Characterization of QTLs for diameter in panicle neck and substitution mapping of qDPN5/qVBN5.2 and qVBN6 in rice (Oryza sativa L.)

Abstract

The vascular bundle system in the panicle neck of rice (Oryza sativa L.) connects the culm to the panicle and transports assimilates. The number of vascular bundles in the panicle neck (VBN) is correlated with the diameter of the panicle neck (DPN), but there are few reported QTLs for DPN. We conducted quantitative trait locus (QTL) analysis using recombinant inbred lines (RILs) derived from a cross between ‘Asominori’ and ‘IR24’ and detected three QTLs—qDPN5, qDPN6, and qDPN11—on chromosomes 5, 6, and 11. The qDPN5, qDPN6, and qDPN11 were in the same position as the QTLs for VBN reported in previous studies. Within the RILs, there was a significant positive correlation between DPN and VBN. In segregating populations, each QTL had a distinct effect on both values. Analysis of chromosome segment substitution lines showed that qDPN5 and qDPN11 affected DPN and qDPN6 affected VBN. Through substitution mapping, we narrowed down the region of qDPN5 and qVBN5.2 to 960 kbp between KNJ8 Indel385 and RM18926, and the region of qVBN6 to 750 kbp between C5 Indel5756 and KNJ8 Indel493. Due to the weak effect of qDPN6 in the ‘IR24’ genetic background, the location of qDPN6 could not be determined.

Introduction

Vascular bundles in rice (Oryza sativa L.) play an important role in transporting photosynthates, nutrients, and water throughout the plant (Lucas et al. 2013). Vascular bundles in the panicle neck of rice are critical for the transport of photosynthates from source to sink (Cui et al. 2003, Fukuyama et al. 1999). To improve rice production, it is essential to increase both source and sink sizes, as well as the capacity for translocation. When the vascular bundle number in the panicle neck (VBN) is limited, assimilate transport and grain filling are restricted (Peng et al. 1999, Xu et al. 2005). The diameter of the panicle neck (DPN), which is highly correlated with VBN, strongly affects source-to-sink matter transfer (Lee et al. 1992, Xu et al. 2005). In addition, DPN is positively correlated with sink size in the form of primary and secondary branch numbers and spikelet number (Lee et al. 1992, Liao et al. 2021, Liu et al. 2008, Xu et al. 2005).

Subspecies indica rice generally has a higher VBN than japonica rice (Fukuyama and Takayama 1995, Fukuyama et al. 1999, Liu et al. 2016, Zhai et al. 2018), with a varying ratio of VBN to primary branch number (Fukuyama and Takayama 1995, Fukuyama et al. 1999). VBN is influenced by both genetic and environmental factors such as nitrogen availability and plant density. Through the use of segregating populations such as recombinant inbred lines (RILs) and doubled haploid lines (DHs), genetic factors for VBN have been detected, and genome-wide association studies (GWAS) have identified associated regions (Bai et al. 2012, Cui et al. 2003, Liao et al. 2021, Nguyen et al. 2023, Sasahara et al. 1999, Zhai et al. 2018, Zhang et al. 2002). Some of the genes responsible for VBN have been cloned and isolated including ABERRANT PANICLE ORGANIZATION 1 (APO1) on chromosome (Chr.) 6 (Terao et al. 2010), NARROW LEAF 1 (NAL1) on Chr. 4 (Fujita et al. 2013, Qi et al. 2008), and LVB9/DENSE AND ERECT PANICLE 1 (DEP1) on Chr. 9 (Fei et al. 2019).

QTLs for DPN have been detected on Chrs. 4, 6, and 11 in RILs derived from ‘Zhenshan 97B’ and ‘IRAT109’ upland rice (Liu et al. 2008). Liao et al. (2021) detected QTLs for DPN on Chrs. 1–7, 11, and 12 by GWAS of introgression lines derived from combinations of ‘Zhenshan 97’, ‘Nipponbare’, ‘Huanghuazhan’, ‘Uz-Rosz 275’, and ‘Celiaj’. Jeon et al. (2018) narrowed down FID2 for DPN on Chr. 2 in a segregating population derived from an interspecific hybrid between japonica ‘Hwaseongbyeo’ and Oryza grandiglumis.

