Introduction

Rice (Oryza sativa L.) is an important staple crop that feeds over one-half of the world’s population. However, due to the ever-increasing population, food shortages have become a serious problem. Thus, it is essential to increase grain yield to meet the food demands required by a growing population (Godfray et al. 2010; Huang et al. 2013). Grain yield in rice is characterized by four components, tiller number (TN), thousand grain weight (TGW), seed setting ratio (SSR), and grain number per panicle (GN). Of these, GN is traditionally considered to be one of the most important factors. At present, many studies have reported that some morphological component of panicle architecture such as panicle length (PL) and the number of primary and secondary branches (PBN/SBN) are important agronomic traits that affect GN (Luo et al. 2009; Mei et al. 2006; Wang et al. 2017). Hence, it is of great theoretical and practical value to identify key genes for GN and GN-related traits.

GN is a canonical complex trait controlled by a number of major and minor quantitative trait loci (QTLs), which presents a challenge in characterizing this trait (Xing et al. 2002). Although researchers have already extensively investigated the QTLs contributing to the regulation of GN in rice, favorable genes/QTLs that could be directly used in rice breeding remain rare. At present, at least 900 QTLs for GN have been identified on the 12 rice chromosomes (http://www.gramene.org), and some of them have been fine mapped. GPA7, a QTL for grain number, was delimited to a small region of ~ 35 kb located on chromosome 7 (Tian et al. 2006). qGY2–1, a QTL for gain yield, was fine mapped to an interval of ~ 102.9-kb on chromosome 2 (He et al. 2006). SPP1, a QTL that controls spikelet number, was fine mapped to a region of ~ 107-kb on the short arm of chromosome 1 (Liu et al. 2009). qTSN12.2, a QTL for total spikelet number per panicle, was fine mapped to an approximately 92 kb region on chromosome 12 (Sasaki et al. 2017). GN4-1, a major QTL responsible for GN on chromosome 4, was delimited to an interval of ~ 190-kb (Zhou et al. 2018). The QTL qSPP2.2, which controls spikelet number per panicle, was localized to an ~ 167.1-kb region on the long arm of chromosome 2 (Kaur et al. 2018). Several QTLs involved in the regulation of GN have already been cloned.

The first QTL isolated for GN was Gn1a that is located on the short arm of chromosome 1 and encodes a cytokinin oxidase/dehydrogenase (CKX) (Ashikari et al. 2005). CKX is the sole enzyme that catalyzes the irreversible degradation of cytokinins (CKs). Hence, decreased expression of Gn1a in inflorescence meristems can lead to the accumulation of CK, which finally results in increased GN (Hwang et al. 2012). A gain-of-function mutation in DEP1 truncates a phosphatidylethanolamine-binding protein-like domain protein and increases the grain number, although the panicle length is reduced (Huang et al. 2009). The IPA1/OsSPL14 gene encodes squamosa promoter binding protein-like 14 and is a target of OsmiR156 (Jiao et al. 2010). The mutant allele that disrupts the OsmiR156-binding site causes increased GN. Further research has demonstrated that IPA1 may be a positive regulator of DEP1 (Lu et al. 2013). NOG1 encodes an enoyl-CoA hydratase/isomerase and can increase grain yield by enhancing GN with no negative effects (Huo et al. 2017). DTH7/Ghd7.1, Ghd7, and DTH8/Ghd8 are three major QTLs for heading date (HD) that can also lead to increased GN under long-day conditions (Gao et al. 2014; Liu et al. 2013; Wei et al. 2010; Weng et al. 2014; Yan et al. 2011). GN2 exhibits pleiotropism and not only influences grain number but also affects plant height and heading date (Chen et al. 2017). GNP1, a gene that encodes gibberellin 20-oxidase, can enhance GN (Wu et al. 2016), suggesting that GA is an important hormone in the regulation of GN.

As mentioned above, although some QTLs for GN-related traits have been mapped or cloned, the underlying genetic and molecular mechanisms that regulate GN are still unknown. It is therefore worthwhile to exploit new genetic loci that could increase GN. In this study, we first identified QTLs responsible for GN regulation using a set of chromosomal segment substitution lines (CSSLs) derived from a cross between the japonica cultivar ‘Nipponbare’ (‘NP’) and the indica cultivar 9311. Among the identified QTLs, a major QTL, qGN1c, located on chromosome 1 (Chr. 1) was fine mapped to a chromosomal interval of ~ 379 kb delimited by flanking marker loci STSN2-17 and STSN2-18 using advanced BCnF2–5 segregating populations. The genetic effect of qGN1c was also evaluated using near-isogenic lines (NILs). The results of our study provide an opportunity to clone qGN1c, which will be useful for improving grain yield in rice breeding.

