INTRODUCTION

Rice (Oryza sativa L.) is mainly divided into subgroups namely indica, and japonica. Indica rice is most prominent in tropical to sub-tropical regions and characteristically its grains are long and have low amylose content. In contrast, japonica is dominantly grown in temperate regions and has short to medium grains with low amylose content (Juliano & Villareal, 1993;Kim, 2019). In general, rice is a short-day (SD) plant wherein flowering is promoted by a reduced day-length. Days to heading, also called as heading date or flowering time in rice, is an important trait of rice for local domestication and is also one of the priority traits in rice breeding programs because it influences rice grain yield (Wei et al., 2010;Yan et al., 2011;Gao et al., 2014;Liu et al., 2015) and other traits such as plant architecture and duration of vegetative phase (Xue et al., 2008; Hori et al., 2015) as well as cropping systems (Toriyama,1970).

The genetics and molecular biology of flowering has been extensively studied due to its economic importance and there are 40 identified genes in rice (Kim et al., 2018). The regulatory networks among the flowering associated genes are monitoring environment factors (mostly daylength in rice) and eventually activating the expression of the florigen genes Heading date 3a (Hd3a) and/or RICE FLOWERING LOCUS T 1 (RFT1) for rice plants to flower (Tamaki et al., 2007;Komiya et al., 2008). The expression of the florigen genes is tightly regulated by several upstream genes including Hd1 (Yano et al., 2000), Ehd1 (Doi et al., 2004), Ghd7 (Xue et al., 2008), and DTH8 (Wei et al., 2010).

Japonica rice is sought for due to a market that favors sticky and soft texture. However, its expansion is limited by temperate conditions, and so regional adaptation in the tropics could be a potential driver. However, tropical adaptation of temperate japonica exhibits extreme early flowering. Temperate japonica varieties typically exhibit very early flowering (~4 weeks after seeding) in short day conditions compared to indica lines. This phenotype at extreme is correlated to reduced yield by preventing overall biomass increase, and normal panicle development. Varied interplay of flowering factors under different photoperiodic conditions limit the speed of japonica rice to adapt to tropics and is one of the main challenges of the Grain Utilization Value Added (GUVA) Project. GUVA is a collaborative work between South Korea and the International Rice Research Institute (IRRI) to breed high quality and high-yielding temperate japonica rice varieties in the tropics at IRRI since 1992. To efficiently develop lines with optimal flowering, it is important to identify, and interdependently use heading date genetic markers on GUVA’s germplasm on top of its conventional breeding strategy.

This study used RILs (Recombinant Inbred Lines) of a cross between Ilpum (temperate japonica) and Zenith (indica) to identify possible QTLs to study flowering time genes under SD and LD. Putative QTLs for heading date were identified using SSR markers under the natural short day conditions of Philippines and natural LD conditions of South Korea. It will serve as a case population to identify putative heading QTLs that determine japonica rice adaptation in the tropics.

METHODS

Plant material and crop establishment

One hundred eighty RILs (F_2:9) were developed from a cross between Ilpum, a popular South Korean japonica variety, and Zenith, an indica variety. The population was developed in the experimental fields of the National Institute of Crop Science in Milyang, Korea. Ilpum, Zenith and RILs were established in the National Institute of Crop Science (NICS), Milyang City, Republic of Korea in 2018 and in the International Rice Research Institute (IRRI) headquarters, Philippines in the 2021 Dry Season. Day length during the rice cropping seasons in Korea was about 12.5 hours in June, and that in the Philippines was around 11.5 hours in January (dry season).

One row of twelve plants was transplanted 21 days after seeding, with row-plant spacing of 30 × 15 cm. Appropriate field practices and management was employed according to the standard cultivation method of NICS and IRRI.

Measurement of flowering time and panicle traits, and statistical analysis

Heading date was observed and recorded when 50% of the plot’s panicles had emerged (50%HD). Days to heading (dates from sowing to flowering time) under LD and SD were observed in South Korea and Philippines, respectively. Twenty four panicles were collected for each RIL and for each panicle, number of spikelets, and number of filled grains was counted.