We identified three stable QTLs for VBN on Chrs. 5, 6, and 11 in RILs derived from a cross between japonica ‘Asominori’ and indica ‘IR24’ (Nguyen et al. 2023) and evaluated their effects alone and in pairs using chromosome segment substitution lines (CSSLs) and pyramided lines (PYLs) with an ‘IR24’ genetic background. We could not identify their precise locations or clarify the correlation between VBN and DPN. Here, we had three objectives: (1) to reveal genetic factors that control DPN; (2) to reveal the relationship between VBN and DPN; and (3) to delimit the regions of the QTLs on Chrs. 5 and 6. We conducted QTL analysis for DPN using RILs and characterized QTL effects using CSSLs and PYLs; revealed the relationship between VBN and DPN using RILs and an F₂ population; and delimited the region of QTLs for VBN and DPN by substitution mapping.

Materials and Methods

Plant materials

To conduct QTL analysis for DPN, we used 71 RILs derived via single-seed descent from a cross between japonica ‘Asominori’ and indica ‘IR24’ (Tsunematsu et al. 1996). To evaluate the effect of the three QTLs in 2020 and 2021, we used three CSSLs—IAS14 (qDPN11), IAS30 (qDPN5) and IAS39 (qDPN6)—carrying target chromosomal segments of ‘Asominori’ in the ‘IR24’ genetic background (Kubo et al. 2002). Since the locations of these QTLs for DPN coincided with those of QTLs for VBN, we used PYLs developed by Nguyen et al. (2023) to evaluate DPN for each combination. For evaluating the effects of DPN on PYL with three QTLs, F₁ derived from the crosses between PYLs with two QTLs were generated (Nguyen et al. 2023) and was self-pollinated to produce F₂. Using SSR markers (Supplemental Table 1), one plant with the ‘Asominori’ homozygous allele at three QTLs was selected among the 96 F₂ plants. In 2022, the F₃ line from selected F₂ plant was assessed for VBN and DPN. To evaluate the effect of ‘IR24’ alleles of QTLs with ‘Asominori’ genetic background, we used three CSSLs—AIS38 (carrying qDPN5), AIS49 (carrying qDPN6), and AIS76 (carrying qDPN11)—carrying target chromosomal segments of ‘IR24’ in the ‘Asominori’ genetic background (Kubo et al. 2002).

Materials for substitution mapping

We used two F₂ populations (168 plants each) derived from IAS30/IR24 and IAS39/IR24 to confirm QTLs for DPN (Supplemental Fig. 1). To conduct substitution mapping, we selected plants with recombination around qDPN5 and qDPN6 from each F₂ population and self-pollinated them to produce F₃ lines. For substitution mapping of qDPN5, we analyzed 54 F₃ lines using 10 additional markers and selected 19 plants with homozygous recombination around qDPN5. Following self-pollination, we developed 19 F₄ lines and measured VBN and DPN. Finally, we measured VBN of 19 F₅ lines to delimit the location of qDPN5. For substitution mapping of qDPN6, we analyzed 44 F₃ lines using 9 additional markers, and selected 40 plants with homozygous recombination around qDPN6. Following self-pollination, we developed 40 F₄ lines and measured VBN and DPN. We measured VBN of 12 F₅ lines to delimit the location of qDPN6.

Evaluation of VBN and diameter of the panicle neck

The plants were grown in a paddy field at Saga University (33°14ʹ32ʺN 130°17ʹ24ʺE) from May to October. Seedlings were transplanted 28 days after germination at one plant per hill, at 20 cm between hills and 25 cm between rows. Inorganic fertilizer was applied at 40 kg/ha N, 17.5 kg/ha P, and 33 kg/ha K. Each F₃, F₄, and F₅ line consisted of 16–24 plants in 3 rows. Five plants were harvested from the second or third rows (excluding plants at the end of each row) 2 weeks after flowering of the tallest panicle. Fresh peduncles were cut less than 1 mm about 1 cm below the panicle base node and observed VBN and DPN under a microscope. Microscope was used to capture images of the oval cross-section of the panicle neck node. The maximum width of the oval cross-section was measured as DPN using Image J (National Institutes of Health, Bethesda, MD, USA) software. For each line, 5–10 samples of VBN and DPN were recorded.