Materials and methods

Plant materials and population development

A set of 125 CSSLs that together cover 95.6% of the 9311 genome were generated by using 9311, an indica cultivar, as the segment donor, and ‘Nipponbare’ (‘NP’), a japonica cultivar, as the recipient. Each of the CSSLs carries one or a few donor chromosomal segments (Xu et al. 2017; Zhang et al. 2011). In the summer of 2014 and 2015, the CSSLs and their parental lines were grown and phenotyped to identify QTLs for panicle-related traits. One CSSL line, N2, was selected to cross with ‘NP’ to develop the segregating population for qGN1c validation. In the summer of 2016, an F2 population containing 300 plants was developed. Of the F2 plants, 282 individuals were selected to genotype and phenotype for genetic analysis. In the winter of 2016, the recombinant individuals (F3 plants) were grown and genotyped to identify individual plants with homozygous genotypes (9311 type). In the summer of 2017, six types of recombinant individuals (F4 generation) with homozygous genotypes (9311 type) were planted and phenotyped for the initial mapping of qGN1c. In addition, 25 individual plants that were found to be heterozygous in the mapped region were selected from the F2 population to produce a large segregating population for the fine mapping of qGN1c. In the summer of 2017, 2820 plants were chosen for the detection of recombination events using the molecular markers STSN2-17 and STSN2-20. In the winter of 2017, six types of recombinant individuals (F4 generation) were planted and genotyped. In the summer of 2018, six types of recombinant individuals (F5 generation) with homozygous genotypes (9311 type) were planted and phenotyped for fine mapping qGN1c. A set of near-isogenic lines (NILs) carrying qGN1c in the ‘NP’ background (NIL-qGN1c9311) were also developed using a backcross scheme with marker-aided selection. In the summer of 2018, 200 individual plants from the NIL-qGN1c9311 population and NP were selected to evaluate the effect of qGN1c on yield in two planting locations.

Field experiments

The rice lines were planted in the experimental fields at Xiao Nei (XN; 32° 23′ 24″ N, 119° 25′ 3″ E) and Jiu Dian (JD; 32° 29′ 54″ N, 119° 26′ 4″ E), Yangzhou, Jiangsu Province, China, in the summer. In the winter, the lines were grown in the fields at Lingshui (18° 31′ 52″ N, 110° 10′ 53″ E), Hainan Province, China. The growing density was 15 × 25 cm for line × row.

Measurement of grain yield–related traits and data analysis

We evaluated a total of 14 agronomic traits, including plant height (PH), heading date (HD), tiller number per plant (TN), seed setting ratio (SSR), grain number per panicle (GN), panicle length (PL), primary branch number (PBN), secondary branch number (SBN), primary branch grain number (PBGN), secondary branch grain number (SBGN), thousand grain weight (TGW), grain length (GL), grain width (GW), and grain yield per panicle (GYPP). These agronomic traits were analyzed using a previously described method (Liu et al. 2016b; Wang et al. 2017; Xu et al. 2017; Yu et al. 2017). In brief, panicles in the tallest tillers were harvested and measured to evaluate the panicle-related trait phenotypes. We chose 10–15 individuals from the middle of the row in each line for the evaluation of all phenotypes except for GYPP. GYPP was measured as the mean weight of grain from 200 individual plants. SPSS software (version 15.0) was used to analyze the correlation coefficients among grain yield–related traits, and the other data for each trait were analyzed using an independent sample t test program.

DNA extraction and molecular analysis

Genomic DNA was extracted from fresh rice leaves of one plant using the CTAB method (Murray and Thompson 1980). PCR was performed and PCR products separated using a previously described method (Wu et al. 2009). The novel InDel molecular markers that were used for validation and fine mapping of qGN1c in this study were developed using PRIMER 5.0 software based on the publicly available rice genome sequence (http://www.ncbi.nlm.nih.gov/). DNA sequences of primers used to amplify the novel InDel markers are given in Supplemental Table 1; Gn1a-M is a gene-tagged marker based on a 16-bp deletion in the 5′-untranslated region of Gn1a in the cultivars ‘Habataki’ and 5150 (Yan et al. 2009).

