Transcriptome analysis of flowering regulation by sowing date in Japonica Rice (Oryza sativa L.)

Hybrid japonica cultivars, such as the Yongyou series, have shown high yield potential in the field in both the early and late growing seasons. Moreover, understanding the responses of rice flowering dates to temperature and light is critical for improving yield performance. However, few studies have analyzed flowering genes in high-yielding japonica cultivars. Based on the five sowing date experiments from 2019 to 2020, select the sensitive cultivar Yongyou 538 and the insensitive cultivar Ninggeng 4 and take their flag leaves and panicles for transcriptome analysis. The results showed that compared with sowing date 1 (6/16), after the sowing date was postponed (sowing date 5, 7/9), 4480 and 890 differentially expressed genes (DEGs) were detected in the leaves and panicles in Ninggeng 4, 9275 and 2475 DEGs were detected in the leaves and panicles in Yongyou 538, respectively. KEGG pathway analysis showed that both Ninggeng 4 and Yongyou 538 regulated rice flowering through the plant circadian rhythm and plant hormone signal transduction pathways. Gene expression analysis showed that Os01g0566050 (OsELF3-2), Os01g0182600 (OsGI), Os11g0547000 (OsFKF1), Os06g0275000 (Hd1), and Os09g0513500 (FT-1) were expressed higher and Os02g0771100 (COP1-1) was expressed lower in Yongyou 538 compared with Ninggeng 4 as the climate conditions changed, which may be the key genes that regulate the flowering process with the change of temperature and light resources in sensitive cultivar Yongyou 538 in the late season.

www.nature.com/scientificreports/ The molecular mechanisms of plant flowering have consistently been a global research hotspot. In recent years, scientists have discovered that many genes influence the rice flowering stage [23][24][25] . Moreover, breeders have developed a large number of high-quality and high-yielding varieties. However, few connections have been made between flowering genes and high-yielding varieties. As the most widely used sequencing technology in genomics research, transcriptome sequencing (RNA-seq) technology can be used to quickly and comprehensively obtain almost all transcript and sequence information for a specific tissue or organ of a species in a specific state. Therefore, in this study, the widely planted cultivars Ninggeng 4 and Yongyou 538 were planted on five sowing dates and two sowing dates were selected as materials for transcriptome analysis. The main objectives of this study were to (1) analyze japonica flowering adaptation in response to climate conditions change and (2) select the key genes regulating flowering adaptability to facilitate the spread of japonica rice to a wider planting area.

