Fine mapping and candidate gene analysis of qHD1b, a QTL that promotes flowering in common wild rice (Oryza rufipogon) by up-regulating Ehd1

Ling Liu, Yingxin Zhng, Zhengfu Yng, Qinqin Yng, Yue Zhng, Peng Xu, Jixin Li,Anowerul Islm, Liqt Shh, Xiodeng Zhn, Liyong Co,c, Shihu Cheng,, Weixun Wu,

a China National Center for Rice Improvement and State Key Laboratory of Rice Biology, China National Rice Research Institute, Hangzhou 310006, Zhejiang, China

b State Key Laboratory of Subtropical Silviculture, Zhejiang A&F University, Hangzhou 311300, Zhejiang, China

c Northern Center of China National Rice Research Institute, Hangzhou 310006, Zhejiang, China

Keywords:Rice Heading date Quantitative trait locus qHD1b Fine mapping

A B S T R A C T Heading date (flowering time) determines the adaptability of cultivars to different environments. We report the fine mapping and candidate gene analysis of qHD1b,a quantitative trait locus(QTL)responsible for early flowering that was derived from common wild rice(O.rufipogon)under both short-day and longday conditions. The introgression line IL7391, which carried segments from common wild rice in a Zhonghui 8015 (ZH8015) background, exhibited early heading compared to the background and was crossed with ZH8015 to generate BC5F2:3 families for QTL analysis. This enabled the identification of two heading-date QTL, named qHD1b and qHD7, of which the first was selected for further research.High-resolution linkage analysis was performed in BC5F4:5 and BC5F6 populations, and the location of qHD1b was confined to a 112.7-kb interval containing 17 predicted genes. Five of these genes contained polymorphisms in the promoter or coding regions and were thus considered as candidates. Expression analysis revealed a positive association between LOC_Os01g11940 expression and early heading. This locus was annotated as OsFTL1, which encodes an ortholog of Arabidopsis Flowering Locus T and was the most likely candidate gene for qHD1b. Our study revealed that qHD1b acts as a floral activator that promotes flowering by up-regulating Ehd1, Hd3a, RFT1, OsMADS14, and OsMADS15 under both shortday and long-day conditions. Field experiments showed that qHD1b affected several yield-related agronomic traits including 1000-grain weight and grain length. qHD1b could be useful for marker-assisted selection and breeding of early-maturing cultivars.

1. Introduction

Heading date represents the transition from the vegetative to the reproductive stage in plants.The flowering process is regulated by numerous internal and external signals including photoperiod,temperature, and levels of phytohormones. The ability of plant species to initiate flowering at the appropriate time during their reproductive process depends on the precise measurement of seasonal changes in day length and temperature[1,2].Plants integrate cellular timekeeping systems by sensing light signals and biological clock changes to identify dynamic changes in photoperiod and accurately control flowering time [3]. For example, in rice,heading occurs earlier under short-day (SD) than under long-day(LD) conditions.

O. rufipogonis the progenitor ofO. sativaand is widely distributed in southern China. During the domestication of rice,approximately 30%-40% of the genetic variation was lost [4,5]. As an important agronomic trait,heading date has undergone natural and artificial selection. The exploitation of novel alleles from wild rice that were lost in cultivated rice could advance rice breeding and evolutionary studies [6]. The only heading-date QTL cloned from wild rice to date isqHD7.2, which is also namedDTH7/Ghd7.1/OsPRR37and known to suppress flowering under both SD and LD conditions[7].Heading date in rice is quantitatively inherited,and QTL mapping of the trait has been conducted in mapping populations derived from interspecific crosses [2,3,8]. More than 191 heading-date QTL or genes have been identified in rice(http://www.ricedata.cn/ontology/). Among them, at least 14 QTL associated with natural variation among rice cultivars have been cloned,includingHd1,Ehd1,Ghd7,DTH8/Ghd8,Hd6,Hd16/EL1,DTH3/OsMADS50,Hd3a,Hd17/OsELF3,DTH2,Ehd4,RFT1,DTH7/Ghd7.1/OsPRR37, andHd18[8]. These molecular genetic analyses revealed two main rice photoperiodic flowering pathways,Hd1-Hd3aandGhd7-Ehd1-Hd3a/RFT1[8].