QTL analysis

The QTLs identified in this study were mapped using the software ICI-Mapping 4.0 with the aid of the Bin-map, in which the 9311 and ‘NP’ genotypes were assigned the values 2 and 0, respectively (Meng et al. 2015; Xu et al. 2017). QTL analysis was performed using the inclusive composite interval mapping (ICIM) method, and the threshold of the logarithm of the odds ratio (LOD) was 2.5. The QTLs were named using the nomenclature as given in McCouch and CGSNL (2008).

Results

Variation in panicle characteristics between the parental lines and the CSSL population

GN is one of the most prominent differences between ‘NP’ and 9311. GN in 9311 is approximately double that in ‘NP’ (Fig. 1a, b). GN is affected by panicle length (PL), primary branch number (PBN), and secondary branch number (SBN) (Luo et al. 2009; Mei et al. 2006; Wang et al. 2017). Therefore, to address this difference in GN, PL, PBN, and SBN were further investigated in the parental lines. The values for PL, PBN, and SBN were also significantly higher in 9311 compared with ‘NP’ (Fig. 1c–e).

Fig. 1
figure 1

Phenotypic differences in panicles of 9311 and ‘NP’ (‘Nipponbare’). a Panicle morphology, scale bar = 2 cm. b–e Comparisons of four panicle-related traits between 9311 and ‘NP’. Error bars represent the mean value ± SE (n = 10 plants). ** and * indicate significant differences at the P < 0.01 and P < 0.05 levels, respectively, as determined by the two-tailed t test

Full size image

To understand the molecular basis of the regulation of GN, we investigated the behaviors of the three factors in the CSSL population and performed QTL analyses. A continuous distribution frequency of these factors confirmed that GN is a complicated trait controlled by multiple QTLs (Fig. 2). Except for the distribution frequency of PBN (Fig. 2c), the values for GN, PL, and SBN were comparable in 2 years, implying that PBN is more sensitive to environmental effects. Although these GN-related factors in the majority of the lines were similar to those in ‘NP’, several lines exhibited improved phenotypes, indicating that some of the introgressed segments harbor elite loci. Correlation analysis showed that the measured traits in this study were significantly positively correlated with each other except for PBN and SBN (Table. 1). A striking correlation (correlation coefficient > 0.88) was found between SBN and GN, indicating that SBN is tightly associated with GN.

Fig. 2
figure 2

Frequency distribution of panicle-related traits in the CSSLs and the parental inbred lines (‘NP’ and 9311) in field seasons 2014 and 2015. a Grain number per panicle, b panicle length, c primary branch number per panicle, and d secondary branch number per panicle. The values for the two parental lines (‘NP’ and 9311) are indicated by black arrows in each panel

Full size image
Table 1 Correlation coefficients among four panicle-related traits
Full size table

QTL mapping

The ICIM approach was used to detect QTLs for panicle-related traits in the CSSL population based on 2 years of phenotypic investigations. Twenty-five QTLs were found to be distributed on eight chromosomes; chromosomes 4, 10, 11, and 12 did not have any QTLs for GN (Table 2; Supplemental Fig. 1). In 2014, a total of 12 QTLs, six for PL, three for SBN, and three for GN, were found on six chromosomes and explained 64.45%, 32.60%, and 42.21% of the phenotypic variance, respectively. Among these QTLs, qPL5, qPL6-3, qPL9-1, qSBN6, qGN1c, and qGN6 were considered to be major QTLs because they could explain > 10% of the phenotypic variance. Interestingly, three QTL clusters located on Chr. 1, 6, and 9 were simultaneously identified to control other panicle-related traits (Table. 2; Supplemental Fig. 1), which is in agreement with the correlation analysis (Table 1). In fact, the regions containing the clustered QTLs mapped on Chr. 6 and 9 include two previously reported QTLs for PL, qPL6 and Lp1, respectively (Liu et al. 2016a; Sun et al. 2017) and the region containing qPL7-1 contains the Ghd7 gene (Xue et al. 2008). The mapped region on Chr. 1 containing another QTL cluster is near Gn1a but does not include Gn1a (Ashikari et al. 2005), because three gaps were found on Chr. 1 and the Gn1a locus is just inside gap 1 (Supplemental Fig. 1; Supplemental Fig. 2). Therefore, we speculate that this mapped region may harbor novel genes for determining GN. Of the 13 QTLs identified in 2015, six for PL, one for PBN, three for SBN, and three for GN were found to explain 53.86%, 8.47%, 38.29%, and 41.84% of the phenotypic variance, respectively (Table 2). In particular, we also identified four major QTLs, including qPL6-4, qSBN6, qGN1c, and qGN6, as well as three QTL clusters located on Chr. 1, 2, and 6 (Supplemental Fig. 1). Notably, the mapped QTL cluster located on Chr. 2 is near DTY2.1, a previously-reported locus for grain yield (Venuprasad et al. 2009).

