Functional divergence of FTL9 and FTL10 in flowering control in rice

Abstract

Background

Floral transition in cereals is a critical phenomenon influenced by exogenous and endogenous signals, determining crop yield and reproduction. Flowering Locus T-like (FT-like) genes encode a mobile florigen, the main signaling molecule for flowering.

Results

In this study, we characterized two FT-like genes, FTL9 and FTL10, to study their functional diversity in flowering control in rice. We compared independent mutant lines of ftl10 with WT and observed negligible differences in the flowering phenotype, or agronomic traits implying potentially redundant roles of FTL10 loss-of-function in flowering control in rice. Nevertheless, we found that overexpression of FTL10, but not FTL9, substantially accelerated flowering, indicating the flowering-promoting role of FTL10 and the divergent functions between FTL9 and FTL10 in flowering. Besides flowering, additive agronomic roles were observed for FTL10-OE regulating the number of effective panicles per plant, the number of primary branches per panicle, and spikelets per panicle without regulating seed size. Mechanistically, our Y2H and BiFC analyses demonstrate that FTL10, in contrast to FTL9, can interact with FD1 and GF14c, forming a flowering activation complex and thereby regulating flowering.

Conclusion

Altogether, our results elucidate the regulatory roles of FTL9 and FTL10 in flowering control, unveiling the molecular basis of functional divergence between FTL10 and FTL9, which provides mechanistic insights into shaping the dynamics of flowering time regulation in rice.

Introduction

Flowering in plants involves transitioning from the vegetative to the reproductive phase to ensure reproduction. Several endogenous and exogenous cues are involved in flowering time regulation; however, appropriate time and day length are critical for flowering induction [1]. Flowering is timely induced by a flowering activation complex (FAC), organized by FLOWERING LOCUS T (FT), a small phosphatidylethanolamine binding protein (PEBP), and FD, a bZIP transcription factor with the assistance of scaffold protein 14-3-3 (GF14 in rice), when plants sense to favorable flowering environmental conditions [2]. In angiosperms, the PEBP gene family can be divided into three major clades: MOTHER OF FT AND TFL1-like (MFT-like), TERMINAL FLOWER1-like (TFL1-like), and FT, which are involved in regulating several growth and developmental processes, including seed germination, flowering time, and plant architecture [3]. Several FT-like genes have been identified in various plant genomes, including O. sativa (13 FT-like homologs), Z. maize (15 FT-like homologs), and T. aestivum and barley (12 FT-like homologs each), mostly involved in flowering time regulation [4,5,6].

Moreover, among the thirteen FTLs, two florigen genes, Hd3a and RFT1, regulate flowering in rice under short days (SDs) and long days (LDs), respectively [7]. Both Hd3a and RFT1 are produced and expressed in the vascular tissues and transported to the shoot apical meristem (SAM) through the phloem. Suppression of Hd3a results in increased RFT1 expression, potentially substituting Hd3a as a floral activator by delaying flowering [7]. Additionally, activation of RFT1 in Hd3a RNAi plants correlates with higher levels of H3k9 acetylation at the RFT1 locus, potentially enhancing RFT1 expression [8, 9]. Interestingly, the knockdown of both Hd3a and RFT1 completely blocked flowering in rice under SDs, depicting the necessary and overlapping functions of these two genes for flowering in rice [7]. The genome-wide analysis reveals the overlapping functions of Hd3a and RFT1 in identifying a common set of targets crucial for floral transition as well as for other interconnected developmental processes in rice [10].

Recently, a study revealed that Hd3a and RFT1 activated FLOWERING LOCUS T-LIKE 1 (FT-L1), a florigen-like protein with unique characteristics, collaborates with Hd3a and RFT1 to facilitate vegetative-to-inflorescence meristem transition and regulate panicle branching by increasing determinacy in distal meristems [11]. Several positive and negative regulators control the expression of the florigens in leaves. Among them, the zinc finger transcription factor Heading date 1 (Hd1), a homolog of Arabidopsis CONSTANS, promotes expression of Hd3a and RFT1 under SD conditions, while it represses it under LD conditions [12, 13]. It was observed previously that Hd3a/RFT1 interact with 14-3-3 proteins in the SAM of cytoplasm and enter the nucleus to combine with OsFD1, forming FAC that induces the expression of downstream flowering-related genes OsMADS14 and OsMADS15 [2].

However, RICE CENTRORADIALIS (RCN), rice TFL1-like proteins, compete with Hd3a for 14-3-3 binding and represses florigenic activity by forming florigen repression complex (FRC) with OsFD1. The balance between the FRC and FAC is regulated depending on the ratio of Hd3a to RCN in the cell. The FRC-FAC balance regulates SAM development, where under inductive SDs Hd3a accumulation in the SAM competes with RCN for FAC formation, initiating the reproductive transition while RCN knockdown reduces FRC formation allowing Hd3a to form FAC, ultimately hindering inflorescence development [2].

Hd3a and RFT1 proteins can engage in both florigen activation and repression complexes [14]. Activation complexes to promote flowering depend on OsFD1. However, closely related additional bZIPs, including Hd3a BINDING REPRESSOR FACTOR1 (HBF1) and HBF2, form repressor complexes that reduce Hd3a and RFT1 expression to delay flowering. The preference of RFT1 and Hd3a to interact with OsFD1 or the HBFs can be driven by relative expression patterns or modifications of OsFD1 and the HBFs under different growing conditions [14].

