1. Introduction

Rice (Oryza sativa L.) is one of the most important crops, feeding more than half of the world's population. In the 1960s, the first global Green Revolution was achieved by reducing plant height, which enhanced the lodging resistance of rice and ultimately led to a breakthrough in rice yield, greatly stabilizing global food security. However, the gene (i.e., SD1) that triggered the first global Green Revolution was not cloned until 2002, revealing its role in the biosynthesis of bioactive gibberellins (GAs) (Sasaki et al., 2002). Given the key role of the SD1 gene, it is highly necessary to conduct continuous and in-depth research on GA biosynthesis.

Gibberellins are a class of plant hormones that are crucial for many developmental processes in plants, including seed germination, stem elongation, leaf expansion, trichome development, pollen maturation, and flowering induction (Achard and Genschik, 2009). The importance of GAs in regulating plant growth is demonstrated by the severe dwarf phenotype of GA-deficient plants (Sasaki et al., 2002). Interestingly, naturally occurring or induced loss-of-function mutations of the GA biosynthesis genes reduced plant height but enhanced seed dormancy (Ye et al., 2015). Since their first discovery, more than 130 kinds of GAs have been identified in plants, fungi, and bacteria, but only a few GAs are biologically active. Therefore, there are many non-biologically active GAs in plants, which are precursors of biologically active forms or inactivated metabolites (Yamaguchi, 2008). The major biologically active GAs, including GA₁, GA₃, GA₄, and GA₇, are derived from a basic diterpene carboxylic acid skeleton and usually have a C₃-hydroxyl group (Yamaguchi, 2008). The SD1 gene that this study focuses on is a GA-biosynthetic enzyme that can affect both the GA₁ and GA₄ pathways (Su et al., 2021). GAs are key hormones that regulate the growth of the intercalary meristem. They significantly increase internode length by activating cell division activity in the intercalary meristem and promoting cell elongation. Stem elongation is caused by the repeated division and subsequent elongation of cells produced by the apical meristem and the intercalary meristem (Davière et al., 2014). Intercalary meristems are promeristems located between mature tissues and are derived from the remnants of the apical meristem. There are obvious intercalary meristems at the base of stem internodes, leaves, and leaf sheaths. Their activity promotes rapid growth of the plant stem (i.e., intercalary growth), ultimately leading to flowering. Typical intercalary meristems are present in many monocotyledonous plants, especially in gramineous plants such as rice, wheat, and maize (Ayano et al., 2014).

The TCP4 transcription factor obtained by screening the library with the SD1 promoter is the core of this study. Initially, TCP was a plant-specific transcription factor family composed of maize TEOSINTE-BRANCHED1 (TB1), snapdragon CYCLOIDEA (CYC), and rice PROLIFERATING CELL FACTORS 1 and 2 (PCF1 and PCF2) transcription factors (TFs) (Martín-Trillo and Cubas, 2010). TCP domain is a basic helix-loop-helix (bHLH) motif that enables DNA binding and protein–protein interactions for TCP TFs. It is located almost in the middle of the primary amino acid sequence (Martín-Trillo and Cubas, 2010). According to the sequence differences in TCP domain, TCP family can be divided into two classes. Class Ⅰ TCP TFs contain 55 amino acids in the TCP domain, and their DNA-binding sequence is 'GGNCCCAC', while Class Ⅱ TCP TFs have 59 amino acids in the TCP domain, whose DNA-binding sequence is 'GTGGNCCC' (Kosugi and Ohashi, 2002; Martín-Trillo and Cubas, 2010).