DNA extraction and genotyping

Fresh leaf (2–4 cm) was freeze-dried for 48 h, and DNA was extracted by the potassium acetate method (Dellaporta et al. 1983). GoTaq Master Mix (Promega) was used for PCR amplification in 35 cycles of 30 s at 96°C, 30 s at 55°C, and 30 s at 72°C, followed by a final extension at 25°C for 1 min. The PCR products were separated by electrophoresis in 4% agarose gel containing 0.5 μg/mL ethidium bromide in 0.5 TBE buffer at 200 V for 1–2 h. Using genotyping data from RFLP markers (Tsunematsu et al. 1996), we conducted QTL analysis for DPN. QTLs were confirmed using the same markers used for identifying the VBN QTL in the F₂ plants of IAS30/IR24 and IAS39/IR24 (Nguyen et al. 2023). For substitution mapping in F₃ recombinant lines, we used polymorphic SSR and indel markers (McCouch et al. 2002, Yonemaru et al. 2015). The F₃ plants derived from 54 F₂ plants recombinant around qDPN5 between RM7081 and RM7446 were genotyped using 10 markers. Similarly, the F₃ plants derived from 44 F₂ plants recombinant around qDPN6 near RM6395 and RM20546 were genotyped using nine markers (Supplemental Table 2).

QTL analysis

Composite interval mapping (CIM) was performed in Windows QTL Cartographer v. 2.5 software (Wang et al. 2012). The significance threshold was determined by 1000 permutation tests, with a logarithm of odds (LOD) score of 3.0 at P < 0.05.

Statistical analysis

One-way ANOVA was conducted to compare the mean values of VBN and DPN of homozygous recombinants. Dunnett’s test was used to analyze the phenotypic differences between homozygous recombinant lines and ‘IR24’. The Tukey–Kramer comparison test was conducted to compare differences in DPN among lines.

Results

Identification of QTLs for diameter of panicle neck in RILs

DPN was 1.70 mm in ‘Asominori’ and 2.55 mm in ‘IR24’ (Supplemental Fig. 2). The frequency distribution of DPN in RILs ranged from 1.7 to 2.7 mm and was continuous, indicating that multiple genetic factors control DPN. QTL analysis detected three QTLs for DPN: qDPN5 on Chr. 5, qDPN6 on Chr. 6, and qDPN11 on Chr. 11 (Table 1). The ‘Asominori’ alleles at those QTLs decreased DPN, with proportion of variance explained (PVE) of 17.3% in qDPN5, 38.1% in qDPN6, and 11.7% in qDPN11 (total = 67.1%).

Table 1. Detection of QTLs for diameter of panicle neck in RILs.

QTL	Chr.	Marker interval	Physical distance (Mbp)a	LOD	Additive effectb	R2 (%)
qDPN5	5	RM18897–RM18910	23.88–24.13	6.3	–0.11	17.3
qDPN6	6	C962–Ky11	28.64	9.5	–0.17	38.1
qDPN11	11	Xnpb189-1–C718	2.03–1.89	4.7	–0.10	11.7

^a On ‘Nipponbare’ genome sequence.

^b Effect of ‘Asominori’ allele.

Validation of single-QTL effect on DPN in ‘IR24’ genetic background

To assess the effect of single QTLs for DPN, we measured the DPN of CSSLs carrying the ‘Asominori’ alleles, IAS30 (qDPN5), IAS39 (qDPN6) and IAS14 (qDPN11) in 2 years. In 2020, the ‘Asominori’ alleles significantly decreased DPN by 0.41 mm in IAS14, by 0.46 mm in IAS30 and by 0.29 mm in IAS39, relative to ‘IR24’ (2.74 mm, Table 2). In 2021, they significantly decreased DPN by 0.38 mm in IAS14, by 0.56 mm in IAS30 and by 0.19 mm in IAS39, relative to ‘IR24’ (2.57 mm, Table 2). The effects were consistent between years. Additionally, the VBN of IAS30, IAS39, and IAS14 was characterized and ‘Asominori’ alleles significantly decreased VBN by 5.2 in IAS14, by 6.4 in IAS39, and by 1.6 in IAS30, relative to ‘IR24’ (22.2, Table 2).