Furthermore, different FT-like members can fine-tune heading date and plant architecture by regulating the balance of FAC and FRC in rice. For example, OsFTL12 interacts with GF14b and OsFD1 to form the FRC by competing with Hd3a for binding [15].

Moreover, several members of the FT-like gene family have been identified to regulate flowering time in different plant species. Orthologs of FTL9 in B. distachyon (BdFTL9) and maize (ZmFTL9) are reported to be involved in flowering [16,17,18]. Additionally, FTL10, the closest paralog of FTL9, is involved in flowering control in rice [19,20,21]. However, a recent study has demonstrated that FTL9 does not affect flowering in both loss-of-function and gain-of-function experiments and therefore more directly affects grain size and number in rice [22]. OsFTL9 protein from Nipponbare (subspecies japonica) contained a stop codon in the PEBP domain, predicting its non-functionality [23]. During rice evolution and domestication, the loss-of-function allele of FTL9 emerged and became fixed, with a non-sense mutation identified in the second exon among O. rufipogon accessions, rendering FTL9 as a non-functional protein. The mechanism highlights a strategy that when maternal resources are sufficient, wild plants increase their offspring number and prevent the increase of offspring size by the action of FTL9, which helps expand their habitats in a fluctuating environment. Furthermore, it also suggests the functional diversification of FTL9 and FTL10 in rice, where FTL10 regulates flowering like other grasses, while FTL9 switched its function to regulate grain size and number [22]. As a dietary staple for over half of the global population, rice consistently holds a central position in the realm of plant science. Flowering time regulation in rice is a well-studied pathway involving FT-like genes that are particularly important in response to photoperiodic cues. FT-like and its homolog proteins act as a mobile signal that moves from leaves to SAM where it triggers the activation of floral meristem identity genes and initiates flowering. Based on the studies of Hd3a and RFT1, the functions of FT-like genes have been well characterized in flowering induction. However, the involvement of florigens in varied day lengths in flowering control and other agronomic characteristics is largely unclear. Therefore, this study investigates the response of rice FT-like genes, FTL9 and FTL10, under SDs. Our results indicated that ftl10 exhibits redundant roles with other FTLs in regulating flowering time in rice under SDs. Critically, overexpression of FTL10 but not FTL9 accelerated flowering in rice under SDs which provides further insights into the molecular basis and functional divergence between FTL10 and FTL9 in the regulation of flowering in rice.

Methods

Plant materials and growth conditions

The Nipponbare (O. sativa L. ssp. japonica) cultivar seeds provided by the Zhejiang Provincial Key Laboratory of Crop Genetic Resources, college of Agriculture and Biotechnology, Zhejiang University was used as plant material in this study. The procedures used for plant material collection complied with relevant institutional, national, and international guidelines and legislation. Mature seeds were soaked in deionized water and transferred to a net floating on a 0.5mM CaCl₂ solution. Afterward, seeds were germinated for 2 days at 28 °C in the dark and were transferred to soil. Plants were grown in growth chambers under SD conditions (30 °C 10 h light/25°C 14 h dark) and LD conditions (30 °C 14 h light/25°C 10 h dark). Light was provided by fluorescent white light tubes (700 mmol/m²/s). Sampling was done using seedlings or leaves used in the experiments harvested at Zeitgeber 4. All samples (each containing three biological replicates) were immediately frozen in liquid nitrogen and processed or stored at -80 °C for further RNA extraction.

Vectors construction and plant transformation

To generate OsFTL9 and OsFTL10 overexpression lines, the full-length coding sequences OsFTL9 and OsFTL10 were amplified from Nipponbare (O. sativa L. ssp. japonica) cDNA with the primers set Fw-OsFTL9-EcoR1/Re-OsFTL9-EcoR1 and Fw-OsFTL10-EcoR1/Re-OsFTL10-EcoR1 (Supplementary Table 1), and cloned into the pCAMBIA1390 with the control of the maize (Zea mays) UBIQUITIN (UBI) promoter. OsFTL10, OsHd1, and OsGI target sequences in sgRNA expression cassettes were cloned into CRISPR-Cas9 system constructs with U3 promoter or U6a promoter to generate mutants. The primers used for vector construction are listed in Supplementary Table 1. Transgenic rice plants were generated by A. tumefaciens strain AGL1 using rice embryogenic callus.

RNA expression analysis

Full-expanded leaves were harvested at the lighting point around 5 weeks under SDs and LDs for gene expression analysis. Total RNA was extracted from leaves using RNAiso Plus (Takara). cDNA was synthesized from 1ug of total RNA with oligo (dT) primers by reverse transcriptase (Promega). The quantitative analysis of gene expression was performed on the Step-One Plus real-time PCR system with the Power Up SYBR Master Mix (ABI) with three biological replicates according to the manufacturer instructions as described [24]. The following cycling conditions were used for Real-time qPCR: 2 min at 95 °C, 40 cycles of 10 s at 95 °C and 40 s at 65 °C, and a final step for melting curve determination (15 s at 95 °C, 1 min at 60 °C and 15 s at 95 °C). GAPDH was used as an internal control. Gene expression was calculated using the 2^−ΔΔCt method following [25]. All primers used in RT-qPCR experiments are listed in Supplementary Table 1.