Table 2 Identification of QTLs for four panicle-related traits by composite interval mapping in 2014 and 2015
Full size table

Taken together, the majority (18/25) of the identified QTLs had positive additive effects, implying that these alleles from 9311 may be directly used in the design of molecular breeding strategies for high grain yield (Table 2). Moreover, based on the mapped region, five QTLs, qPL5, qSBN1, qSBN6, qGN1c, and qGN6, were found to be co-localized. In addition, two QTL pairs (qPL6-1 and qPL6-2 as well as qPL6-3 and qPL6-4) for PL were also identified as being in close proximity to each other with similar additive effects (Table 2; Supplemental Fig. 1). We therefore considered that these QTLs were worthy of further analysis. Only one QTL for PBN was detected in 2 years, suggesting that this trait is sensitive to environmental effects.

Validation and fine mapping of qGN1c

Given that qGN1c is a new locus that has a positive contribution to GN, we next performed fine mapping to validate this QTL. Re-sequencing data of the CSSL population showed that CSSL line N2 carrying qGN1c had a single 9311-derived fragment introgression (2,141,284 to 5,084,973 bp) on Chr. 1 (Fig. 3a). Seven polymorphic marker loci distributed in the region from 2.1 to 5.2 Mb were developed, and the physical location of the 9311 chromosomal segment was confirmed in the N2 line (Supplemental Table.1; Supplemental Fig. 3). N2 plants had higher GN than did ‘NP’ plants (Fig. 3c, d). We next investigated five GN-related traits, including PL, PBN, SBN, PBGN (primary branch grain number), and SBGN (secondary branch grain number), in ‘NP’ and N2. SBN and SBGN were significantly increased in N2 (Fig. 3f, g), resulting in higher GN in N2 than in ‘NP’, whereas PL, PBN, and PBGN, as well as plant height (PH) and heading date (HD), were identical in N2 and ‘NP’ (Fig.3b, c, e–i). Taking all of these findings together, we chose N2 as a parent in a cross with ‘NP’ to generate a fine-mapping population.

Fig. 3
figure 3

Phenotypic differences between ‘NP’ and the CSSL line N2 in 2014. a Graphical genotype of CSSL line N2. The position of the introgressed chromosomal segment from 9311 is shown in red. The location of Gn1a is indicated by an arrow. b Mature plants of ‘NP’ and N2. c Panicle phenotypes. d–i Comparisons of GN, PL, PBN, SBN, PBGN (primary branch grain number), SBGN (secondary branch grain number), PH (plant height), and HD (heading date) between ‘NP’ and N2. Error bars represent the mean value ± SE (n = 10 plants). ** and * indicate significant differences at the P < 0.01 and P < 0.05 levels, respectively, as determined by the two-tailed t test. Scale bar = 5 cm in b and 2 cm in c

Full size image

After phenotyping and genotyping the 282 individual plants of an F2 population derived from N2 and ‘NP’ (Supplemental Table 1; Fig. 4A), we found that the plants with 9311 homozygous genotype (group III) yielded more grain numbers, and the heterozygous genotype (group II) had grain numbers between ‘NP’(group I) and the plants that were homozygous for qGN1c (Fig. 4b, c). These results showed that qGN1c is a real QTL for GN and that it is semi-dominant. We also identified six types of recombinants in the F2 families based on the positions of recombination breakpoints determined using molecular markers, and individual recombinant F4 plants with homozygous 9311 genotypes were selected for quantification of the GN phenotype (Fig. 4d). Compared with ‘NP’, plants in groups III, IV, VI, and VII had significantly increased GN. In contrast, no significant differences were observed between ‘NP’ and plants in groups II and V (Fig. 4c). Therefore, combining the genotypes and phenotypes of the six groups allowed us to map qGN1c to a ~ 1.14-Mb genomic region flanked by marker loci STSN2-17 and STSN2-20 (Fig. 4d).