Statistical differences between means were analyzed using Duncan’s multiple range test after one-way analysis of variance (ANOVA). The level of significance was designated as p < 0.05 and was determined using the SAS Enterprise Guide 4.3 programs (SAS Institute Inc., Cary, NC, USA).

DNA extraction and polymerase chain reaction

Genomic DNA from young leaf tissue was prepared according to the CTAB method (Murray and Thompson, 1980) with minor modifications. Polymerase chain reac-tion (PCR) was performed in 25-μL reaction mixture containing 25 ng template DNA, 10 pmol of each primer, 10 × e-Taq reaction buffer, 25 mM MgCl₂, 10 mM dNTP mix and 0.02 U of SolGent e-Taq DNA polymerase (Sol- Gent, Daejeon, South Korea). The reaction conditions were set as follows: initial denaturation at 94°C for 2 min; 35 cycles of denaturation at 94°C for 20 s, annealing at 57 °C for 40 s and extension at 72°C for 40 s; and a final extension at 72°C for 7 min. The amplification products were electrophoresed on a 3% (w/v) agarose gel and visualized by ethidium bromide staining.

QTL identification for heading date

To identify QTL positions that influence heading, RM (Rice Microsatellite) Markers available at GRAMENE database (http://www.gramene.org) were employed. Heading date sets recorded under SD and LD conditions were separated for the analysis. Initial QTL analysis was performed as described by Lee et al. (2021). Polymorphic SSR markers (n = 164) that were evenly distributed on rice chromosomes were selected. These markers were used to construct a linkage map and for QTL analysis of the F_2:9 populations. The linkage map was constructed using Mapmaker/ Exp v.3.0, and the genetic distance was obtained using the Kosambi map function (Lander et al., 1987). Putative QTLs were detected using the composite interval mapping (CIM) function in WinQTLcart v.2.5 (WinQTL cartographer software) (Wang et al., 2007). A logarithm of the odds (LOD) ratio threshold of 3.0 was used to confirm the significance of a putative QTL.

Development of InDel markers near the qHD6-SD

To narrow down the chromosomal region of the qHD6-SD locus, we developed new InDel markers by comparing the sequences of the corresponding chromosomal regions followed by the whole genome sequencing of six rice varieties (five indica varieties: Habataki, ST12, NSIC Rc158, NSIC Rc222, NSIC Rc238 and one japonica: ST6) generated by Kim et al. (2016). The row sequencing reads were compared with the rice reference genome sequence (https:/ /rapdb.dna.affrc.go.jp/) using the IGV software (Robinson et al., 2011) and the eight potential polymorphic InDels markers between japonica and indica varieties were developed. Consequently, total of four markers out of eight showed polymorphism between Ilpum and Zenith.

RESULTS

Days to heading of the parental rice varieties

Days to heading on Ilpum and Zenith under SD (short day, Philippines) conditions were 48 and 79 respectively, while those of under LD (long day, Republic of Korea) were 98 and 102 (Figure 1, Figure 2). Difference of days to heading of Ilpum between under SD and LD was 50 days, while that of Zenith was 29 day.

QTL analysis and mapping of heading date using 180 F_2:9 RILs

A cross between these two contrasting varieties and 180 F_2:9 RILs and used to identify QTLs that could explain varied days to heading patterns under SD and LD. Days to heading in the 180 F_2:9 RILs under SD was 64.9 ± 6.1 days (ranging from 45 to 80 days), which was significantly different from that of under LD (103.8 ± 11.0 days, ranging from 83 to 125 days) (Figure 2).

For the QTL analysis, we selected 164 markers showing polymorphism between Ilpum and Zenith in 3% agarose gel from 1,150 RM markers (http://gramene) as described by Lee et al. (2021). Primary QTL mapping using the 180 F_2:9 RILs showed that a significant QTL associated with days to heading under SD was located between the SSR markers, RM19850 and RM19912 on chromosome 6, and it was designated qHD6-SD (Table 1, Figure 3).