Results
Differences in yield and biomass of japonica cultivars among sowing dates. There was significant interaction effect of sowing date and year for the yield (Table 1). With the postponed sowing date, the yield of Ninggeng 4 remained unchanged or slightly increased, and the yield changed 9.5% (2019) and 7.2% (2020) among the five sowing dates. However, Yongyou 538 showed a trend of rising first and then decreasing or decreasing all the time. This indicated that the yield of Yongyou 538 was more sensitive to sowing date, while Ningjing 4 was relatively slow.
There was significant interaction effect of sowing date and year for the biomass during the different stages, too ( Table 1). The biomass of Ninggeng 4 decreased slightly at the vegetative stage, except for the sowing date 2 in 2019 and the sowing date 5 in 2020, and the coefficient of variation was 9.1% and 7.0% in 2019 and 2020, respectively. While Yongyou 538 showed a decreasing trend, and the coefficients of variation was 18.8% and 17.4% in 2019 and 2020, respectively. During the reproductive and grain filling period, the biomass of Ninggeng 4 and Yongyou 538 had no obvious trend among the five sowing dates, and the coefficients of variation was 39.0% and 31.4% at the reproductive stage, 25.2% and 32.2% at the grain filling stage for Ninggeng 4 and Yongyou 538, respectively. It indicated that the biomass was affected by the sowing dates, and the biomass of Yongyou 538 had a significantly higher variation range than that of Ninggeng 4.
Differences in growth stage of japonica cultivars among sowing dates. In general, with the postponed sowing dates, the vegetative stage of Ninggeng 4 and Yongyou 538 was shortened, while the grain filling period was extended, and the reproductive stage was slightly different (Table 2). Compared sowing date 5 with sowing date 1, the vegetative stage of Ninggeng 4 was shortened by 3 days and 9 days in 2019 and 2020, respectively, and the effective accumulated temperature was changed by 45 °C and 47 °C in 2019 and 2020, respectively. While Yongyou 538 was shortened by 14 days and 15 days in 2019 and 2020, respectively, and the effective accumulated temperature was reduced by 173 °C and 166 °C in 2019 and 2020, respectively. The results showed that the response of different japonica cultivars to sowing dates was significantly different, and the difference in the vegetative stage of Ninggeng 4 was significantly lower than that of Yongyou 538, which indicated that Ninggeng 4 was insensitive to temperature and light, while Yongyou 538 was sensitive. www.nature.com/scientificreports/ Correlation analysis. Table 3 showed the correlation analysis between the yield and the temperature and light resources (effective accumulated temperature, accumulated ligth hours and solar radiation) of Ninggeng 4 and Yongyou 538 during the vegetative stage from 2019 to 2020. The results showed that the yield of Ninggeng 4 was significantly positively correlated with the effective accumulated temperature and light hours at this stage, and the correlation coefficients were 0.7 and 0.8, respectively. While Yongyou 538 was significantly positively correlated with the accumulated light hours at this stage, and the correlation coefficient was 0.7. This indicated that the stable yield of Ninggeng 4 was due to its insensitivity to the vegetative stage, while the yield of Yongyou 538 decreased with the shorted vegetative period.
Sequencing quality analysis of japonica rice tissue. The clean reads from the leaves and panicles of the japonica cultivars Ninggeng 4 and Yongyou 538 sown on the two sowing dates (CK vs. T) are presented in Table 4 and Table S1. In the leaves, the number of clean reads from Ninggeng 4 in the T treatment was the highest and was 0.43%, 12.85% and 16.37% higher than those from Ninggeng 4 in the CK treatment and Yongyou 538 in the CK and T treatment, respectively. In addition, in the panicles, the number of clean reads from Ninggeng 4 in the T treatment was the highest and was 4.76%, 8.68% and 2.11% higher than those from Ninggeng 4 and Yongyou 538 in the CK treatment and Yongyou 538 in the T treatment, respectively. The number of bases with a mass greater than 20 or 30 in the leaves and panicles accounted for more than 95% of the total number of bases,  The results of a functional query with the keyword "flowering" for the DEGs showed that when the sowing date was delayed 23 days, there were 7 upregulated and 8 downregulated flowering DEGs in the leaves and 4 downregulated flowering DEGs in the panicles of Ninggeng 4. In Yongyou 538, there were 16 upregulated and 22 downregulated flowering DEGs in the leaves and 3 upregulated and 8 downregulated flowering genes in the panicles (Fig. 1C,D). The Venn diagram of the flowering DEGs indicated that 6 and 1 flowering DEGs were found in the leaves and panicles of the two japonica cultivars, respectively; 9 and 3 flowering DEGs were found only in the leaves and panicles of Ninggeng 4, respectively; and 32 and 10 flowering DEGs were found only in the leaves and panicles of Yongyou 538, respectively (Fig. 1E,F).

Analysis of flowering DEGs by KEGG pathway. KEGG pathway analysis was carried out to further
analyze the function of flowering DEGs in the leaves and panicles of Ninggeng 4 and Yongyou 538 sown on the two sowing dates (Fig. 2). In Ninggeng 4, a total of 6 flowering DEGs (2 upregulated genes Os11g0547000 and Os06g0552900, 4 downregulated genes Os09g0513500, Os06g0157500, Os03g0856700 and Os10g0463400) in the leaves from NCK vs NT were assigned to 3 KEGG pathways, including the plant circadian rhythm, diterpenoid biosynthesis and plant hormone signal transduction pathways, but no pathway was detected for the flowering DEGs in the panicles. On the other hand, in Yongyou 538, a total of 22 flowering DEGs (13 upregulated and 9 downregulated) in the leaves under YCK vs YT were assigned to 14 KEGG pathways, while only 2 upregulated genes (Os01g0182600 and Os11g0547000)in the panicles were assigned to the plant circadian rhythm pathway. Compared with the plant circadian rhythm pathway mapped in the panicles, the upregulated genes in the leaves were significantly enriched in the plant circadian rhythm, glycerolipid metabolism, glycosphingolipid biosynthesis-globin and isoglobin series pathways, while the downregulated genes were involved in porphyrin and chlorophyll metabolism, brassinosteroid biosynthesis, glutathione metabolism, cysteine and methionine metabolism, plant hormone signal transduction, plant circadian rhythm and ubiquitin mediated proteolysis pathways; of these, the porphyrin and chlorophyll metabolism pathways were significantly enriched.  Table S2. The results showed that a total of 12 DEGs were involved. In the leaves, CRY and COP1 were significantly downregulated, while FKF1 and FT were significantly upregulated in both Ninggeng 4 and Yongyou 538, and the DEGs Os09g0513500 (FT-3) and Os06g0157500 (RFT1) were both downregulated in Ninggeng 4. In

qRT-PCR for DEGs. Four DEGs involved in the plant circadian rhythm pathway were analyzed by qRT-
PCR. The results showed that the expression patterns detected in the qRT-PCR were similar to those detected with RNA-seq (Fig. 5), indicating that the RNA-seq data used in this experiment were reliable.