The TCP TFs were initially discovered to be involved in regulatory areas including branching, floral symmetry, leaf development, and cell cycle-mediated growth regulation (Doebley et al., 1995; Luo et al., 1996; Kosugi and Ohashi, 1997; Palatnik et al., 2003). For example, Class Ⅱ type TCP mutants brc1 in Arabidopsis thaliana exhibit excessive stem branching, indeed, BRC1 promotes reversible growth inhibition in the leaf axils, so once BRC1 is downregulated, buds regain normal growth and produce branches (Finlayson, S.A., 2007). Similarly, Class Ⅱ type TCP cyc mutants in snapdragon show reduced bilateral symmetry in flowers, indicating the functional role of CYC gene in reducing cell proliferation in the dorsal region of the young floral meristem (Luo et al., 1999). Additionally, Class Ⅱ type TCP mutants cin in tomato have defects in lateral organs, whose leaf cells remain divided for a long period, ultimately producing larger leaves (with altered leaf-shape and wrinkled leaf-surfaces), suggesting CIN-type genes function in restricting cell proliferation at the margins of developing leaf primordia (Ori et al., 2007). Fusion of TCP3 (a Class Ⅱ type TCP that belongs to the CIN subclass) with a dominant repressor led to severe developmental defects in all organs of Arabidopsis, including ectopic stem formation, serrated leaves, modified sepals and petals, and wavy siliques. These defects were attributed to the aberrant expression of the boundary-specific genes CUC and LATERAL ORGAN BOUNDARIES (Koyama et al., 2007). From an evolutionary perspective, TCP TFs first emerged in freshwater algae, Charophyta (Navaud et al., 2007). In the moss Physcomitrella patens, knockout of PpTCP5 resulted in an increased number of sporangia attached to a single seta, reminding us the branching phenotype of tcp mutants in higher land plants (Ortiz-Ramírez et al., 2016). Overexpression of OsTB1 strongly reduces tiller number in rice, with no effect on axillary bud number but impairing axillary bud growth (Takeda et al., 2003). Similar phenomena have been observed in pea (Braun et al., 2012), poplar (Muhr et al., 2016), Arabidopsis (Aguilar-Martínez et al., 2007; Poza-Carrión et al., 2007), and potato (Nicolas et al., 2015), demonstrating that TCP TFs influence tillering by affecting the subsequent growth of axillary meristems rather than the initial formation of axillary meristems. Therefore, TCP TFs exhibit a conserved function in controlling axillary meristem activity. Two Class Ⅰ TCP TFs in Arabidopsis, TCP14 and TCP15, were found to be functionally redundant in modulating cell proliferation during leaf development and in other tissues. The most striking effect was their impact on internode length: the tcp14 tcp15 double mutant showed shortened internodes, ultimately resulting in dwarf phenotype (Kieffer et al., 2011). Furthermore, in Arabidopsis, ectopic expression of AtTCP20 fused with the EAR repressor domain (i.e., AtTCP20 overexpression plants) results in severe dwarfism and late flowering, suggesting that Class Ⅰ type TCP AtTCP20 plays a role in cell division, cell expansion, and cell differentiation (Hervé et al., 2009). The above researches show that TCP TFs can play regulatory roles in various ways, such as branch regulation, flower symmetry regulation, leaf morphology regulation, and flowering period regulation, etc.

By 2009, 26 TCP genes had been discovered in rice (Oryza sativa), which belong to different types (Class Ⅰ or Class Ⅱ) (Martín-Trillo and Cubas, 2010). Several of them have been reported to regulate tillering. It is known that both OsTb1 and OsTb2 belong to the TCP gene family (Lyu et al., 2020). OsTb2 is a novel gene produced by duplication of OsTb1, OsTb2 was artificially selected during the adaptation of upland rice to terrestrial environments. Transgenic rice overexpressing OsTb2 exhibits increased tillering, whereas OsTb1 suppresses tillering, which means the two genes have exactly opposite effects on rice tillering. Furthermore, Lyu, J. et al. found that the OsTb2 protein can interact with the OsTb1 protein, thereby counteracting OsTb1's inhibitory effect on tillering (Lyu et al., 2020). OsSPL14 TF can inhibit tillering by directly activating the expression of OsTb1 (Jiao et al., 2010; Miura et al., 2010; Lu et al., 2013). Additionally, both OsPIL11 and OsPIL12 TFs can inhibit rice tillering, and OsPIL11 inhibits tillering by activating the OsTb1 gene (Zhang et al., 2022a). Through analysis of 110 germplasm materials around the world, Liu et al. discovered that as nitrogen concentration increases, expression level of OsTCP19 gradually decreases (Liu et al., 2021). It is known that DLT promotes tillering through regulating strigolactone signaling (Tong et al., 2009, 2012). OsTCP19 negatively regulates tillering by directly inhibiting DLT's expression. When nitrogen increases, LBD responds to it by inhibiting the transcription of OsTCP19, thereby positively regulating tillering (Liu et al., 2021).