The LOD score of qHD6-SD was 8.8, which accounted for 20.2% of the total phenotypic variation. Another QTL namely qHD6-SD those of under LD was located between RM2615 and RM527, which overlaps the position of the Hd1 gene (Yano et al., 2000) (Figure 4).

Secondary QTL analysis and sequence analysis of Hd1 gene

A finer location of qHD6-SD and qHD6-LD was determined by secondary QTL analysis with an additional four InDel (insertion/deletion) markers designed by using whole genome sequences of six varieties with the rice reference genome sequence and three RM markers in the qHD6-SD region (Table 2). Both qHD6-SD and qHD6-LD were shifted and delimited to the similar region between the InDel markers, HD1-RCM2 (9.263 Mbp) and HD1- RCM4 (9.361 Mbp), which harboring Hd1 gene (Yano et al., 2000) (Figure 5).

Effects of qHD6-SD on days to heading and spikelet development

Days to heading in the RILs harboring qHD6-SD^ZE (Zenith allele type of qHD6-SD, n=97) under SD was 68.6 ± 4.2 days (ranging from 53 to 68 days), which was significantly longer than those of harboring qHD6-SD^IL (Ilpum allele type of qHD6-SD) (60.5 ± 4.9 days, ranging from 45 to 71 days, n=81) (Figure 6). On the contrast, days to heading in the RILs harboring qHD6-SD^ZE under LD was 95.4 ± 5.1 days (ranging from 86 to 115 days), which was significantly shorter than those of harboring qHD6-SD^IL (113.8 ± 7.1 days, ranging from 83 to 125 days). Days to heading under SD condition showed significant correlation with no. of spikelet (r=0.699^**) and filled grains (r=0.685^**) per panicle. No. of spikelet per panicle in the RILs harboring qHD6-SD^ZE under SD was 131 ± 22.5, which was significantly greater than those of harboring qHD6-SD^IL (92 ± 24.3) (Figure 7). No. filled grain per panicle of the RILs harboring qHD6-SD^ZE was 106 ± 25.0, which was significantly greater than those of harboring qHD6-SD^IL (57 ± 22.2).

DISCUSSION

Japonica rice cultivation is majorly limited to temperate regions and this is due to its inferior adaptation in tropical areas. Tropics usually impose short day conditions that typically cause temperate japonica to exhibit extremely early flowering, low biomass accumulation, and finally inferior yields. It was reported that Korean temperate japonica flowered very quickly (~45 d), and the plants became stunted at the reproductive stage and showed very low grain yield, which caused by the reduced day-length according to the geographical transition from a high-latitude location, Korea, to a low-latitude location, the Philippines (Kim et al., 2018;Jeong et al., 2020). Therefore, prolonging days to heading is the foremost breeding objective to overcome these challenges in developing tropically adapted japonica rice. In rice, flowering time is regulated by the complex genetic mechanisms involving hundreds of QTLs and at least 14 cloned genes (Hori et al., 2016;Matsubara et al., 2018).

The regulatory networks among the flowering associated genes are monitoring environment factors (mostly daylength in rice) and eventually activating the expression of the florigen genes; Heading date 3a (Hd3a) and/or RICE FLOWERING LOCUS T 1 (RFT1) for rice plants to flower. Both Hd3a and RFT1 are mobile proteins that eventually cause the transition from the vegetative shoot apical meristem to the reproductive meristem which is called panicle initiation in rice (Tamaki et al., 2007;Komiya et al., 2008). The expression of the florigen genes is tightly regulated by several upstream genes. Heading date 1 (Hd1) promotes flowering by activating the transcription of the florigen genes under SD condition, but it suppresses the florigen genes under LD condition (Yano et al., 2000;Hayama et al., 2003). Early heading date 1 (Ehd1) encoding a B-type response regulator protein promotes the expression of Hd3a and RFT1 under both SD and LD conditions (Doi et al., 2004). DTH8 (QTL for days to heading) also delays flowering in LD condition by suppressing the expression of Ehd1 (Wei et al., 2010).