For transcriptome analysis, RNA was extracted from osftl10 and WT with three biological replicates under SDs. Library preparation and RNA-sequencing were performed at Novogene Bioinformatics Company (Beijing, China) with illumina NovaSeq 6000 platform (PE150). The detailed information for RNA-seq data is presented in Supplementary Table 2. Clean reads were aligned to the Rice genomic sequence (O. sativa_323_v7.0) and downloaded from the Phytozome database (Version 12.0) using Hisat2 [26]. Differentially expressed genes were selected with Hochberg-adjusted P-values (FDR) < 0.05 and |log2 (Fold change)| >1 setting as the threshold. The GO terms pathways were carried with FDR < 0.05 and were considered as significantly altered and get the GO annotations based on biological process.

Phylogenetic analysis

The phylogenetic tree was constructed by MEGA software using the Neighbor-Joining method. Alignments of the protein sequences were produced by ClustalW. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branch. Accession numbers for the Phytozome database website (https://phytozome-next.jgi.doe.gov/) genes referred to phylogenetic trees were as follows: FT (AT1G65480), TSF (AT4G20370), TFL1 (AT5G03840), BdFTL4 (Bradi1g38150), BdFTL1 (Bradi1g48830), BdFTL2 (Bradi2g07070), BdFTL10 (Bradi2g19670), BdFTL9 (Bradi2g49795), BdFTL6 (Bradi3g08890), BdFTL12 (Bradi3g48036), BdFTL4 (Bradi4g35040), BdFTL5 (Bradi5g14010), FTL8 (LOC_Os01g10590), FTL1 (LOC_Os01g11940), FTL9 (LOC_Os01g54490), FTL13 (LOC_Os02g13830), FTL5 (LOC_Os02g39064), FTL6 (LOC_Os04g41130), FTL10 (LOC_Os05g44180), RFT1 (LOC_Os06g06300), Hd3a (LOC_Os06g06320), FTL12 (LOC_Os06g35940), FTL4 (LOC_Os09g33850), FTL11 (LOC_Os11g18870), FTL7 (LOC_Os12g13030).

Yeast two-hybrid assays

The full-length coding sequences of OsFTL9 and OsFTL10 were amplified from Nipponbare cDNA and cloned into pGBKT7. The full-length coding sequences of OsFD1 and OsGF14c were amplified from Nipponbare cDNA and cloned into pGADT7. Primers for the constructs are listed in Supplementary Table 1. The recombinant plasmids were co-transformed into yeast strain AH109 using the lithium acetate-based transformation protocol. Positive colonies grown on SD/-Leu/-Trp (-LT) medium were transferred onto selective SD/-Leu/-Trp/-His/-Ade (-LTHA) medium. Positive interactions on -LTHA medium were observed after incubation at 30 °C for approximately 4 days.

BiFC assays

The full-length coding sequences of OsFTL9, OsFTL10, OsFD1, and OsGF14c were amplified and constructed into vectors containing either N- or C-terminal-enhanced yellow fluorescence protein fragments. Gene sequences were amplified by PCR and restriction cloned into the (Bam HI and/or Eco RI). Primers for the constructs are listed in Supplementary Table 1. The recombinant plasmids were transformed into A. tumefaciens strain EHA105, and transiently co-expressed in N. benthamiana leaves through Agrobacterium-mediated transformation adjusted to OD600 = 0.8. YFP Fluorescence signals were observed after 2–3 days using a confocal laser scanning microscope (Olympus FV3000).

Results

Phylogeny of FT-Like genes inA. thaliana, B. distachyon, and O. sativa

Flowering time control is a critical agronomic trait that determines the fate of crop production. Various flowering genes, including FT, GI, and HD, orchestrate the flowering activation complex to regulate flowering time. Unlike the two FT-like protein homologs in A. thaliana, there are 13 FT orthologs in rice (O. sativa) and 6 FT-like orthologs in B. distachyon [17, 27, 28]. Due to the close sequence alignment of OsFTL9 and OsFTL10 with BdFTL9, which has strongly influenced the flowering transition when overexpressed in B. distachyon [17], we asked whether FTL9 and FTL10 can influence the flowering transition in rice. Therefore, we aligned the protein sequences of FTs in A. thaliana, B. distachyon, and rice and found a similarity, illustrating that B. distachyon and rice FTs share the same protein sequences and will function in flowering control (Fig. 1a). Thus, we further investigated the biological significance and molecular basis of FTL9 and FTL10 in rice by investigating them with CRISPR-cas9 and OE lines to study them in detail.