Mukhopadhyay and Tyagi (2015) discovered that abiotic stresses universally upregulate the expression of OsTCP19. Within hours of stress onset, rice genotypes with better tolerance possess more abundant OsTCP19 transcripts compared to sensitive genotypes. Moreover, Zhang et al. discovered overexpression of OsTCP4 leads to an increase in grain length, while inhibition of OsTCP4 expression results in an increase in grain width (Zhang et al., 2024). It can be seen that as of now, there have been no reports on the regulation of rice plant height by TCP TFs, which is also a gap that this study aims to fill.

In summary, TCP TFs are key regulatory factors mediating plant growth. However, to date, there is still a gap in understanding how TCP TFs regulate rice plant height. In this study, we found that OsTCP4 acts as a transcriptional repressor of the Green Revolution gene SD1 and ultimately regulates rice plant height by modulating internode length. Therefore, our study provides a new mechanism for transcriptional regulation of the Green Revolution gene SD1 in rice and reports the molecular mechanism by which TCP TFs regulate rice plant height for the first time.

2. Materials and methods 2.1. Yeast one hybrid (Y1H) screen

Y1H library screening was performed as previously described (Zhang et al., 2022b). Briefly, a rice cDNA library containing over 1500 TFs of rice was constructed by OE-biotech® (Shanghai, China), and Y1H library screening was conducted according to the manufacturer's instructions. The PA-1 fragment of the SD1 promoter were cloned and inserted into the pHisi-1 vector to generate pHisi-1-PA-1, which were used for the Y1H library screening (Figs. 1A and S1A). The pHisi-1-PA-1 + pGADT7 yeasts was spread on SD/-Trp/-Leu/-His solid medium with different concentrations of 3-AT to determine the minimum concentration of 3-AT that can inhibit the self-activation of pHisi-1-PA-1. Finally, the cDNA library was transformed with bait reporter (pHisi-1-PA-1) yeasts and then spread on SD/-Trp/-Leu/-His solid medium with the minimum concentrations of 3-AT that can inhibit the self-activation. After incubation at 30 ℃ for 4–6 d, colonies that grew normally on SD/-Trp/-Leu/-His + 3-AT plates were identified as primary positive clones. The primary positive clones were then transferred to selection plates under the same conditions for secondary library screening. Clones that continued to grow normally were considered as true interactors.

2.2. Transgenic plant construction

The overexpression construct was acquired by amplifying full-length coding sequence (CDS) of OsTCP4 (LOC_Os02g42380.1) that starts from 'ATG' and doesn't contain termination codon from the cDNA of Nipponbare, followed by cloning the CDS of OsTCP4 into pCAMBIA-1301-3MYC, which contains a 3*Myc-tag at the C-terminal end. The mutation construct was obtained using the CRISPR/Cas9 system. The CRISPR-GE (http://skl.scau.edu.cn/home/) was used to seek the ideal target sequence that possesses specific binding ability and lower maximum off-target rate (Xie et al., 2017). One single guide RNA (sgRNA) with one target sequence and U6a promoter was amplified from U6a vector and then inserted into the pMH-SA vector (Liang et al., 2016). PCR was carried out in a 50 μL reaction volume using Phanta Max Super-Fidelity DNA Polymerase (P505; Vazyme®, Nanjing, China). Amplification procedures consisted of 35 cycles with 95 ℃ for 15 s, 60 ℃ for 15 s, and 72 ℃ for X s (1 kb/15 s). The overexpression construct and mutation construct were transformed into rice plants by Edgene® Biotech (Wuhan, China). The primers used in vector construction are listed in Table S2.