Hd1,the first cloned heading-date QTL in rice,encodes an ortholog of theArabidopsis CONSTANSgene that promotes and represses flowering under SD and LD conditions, respectively, by regulatingHd3a[9].Ghd7encodes a CCT-domain protein, which is a strong repressor ofEhd1expression,causing late flowering under LD conditions [10].DTH8/Ghd8encodes a putative HAP3 subunit of the CCAAT-box-binding transcription factor that down-regulatesEhd1andHd3atranscripts under LD conditions [11].Hd6encodes the α subunit of casein kinase II,which causes a strong photoperiodic response by indirectly repressing flowering viaHd1under LD conditions[12].Hd16represses flowering by specifically phosphorylatingGhd7[13], whileHd17/OsELF3functions as a floral promoter by repressing theGhd7transcripts under both SD and LD conditions[14].Ehd1,which lacks anArabidopsisortholog,encodes a B-type response regulator that promotes flowering by activating florigen expression especially under SD conditions [15].Ehd1is both positively and negatively regulated by several genes, being down-regulated byGhd7,DTH8/Ghd8, andHd16[10,11,13] and up-regulated byDTH3/OsMADS50,Ehd2,Ehd3,Ehd4, andqHd1/OsMADS51[16-20].Several regulators of theEhd1pathway,includingDTH7/Ghd7.1/OsPRR37,DTH8/Ghd8, andHd16, require a functionalHd1to repress flowering under LD conditions [21]. A recent study [22] showed thatHd1interacts directly withGhd7andDTH8to regulate rice flowering. However, these studies are insufficient to explain the regulatory mechanisms of heading date in rice. Recently discovered QTL or genes includeqHD19, a QTL located in chromosome 5 that is responsible for delaying heading[23];qHd2-1, which induces flowering under LD conditions and was mapped to a 105-kb interval on chromosome 2 [24]; andEfcd,which encodes a long noncoding RNA transcribed from the antisense strand of the flowering activatorOsSOC1locus and promotes flowering with no yield penalty [25].

InArabidopsis,FThas been thoroughly investigated and identified as a florigen that acts as a floral activator under LD conditions[26].In rice,13FT-likehomologs have been described[27],includingHd3a/OsFTL2andRFT1/OsFTL3,which are recognized as rice SD and LD florigens, respectively [28,29]. In the present study, we identifiedqHD1b, a QTL induces heading under four distinct environmental conditions. Through fine mapping using homozygous recombinant analysis,we resolved its location to a 112.7-kb interval.After performing DNA sequencing and expression analyses,we identified anFT-like gene as the most likely candidate forqHD1b,and found theO.rufipogonallele ofqHD1bincreased the expression ofEhd1which led to early heading.These results provide a theoretical basis for cloning theqHD1bgene to breed early-maturing rice cultivars.

2. Materials and methods

2.1. Plant materials

A set of BC4F6introgression lines were developed by introgression of chromosomal segments from theO. rufipogonaccession BJ194 (donor parent) into anindicarestorer line ZH8015 (using as recipient parent a male plant from China’s super hybrid rice Nei5you 8015) based on a single cross, four backcrosses, and five selfing generations (Fig. S1). The early-heading introgression line IL7391 was crossed with ZH8015 to generate BC5F1plants(henceforth F1plants). These were inbred to generate F2:3families in order to perform heading-date QTL analysis and initial mapping of theqHD1bregion. A large F4:5population was then developed for further mapping and isolation of a nearly isogenic line (NIL),NIL(qHD1b). Finally, F6populations and F7families were used for fine mapping and progeny verification, respectively. A schematic description of this procedure is shown in Fig. S1.

2.2. Plant growth conditions

The ZH8015, IL7391, NIL(qHD1b) plants, and other populations used for QTL analysis and mapping of theqHD1bgene were cultivated in experimental fields of the China National Rice Research Institute.The fields are located in the natural long-day(NLD)environment of Hangzhou, Zhejiang province, eastern China,119°57′E/30°03′N, day length ＞14 h during the vegetative period;and the natural short-day (NSD) environment of Lingshui, Hainan province,southern China,110°02′E/18°30′N,day length ＜12 h during the vegetative period (Table S1).

To investigate the photoperiodic response ofqHD1b, the ZH8015, IL7391, and NIL(qHD1b) plants were also grown under controlled short-day (CSD, 10 h light, 30 °C/14 h dark, 25 °C) and controlled long-day(CLD,14 h light,30°C/10 h dark,25°C)conditions in a growth cabinet with 70% relative humidity and a light intensity of 300 μmol m-2s-1. Days to heading was measured as described in a previous study [30].

2.3. DNA extraction and DNA marker analysis

Total DNA was extracted from the leaves of rice plants by the CTAB method [31] with slight modifications. Fourteen polymorphic DNA markers between ZH8015/IL7391 were screened from 525 simple sequence repeat (SSR) markers uniformly distributed along the rice genome[32].The InDel(insertion/deletion)markers used for fine mapping were designed with the software Primer Premier version 5.0(Premier Biosoft International,Palo Alto,CA,USA).These sequences are listed in Table S2. The PCR reaction system included a 12 μL reaction volume containing 2 μL DNA (100 ng),2 μL of 10 pmol μL-1primers (1 μL of forward and reverse primers), 5 μL of 2× Rapid Taq Master Mix (Vazyme Biotech Co.,Ltd., Nanjing, Jiangsu, China) and 3 μL ddH2O. PCR amplification began with a pre-denaturing step of 3 min at 95 °C, followed by 30 amplification cycles (denaturing at 95 °C for 15 s, annealing at 55 °C for 15 s, extension at 72 °C for 5 s), and a final extension at 72°C for 10 min.The PCR products were visualized on 8%polyacrylamide gels using silver staining [30].