Fig. 1

Phylogenetic analysis of FT-like proteins and expression patterns of FTL-9 in rice. (a) A phylogenetic tree of FT-like proteins in A. thaliana, B. distachyon, and O. sativa. The tree was constructed by MEGA software using the Neighbor-Joining method produced by ClustalW with bootstrap test 1000 replicates. The evolutionary distances were computed using the Poisson correction method. (b) Expression levels of FTL9 in different rice tissues. (c) Diurnal expression patterns of FTL9 under LDs and SDs. Abscissas represent the zeitgeber time (ZT) point. The white and black bars parallel to the horizontal axis represent light and dark photoperiods respectively. (d) Schematic diagram of GI and construction of gi knockout mutants using CRISPR-cas9 approach to develop transgenic rice lines. Blue boxes represent exons and black lines represent introns. The sgRNA sequence that specifically targets GI is indicated in blue. The PAM sequence is underlined in red. The red dashes indicate the deletion of base pairs and the insert base pairs are marked in red. (e) Schematic diagram of Hd1 and construction of oshd1 knockout (CRISPR-Cas9) transgenic rice lines. Blue boxes represent exons and black lines represent introns. The sgRNA sequence that specifically targets Hd1 is indicated in blue. The PAM sequence is underlined in red. The red dashes indicate the deletion of base pairs, and the insert base pairs are marked in red. (f) qRT-PCR analysis shows the relative expression levels of FTL9 in WT and gi cas9 lines under SDs. (g) qRT-PCR analysis of FTL9 relative expression levels in WT and hd1 cas9 lines under SDs. Leaf samples were harvested from plants grown for 5 weeks. Data are means ± standard deviation. Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

Tissue-specific and diurnal expression pattern of FTL9 in rice

Tissue-specific expression of a gene is also very important which allows efficient utilization of resources and adaptation to diverse conditions. To study the tissue-specific expression pattern of FTL9, we performed qRT-PCR to study the relative expression patterns of FTL9 in different plant tissues. We observed significantly higher FTL9 transcription in leaves (Fig. 1b). Further, we noticed that in leaf blade III, the expression of FTL9 was higher compared to other leaf blades, indicating that FTL9 accumulates in leaves, specifically in leaf blade III.

In B. distachyon, FTL9 is a short-day-specific gene [17, 29]. Since FT genes usually have diurnal expression patterns, we examined the oscillation of FTL9 every 4 h during the day-night cycle. Accordingly, we observed a high relative expression of FTL9 at the end of the dark and low expressions in light time when plants were grown in SDs. However, the relative expression of FTL9 was significantly repressed at any time point when plants were grown in LDs (Fig. 1c). Altogether, these results indicated the tissue-specific behavior of FTL9 specifically induced under SDs in rice.

GI and Hd1 positively regulate FTL9 expression in rice under SDs

Flowering initiation is a complex agronomic trait that defines the transition between vegetative and reproductive stages and is controlled by various flowering-related genes and internal cues, such as circadian rhythm and photoperiod. Many factors are involved in the control of photoperiodic flowering, including genes like GIGANTEA (GI) and Heading date 1 (Hd1) [30, 31]. GI plays a central role in photoperiodic flowering, both dependent and independent of the gene CONSTANS (CO), the key gene in the photoperiod pathway that triggers FT in leaves [32]. Therefore, on the first note, we prepared gi loss-of-function mutants with the help of CRISPR-cas9 technique (Fig. 1d).

Similarly, Hd1, the CO ortholog in rice, plays a dual role in flowering control, promoting flowering under SDs through the expression of Hd3a, the FT ortholog, and suppressing flowering under LDs by the repression of Hd3a [33]. Therefore, we prepared hd1 mutants using CRISPR-Cas9 (Fig. 1e).

Since the accumulation of these flowering-related genes in SDs and their role in flowering time control, we observed that the relative expression of FTL9 in GI loss-of-function was significantly reduced, indicating the involvement of GI in the expression regulation of FTL9 (Fig. 1f). Interestingly, when we measured the expression of FTL9 in HD1 loss-of-function under SDs, we observed that the expression of FTL9 was obviously decreased compared with WT (Fig. 1g), suggesting that hd1 positively regulates the expression of FTL9 in rice under SDs.

Tissue-specific and diurnal expression pattern of FTL10 in rice

To study the tissue-specific expression pattern of FTL10, we performed qRT-PCR of FTL10 in different tissues of rice. We observed significantly higher expression of FTL10 accumulation in leaves (Fig. 2a). Further, we noticed that in leaf blade III, the expression of FTL10 was higher compared to other leaf blades, following the same trend as observed in FTL9, indicating that FTL10 accumulates in leaves, specifically leaf blade III.

Fig. 2

Expression pattern of FTL10 in different tissues and day lengths. (a) Expression levels of FTL10 in different rice tissues. (b) Diurnal expression patterns of FTL10 under LDs and SDs. Abscissas represent the zeitgeber time (ZT) point. The white and black bars parallel to the horizontal axis represent light and dark periods, respectively. (c) qRT-PCR analysis shows the relative expression levels of FTL10 in WT and osgi cas9 lines under SDs. (d) qRT-PCR analysis of FTL10 relative expression levels in WT and hd1 cas9 lines under SDs. Samples were harvested from plants grown for 5 weeks. Data are means ± standard deviation. Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

Since FT genes usually have diurnal expression patterns, we examined the oscillation of FTL10 every 4 h during the day-night cycle. Accordingly, we observed a high relative expression of FTL10 at the end of the dark and low expressions in light time when plants were grown in SDs. However, the relative expression of FTL10 was significantly repressed at any time point when plants were grown in LDs (Fig. 2b). Altogether, these results indicated the tissue-specific behavior of FTL10 specifically induced under SDs in rice.