2.3. Electrophoretic mobility shift assay (EMSA)

Full-length CDS of OsTCP4 was amplified and cloned into the pGEX-4T-1 vector (GE® Healthcare, Chicago, USA) to generate the glutathione S-transferase (GST)-OsTCP4 fusion protein. The recombinant plasmid was transformed into Escherichia coli (BL21 strain). Then GST tag protein and GST-OsTCP4 fusion protein were purified as described by Wang et al. (2016). The EMSA was carried out using a Chemiluminescent EMSA kit according to the manufacturer's instructions (GS009; Beyotime®, Shanghai, China). The biotin-labeled probes and unlabeled probes were synthesized by Sangon® Biotech (Shanghai, China). The biotin-labeled probes were subjected to incubation with GST-OsTCP4 fusion protein at 20–25 ℃ for 20 min. Competition for the biotin-labeled probes was assessed by introducing unlabeled 100× concentrations of WT probes or mutant probes to the 1 x biotin-labeled probes. The biotin-labeled probes were also subjected to incubation with GST tag protein as a negative control. The probes utilized for EMSA and the primers employed for vector construction can be found in Table S2.

2.4. Plant materials and growth conditions

As the TCP4 genomic DNA sequence of Zhonghua 11 (ZH11) from Rice Resource Center (Rice RC) database (Qin et al., 2021) is identical with that of Nipponbare (Nip) from Rice Genome Annotation Project (RGAP) database (Kawahara et al., 2013; Hamilton et al., 2025), which includes 3500 bp promoter, 636 bp CDS, and 1000 bp 3'UTR sequence (Figs. S6 and S7), two rice model varieties Nip and ZH11 were employed simultaneously as the wild type (WT) to confirm the phenotypes of ostcp4 mutants. Ensuring that the promoter sequence and genomic sequence of OsTCP4 in Nip and ZH11 are identical is to eliminate the variable of sequence differences (i.e., control variable method).

Using the CRISPR-Cas9 gene editing technology, we successfully generated different ostcp4 (ZH11), ostcp4 (Nip), ossd1 (Nip), and ossd1 ostcp4 (Nip) mutant T₀ lines. Homozygous T₁ or T₂ transgenic plants were utilized for the analysis of molecular and phenotypic characterization. Using the vector pCAMBIA-1301-3MYC, we successfully got different OsTCP4-OE (ZH11) T₀ lines, from which the T₂ overexpressing lines were planted in 2024 and the T₃ overexpressing lines were planted in 2025. Plant specimens were grown and maintained in Chenggong paddy field (102°47′34.9″E, 24°49′56.9″N, Altitude, 1866 m; Low-latitude plateau monsoon climate) or Chengjiang paddy field (102°55′11.3″E, 24°38′11.9″N, Altitude 1700 m, Subtropical monsoon climate) or Yuanjiang paddy field (101°58′ 23.8″E, 23°35′18.9″N, Altitude 372 m, Tropical monsoon climate) or Yuanmou paddy field (101°50′06.2″E, 25°46′13.0″N, Altitude 1023 m, South asian tropical dry-hot river valley climate) to investigate their phenotypes (Fig. S11). Paddy field management was performed according to normal rice production.

2.5. β-glucuronidase (GUS) staining

The putative 3205-bp promoter of OsTCP4 was amplified from genomic DNA using specific primers (Supplementary Table S2). The proOsTCP4 sequence were cloned into the pCAMBIA-1300-Gus binary vector. Transgenic plants were subjected to GUS staining as described in GUS staining Kit's protocol (SL7160; Coolaber®, Beijing, China). The stained tissue sections were observed and photographed utilizing a Interchangeable Lens Digital Camera (R10; Canon®, Tokyo, Japan).