Fig. 4
figure 4

Fine-mapping of qGN1c. a The physical location of the 9311 donor chromosomal segment on Chr. 1 in CSSL line N2. N2-15 to N2-25 indicate the locations of sequence-tagged site (STS) molecular marker loci. Gn1a-M is a gene-tagged marker specific for Gn1a. The genomic region from 9311 is shown as a black rectangle, and the genome of ‘Nipponbare’ (NP) is shown as white rectangles. B and c Genetic analysis using the subpopulation derived from the cross of ‘NP’ and N2. Representative genotypes of the segregating lines are shown on the left (b) and the GN values for the corresponding plants are shown on the right (c). The black, white, and gray rectangles indicate the homozygous 9311 genotype, the homozygous ‘NP’ genotype, and the heterozygous genotype, respectively. d and e Primary mapping of qGN1c based on the genotypes (left) and phenotypes (right) of recombinants in an F4 population. e Fine mapping of qGN1c using the F5 population. The numbers in parentheses next to the marker loci represent the number of recombinant individuals, and the ~ 379-kb region containing qGN1c is indicated below the figure in e. The homozygous 9311 and homozygous ‘NP’ genotypes are shown as black and white rectangles, respectively. Error bars represent the mean value ± SE (n = 30 individuals in the F2 population, 15 individuals in the substitution line). ** and * indicate significant differences at the P < 0.01 and P < 0.05 levels, respectively, as determined by the two-tailed t test

Full size image

To further narrow down the mapped region, 25 F2 heterozygous plants were used to develop a self-population for fine mapping. By screening 2820 F3 generation individuals, we identified 42 recombinants based on the STSN2-17 and STSN2-20 marker alleles (Fig. 4e). Using four additional polymorphic markers, the 42 recombinants were further classified into six groups based on the positions of the recombination breakpoints (Fig. 4E). The results of a progeny test (F5 generation) of these recombinants revealed that the plants in groups II–V had increased GN, while plants in groups VI and VII had reduced GN (Fig. 4C). Ultimately, the position of qGN1c was narrowed to within a ~ 379-kb region of Chr. 1 flanked by marker loci STSN2-17 and STS N2-18 (Fig. 4e).

Performance of NIL-qGN1c 9311

To evaluate the genetic effects of qGN1c, we identified a NIL for qGN1c (NIL-qGN1c9311) carrying a ~ 1.51-Mb chromosomal segment introgressed from 9311 flanked by marker loci STSN2-15 and STSN2-20 (Fig. 5a). No significant differences were observed for PL, PBN, PBGN, HD, PH, SSR, TN, and grain width (GW) between NIL-qGN1c9311 and ‘NP’ plants. Similar to the CSSL line N2, NIL-qGN1c9311 plants produced higher SBN and SBGN (Fig. 5c; Table 3), resulting in GN increases of 15.92% and 18.61% in the Xiao Nei (XN) and Jiu Dian (JD) environments, respectively. Compared with ‘NP’, grain yield per plant (GYPP) in NIL-qGN1c9311 was increased by 13.34% and 14.46% in the two locations (Table 3). Due to a reduction in grain length (GL), the TGW of NIL-qGN1c9311 was significantly decreased by an average of 6.43% and 5.51% compared with ‘NP’ (Fig. 5d; Table 3), but this did not affect GYPP. These results indicate that qGN1c from 9311 can enhance grain yield with no significant negative effects on other agronomic traits. Therefore, qGN1c may be a favorable allele for breeding high-yield rice cultivars.

Fig. 5
figure 5

The genotype and phenotype of NIL-qGN1c9311. a Graphical karyotype showing the physical location of the 9311 donor chromosomal segment on Chr. 1 in NIL-qGN1c9311. Comparisons of gross morphology of mature plants (b) and panicles (c), grain length (d), and grain width (e) in ‘NP’ and NIL-qGN1c9311. Scale bars = 10 cm in b, 2 cm in c, and 1 cm in d and e

Full size image
Table 3 Agronomic traits for ‘NP’ and NIL-qGN1c9311 plants grown in two locations in 2018
Full size table