We identified two major QTLs, qHD6-SD and qHD6- LD, conferring days to heading under SD and LD conditions, respectively. Fine mapping revealed that both qHD6- SD and qHD6-LD was narrowed down into the similar 98 kb region between the InDel markers, HD1-RCM2 (9.263 Mbp) and HD1-RCM4 (9.361 Mbp), which harbors Hd1 gene (Yano et al., 2000). Ilpum (temperate japonica) showed extreme early heading and stunted plants under SD condition (Philippines), while it showed late heading and vigorous vegetative growth under LD (Republic of Korea), while Zenith (indica) exhibited similar growth under different day length. Furthermore, days to heading in the RILs harboring Zenith allele type of qHD6-SD under SD is significantly longer than those of harboring Ilpum allele type. On the contrary, days to heading in the RILs harboring Zenith allele type of qHD6-SD under LD was significantly shorter than those of harboring Ilpum allele type. These bi-functionality of qHD6-SD upon heading tendency strongly support that both qHD6-SD and qHD6- LD might be the Hd1 gene. Lin et al. (2000) reported that the near-isogenic line (NIL) having non-functional hd1- Kasalath allele showed late and early flowering under controlled SD and LD conditions, respectively, compared to the background cultivar Nipponbare, indicating that Hd1 promotes flowering under SD but inhibits it under LD. Kim et al. (2018) reported that most of Korean temperate japonica varieties including Ilpum possess functional Hd1 alleles of temperate japonica varieties induced extremely early flowering in the tropics and the non-functional hd1 alleles brought about the adaptation of temperate japonica rice to tropical regions. Also, temperate japonica breeding lines adapted to the tropics possessed the loss-of-function alleles of Hd1 with no change of other flowering genes compared to common Korean temperate japonica varieties.

We also investigated that no. of spikelet per panicle in the RILs harboring Zenith allele type of qHD6-SD under SD was greater than those of harboring Ilpum allele type. No. filled grain per panicle of the RILs harboring Zenith allele type of qHD6-SD was also greater than those of harboring Ilpum allele type. These results suggested that Zenith allele type of qHD6-SD (loss-of-function alleles of Hd1) causes a delayed heading date which is associated with increased spikelet number and grain filling which likely due to increased biomass under prolonged vegetative stage. Endo-Higashi & Izawa (2011) reported that functional allele of Hd1 not only caused early SD flowering, but also reduced spikelet number due to decreased primary branches and number of spikelet per branch.

Regional adaptability of japonica rice requires complex strategies wherein in temperate regions, reduction of flowering time is preferred but the opposite could be said for most tropical regions due to photoperiodic differences. Our findings further support the observed function of Hd1 as a central player in flowering time, a functional allele type which delays long-day heading and promotes extremely early short-day heading. A non-functional Hd1 type is critical to tropical adaptation of japonica rice since it delays days to heading which is essential to attain prolonged vegetative state in order to achieve optimum biomass, increased spikelet number and grain filling capacity.

적 요

인디카 벼는 일조시간이 짧은 열대지역에서 주로 재배하며, 자포니카 벼는 일조시간이 긴 한국, 일본, 및 중국(동북부 지 역)을 포함하는 온대지역에서 재배한다. 최근 동남아 열대지역 에서 자포니카 쌀에 대한 수요가 증가함에 따라 농촌진흥청은 필리핀 국제미작연구소(IRRI)와 공동으로 열대지역에 적응하 는 다수성 온대 자포니카 벼 품종을 개발하고 있다. 일반적으 로 자포니카 벼를 단일조건인 열대지역에 재배하면 이앙 후 바로 개화가 촉진되어 극조기 출수가 유도된다. 따라서 열대지 역에 적응하는 자포니카 벼를 개발하기 위해서는 단일조건에 서도 충분히 생장한 후 출수를 하는 특성이 우선적으로 필요 하다.

본 연구는 국내 자포니카형 벼 품종인 ‘일품’과 미국에서 육 성한 인디카형 벼인 ‘Zenith’를 교배한 F₉ RIL 180 계통을 이용하여 필리핀의 단일조건하에서 재배하면서 출수기에 관련 한 양적형질유전자좌(QTL) 분석을 수행한 바, 그 결과는 다음 과 같다.