2.6. Subcellular localization

Full-length CDS of OsTCP4 (the stop codon of which has been removed) was amplified and cloned into pCAMBIA30-YFP vector, which is driven by the powerful CaMV35S promoter. Moreover, a blank pCAMBIA30-YFP vector was used as a negative control. Two methods for subcellular localization were simultaneously employed: one involved transforming rice protoplasts with the high-concentration plasmid, the other entailed infiltrating Nicotiana benthamiana (tobacco) leaves with Agrobacterium solution harboring the target plasmid. For the first method, the experiment was conducted as described by Zhang et al. (2011). For the second method, the vectors were transformed into Agrobacterium tumefaciens strain GV3101 and infiltrated into tobacco leaves as described by Wang et al. (2015). After 48–72 h, the infected leaves were sectioned for observation. YFP and DAPI fluorescence were observed under a Laser Scanning Confocal Microscope (LSM 980; Zeiss®, Oberkochen, Germany). The signal patterns were detected at excitation wavelengths of 515 nm (YFP: yellow) and 358 nm (DAPI: blue) through confocal imaging.

2.7. RNA extraction and reverse transcription quantitative PCR (RT-qPCR) analysis

Total RNA of all the tissues was isolated using a TRIzon Reagent Kit (CW0580S; CWBIO®, Taizhou, China) according to the manufacturer's protocol. The cDNAs were synthesized from 1000-ng total RNA using HiScript Ⅲ RT SuperMix for qPCR (+ gDNA wiper) (R323-01; Vazyme®, Nanjing, China) according to the manufacturer's protocol. 50-ng resultant cDNA was quantified with a qPCR Instrument (LC480; Roche®, Basel, Switzerland) with the FastReal qPCR PreMix (SYBR Green) (FP217; TIANGEN®, Beijing, China) according to the manufacturer's instructions. The transcripts of OsActin1 (LOC_Os03g50885) gene was used as an internal reference. RT-qPCR analysis was performed with at least 3 technical replicates per template. Each template was analyzed with at least 3 independent biological replicates, and the average value represented the relative expression level. The primers used for RT-qPCR analysis are listed in Supplementary Table S2.

2.8. Measurements of agronomic traits

Plant height was measured and recorded at the tillering stage, heading stage, and maturity stage of rice. Internode length, effective panicle number per plant, filled grain number per panicle, 1000-grain weight, and yield per plant were measured and recorded at the maturity stage of rice. Plant height and internode length were measured by Mayi® standard steel tape (Henan, China) manually. Effective panicle number per plant and filled grain number per panicle were counted manually. Yield per plant was weighed with a Puchun® standard balance (JE203; Shanghai, China). All rice seeds for 1000-grain weight analysis were air-dried before testing by seed trait analyzer (TPKZ-3; Tuopuyunnong®, Hangzhou, China), and the values were averaged and used as the measurement result for the plant.

2.9. Microscopic observations and measurements

The internodes' sections were collected and fixed in FAA solution (5% formaldehyde, 5% acetic acid, 50% alcohol, and 40% distilled water) at 4 ℃ for 2 d. After dehydration with a graded series of alcohol, the samples were embedded in paraffin wax. Approximately 50 μm to 100 μm-thick longitudinal sections were cut using a vibratome (VT1200S; Leica®, Nussloch, Germany). Images of sections were taken using a microscope (DM4B; Leica®, Nussloch, Germany), and individual cell length was measured using the ImageJ software.

2.10. Dual luciferase reporter gene assay

A reporter construct was generated by cloning 3000 bp of SD1 promoter DNA into the pGreenII-0800 vector. Full-length CDS of OsTCP4 was amplified and inserted into pGreenII 62-SK vector and used as the effector construct. Assays for transient expression in rice protoplasts were then incubated in dark for 3–6 h. Luciferase activity was quantified using a Dual-Luciferase reporter assay kit (E1910; Promega®, Wisconsin, USA). Three independent technical repeats for each sample were performed, and the relative luciferase activity was calculated as the ratio (LUC/REN) of firefly luciferase (LUC) to renilla luciferase (REN). Moreover, pGreenII 62-SK-YFP vector was used as a negative control. The sequence of SD1pro is shown in Dataset S2. The primers used for this assay are listed in Supplementary Table S2.