Table 2. Validation of effect of a single QTL for diameter of panicle neck in CSSLs with ‘IR24’ genetic background.

Line	QTL	Diameter of panicle neck (mm)		Vascular bundle number
		2020	2021	2021
IR24		2.74 ± 0.17	2.57 ± 0.12	22.2 ± 2.4
IAS30	qDPN5	2.28 ± 0.10**	2.01 ± 0.06***	20.6 ± 4.4
IAS39	qDPN6	2.45 ± 0.21*	2.38 ± 0.10*	15.8 ± 0.8**
IAS14	qDPN11	2.33 ± 0.19**	2.19 ± 0.05***	17.0 ± 2.8*

Asterisks: P-values at *5%, **1%, and ***0.1% by Dunnett’s test.

Evaluation of pyramiding effects of QTLs for DPN in ‘IR24’ genetic background

To assess the interactions of QTLs associated with DPN, we used PYLs with multiple QTLs. Table 3 indicates that DPN of PYL1 (qDPN5 + qDPN11) was 2.02 mm in 2020 and 1.89 mm in 2021, similar to that of parental lines IAS30 (qDPN5) and IAS14 (qDPN11), and ‘Asominori’ but significantly smaller than that of ‘IR24’ in both years. DPN of PYL2 (qDPN6 + qDPN11) was 2.26 mm in 2020 and 2.04 mm in 2021, similar to that of IAS14 but significantly smaller than that of ‘IR24’ in both years, and similar to that of IAS39 (qDPN6) in 2020 but significantly smaller in 2021. DPN of PYL3 (qDPN5 + qDPN6) was 2.06 mm in 2020 and 1.98 mm in 2021, similar to that of IAS30 and ‘Asominori’ but significantly smaller than those of IAS39 and ‘IR24’ in both years, and significantly larger than that of ‘Asominori’ in 2020 but similar in 2021. In 2022, the DPN tendency in PYL1-2 was consistent with 2020 and 2021, while PYL3 showed the same DPN as the parental line IAS39. In 2022, DPN of PYL4 carrying three QTLs for DPN (qDPN5+6+11) was 1.79 mm, which was similar to that of other PYLs carrying qDPN5+11, qPDN6+11, and qDPN5+6 (Table 3). DPN of PYL4 was significantly different from IAS39, IAS14, and ‘IR24’. Additionally, VBN of PYL4 (qDPN5+6+11) was 11.2, similar to that of PYL2 (qDPN6+11) and ‘Asominori’ (Supplemental Fig. 3). The VBN of PYL4 was significantly lower than that of IAS30, IAS39, IAS14, PYL1 (qDPN5+11), and PYL3 (qDPN5+6).

Table 3. Effects of diameter of the panicle neck in pyramided lines with ‘IR24’ genetic background in 2020, 2021 and 2022.

Line	QTL	Diameter of the panicle neck (mm) (Mean value ± SD)
		2020	2021	2022
IAS30	qDPN5	2.23 ± 0.11bc	1.94 ± 0.15ab	1.97 ± 0.15bcd
IAS39	qDPN6	2.37 ± 0.15c	2.38 ± 0.11c	2.18 ± 0.19de
IAS14	qDPN11	2.25 ± 0.07bc	2.13 ± 0.07b	2.06 ± 0.09cde
PYL1	qDPN5+11	2.02 ± 0.16ab	1.89 ± 0.13ab	1.87 ± 0.11bc
PYL2	qDPN6+11	2.26 ± 0.13bc	2.04 ± 0.14b	1.83 ± 0.09abc
PYL3	qDPN5+6	2.06 ± 0.11ab	1.98 ± 0.06ab	1.98 ± 0.09bcd
PYL4	qDPN5+6+11	–	–	1.79 ± 0.14ab
Asominori		1.81 ± 0.09a	1.78 ± 0.17a	1.61 ± 0.13a
IR24		2.74 ± 0.17d	2.57 ± 0.12c	2.23 ± 0.12e

Values with the same letter are not significantly different between genotypes in each year 2020, 2021 and 2022 by Tukey–Kramer multiple comparison test (P < 0.05) in 2020, 2021 and 2022.