Discussion
In our study, when the sowing date was delayed by 23 days, the days-to-heading of Ninggeng 4 and Yongyou 538 were occurred 6 and 15 days earlier than that in the sowing date 1 (CK), respectively. In the different growth stages, the vegetative and reproductive periods of Ninggeng 4 were both decreased by 3 days, while the vegetative and reproductive stages of Yongyou 538 were decreased by 14 and 1 day, respectively ( Table 2). This means that the change in the days-to-heading occurred mainly in the vegetative growth period, which is consistent with the results of a previous study 27,28 . During the vegetative stage, the average daily temperatures for Ninggeng 4 and Yongyou 538 under the T treatment were 29.5 °C, which was 7.4% and 5.6% higher than those in the CK treatment and close to the optimum temperature for japonica growth and development (30 °C) (Fig. 6). Duan et al. 29 pointed out that the growth rate of japonica was accelerated and its growth stage was shortened when the average daily temperature increased during the vegetative growth period and that japonica showed different degrees of temperature sensitivity. Similarly, the sensitivity of the responses of Ninggeng 4 and Yongyou 538 to temperature and light changes were not consistent, and Yongyou 538 was more sensitive than Ninggeng 4. The greater number of DEGs in Yongyou 538 than in Ninggeng 4 also supports this idea (Fig. 1A-D). In addition, more DEGs were detected in the leaves than in the panicles (Fig. 1A-D), indicating that leaves, rather than panicles, are the organs that perceive climate conditions change. This finding is similar to that in a previous study showing that plants perceive the change in light hours with their leaves and initiate stem apex growth and flower formation based on these perceptions 30 .
The biological clock of plants plays an important role in regulating their flowering process. Under different photoperiod treatments, the plant flowering stages can be affected by the genes involved in the plant circadian rhythm pathway. In our study, as the climate conditions changed, DEGs under CK versus T in the circadian rhythm pathways were observed in both Ninggeng 4 and Yongyou 538 (Fig. 2). The meteorological data during this period showed that the average daily light hours decreased by 25.4% and 6.3% and the accumulated light hours decreased by 36.6% and 10.2% from panicle initiation to heading in Ninggeng 4 and Yongyou 538, respectively (Fig. 6). A previous study showed that when the blue light receptor CRY in plants receives blue light signals, the activity of its downstream gene COP1 is inhibited 31 . As a result, COP1 was significantly downregulated both in Ninggeng 4 and Yongyou 538 in the sowing date 5 (T) compared with sowing date 1 (CK) (Fig. 4). COP1 can interact with its downstream gene ELF3 and regulate the biological activity of light signals input by the downstream target GI to regulate the circadian rhythm pathway and the flowering stage 32 . In addition, OsELF3-2 was significantly upregulated in Yongyou 538 (Fig. 3). Yu et al. 32 pointed out that the function of ELF3 in regulating plant circadian rhythms required complete COP1 activity. Therefore, the downregulation of COP1 in Yongyou 538 resulted in increased expression of GI (Fig. 4). As a gate switch for light signal inputs, the homologous gene of GI in Oryza sativa, OsGI, activates the transcription of Hd1 (the homologous gene of CO in Oryza sativa) under short-day light conditions 7 . In addition, GI can combine with FKF1 to form a blue-light-dependent protein complex, which can indirectly promote the transcription of CO by degrading CDFs (the transcription repressors for CO) [33][34][35] . Subsequently, CO can promote the differentiation of flower primordia to form flower tissues www.nature.com/scientificreports/ by inducing the transcription of FT 36 . Therefore, OsFKF1, Hd1 and FT-1 were upregulated with the upregulation of the upstream gene GI in Yongyou 538 in our study (Fig. 3). In Ninggeng 4, OsFKF1 and FT, including FT-2, FT-3 and RFT1, expressed differentially compared CK and T, while OsGI was not differentially expressed ( Fig. 4), resulting in a small change in the number of protein complexes formed by FKF1 and GI. As a result, the interaction of upregulated genes and downregulated genes in FT may be the main reason for the difference in the flowering stage of Ninggeng 4 under the two sowing dates. In addition to their biological clocks, plants can also regulate the flowering process with hormones. Our research found that B-ARR was significantly downregulated in Ninggeng 4 through the CTK transduction pathway. Previous studies have shown that cytokinins can regulate the rice flowering process through genes such as AHK2 and AHK3 37 , B-ARR and A-ARR 38 . In Oryza sativa, Ehd1 is highly homologous to B-ARR and can interact with the A-ARR protein OsRR1 to form a heterodimer and inhibit the activity of Ehd1 38 . In terms of photoperiod regulation, Ehd1 can directly regulate its downstream gene FT and promote the flowering stage under short-day light conditions 16 . Therefore, two downregulated genes in FT were detected in Ninggeng 4 (Fig. 4). In addition, BRI1 was significantly downregulated in Yongyou 538 through the BR transduction pathway. In this pathway, BRI1 is a receptor gene on the membrane and can promote dark morphogenesis and early flowering 39 . However, the effect of BRI1 in Yongyou 538 may have been weakened by other flowering genes in our study. In addition, DEGs also existed in the signal transduction pathways of IAA, GA, ABA, Eth, JA, and SA (Table 5), which have been confirmed to play a role in rice flowering [40][41][42][43][44] . The number of hormone-related DEGs, especially those related to IAA, was higher in the sensitive cultivar Yongyou 538 than in the insensitive cultivar Ninggeng 4. This result indicates that hormones play an important role in japonica flowering in response to climate conditions change.
Two different subspecies of rice were selected as the materials for this research. A previous study showed that the functional diversity of flowering-time gene sequences has no special relationship between rice subspecies 45 . Gene expression analysis showed that the expression of Os01g0566050 (OsELF3-2), Os11g0547000 (OsFKF1), Os01g0182600 (OsGI), Os06g0275000 (Hd1) and Os09g0513500 (FT-1) was slightly or significantly higher in the sensitive cultivar Yongyou 538 after the sowing date delay (YT) than other treatments (NCK, NT and YCK), while the expression of Os02g0771100 (COP1-1) was significantly lower, these may be the key genes that regulate