Discussion

Grain yield is a complex trait that is controlled by a combination of GN, TGW, SSR, and TN; GN is typically the largest and most important factor for improvement of grain yield in rice breeding. Similarly, GN is also a complex trait that is likely to be affected by PBN, SBN, and PL. In addition to the influences of genetic variations, environmental conditions may also have a significant effect on this trait, which makes it challenging to characterize GN (Xing et al. 2002). Consistent with the complexity of GN, a large number of genes/QTLs have been inferred to be involved in the regulation and control of GN. To date, several different mapping populations consisting of F2, recombinant inbred lines (RILs), backcross inbred lines (BILs), and CSSLs, have been used to detect QTLs for complex traits, and CSSL populations have been shown to be ideal permanent genetic resources for identification, fine mapping, and cloning of QTLs underlying complex traits (Ali et al. 2010; Qi et al. 2018; Tang et al. 2018; Xing et al. 2008; Zhou et al. 2009). It is widely accepted that mapping populations developed from parental lines with significant differences in the targeted traits are more effective for QTL identification (Ashikari et al. 2005). Oryza sativa indica and Oryza sativa japonica are two subspecies of rice. Genetic variations between cultivars of the two subspecies have been widely used in rice breeding (Qi et al. 2018). ‘NP’ and 9311 are representative japonica and indica rice cultivars, respectively, which display characteristics typical of the two subspecies. Hence, the CSSL population produced by introgressing 9311 genomic fragments into ‘NP’ was ideal for identifying and characterizing QTLs for GN and GN-related traits in our study. In fact, this population was used previously to dissect other grain yield related traits, such as TGW, GL, and GW (Xu et al. 2017).

In our study, we used the CSSL population to dissect the QTLs for panicle-related traits. A total of 25 QTLs for panicle-related traits were identified based on data collected over 2 years. A comparison of the location of QTL regions identified in the 2 years should clearly show which QTLs could be stably inherited. qPL5, qSBN1, qSBN6, qGN1c, and qGN6 were found to be co-localized; qPL6-1 and qPL6-2 or qPL6-3 and qPL6-4 were close to each other with similar additive effects, indicating that these QTLs are stably inherited loci. qSBN1 and qGN1c, as well as qSBN6, qGN6, and qPL6-4, were found to be located in the same regions, which is consistent with the correlation analysis and indicates that the corresponding gene may regulate several panicle-related traits. Three QTLs (qSBN6, qGN6, and qPL6-4) co-localize with the previously reported QTL qPL6 (Sun et al. 2017), suggesting that the QTLs identified in this study are reliable. Moreover, several new QTLs, such as qPL5, qPL6-1, and qPL6-2, were identified in this study, and these will further our understanding of the mechanisms that control GN in rice.

qGN1c identified here is a major QTL for GN, and it mapped to the distal region of the genome containing the cloned gene Gn1a (Ashikari et al. 2005). Genotyping using the gene-tagged marker (Gn1a-M) for Gn1a revealed that the CSSL line N2 carrying qGN1c does not contain the 9311-type Gn1a. By using the population generated from N2 and ‘NP’, the qGN1c for GN was fine mapped to an interval of ~ 379-kb on Chr. 1, which is in the upper region of Gn1a. To date, several QTLs located near Gn1a have been reported, such as SPP1 and Gn1b. SPP1 has been fine mapped to the down region of Gn1a indicating that qGN1c and SPP1 are two different genes (Liu et al. 2009). Gn1b is located in the upper region of Gn1a, but the actual physical location of Gn1b is ambiguous (Ashikari et al. 2005). Thus, further study is needed to determine whether or not qGN1c and Gnlb are identical genes.

Currently, although a number of QTLs/genes for GN have been mapped and/or cloned, many of them show pleiotropic effects on other important agronomic traits. For example, The QTL qGN4.1 for GN has pleiotropic effects on agronomic traits such as the number of tillers per plant and flag leaf length and width (Deshmukh et al. 2010). The QTL qSPP2.2, derived from O. longistaminata, showed a positive additive effect on GN, but decreased the number of tillers per plant (Kaur et al. 2018). The QTL TSN12.2 for total spikelet number affects the morphology of both leaves and internodes (Sasaki et al. 2017). qSPP7/DTH7/Ghd7.1, Ghd7, and DTH8/Ghd8, three major cloned QTLs for grain yield, have also been reported to affect heading date and leaf size (Gao et al. 2014; Liu et al. 2013; Tang et al. 2018; Wei et al. 2010; Weng et al. 2014; Xing et al. 2008; Yan et al. 2011). IPA1, DEP1, and the fine-mapped QTL qGN4 for grain number simultaneously regulate panicle size (Huang et al. 2009; Jiao et al. 2010; Zhou et al. 2018). The genome of NIL-qGN1c9311 harbors a ~ 1.51-Mb chromosomal fragment that was introgressed from 9311 into the ‘NP’ background, and this NIL shows increased GN without significant negative effects on PH, HD, SSR, and TN. The GYPP of NIL-qGN1c9311 was thereby increased by ~ 13% compared with ‘NP’, suggesting the potential of this locus for grain yield improvement in rice breeding. Taken together, these findings extend our understanding of the factors that control GN and also provide a tool to improve grain yield in rice production.