Confirmation of QTLs for DPN

To confirm detected QTLs for DPN in RILs, we conducted QTL analysis using populations segregating at a single QTL. The frequency distribution of DPN in the F₂ population derived from IR24/IAS30 ranged from 1.8 to 3.0 mm (Supplemental Fig. 4A). A single QTL, qDPN5, was detected on Chr. 5, with a PVE of 15.9%. The ‘Asominori’ allele decreased DPN by 0.04 mm (Table 4). Similarly, the frequency distribution of DPN in the F₂ population derived from IR24/IAS39 ranged from 2.0 to 3.1 mm (Supplemental Fig. 4B). A single QTL, qDPN6, was detected on Chr. 6, with a PVE of 33.7%. The ‘Asominori’ allele decreased DPN by 0.06 mm (Table 4). These results confirm the presence of these two QTLs for DPN, and the ‘Asominori’ alleles decreased DPN. Based on the nearest markers to qDPN5, the DPN of plants with homozygous for ‘Asominori’ was 2.21 mm and the DPN of plants with homozygous for ‘IR24’ was 2.45 mm (Supplemental Fig. 4A). Based on the nearest markers to qDPN6, DPN of plants with homozygous for ‘Asominori’ was 2.23 mm and the DPN of plants with homozygous for ‘IR24’ was 2.58 mm (Supplemental Fig. 4B). Since these QTLs overlap with QTLs detected for VBN in our previous study (Nguyen et al. 2023), suggesting a relationship between DPN and VBN, we conducted correlation analysis between VBN and DPN in the RILs and F₂ populations. In the RILs derived from a cross between ‘Asominori’ and ‘IR24’, there was a positive correlation (r = 0.55, P < 0.001; y = 6.04x + 3.00; Supplemental Fig. 5). In the F₂ populations derived from IAS30/IR24 and IAS39/IR24, there were also positive correlations (r = 0.44 and 0.6, P < 0.001).

Table 4. Confirmation of QTLs for diameter of panicle neck in F₂ populations derived from CSSLs.

Population	QTL	Chr.	Marker interval	Physical distance (Mbp)a	LOD	Additive effectb	Dominant effect	R2 (%)
F2 (IAS30/​IR24)	qDPN5	5	RM7081–RM7446	24.59–25.02	5.5	–0.12	–0.04	15.9
F2 (IAS39/​IR24)	qDPN6	6	RM6395–RM20546	26.00–27.40	13.4	–0.18	–0.06	33.7

^a On ‘Nipponbare’ genome sequence.

^b Effect of ‘Asominori’ allele.

Substitution mapping of qDPN5 and qVBN5.2

In our previous study, we identified two QTLs, qVBN5.1 and qVBN5.2, in F₂ populations derived from IAS30/IR24 (Nguyen et al. 2023). qVBN5.2 was co-located with qDPN5 between RM7081 and RM7446. To delimit qDPN5 and qVBN5.2, we conducted substitution mapping. VBN of lines carrying the ‘Asominori’ segment varied from 19.2 to 23.9 in F₄ and from 16.2 to 22.6 in F₅ (Fig. 1). DPN of F₄ lines carrying the ‘Asominori’ segment varied consistently with VBN, from 1.91 to 2.25 mm. Among them, VBN of lines 4 and 23 did not differ from that of ‘IR24’ in F₄ or F₅, but DPN of these lines was significantly smaller than that of ‘IR24’. The types of ‘IR24’ and ‘Asominori’ of each line in Fig. 1 were classified based on the VBN. On the basis of the recombination between KNJ8 Indel385 and RM18926 in lines 4 and 23, we delimited qDPN5 and qVBN5.2 to ~960 kbp between these markers on the ‘Nipponbare’ genome sequence.

Fig. 1.