2.11. Measurements of endogenous GAs

Levels of GA₁, GA₃, GA₄, and GA₇ were determined by HPLC–MS/MS. The samples were collected from the rice plants growing in paddy field and then immediately put into liquid nitrogen. Approximately 0.5 g of internode tissue that includes the first node was ground in a pre-cooled mortar containing 5 mL of extraction buffer composed of isopropanol and hydrochloric acid. The extract was shaken at 4 ℃ for 30 min. Then, 10 mL of dichloromethane was added, and the sample was shaken again at 4 ℃ for 30 min. The sample was then centrifuged at 18,000×g for 5 min at the same temperature, and the lower, organic phase was extracted. The organic phase was dried under N₂, dissolved in 150 μL methanol (0.1% methane acid) and filtered through a 0.11-μm filter membrane. The purified product was then subjected to HPLC–MS/MS analysis. The HPLC analysis was performed using a ZORBAX SB-C18 (Agilent Technologies®, United States) column (2.1 mm × 150 mm; 3.5 mm). The mobile phase A solvent consisted of methanol and 0.1% methanoic acid, and the mobile phase B solvent consisted of ultrapure water and 0.1% methanoic acid. The injection volume was 2 μL. MS conditions were as follows: a spray voltage of 4500 V, air curtain, nebulizer, and GA gas pressures of 15, 65, and 70 psi, respectively. The atomizing temperature was 400 ℃.

The external standard refers to the types of hormones being measured, and the internal standard is D - GA₄ (i.e., deuterated gibberellin 4). Simply speaking, the calibration process is as follows: First, use the ratio of the peak area of the external standard to that of the internal standard as the ordinate and the concentration of the external standard as the abscissa to obtain the standard curve. Then, divide the external standard of sample (i.e., the endogenous background) by the internal standard of sample (i.e., the exogenous internal standard), and substitute this ratio into the standard curve to obtain the actual GA concentration of the sample for analysis. Specifically, first, prepare mixed standard working solutions with concentrations of 0.1 ng/mL, 0.2 ng/mL, 0.5 ng/mL, 1 ng/mL, 2 ng/mL, 5 ng/mL, 10 ng/mL, 20 ng/mL, 50 ng/mL, 100 ng/mL, and 200 ng/mL respectively from the external standard stock solution using ultrapure water (where the concentration of the internal standard solution is 10 ng/mL). Then, use the area ratio of the external standard peak to the internal standard peak in the measured characteristic ion mass chromatogram as the ordinate and the corresponding concentration of the external standard solution as the abscissa. In this way, the standard curve can be drawn, and the regression equation and correlation coefficient can be obtained. Each sample consisted of three biological replicates. GA measurements were accomplished by Webiolotech® Testing Technology Co., Ltd.

2.12. Yeast two hybrid (Y2H) experiment

For protein interaction tests, full-length CDS of OsTCP4 was cloned into pGBKT7, and full-length CDS of OsTCP4 was cloned into pGADT7. Vectors were co-transformed into yeast strain Y2HGold (Clontech®, California, USA). Growth was determined as described in the Y2H System User Manual (Clontech®, California, USA). Experiments were repeated at least three times. Primers used for the vector construction are listed in Supplementary Table S2.

2.13. Accession numbers

The ZH11 GeneID of OsTCP4 in Rice RC database (Qin et al., 2021) is OsZH11G0203185500.01, the Nip GeneID of OsTCP4 in RGAP database (Kawahara et al., 2013; Hamilton et al., 2025) is LOC_Os02g42380, the GeneID of OsTCP4 in National Center for Biotechnology Information (NCBI) database (O'Leary et al., 2016) is LOC4330080 that annotated it as 'transcription factor TCP7', and the GeneID of OsTCP4 in Rice Annotation Project Database (RAP-DB) (Sakai et al., 2013) is Os02g0635800.