2.4. QTL analysis

QTL analysis was conducted using the genotype and phenotype of IL7391/ZH8015 F2:3,F4,and F4:5populations with Windows QTL Cartographer 2.5 [33]. Genetic distances between markers in centiMorgans (cM) were calculated using the Kosambi’s function[34]. QTL analysis was performed by composite interval mapping with a LOD threshold ≥2.5 indicating the presence of a QTL (selected by permutation tests based on 1000 runs,P＜0.05) [35].The LOD scores, additive effect (A), and proportion of phenotypic variance explained by the QTL effect (PVE) were obtained.

2.5. Sequence analysis

The 17 open reading frames(ORFs)and their putative functions in the 112.7-kb region between INDEL4 and RM10390 were annotated using the RGAP(Rice Genome Annotation Project,http://rice.uga.edu/)database.The sequencing primers(Table S3)for the ORFs were designed according to the Nipponbare sequences available in Rice Annotation Project Database(RAP-DB,https://rapdb.dna.affrc.go.jp/).The PCR reaction mixture(50 μL reaction volume)was prepared with the KOD-FX kit (Toyobo Co. Ltd., Osaka, Japan). The amplified genomic sequences of ZH8015 and NIL(qHD1b) were aligned with each other using the software Geneious v4.7 [36].

2.6. RNA extraction and real-time quantitative RT-PCR

Total RNA was extracted with the RNAprep pure Plant kit(Tiangen Biotech Co. Ltd., Beijing, China) and cDNA was synthesized using ReverTra Ace qPCR RT Master Mix with gDNA Remover kit(Toyobo). Real-time quantitative RT-PCR (RT-qPCR, 20 μL reaction volume) was performed using the TB Green Premix Ex Taq II kit(Takara Bio, Inc., Kusatsu, Shiga, Japan)in a QuantStudio 12 K Flex Real-time PCR system (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA). TheUBQgene (Os03g0234350) was used as endogenous control.The relative expression level was calculated by the relative quantification method[37]with two biological and three technical replicates. The primers used are listed in Table S4.

2.7. Data analysis

Student’st-tests were used to evaluate the phenotypic differences between ZH8015 and IL7391, ZH8015 and NIL(qHD1b), and various recombinant groups with either ZH8015 or NIL(qHD1b).

3. Results

3.1. Phenotypic analysis of ZH8015 and IL7391

To identify new heading date QTL,we constructed an introgression line IL7391 that carried six homozygousO. rufipogonintrogression segments in the genetic background of ZH8015 (Fig. 1).IL7391 showed earlier heading than its recurrent parent ZH8015 under multiple conditions. Under Hangzhou NLD conditions,IL7391 (85.3 ± 3.3 d) flowered 7.5 days earlier than ZH8015(92.8±2.3 d;Fig.2 and B).Similarly,IL7391(105.4±1.7 d)headed 8.8 days earlier than ZH8015(114.2±1.9 d;Fig.2B)under Lingshui NSD conditions.These patterns were replicated in growth cabinets under CLD and CSD conditions, where days to heading in IL7391 was respectively 13.0 and 18.1 days earlier than that in ZH8015(Fig. 2B).

3.2. QTL analysis of heading date

To perform QTL analysis and determine the genomic regions associated with the regulation of the heading date, a total of 150 IL7391/ZH8015 F2:3families showing heading-date variation were grown under Hangzhou NLD conditions in 2017.Fourteen markers differentiating IL7391 and ZH8015 were used to screen 300 plants from the F2:3families (two plants per family). Two heading-date QTL under NLD were detected:qHD1bon chromosome 1 andqHD7on chromosome 7. No heading-date QTL were detected in the other four introgression regions (Fig. 1). The LOD scores for theqHD1bandqHD7QTL were 38.0 and 2.8,and explained respectively 47.7% and 3.8% of phenotypic variation (Table 1). The negative sign of the additive effect indicates that theqHD1bO. rufipogon(ruf)allele leads to early flowering (additive effect of -3.22), while theqHD7 rufvariant causes late flowering (additive effect of 0.14).qHD1bwas accordingly selected for further analysis owing to its larger estimated effect on heading date thanqHD7. TheqHD1bQTL was initially mapped within a 1500-kb interval flanked by the markers RM10316 and RM10394 (Fig. 3A). After interference byqHD7was eliminated (qHD7locus with homozygous ZH8015 background), two F3plants heterozygous forqHD1bwere selected to develop an F4population (just like F2) with 531 plants. Genetic analysis of this population showed a 3:1 segregation ratio of earlyand late-flowering plants(397:134;χ2=0.02,P=0.90).This result suggested that theqHD1blocus corresponded to a single Mendelian factor and that theO. rufipogonallele is more functional.Furthermore, eleven F3plants heterozygous forqHD1bwere selected for development of 11 F4populations(4360 plants)under Lingshui NSD conditions in 2018.QTL analysis showed thatqHD1bexplained 44.2% of the phenotypic variation observed in heading date (Table 1). These results suggested thatqHD1bwas the target gene involved in the control of heading date.