GI and Hd1 positively regulate FTL10 expression in rice under SDs

Furthermore, since a higher accumulation of FTL10 in rice under SDs, we observed that the relative expression of FTL10 in gi loss-of-function was significantly reduced, indicating the involvement of GI in the expression regulation of FTL10 (Fig. 2c). Interestingly, when we measured the expression of FTL10 in HD1 loss-of-function under SDs, we observed that the expression of FTL10 was obviously decreased compared with WT (Fig. 2d), suggesting that HD1 positively regulates the FTL10 in rice under SDs.

Overlapping functions of OsFTL10 in flowering control in rice

Since, OsFTL9 from Nipponbare (subspecies japonica) is predicted to be non-functional and does not affect flowering in both loss-of-function and gain-of-function experiments [22, 23], we asked whether FTL10 can perform any role in flowering time control in rice. Therefore, we generated ftl10 loss-of-function mutants using the CRISPR/Cas9 approach to analyze the FTL10 deficiency on flowering in rice (Fig. 3a). However, we observed no significant phenotypic differences and days to heading between the WT and FTL10, indicating that FTL10 has redundant roles with other FTL genes in the biological control of flowering in rice (Fig. 3b, and c). Furthermore, the relative expression levels of MADS14 and MADS15 observed in ftl10 showed no significant differences between WT and ftl10 loss-of-function, indicating FTL10 has no influence over flowering control in rice under SDs (Fig. 3d and e).

Fig. 3

Regulatory roles of ftl10 loss-of-function in flowering time control in rice under SDs. (a) Schematic diagram of FTL10 and construction of ftl10 knockout transgenic rice lines using CRISPR-Cas9. Blue boxes represent exons and black lines represent introns. The sgRNA sequence that specifically targets FTL10 is indicated in blue. The PAM sequence is underlined in red. The red dashes indicate the deletion of base pairs and the inserted base pairs are marked in red. (b) Representative flowering phenotypes of WT and ftl10 under SDs. White arrows point to spikes. Scale bars, 4 cm. (c) Measurement of days to flowering time in WT and osftl10 under SDs. (n = 15 plants). (d) and (e) qRT-PCR analysis of MADS14 and MADS15 expression in WT and ftl10 lines under SDs. Data are means ± standard deviation. Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

Transcriptome profiling and agronomic characterization of FTL10 loss-of-function

Transcriptome sequencing of the FTL10 loss-of-function showed that a significant number of genes and pathways were differentially regulated. Several genes related to plasma membrane and protein kinase activity were differentially expressed among ftl10 and WT. A higher number of differentially expressed genes were observed for the plasma membrane, chloroplast regulation, and ATP binding followed by protein phosphorylation, protein serine activity, and kinase activity (Fig. 4a and b). However, similar to the phenotypic characterization of ftl10 in flowering control, our transcriptome analysis did not observe any significant changes in the gene expression profile. Since no noticeable changes were observed in the flowering phenotype, we shifted our observations to agronomic traits to explore any potential alterations in other aspects of plant growth and development. However, we failed to observe any significant changes in plant height, the number of effective panicles, the number of primary branches per panicle, and the number of spikelets per panicle (Fig. 4c, d, e, and f). We also compared seed length and seed width of ftl10 with WT, however, we did not observe any noticeable changes in seed measurements (Fig. 4g and h), indicating that FTL10, although it belongs to a flowering control gene family, does not influence flowering control or other agronomic characterization in rice under SDs.

Fig. 4

Regulatory roles of ftl10 loss-of-function in flowering time control in rice under SDs. (a) Volcano plot of significant up-and down-regulated differentially expressed genes in osftl10 vs. WT in rice leaves under SDs. (b) Gene Ontology (GO) enrichment analysis for differentially expressed genes in osftl10 compared with WT in rice leaves under SDs. (c) Measurement of plant height of WT and osftl10 under SDs (n = 10 plants). (d, e, and f) Measurement of the number of effective panicles per plant, the number of primary branches per panicle, and spikelets per panicle (n = 10 plants). (g and h) Measurement of seed length and seed width (mm) among WT and osftl10 (n = 10 plants). Data are means ± standard deviation. Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

Over-expression of FTL10 leads to early flowering in rice under SDs

Next, we observed the overexpressed lines of FTL10 in rice under SDs. Interestingly, we witnessed a 300-600-fold increase in the relative expression level of FTL10 overexpression (FTL10-OE) in several independent rice lines (Fig. 5a). Moreover, we observed significant differences in the phenotype of WT and FTL10-OE plants under SDs. There were significant differences in plant height between FTL10-OE and WT, indicating FTL10 to be an ideal growth-regulating gene in rice (Fig. 5b). More interestingly, there was a highly significant difference in time to flowering between WT and FTL10-OE plants under SDs. FTL10-OE plants flowered approximately 9.8 days earlier than their respective WT (Fig. 5c). Furthermore, we also observed the relative expression of MADS14 and MADS15 in FTL10-OE to check their involvement in flowering time control in rice under SDs. Most interestingly, we found a highly significant relative expression level of MADS14 and MADS15 in FTL10-OE compared with WT (Fig. 5d and e). The relative expression level of FTL10 was 35–70 folds higher in FTL10-OE compared with WT; however, a 2-15-fold increase was observed for MADS14 and MADS15 in FTL10-OE compared to WT. Considering agronomic traits other than flowering, we observed significant differences in plant height, the number of effective panicles per plant, the number of primary branches per panicle, and the number of spikelets per panicle among FTL10-OE and WT (Fig. 5f, g, h, and i). FTL10-OE plants exhibited a significantly higher number of effective panicles with a significant reduction in primary branches and spikelets per panicle. However, there were no observable changes noticed in seed length and seed width among FTL10-OE and WT (Fig. 5j and k), maybe following the mechanism that when maternal resources are sufficient, wild plants increase their offspring number and prevent the increase of offspring to help expand their habitats in a fluctuating environment. Altogether, these results suggested that FTL10, when overexpressed in rice, can influence the relative expression levels of MADS14 and MADS15, alternatively playing a significant role in flowering time regulation and other agronomic traits in rice under SDs. However, further investigation is warranted to elucidate the roles of FTL10 in regulating rice agronomy and development.

