIL-G qGW7 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of six cell expansion related-genes between NILs of qGW7.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 5 .
Figure 5. Analysis of qGW7 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW7.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW7 and NIL-G qGW7 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of six cell expansion related-genes between NILs of qGW7.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 6 .
Figure 6.Analysis of qGL9 influence grain length.(A) Grain morphology.Scale bar: 5 mm.(B) Grain length difference among three haplotypes in 2020.(C) Fine mapping of qGL9.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGL9 and NIL-G qGL9 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of fifteen cell cycle related-genes between NILs of qGL9.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 6 .
Figure 6.Analysis of qGL9 influence grain length.(A) Grain morphology.Scale bar: 5 mm.(B) Grain length difference among three haplotypes in 2020.(C) Fine mapping of qGL9.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGL9 and NIL-G qGL9 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of fifteen cell cycle related-genes between NILs of qGL9.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a, b and c indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 7 .
Figure 7. Analysis of qGW12 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW12.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW12 and NIL-G qGW12 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of six cell cycle related-genes and three cell expansion related-genes between NILs of qGW12.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a and b indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 7 .
Figure 7. Analysis of qGW12 influence grain width.(A) Grain morphology.Scale bar: 5 mm.(B) Grain width difference among three haplotypes in 2020.(C) Fine mapping of qGW12.The numbers below the bar are physical distance (Mb).(D,E) Scanning electron microscopy of the outer epidermal cells of NIL-W qGW12 and NIL-G qGW12 .Scale bar: 100 µm.(F) Cell length.(G) Total number of longitudinal cells.(H) Cell width.(I) Total number of transverse cells.(n = 10).(J) qRT-PCR analysis of six cell cycle related-genes and three cell expansion related-genes between NILs of qGW12.Data are represented as mean ± s.e.m. (n = 9).Duncan's multiple range tests were used to conduct statistical analysis (a and b indicate p < 0.01).The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 8 .
Figure 8. Phenotypes of rice yield and quality comparison among qGW1, qGW5, qGW7, qGL9, and qGW12 identified in this study.(A-H) Grain length, grain width, length to width ratio, 1000-grain weight, plant height, number of tillers per plant, number of filled grains per panicle, and grain yield per plant in five NILs.(n = 12).(I-P) Albumin content, globulin content, prolamin content, glutenin content, total starch content, amylose content, gel consistency, and taste score.(n = 6).All phenotypic data in (A-P) were measured from paddy-grown NIL plants grown under normal cultivation conditions.Data are represented as mean ± s.e.m.The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 8 .
Figure 8. Phenotypes of rice yield and quality comparison among qGW1, qGW5, qGW7, qGL9, and qGW12 identified in this study.(A-H) Grain length, grain width, length to width ratio, 1000-grain weight, plant height, number of tillers per plant, number of filled grains per panicle, and grain yield per plant in five NILs.(n = 12).(I-P) Albumin content, globulin content, prolamin content, glutenin content, total starch content, amylose content, gel consistency, and taste score.(n = 6).All phenotypic data in (A-P) were measured from paddy-grown NIL plants grown under normal cultivation conditions.Data are represented as mean ± s.e.m.The student's t-test was used to produce p values (*, ** indicate significance at p < 0.05 and p < 0.01, respectively).

Figure 9 .
Figure 9.The differential expression analysis of (A,B) cell number and size genes.

Figure 9 .
Figure 9.The differential expression analysis of (A,B) cell number and size genes.

Table 1 .
QTLs for grain size in the BC 1 F 2 and BC 1 F 2:3 populations.
qGL, QTL for grain length; qGW, QTL for grain width; qLWR, QTL for length to width ration; qGT, QTL for grain thickness.LOD, logarithm of odds.Add, additive effect of QTL.Positive value and negative value of additive effects indicated the W240 and GZ63-4S alleles, respectively.PVE, phenotypic variance explained by the QTL.Int.J. Mol.Sci.2024, 25, x FOR PEER REVIEW 6 of 21