3.3. Isolation of NIL(qHD1b) and fine mapping of qHD1b

To further mapqHD1b, we cultivated F4:5families containing 52,320 plants under Hangzhou NLD conditions in 2018. A set of 200 randomly sampled plants belonging to the F4:5families were then used for QTL analysis using six SSR markers located between markers RM10316 and RM10394. Consistently,qHD1bwas found to account for 49.0% of phenotypic variation (Table 1). We further investigated whetherqHD1bwas involved in the photoperiodic response in rice by isolating NIL(qHD1b) plants from F4:5populations that were homozygous forO. rufipogonalleles in the RM10316-RM10394 region (Fig. 1). The photoperiodic responses in the field and growth chambers showed that NIL(qHD1b) plants consistently flowered earlier than ZH8015 under four distinct conditions (Fig. 2C and D; Fig. S2). This finding again suggested that theqHD1brufallele promotes flowering irrespective of day length.

At the same time, we selected 2578 extremely late-heading plants in the F4:5populations and mappedqHD1bto a 425.8-kb region between the markers RM3746 and RM10390 (Fig. 3B).Sequential residual heterozygotes (SeqRHs) which carrying sequentially arranged heterozygous segments ofqHD1bwere used for fine mapping.Six F5SeqRHs were used to develop six F6populations and these were genotyped with 13 markers located between RM10316 and RM10394, revealing 118 recombinants.According to the marker genotypes,the homozygous recombinants were classified into six groups (G1-G6, Fig. 3C). The mean days to heading of each homozygous recombinant group was compared with those of either ZH8015 or NIL(qHD1b) to localizeqHD1b. G1 recombinants showed later heading than NIL(qHD1b), whereas G3 recombinants with the opposite genotype showed a phenotype similar to that of NIL(qHD1b).This delimited the location ofqHD1bto the region of RM3746-RM10394(Fig.3C).Using a similar procedure,G2 and G5 recombinants placedqHD1bin a region of INDEL4-RM10394. The G6 recombinants showed late flowering, indicating thatqHD1bis not located in the region of RM10390-RM10394.Finally,G4 recombinants also showed an early-heading phenotype compared to ZH8015,suggesting thatqHD1bwas located between RM10376 and RM10394. Thus,qHD1bwas finally localized to a 112.7-kb interval between markers INDEL4 and RM10390(Fig. 3C). The recombinants carrying theqHD1brufallele between these markers showed earlier heading than those carrying theqHD1bZH8015allele. Five F7homozygous recombinant groups were used for progeny testing (Table S1), and yielded results consistent with those from the previous generation (Fig. 3C).

3.4. Candidate genes and sequencing analysis

According to RGAP,there were 17 predicted genes in theqHD1bgenomic region (Fig. 3D; Table 2). Among these 17, nine were not involved in flowering-time regulation, encoding four expressed proteins, two endoglucanase precursors, one retrotransposon protein,one transposon protein,and one hypothetical protein.We accordingly focused on the remaining eight genes.

To identify sequence differences between candidate genes, we performed sequence comparisons between the coding sequences(CDS) of the eight genes with functional domains in the two parents ZH8015 and NIL(qHD1b) (Table 2).LOC_Os01g11910encodes a helix-loop-helix DNA-binding protein that harbors two synonymous single-nucleotide polymorphisms (SNPs) distinguishing the two parents.The geneLOC_Os01g11980encodes an SFT2-like family protein and shows no differences between parents.LOC_Os01g12020encodes an LTPL18-protease inhibitor and contains a single synonymous SNP.These results allowed us to further limit the candidate genes to the remaining five genes.Among these five,LOC_Os01g11940encodesOsFT-Like1(OsFTL1), a gene that is orthologous toArabidopsis FTand showed no CDS differences between the two parents.LOC_Os01g11946encodes an ATPbinding protein and harbors two nonsynonymous and one synonymous SNPs.LOC_Os01g11952is annotated as SET-domain group protein 721 (SDG721), which encodes a TRITHORAX-like protein and contains seven nonsynonymous and two synonymous SNPs.LOC_Os01g11960encodes a PHD-type zinc finger domaincontaining protein and harbors two nonsynonymous and one synonymous SNPs.LOC_Os01g12080encodes a plant-specific domain from the TIGR01589 protein family and contains three nonsynonymous SNPs (Table 2; Fig. S3). Among these five genes, the FT-like protein, the SET-domain protein, and the PHD-type zinc finger family protein have all been reported [18,38,39] to be involved in photoperiodic flowering-time regulation. This observation,together with the presence of nonsynonymous SNPs in the other two genes, led us to sequence the promoter regions of these five genes in the two parental lines. ForLOC_Os01g11940, we found 16 SNPs; forLOC_Os01g11946, 3 SNPs and 1 InDel; forLOC_Os01g11952, 20 SNPs and 5 InDels; forLOC_Os01g11960, 16 SNPs and 3 InDels; and forLOC_Os01g12080, 2 SNPs and 1 InDel(Table 2; Fig. S4).