Fig. 5

Over-expression of FTL10 leads to early flowering in rice under SDs. (a) qRT-PCR analysis of FTL10 expression in WT and FTL10-OE lines in rice leaves. (b) Representative flowering phenotypes of WT and FTL10-OE lines under SDs. White arrows point to spikes. Scale bars, 4 cm. (c) Measurement of days to flowering in WT and FTL10-OE lines under SDs. (n = 15 plants). (d) and (e) qRT-PCR analysis of MADS14 and MADS15 expression in WT and FTL10-OE lines under SDs. Data are means ± standard deviation. (f) Measurement of plant height (cm) of WT and FTL10-OE lines under SDs (n = 10 plants). (g, h, and i) Measurements of the number of effective panicles per plant, the number of primary branches per panicle, and spikelets per panicle (n = 10 plants). (j and k) Measurements of seed length and seed width (mm) among WT and FTL10-OE (n = 10 plants). Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

Over-expression of FTL9 has negligible effects on flowering and agronomic traits in rice under SDs

Considering the regulatory roles of FTL10 in flowering control in rice, we investigated whether overexpression of FTL9 could influence rice development. Therefore, we constructed overexpressed lines of SD-induced FTL9 (FTL9-OE) in rice. The relative expression level of FTL9 was highly increased, ranging from 10k to 30k-fold in our several independent FTL9-OE lines (Fig. 6a). However, it was very surprising to observe no significant differences in the phenotype and days to flowering between the WT and FTL9-OE plants under SDs (Fig. 6b and c). Therefore, we also observed the relative expression of MADS14 and MADS15 in FTL9-OE. We found negligible variations in the expression patterns of MADS14 and MADS15 in FTL9-OE compared with WT (Fig. 6d and e), suggesting that FTL9, when overexpressed in rice, exerts no discernible influence on flowering regulation in rice. Since we did not observe any changes in flowering, we investigated other agronomic traits to explore any potential alterations in rice growth and development. However, there were no observable changes in plant height, the number of effective panicles, the number of primary branches per panicle, and the number of spikelets per panicle among FTL9-OE and WT (Fig. 6f, g, h, and i). We also compared seed length and seed width of FTL9-OE with WT, however, we did not observe any noticeable changes in seed measurements (Fig. 4j and k), indicating that FTL9-OE does not exert any changes in flowering control or other agronomic characteristics in rice under SDs.

Fig. 6

Over-expression of FTL9 in rice under SDs. (a) Relative expression of FTL9 in WT and FTL9-OE lines in leaves. (b) Representative flowering phenotypes of WT and FTL9-OE lines under SDs. White arrows point to spikes. Scale bars, 4 cm. (c) Days to flowering in WT and OsFTL9-OE lines under SDs. (n = 15 plants). (d) and (e) qRT-PCR analysis of MADS14 and MADS15 expression in WT and OsFTL9-OE lines under SDs in rice. (f) Measurement of plant height (cm) of WT and FTL9-OE lines under SDs (n = 10 plants). (g, h, and i) Measurements of the number of effective panicles per plant, the number of primary branches per panicle, and spikelets per panicle (n = 10 plants). (j and k) Measurements of seed length and seed width (mm) among WT and FTL9-OE (n = 10 plants). Data are means ± standard deviation. Asterisks indicate the P-value calculated using ANOVA: ***: P < 0.001; **: P < 0.01; *: P < 0.05; ns = not significant