Conclusions
The flowering stage in the two japonica varieties responds differently to climate change. Yongyou 538 is more sensitive and affected by the delay in the sowing date than Ninggeng 4, as indicated by its higher number of DEGs. Therefore, Yongyou 538 can adapt to climate conditions change by modulating the plant circadian rhythm and hormone signal transduction pathways. Gene expression analysis indicated that Os01g0566050 (OsELF3-2), Os01g0182600 (OsGI), Os11g0547000 (OsFKF1), Os06g0275000 (Hd1), and Os09g0513500 (FT-1) were expressed higher and Os02g0771100 (COP1-1) was expressed lower in the sensitive cultivar as the climate conditions changed, these may be the key genes that regulate the flowering process in japonica cultivars adapting to different planting areas.  cDNA library preparation and RNA-seq. The library was constructed using an Illumina TruSeq TM RNA Sample Prep Kit method. Total RNA was extracted from the leaves and panicles of Ninggeng 4 and Yongyou 538 from the CK and T treatments. Its concentration and purity were detected by a Nanodrop2000. The integrity of the RNA was detected by agarose gel electrophoresis, and the RNA integrity number (RIN) value was detected by Agilent 2100. Then mRNA was enriched by oligo-dT selection, and fragmentation buffer was added to fragment the mRNA. Under the action of reverse transcriptase, a random six-base primer was added, and cDNA was synthesized by reverse transcription using mRNA as a template. The cDNA was then connected to the adaptor. The library was enriched and amplified in 15 PCR cycles. The target band recovered from the 2% agarose gel was quantified by TBS380 Picogreen and mixed in the indicated ratios. Clusters were generated by bridge PCR amplification on cBot. Finally, the PE library was sequenced on the Illumina platform with a read length of 2 × 150 bp. Tables and figures were  . Two-way ANOVA was used to compare the effects of year and sowing date with a general linear model (GLM). One-way ANOVA was performed to evaluate the effects of sowing date on biomass, yield and growth period parameters. Means were subjected to Tukey's honestly significant difference (HSD) test at the 0.05 probability level. The raw reads obtained by Illumina sequencing were filtered to obtain highquality sequencing data (clean reads). TopHat2 was used to compare the clean data with the japonica reference genome Nipponbare (https:// www. ncbi. nlm. nih. gov/ assem bly/ GCF_ 00143 3935.1/) for gene sequence comparison, obtain the mapped data and evaluate the quality of the sequence results. The gene expression amounts were calculated and normalized as transcripts per million (TPM). The thresholds of |log 2 FC|> = 1 and padjust < 0.05 were used to screen for differentially expressed genes (DEGs) with DESeq2. Then, the DEGs were associated with KEGG pathways, which were considered significantly enriched at P value_uncorrected < 0.05. The sequencing and preliminary analysis in this experiment were completed by Shanghai Majorbio Biopharm Technology Co., Ltd.

Information analysis.
Real-time fluorescence quantitative PCR (qRT-PCR) analysis. Four DEGs involved in KEGG pathways were further analyzed by qRT-PCR. A ChamQ SYBR Color Master Mix (2X) (Nanjing Novazan Biotechnology Co., Ltd.) reverse transcription kit was used in the ABI 7500 Real-Time PCR System. The internal reference gene was actin, and the relative expression of the genes was calculated by the 2 −∆ ∆Ct method. The primer sequences are shown in Table 7.   www.nature.com/scientificreports/ Ethics declarations. Experimental research and field studies on cultivars have complied with relevant institutional, national, and international guidelines and legislation. And all cultivars were purchased through formal channels.