Table 1QTL detected for heading date.

Table 2Predicted genes at the qHD1b locus and the polymorphisms between ZH8015 and NIL(qHD1b).

3.5. Expression analysis of the candidate genes of qHD1b

To further identify a candidate gene forqHD1b,we investigated the expression patterns of the five candidate genes(LOC_Os01g11940,LOC_Os01g11946,LOC_Os01g11952,LOC_Os01g11960, andLOC_Os01g12080) using RT-qPCR in leaf tissues from ZH8015 and NIL(qHD1b) under NLD conditions in the field. We found higher expression levels ofLOC_Os01g11940in NIL(qHD1b) than in ZH8015 (Fig. 4A), but no differences in the expression ofLOC_Os01g11946,LOC_Os01g11952,LOC_Os01g11960,andLOC_Os01g12080(Fig. 4B-E).

We investigated the expression pattern ofLOC_Os01g11940under 48 h of continuous light (LL)/dark (DD) after normal CSD/CLD treatment of the two parental lines. As with the observations under NLD conditions, we found a higher abundance ofLOC_Os01g11940transcripts in NIL(qHD1b) than in ZH8015. Moreover, the expression ofLOC_Os01g11940showed a diurnal rhythm in both parents under both CSD and CLD conditions, and its transcript levels were decreased in the light and increased in the dark periods (Fig. 4F-I). The circadian expression pattern ofLOC_Os01g11940was irregular under LL/DD conditions (Fig. 4F-I),indicating that the circadian oscillations ofLOC_Os01g11940needs light/dark alternating conditions.For the remaining four predicted genes,the expression levels were sometimes higher in ZH8015 and sometimes higher in NIL(qHD1b), indicating that negligible differences between the two parental lines under both CSD and CLD conditions(Fig.4J-Q).These results supportedLOC_Os01g11940as the most likely candidate gene ofqHD1b.

3.6. qHD1b promotes flowering upstream of Ehd1

The expression differences ofLOC_Os01g11940between the parents and its diurnal expression pattern in leaf tissues supported it as the most likely candidate gene forqHD1b,which is thus involved in the regulation of flowering time in rice.To investigate the role ofqHD1bon rice heading-date regulation, we performed RT-qPCR assays to compare the expression levels of eight heading-date genes (OsMADS14,OsMADS15,Hd3a,RFT1,Hd1,Ehd1,Ghd7, andDTH8) in leaf tissues from ZH8015 and NIL(qHD1b) plants under both CSD and CLD conditions.OsMADS14andOsMADS15act as key activators of flower development [29]. The expression of these genes was higher in NIL(qHD1b) than in ZH8015 under both CSD and CLD conditions(Fig. 5A-D). TheqHD1brufallele showed dominance to theqHD1bZH8015allele, suggesting thatqHD1bpromotesOsMADS14andOsMADS15expression.Hd3aandRFT1also showed higher expression levels in NIL(qHD1b) than in ZH8015 under both CSD and CLD conditions(Fig.5E-H).This observation is consistent with the phenotypic differences observed between ZH8015 and NIL(qHD1b) (Fig. 2C and D).Hd1andEhd1are floral integrators that both regulateHd3aandRFT1[8].We also compared the expression of both genes in the parents to investigate whether the regulation byqHD1bofHd3aandRFT1is mediated byHd1,Ehd1,or both.The expression levels ofEhd1were also higher in NIL(qHD1b) than in ZH8015 under both conditions (Fig. 5K and L). In contrast, the expression ofHd1showed no significant differences between the two parents (Fig. 5I and J). These results suggest thatqHD1bacts upstream ofEhd1but has no effect onHd1transcripts.Ghd7andDTH8are two long-day flowering repressors located upstream ofEhd1in theEhd1-Hd3a/RFT1pathway[10,11].There was no significant difference in the expression level ofGhd7between ZH8015 and NIL(qHD1b) (Fig. 5M and N). A similar observation was found inDTH8(Fig. 5O and P), suggesting thatGhd7andDTH8act either upstream or parallel toqHD1b.These results indicated thatqHD1bis a positive regulator of heading date that up-regulates the expression ofEhd1,Hd3a,RFT1,OsMADS14,andOsMADS15under CSD and CLD conditions.