The mechanism of FTL9 and FTL10 in the flowering control in rice

Flowering initiation is a complex process that requires the interaction of several candidate genes involved in the activation of the flowering. Therefore, to check the interaction of FTL9 and FTL10 with other flowering candidate genes, we performed yeast two-hybrid (Y2H) and bimolecular fluorescence complementation (BiFC). Our assay for Y2H between FTL9 and FD1 showed no interaction, while there was a strong interaction noticed between FTL10 and FD1 (Fig. 7a and b). This conspicuous difference in interaction patterns elucidates that FTL9 may not possess direct implications to regulate flowering. However, FTL10 can interact with other candidates to activate the flowering complex for flowering initiation. Furthermore, to confirm our results through BiFC assay, there was still no interaction between FTL9 and FD1, while there was a solid interaction noticed between FTL10 and FD1 (Fig. 7c and d), depicting that FTL10 and FD1 can interact with other flowering genes to activate the flowering complex in rice. We expanded our investigation to GF14c to observe its regulatory roles in the flowering pathway in rice and found no interaction between FTL9 and GF14c (Supplementary Fig. 1), indicating that FTL9 may have different regulatory mechanisms independent of GF14c. However, there was a strong interaction observed between FTL10 and GF14c (Supplementary Fig. 1), suggesting that FTL10 may interact with GF14c to regulate flowering in rice, indicating a direct functional relationship between these two genes in the flowering pathway. Altogether, these results elucidate that, unlike FTL9, the strong interactive signals observed in Y2H and BiFC assays between FTL10, FD1, and GF14c suggest that FTL10, might be directly involved in the flowering control pathway. These findings highlight the specific molecular interactions in shaping the dynamics of flowering time regulation in rice.

Fig. 7

Interactions of FTL9 and FTL10 with FD1 in yeast and tobacco. (a) and (b) Interaction between FTL9/FTL10 and FD1 proteins in yeast cells. Yeast cells grown on selective media without Leu, Trp (-LT), His and Ade (-LTHA) represent positive interactions. (c) and (d) Bimolecular fluorescence complementation (BiFC) assay of the interactions between FTL9/FTL10 and FD1 in N. benthamiana leaves. Scale bar, 20 μm. N-YFP and C-YFP clones containing empty vectors were used as negative controls

Discussion

Flowering initiation is a complex developmental process that ensures the reproductive success of plants, critically influenced by environmental cues. Day length influences the expression levels of several flowering-related genes that interact in the flowering initiation network. Under LD plants, CONSTANS (CO) interact with the cis-element of Flowering Locus T (FT) to induce flowering [34] illuminates how flowering is induced under LDs; however, how SD plants trigger flowering is still unclear. Therefore, in the present study, we characterized two SD-induced FTs, FTL9 and FTL10, to investigate their role in flowering time control in rice under SDs. FT family proteins play central roles in regulating reproductive development in the flowering control pathway, although several FT homologs repress the flowering transition [35]. The biological significance of FT-like genes is predominant to induce flowering in SD plants. In SD conditions, the loss of function of ftl9 can lead to delayed flowering in B. distachyon. Conversely, with the transition to long photoperiods in spring, winter plants initiate flowering through various mechanisms, including FT1 activation and FTL9 reduction [17]. Hence, FTL9 functions offer valuable insights for SD plants to meet diverse agricultural demands and adapt to climatic conditions. Recent investigations have reported that FTL9 orthologs in B. distachyon and maize and its paralog FTL10 are involved in flowering [16,17,18,19,20,21], however, rice FTL9 does not affect flowering both in gain-of-function and loss-of-function experiments. This investigation further implies the functional diversification of FTL9 and FTL10 in rice. Specifically, FTL10 regulates flowering in rice like other grasses, while FTL9 shifts its function to regulated grain size and number [22].

Appropriate flowering time in rice is imperative for enhancing environmental adaptability and augmenting grain yield. Photoperiod, a vital environmental signal, governs flowering. The intricate interplay of light stimuli and circadian rhythms orchestrates the regulation of photoperiodic flowering. In rice, Heading date 1 (Hd1), the CO ortholog, exerts dual regulation, promoting flowering under SDs by enhancing the FT counterpart Hd3a and retarding flowering under LDs by suppressing Hd3a [33]. A previous study unveiled that Ghd7 forms a complex with Hd1, leading to the suppression of Hd3a transcription and concurrent inhibition of Early Heading date 1 (Ehd1) expression, thereby negatively regulating flowering in rice under LDs [36]. Another study identified a mutation in OsELF3/ELF7, an Arabidopsis ELF3 homolog, which markedly elevates Ghd7 transcription, leading to delayed flowering in rice in both LDs and SDs [37].

Moreover, Hd1 has contrasting regulatory roles dictated by day length, facilitated by DAYS TO HEADING 8, a NUCLEAR FACTOR Y (NF-Y) transcription factor potentially interacting with Hd1 [38]. Further investigation is warranted to explain the specific pathway involving CO1 and NF-Y, enabling the activation of FT1 and the suppression of FTL9 to promote flowering in rice under SDs. However, despite potential regulation by CO1, FTL9 expression remains relatively unaltered in CO1 knock-down lines under LDs [17]. Additional investigations are required to discern alternative upstream regulators of FTL9 in the photoperiodic flowering pathway.

Ehd1 facilitates flowering in rice under both SDs and LDs. It plays a pivotal role in two SD modules, OsGI-OsCO-Ehd1-Hd3a and OsGI-OsMADS15-Ehd1-Hd3a, where Ehd1 acts upstream of Hd3a, directly upregulating Hd3a expression to promote flowering. Enhanced OsFTL10 overexpression, coupled with elevated Ehd1 and Hd3a levels in transgenic plants, is likely to accelerate flowering in rice [39]. Given that Ehd1 plays a pivotal role in integrating flowering signals from both SDs and LDs, its regulation by various genes, including the feedback from FTL10, merits further investigation in rice. In the context of SDs, MADS51, downstream of GI, functions as a flowering promoter. Reduction in the expression level of MADS51 in FTL10-OE lines contradicts the early flowering phenotype, potentially attributed to the elevated Ehd1 expression downstream of MADS51 [19]; however, RFT1 displayed a modest downregulation. Therefore, it is suspected that FTL10 or other genes related to flowering may have modulated the expression of MADS51, analogous to the effects observed with Hd3a and RFT1, where, in response to RNAi-induced Hd3a downregulation, RFT1 was activated to promote flowering under SDs [8]. RFT1 primarily initiates flowering under LDs, whereas Hd3a predominantly promotes flowering under SDs. Hd3a and RFT1 respond to photoperiodic regulation by undergoing transcription, as evidenced by their silencing via RNAi, leading to a complete absence of flowering even 300 days post-sowing [7, 40].