Chromosomal locations of qVBN5.2 and qDPN5 by substitution mapping. Boxes indicate genotypes homozygous for ‘Asominori’ and ‘IR24’, or recombinant. ‘IR’ and ‘Aso’ types mean ‘IR24’ and ‘Asominori’. ‘Asominori’ or ‘IR24’ types were classified based on VBN in each line. Asterisks indicate significant differences from ‘IR24’ at P-values of *5%, **1%, and ***0.1% by Dunnett’s multiple comparison test.

Substitution mapping of qVBN6 (but not qDPN6)

In our previous study, we identified qVBN6 for VBN on Chr. 6 in an F₂ population derived from IAS39/IR24, near marker RM20546 for qDPN6. To delimit qDPN6 and qVBN6, we conducted substitution mapping. VBN of lines carrying the ‘Asominori’ segment varied from 13.2 to 22.4 in F₄ and from 13.8 to 21.2 in F₅, but DPN did not differ significantly among these lines (Fig. 2). Among them, VBN of lines 116, 102, 113, 129, 132, and 139 was significantly lower than that of ‘IR24’, whereas that of lines 133, 112, 103, and 109 was similar to that of ‘IR24’ in F₄ and F₅. Similarly, ‘IR24’ type and ‘Asominori’ type of each lines in Fig. 2 were classified based on the VBN. On the basis of the recombination between C5 Indel5756 and KNJ8 Indel493 in lines 133, 116, 139, and 112, we delimited qVBN6 to ~750 kbp between these markers on the ‘Nipponbare’ genome sequence. Although DPN tended to be reduced in lines with low VBN, we could not estimate the location of qDPN6.

Fig. 2.

Chromosomal locations of qVBN6 by substitution mapping. Boxes indicate genotypes homozygous for ‘Asominori’ and ‘IR24’, or recombinant. ‘IR’ and ‘Aso’ types mean ‘IR24’ and ‘Asominori’. ‘Asominori’ or ‘IR24’ types were classified based on VBN in each line. Asterisks indicate significant differences from ‘IR24’ at P-values of *5%, **1%, and ***0.1% by Dunnett’s multiple comparison test; ns, not significant.

Discussion

Several loci for DPN have been identified in segregating populations and by GWAS (Jeon et al. 2018, Liao et al. 2021, Liu et al. 2008). So far, one QTL for DPN, qFID2, on Chr. 2, has been identified by substitution mapping (Jeon et al. 2018), but its gene function has not been determined. Here, we identified three QTLs for DPN, the locations of which overlap those reported previously: qDPN5 at the same location as a QTL for DPN at 18.38–29.16 Mbp on Chr. 5, qDPN6 at the same location as a QTL for DPN at 28.0–31.2 Mbp on Chr. 6, and qDPN11 near a QTL for DPN at 0.8–2.1 Mbp on Chr. 11 (Liao et al. 2021). The three QTLs had a total PVE of 67.1% (Table 1), so other genetic factors might also contribute. These results will help explain the genetic basis of DPN in rice, and could lead to precise mapping and cloning of related loci.

The locations of the QTLs for DPN detected here correspond to known QTLs for VBN: qVBN5 at ~23.0 Mbp, qVBN6 at ~28.6 Mbp, and qVBN11 at ~3.8 Mbp (Nguyen et al. 2023). The qDPN11 was co-located with qVBN11 at ~2.45 Mpb on Chr. 11 and substitution mapping and characterization of qDPN11 and qVBN11 are currently conducting for future publication. qDPN5 and qDPN6 were confirmed in the F₂ population (Supplemental Fig. 4, Table 4). The results of QTL analysis showed that these QTLs exhibited incomplete dominance effects because the dominance values were smaller than the values of additive effect (Table 4). In addition, the frequency distribution of DPN on F₂ populations based on plants with homozygous for ‘Asominori’ and ‘IR24’ and heterozygous suggests that there might be one QTL controlling DPN in each population (Supplemental Fig. 4). We delimited the regions of qVBN5.2 and qDPN5 to ~960 kbp on the ‘Nipponbare’ genome sequence (Fig. 1). There are no reported genes for DPN here, and this locus might have pleiotropic effects on VBN and DPN. qVBN6 and qDPN6 were delimited to ~750 kbp, near APO1 (Os06g0665400) at ~27.46 Mbp on Chr. 6. APO1 enhances the development of vascular bundles, promoting carbohydrate translocation to the panicle (Terao et al. 2010), so qVBN6 might be APO1. In substitution mapping of qDPN6, we found no significant differences in DPN among F₄ lines (Fig. 2). However, DPN tended to be lower in F₄ lines with low VBN, may be because DPN depends more on environmental conditions whereas VBN depends more on genetic control. Takeoka (1977) mentioned that the histogenesis of the inner VBN or VBN linked to the development of young panicles. Therefore, the gene on this location mainly controlled VBN as the histogenesis of the VBN at development of young panicles. Subsequently, the histogenesis of the VBN might influence to the size of DPN. Therefore, VBN could be clearly characterized in the F₄ and F₅ lines, while DPN was unclear. It will be necessary to confirm effects of qDPN6 as future study.