Fig. 2. Phenotypic characterization of ZH8015, IL7391, and NIL(qHD1b) plants. (A) Heading-date phenotypes of ZH8015 and IL7391. (B) Comparison of days to heading in ZH8015 and IL7391 in four environments. (C) Heading-date phenotypes of ZH8015 and NIL(qHD1b). (D) Comparison of days to heading in ZH8015 and NIL(qHD1b) in four environments.The photos were made under NLD conditions in Hangzhou.Scale bars,20 cm.The days to heading value of each plant is represented by a dot.Values represent mean ± SD; a Student’s t-test was used to generate the P values, *, **, and *** indicate P ＜0.05, P ＜0.01, and P ＜0.001, respectively.

Fig. 3. Fine mapping of the qHD1b locus on chromosome 1. (A) Initial mapping using F2:3 families. (B) Further mapping using F4:5 families. The number of recombinants is shown below each marker.(C)Genotype and phenotype of two parents and six homozygous recombinant groups from F6 populations for fine mapping.White and black bars represent respectively ZH8015 and O. rufipogon homozygotes. Numbers of homozygous recombinant plants in the corresponding groups for fine mapping are shown in brackets.Numbers of days to heading are presented as mean±SD.The superscripted letters(a,b,and c)indicate significant differences in the heading dates of recombinants relative to the parents. (D) Seventeen predicted genes overlap the target region.

Fig. 4. Expression patterns of five candidate genes in ZH8015 and NIL(qHD1b) plants by RT-qPCR. (A-E) LOC_Os01g11940, LOC_Os01g11946, LOC_Os01g11952,LOC_Os01g11960, and LOC_Os01g12080 transcripts under NLD conditions, leaves were collected every 4 h (8:00 to 16:00 for two days) from 72-day-old plants. (F-I)LOC_Os01g11940 transcripts from CSD to LL(F),CSD to DD(G),CLD to LL(H),CLD to DD(I)conditions.(J-Q)LOC_Os01g11946(J,N),LOC_Os01g11952(K,O),LOC_Os01g11960(L,P),and LOC_Os01g12080(M,Q)transcripts under CSD and CLD conditions.For CSD and CLD conditions,leaves were collected from 85 and 100-day-old plants,respectively.The penultimate leaves of three plants were collected every 4 h.The white and black bars represent light and dark periods,respectively.LL and DD indicate continuous light and continuous dark,respectively.The light gray and dark gray bars indicate respectively subjective dark during LL conditions and subjective light during DD conditions.The mean ± SD values were obtained from two biological and three technical replicates.

3.7. qHD1b affects yield-related traits

To investigate the genetic effect ofqHD1b, we performed phenotypic comparisons of several yield-related agronomic traits in ZH8015 and NIL(qHD1b) plants under NLD and NSD conditions(Fig. S5). In addition to increased flowering (Fig. 2D), NIL(qHD1b)plants showed significant reductions in plant height, main panicle length, secondary branch number, and number of grains in the main panicle under NLD conditions (Fig. S5A, C, E and F). The primary branch number, number of grains in the main panicle,seed-setting rate, and grain width were significantly different between ZH8015 and NIL(qHD1b) under NSD conditions(Fig.S5D,F,H and L).There was no significant difference in panicle number,filled grain number per plant,and yield per plant between ZH8015 and NIL(qHD1b) under both NSD and NLD conditions(Fig. S5B, G and I). The 1000-grain weight and grain length of NIL(qHD1b) plants were significantly greater than those of ZH8015 under both conditions (Fig. S5J and K).

4. Discussion

4.1. qHD1b is likely a novel heading date QTL

In the present study, we choseO. rufipogonas the donor and ZH8015 as the recipient and constructed a series of introgression lines (Fig. S1). We used the SeqRHs method to develop a finemapping population and localizedqHD1bto a 112.7-kb region between the markers INDEL4 and RM10390 on the short arm of chromosome 1 (Fig. 3). Despite hundreds of QTL or genes associated with heading date having been identified in rice, only seven genes have been found on chromosome 1, namelyOsGI,OsEF3,OsLFL1,OsABF1,OsCCT01,qHd1/OsMADS51, andSe13[20,40-45].Among these genes,OsGIfunctions as an upstream activator ofHd1[8] and is the closest gene toqHD1b(at a distance of ～2000 kb).

Fig.5. Diurnal expression patterns of flowering-time genes OsMADS14(A,B),OsMADS15(C,D),Hd3a(E,F),RFT1(G,H),Hd1(I,J),Ehd1(K,L),Ghd7(M,N),and DTH8(O,P)in ZH8015 and NIL(qHD1b)plants under CSD and CLD conditions by RT-qPCR.Penultimate leaves of three plants were collected every 4 h throughout a 48-h period from 85 and 100-day-old plants under CSD and CLD conditions,respectively.The white and black bars represent light and dark periods,respectively.The mean±SD values were obtained from two biological and three technical replicates.