The Nip GeneID of SD1 in RGAP database (Kawahara et al., 2013; Hamilton et al., 2025) is LOC_Os01g66100, the GeneID of SD1 in NCBI database (O'Leary et al., 2016) is LOC4325003, and the GeneID of SD1 in RAP-DB (Sakai et al., 2013) is Os01g0883800.

3. Results 3.1. Y1H screening and phylogenetic analysis of OsTCP4

Given the key role of the SD1 gene, we carried out Y1H screening strategy to identify the upstream regulators of SD1 gene. The 200 bp-DNA fragment (i.e., 'PA-1') in the upstream of the SD1's 'ATG' was used as the target promoter sequence to screen key transcriptional regulators (Figs. 1A and S1A). Twenty TFs were identified in Y1H screening, including ZOS-C₂H₂ zinc finger family and TCP family, which took up 85% quantity of the total candidate TFs (Supplementary Table S1). As a Cys₂/His₂ zinc finger protein has been reported to act as a regulator of rice plant height (Duan et al., 2021), while there have been no reports on the regulation of rice plant height by any TCP TF. Therefore, we decided to focus on the TCP TFs we got in Y1H screening (i.e., OsTCP1, OsPCF1, OsTCP4, OsTCP10, and OsTCP19). As intraspecific homology-based phylogenetic analysis indicates that OsPCF1 protein, OsTCP4 protein, and OsTCP10 protein show the highest homology to each other within Oryza sativa (Fig. S2A), we decided to focus on these three TCP TFs. In the first year we plant tcp mutants in paddy field (i.e., 2023 Chenggong), we observed the most obvious plant height difference between ostcp4 and WT (Fig. S3). Therefore, we finally decided to conduct an in-depth study on OsTCP4 TF to reveal its regulatory role on rice plant height.

To verify whether OsTCP4 protein can directly bind to the SD1 promoter in vitro, EMSA was performed (Fig. 1B). Two fragments containing 'CCCAC' or 'GTGGG' within PA-1 were selected in EMSA according to previous study that shows the consensus binding site for class Ⅰ TCP is 'GGNCCCAC' (Kosugi and Ohashi, 2002). The binding of GST-OsTCP4 fusion protein to the biotin-labeled WT fragments was inhibited by the competitor probe with the correct WT sequence, but not inhibited by the competitor probe with the 'm5-1' or 'm6' mutated sequence, which indicated OsTCP4 protein can directly bind to the '5-1' and '6' sites of the SD1 promoter in vitro (Fig. 1B). The conclusion is further supported by the online PlantPan 4.0 prediction result, which shows TCP4's binding site on the SD1 promoter includes '5-1' site (Fig. S4). Therefore, we focused on OsTCP4 for further research.

Corresponding amino acid sequence of OsTCP4 protein was deduced from the gene sequence provided by Y1H screen (Fig. S1A and Supplementary Table S1). Interspecific homology-based phylogenetic analysis indicates that OsTCP4 is a bHLH protein showing homology to many other TCP proteins in Poaceae family, including Oryza (genus), Aegilops, Triticum, Panicum, Phragmites, Setaria, Sorghum, Brachypodium, Hordeum, Zea, Lolium, and Miscanthus, in which the amino-acid sequences of OgTCP4-like and ObTCP4-like from Oryza genus exhibit the highest homology to the amino-acid sequence of OsTCP4 (highlighted with red triangle) (Fig. 1C). Intraspecific homology-based phylogenetic analysis indicates that OsPCF1 protein and OsTCP10 protein show the highest homology to the OsTCP4 protein (highlighted with red triangle) within Oryza sativa (Fig. S2A), which is further supported by the conserved motif distribution analysis of TCP proteins (Fig. S2B).