We verified each detected QTL for DPN and VBN on its respective chromosome in CSSLs with the ‘IR24’ genetic background. qDPN5 and qDPN11 had a strong effect on DPN but a weak effect on VBN; qDPN6 had a weak effect on DPN but a strong effect on VBN (Tables 2, 3, Supplemental Fig. 3). To examine QTL interactions associated with DPN, we used PYLs carrying pairs of QTLs. PYL4 (qDPN5 + qDPN6 + qDPN11) had reduced DPN, similar to that of ‘Asominori’ (Table 3). This result implies additive effects among QTLs. It was similar to the additive effect on VBN between qVBN6 and qVBN11 (Nguyen et al. 2023). A common positive correlation between DPN and VBN is associated with the capacity to transport matter from leaf to stem (Lee et al. 1992, Liao et al. 2021, Liu et al. 2008, Xu et al. 2005). VBN is significantly positively correlated with panicle structure, such as primary branch number, secondary branch number, spikelet number, and yield-related traits (Fukuyama and Takayama 1995, Liao et al. 2021, Nguyen et al. 2023, Sasahara et al. 1999, Zhai et al. 2018). We also found a correlation between DPN and VBN in the RILs (Supplemental Fig. 5) and F₂ populations. The results of our study using the ‘IR24’ genetic background suggest that combining QTLs for DPN can reduce DPN and VBN and the reduction of VBN would be reduced translocation, the filled spikelets on the panicle, and yield potential.

In previous study, Nguyen et al. (2023) found three stable QTLs for VBN, qVBN5, qVBN6 and qVBN11 on RILs of the cross between ‘Asominori’ and ‘IR24’. These QTLs for VBN were evaluated the effects under ‘IR24’ and ‘Asomiori’ genetic backgrounds. To confirm the relationships among VBN and yield-related traits under japonica genetic background, VBN and yield-related traits of three CSSLs carrying ‘IR24’ alleles in the ‘Asominori’ genetic background (AIS38, AIS49, and AIS76) were observed (Nguyen et al. 2023). Spikelet number per panicle and VBN of AIS49 (carrying ‘IR24’ allele at qDPN6) and AIS76 (carrying ‘IR24’ allele at qDPN11) were increased relative to ‘Asominori’. These results correspond to the effects of APO1 in a previous study: the ‘Habataki’ allele of APO1 increased primary branch number and VBN, increasing grain yield per plant by enhancing spikelet number and carbohydrate translocation to panicles (Terao et al. 2010). Additionally, DPN of AIS38 (carrying ‘IR24’ allele at qDPN5), AIS49, and AIS76 in the ‘Asominori’ genetic background slightly increases but no significant difference from ‘Asominori’ (data not shown). These QTLs had tendency of increasing VBN, DPN and spikelet number of CSSLs with the ‘Asominori’ genetic background. The QTLs for VBN plus DPN on Chrs. 5 and 6 could be introduced into other japonica cultivars and used in developing indica–japonica cross cultivars and higher-yielding cultivars to improve yields through marker-assisted breeding.

Author Contribution Statement

All authors contributed to the study conception and design. HTLN and YA prepared the material, collected data, and performed analyses. HTLN, SHZ, and DF wrote the manuscript.