According to the RGAP database,there were 17 predicted candidate genes in the target region ofqHD1b.Five of these genes carry sequence differences between the two parents (Table 2; Figs. S3 and S4).LOC_Os01g11946andLOC_Os01g12080are likely not involved in flowering-time regulation under SD and LD conditions,in view of their putative biological functions (Table 2). As for the remaining three genes,LOC_Os01g11952encodes a SET domain protein namedSDG721that was recently reported to affect the development of floral organs in rice, but has no effect on heading date regulation [46];LOC_Os01g11960encodes a PHD-type zinc finger protein whose member was previously reported to promote flowering [18];LOC_Os01g11940was annotated asOsFTL1and showed 46.4% amino-acid identity with theArabidopsisflorigen FT that induces flowering under LD conditions [26]. Sequencing analysis revealed that there were no differences in theOsFTL1coding region, but we found a total of 16 SNPs in its promoter region(Figs. S3 and S4A). MostFT-like genes are involved in floweringtime regulation[39]and are widely distributed inArabidopsis,rice,barley, wheat, and other species. There are 13FT-like genes in the rice genome, of whichHd3a/OsFTL2was the first to be cloned and encodes an SD-florigen that promotes heading mainly under SD conditions [28].RFT1/OsFTL3is the closest homolog ofHd3aand acts as an LD-florigen that promotes flowering, particularly under LD conditions [29].OsFTL1is the second closest homolog ofHd3a,and plants overexpressing this gene similarly showed early flowering toHd3aorRFT1-overexpressing plants [47,48].BdFT2,HvFT2,andTaFT-A2are orthologs of riceOsFTL1inBrachypodium dis-tachyon, barley, and wheat, respectively. These genes are also involved in flowering-time regulation [49]. These observations suggest thatOsFTL1is functionally conserved among plant species.In a recent study [27], overexpression ofOsFTL10induced flowering in rice. Soybean contains 10FTorthologs, of whichGmFT2aandGmFT5aare flowering activators and integrators [50].

Expression analysis showed thatOsFTL1had higher expression in NIL(qHD1b) than ZH8015 plants under NLD conditions(Fig. 4A). In contrast, four other genes (LOC_Os01g11946,LOC_Os01g11952,LOC_Os01g11960, andLOC_Os01g12080) showed no significant expression differences between parents (Fig. 4B-E).These results were also confirmed under CSD and CLD conditions(Fig. 4F-Q). The observations from our sequencing and expression analyses suggest that the increased expression ofOsFTL1in NIL(qHD1b) plants is the reason for its early heading. In a previous study [51], the heading-date QTLdth1.1was located on the short arm of chromosome 1 and contained the sub-QTLdth1.1aanddth1.1b, andOsFTL1was located between these two sub-QTL. A subsequent study[52]found that thedth1.1aregion containsOsGIandOsFTL8,while thedth1.1bregion containsOsEMF1andPNZIP.In light of these findings, our results suggest thatqHD1bis a novel heading-date QTL in rice that is distinct fromdth1.1, and thatOsFTL1is the most likely candidate gene forqHD1b.

4.2. qHD1b is involved in the flowering-time regulatory network

The finding that NIL(qHD1b) flowered earlier than the background ZH8015 under both LD and SD conditions (Fig. 2D) indicates thatqHD1bacts as a floral activator irrespective of day length. The expression ofOsFTL1showed a marked circadian rhythm and differential expression in the two parental lines, with a peak expression observed at dawn and lowest expression at dusk under both natural and controlled conditions (Fig. 4A, F-I). These observations show thatqHD1bparticipates in the rice photoperiodic flowering pathway. Substitution ofqHD1bZH8015with theqHD1brufallele resulted in up-regulation ofEhd1but notHd1,Ghd7, andDTH8transcripts (Fig. 5I-P).Ehd1was highly expressed in NIL(qHD1b), suggesting thatqHD1bacts upstream ofEhd1;moreover, compared with ZH8015, the expression levels ofHd1,Ghd7, andDTH8in NIL(qHD1b) showed no significant differences.This finding suggests thatqHD1bacts either downstream or in parallel toHd1,Ghd7,andDTH8.Consequently,the expression levels ofHd3aandRFT1, located downstream ofEhd1, increase and subsequently lead to increases inOsMADS14andOsMADS15transcripts in NIL(qHD1b) under CSD and CLD conditions (Fig. 5A-H). Our RT-qPCR analysis indicated thatqHD1bactivates flowering via theEhd1-Hd3a/RFT1-OsMADS14/OsMADS15pathway. The similar expression patterns ofOsFTL1withEhd1,Hd3a,RFT1,OsMADS14,andOsMADS15also indicate their similar function in regulating heading date (Fig. 4F-I, 5A-H, K, L). BecauseEhd1is influenced by many upstream regulatory factors, including activatorsEhd2/3/4,DTH3, andOsMADS51and repressorsHd16,PHYB, andOsLFL1[8],the exact position ofqHD1bin the rice heading date regulation pathway remains unknown.