In Arabidopsis, FD with FD PARALOG (FDP) is redundant. Double mutations of fd/fdp flower later than fd single mutations, and almost completely suppress precocious flowering of 35 S: FT [41]. In addition, by mutating FD3 and FD4 in rice, floral-promoting factors were identified. Mutations in fd4, but not in fd3, can delay flowering and activation of MADS targets. However, flowering was not delayed as much as in fd1 mutants, indicating a less prominent role compared with FD1. Also, fd1/fd4 double mutants did not further delay flowering compared with fd1 single mutants, proposing that FD1 activated FD4 transcription in rice [42]. Thus, this arrangement does not fully exclude redundancy but suggests a complex regulatory network.

Further, MADS62 controls pollen maturation and germination, partially redundantly with MADS63 and MADS68 in rice [43]. FD1 and FD4 promoted OsMADS62 expression, a function they probably perform after meristem commitment, during inflorescence and flower development. It is also proposed that FAC containing FD1 directly targets the relative expression of MADS15 [44]. However, neither FD1 nor FD4 is bound to the MADS14, MADS15 or MADS34/PAP2 promoter in rice [42]. Other proteins could possibly stabilize FD4 and FD1 in vivo and allow binding to these loci. Alternatively, the rice FD1-MADS connection might be indirect. Data in support of a direct connection have not been as thoroughly repeated and validated in rice as they have been in Arabidopsis, in which direct contacts between FD and the C-box element in the AP1 promoter were disproved [45,46,47,48]. Further assessment of in vivo binding is necessary.

Thus far, no FT-like gene with dual functionality in both SD and LD contexts has been identified. Knockout mutants of FTL10 displayed no significant phenotypic alterations under SD conditions in rice. Genetic redundancy in which the effect of mutations in a single gene is strongly enhanced by a second mutation in a related gene has been demonstrated in osfd1-2 and osfd4-1, which do not have completely overlapping functions [42]. However, the overexpression of FTL10, an Hd3a homolog, induced early flowering in rice. A study demonstrated that the overexpression of both genes FTL10 and Hd3a, under the same promoter, yielded varying flowering periods, potentially attributed to amino acid variations between the two proteins [19].

Based on the study, several observations were made differentiating between FTL10 and Hd3a: (1) FTL10 exhibited high expression in young leaves, whereas Hd3a primarily expressed in mature leaves; (2) Both FTL10 and Hd3a proteins were localized in the cytoplasm and nucleus, yet BiFC assay results indicated that FTL10 predominantly interacted with GF14c in the nucleus, while Hd3a interacted with GF14b in the cytoplasm; (3) FTL10-OE led to increased Ehd1 expression, which, in turn, may affect the flowering pathway. The relative expression level of MADS51 was also decreased in FTL10 transgenic plants, suggesting that flowering regulation of FTL10 may exhibit subtle distinctions from the pivotal florigens Hd3a and RFT1 [19].

According to our observations, the expression patterns of MADS14 and MADS15 in FTL9-OE lines showed no significant differences when compared with WT. However, MADS14 and MADS15 expression in FTL10-OE showed highly significant differences between the OE lines and WT, suggesting that FTL10, when overexpressed in rice under SDs, can play significant roles in flowering control. Moreover, we also demonstrated in the current study that FTL10 can interact with FD1 to induce floral transition in rice. FTL10 interacts with FD1 in Y2H and BiFC assays; however, it is not clear whether the interaction of FTL10 and FD1 requires the presence of 14-3-3 or not. Therefore, further investigation is needed to demonstrate the interaction of 14-3-3 with FT-like genes to regulate flowering time in rice under SDs.

Conclusion

In conclusion, this study investigates the functional diversity of FTL9 and FTL10 in rice flowering time regulation under SD conditions, giving rise to a nuanced perspective. While there is a non-sense mutation identified in the second exon of FTL9 among O. rufipogon accessions [22], rendering FTL9 as a non-functional protein, FTL10, the closest homolog of FTL9, exerted no discernible influence over flowering time regulation in rice under SDs. Nevertheless, overexpression of FTL10, but not FTL9, substantially accelerated flowering, and regulated other agronomic traits without regulating seed size. Moreover, the augmented expression patterns of MADS14 and MADS15 in FTL10 overexpression lines proposed an alternative regulatory pathway in flowering regulation in rice under SDs. Collectively, these findings imply the existence of unexplored FT-like genes involved in flowering time regulation in rice under varying environmental cues, encouraging further exploration to study the comprehensive orchestration of flowering time control in rice across diverse environmental settings.