4.3. Potential use of qHD1b in crop adaptation and breeding

With the continuous reduction of arable land area,it is vital for rice breeding to optimize heading date in order to achieve maximum yield. It is thus desirable to identify heading-date QTL and the underlying genes. Compared to ZH8015, NIL(qHD1b) plants promoted heading under both SD and LD conditions, indicating thatqHD1brufis photoperiod-insensitive(Fig.2C,D).Breeders tend to prefer weakly photosensitive cultivars for their adaptability to diverse light and temperature conditions,and weak photosensitivity is a key factor for rice adaptation to high-latitude regions.Wild rice is sensitive to LD conditions and is confined to the tropics and low-latitude subtropics, while cultivated rice is distributed across latitudes ranging from 55°N to 36°S [53]. The genesHd1,Ghd7,DTH8, andDTH7belong to LD-flowering suppressors and their weak alleles are likely important for rice expansion into and adaptation to northern latitudes [9-11,54]. The combination ofHd1,Ghd7, andDTH8alleles showed strong photosensitivity at low latitudes, while their weak alleles were photoperiod-insensitive and were distributed at higher latitudes [22]. Several photoperiodinsensitive heading-date QTL, includingHd3a,Hd17/Hd3b,DTH2,qHd1/OsMADS51, andHd18also have similar functions.This diversity of QTL resources allows fine tuning of heading date and yield traits under various ecological and geographic conditions[14,20,47,53,55]. We showed thatqHD1bis associated with heading date differences between ZH8015 and NIL(qHD1b). Moreover,we observed significant differences between ZH8015 and NIL(qHD1b) plants in many other agronomic traits. In particular, theqHD1brufallele caused shorter plant height, shorter main panicle length, fewer secondary branch numbers, and lower numbers of grains in the main panicle under NLD conditions (Fig. S5A, C, E and F). However, we found no significant difference in yield per plant between the two parents under either NLD or NSD conditions(Fig.S5I), possibly owing to the increase of 1000-grain weight and grain length of NIL(qHD1b) (Fig. S5J and K). Thus,qHD1brufshould be an attractive breeding resource,owing to its photoperiod insensitivity with early flowering and without yield reduction. Positional cloning ofqHD1bintrogressed fromO. rufipogonwill shed light on the molecular mechanisms of rice flowering and earlymaturing rice breeding programs. TheqHD1blocus influences the regional and seasonal adaptation of rice crops to distinct environments. Marker-assisted selection using markers closely linked to theqHD1blocus should thus be helpful in assisting rice breeding programs to fine-tune heading date regulation.

5. Conclusions

We performed fine-mapping analysis of the heading-date QTLqHD1band localized a candidate gene to a 112.7-kb genomic region on the short arm of chromosome 1, between the markers INDEL4 and RM10390. TheO. rufipogonallele at theqHD1blocus promoted heading in rice under both SD and LD conditions.Sequence and expression analyses indicated thatLOC_Os01g11940,which encodes theOsFTL1gene, was the most likely candidate forqHD1b.Expression analysis revealed thatqHD1bup-regulatesEhd1transcripts and then promotes the expression ofHd3a,RFT1,OsMADS14, andOsMADS15. Future studies of this newly described QTL responsible for regulating flowering time in rice will contribute to genetic resources and theoretical foundations that may be useful for marker-assisted breeding.

CRediT authorship contribution statement

Ling Liu:Investigation, Validation, Formal analysis, Visualization, Writing - original draft.Yingxin Zhang:Methodology,Resources.Zhengfu Yang:Investigation, Validation, Writing -review & editing.Qinqin Yang:Investigation, Software.Yue Zhang:Investigation, Software.Peng Xu:Investigation, Software.Jiaxin Li:Investigation, Software.Anowerul Islam:Investigation.Liaqat Shah:Investigation.Xiaodeng Zhan:Resources, Project administration.Liyong Cao:Funding acquisition, Project administration.Shihua Cheng:Conceptualization, Funding acquisition,Project administration, Supervision.Weixun Wu:Conceptualization, Funding acquisition, Project administration, Data curation,Supervision, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank the National Mid-term Genebank for Rice of the China National Rice Research Institute for providing theO. rufipogonaccession. This work was supported by the National Natural Science Foundation of China (31871604, 32071996, and 31961143016), the National Key Research and Development Program of China (2020YFE0202300), the Fundamental Research Funds of Central Public Welfare Research Institutions (CPSIBRFCNRRI-202102),and the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences(CAAS-ASTIP2013-CNRRI).

Appendix A. Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.12.009.