3.2. Tissue expression pattern and subcellular localization of OsTCP4

To investigate the spatial expression patterns of OsTCP4 in rice, total RNA was extracted from the tissues including root, mature leaf, young leaf, node, inter node, tiller base, young panicle, and mature panicle from Nip WT plant. RT-qPCR analysis revealed that OsTCP4 was expressed in various rice tissues, with particularly higher expression levels detected in the node and tiller base than in other tissues (Fig. 2A). To more precisely analyze the tissue-specific expression of OsTCP4 in rice, a vector for the expression of the GUS gene driven by the OsTCP4 promoter was generated. Basically consistent with the result of RT-qPCR analysis, the proOsTCP4-GUS (Nip) transgenic plants showed stronger GUS staining in the node and tiller base (Fig. 2B), implying a potential role for OsTCP4 in the modulation of internode elongation.

To more comprehensively characterize the function of OsTCP4 protein, we analyzed its subcellular localization, which was determined by transient expression in the rice protoplasts and the tobacco epidermal cells. The results showed that the yellow fluorescent signal emitted by OsTCP4-YFP exhibited a merging pattern with the red signal emitted by the nuclear-localized HY5-RFP, while the YFP control was distributed almost throughout the whole rice protoplast (Fig. 2C). HY5-RFP was used as a nuclear-localized marker here. It is known that ELONGATED HYPOCOTYL 5 (HY5) is a famous TF of Arabidopsis thaliana, whose mutant (i.e., hy5 mutant) is deficient in gravitropic response of roots and light-dependent hypocotyl elongation and whose subcellular localization coincides with that of DAPI (Oyama et al., 1997).

The nuclear-localized subcellular localization of OsTCP4 in the rice protoplast was further confirmed by the subcellular localization of OsTCP4 in the tobacco epidermal cell (Fig. S5).

Taken together, OsTCP4 is a nuclear-localized protein that preferentially expressed in the node and tiller base, indicating that OsTCP4 may be involved in the regulation of stem elongation during rice growth and development.

3.3. ostcp4 mutant increased rice plant height, while overexpression of OsTCP4 inhibited rice plant height

It is known that the OsTCP4 genomic DNA sequence of ZH11 in Rice RC database (Qin et al., 2021) is totally identical with that of Nip in RGAP database (Kawahara et al., 2013; Hamilton et al., 2025), which includes 3500 bp promoter, 636 bp exon, and 1000 bp 3′UTR sequence (Figs. S6 and S7).

To study the biological functions of OsTCP4 in rice, the loss-of-function mutants of OsTCP4 was produced by CRISPR-Cas9 gene editing technology in two backgrounds, ZH11 and Nip (Figs. 3A and S8). Specifically, the knockout target of ostcp4 was selected between 'ATG' and genomic DNA coding for TCP domain, which can achieve the frameshift mutation of entire TCP domain that is responsible for DNA binding and protein interaction (Figs. 3A, S7, and S8). The pMH-SA vector modified by Liang et al. (2016) containing the sgRNA sequences under the control of the OsU6a promoter was successfully constructed and introduced into Nip and ZH11 rice varieties. Subsequently, 5 independent T₀-generation homozygous mutant lines of ZH11 background were chosen from more than 20 T₀-generation seedlings for subsequent analysis, namely ostcp4-1 (ZH11), ostcp4-2 (ZH11), ostcp4-6 (ZH11), ostcp4-24 (ZH11), and ostcp4-29 (ZH11) (Figs. 3A and S8). Moreover, 3 independent T₀-generation homozygous mutant lines of Nip background were selected from more than 20 T₀-generation seedlings to study the biological functions of OsTCP4, which were ostcp4-2 (Nip), ostcp4-3 (Nip), and ostcp4-5 (Nip) (Fig. S8). All these 8 mutant lines have frame-shift mutation at the OsTCP4 gene loci (Fig. S9). Furthermore, transgenic plants overexpressing OsTCP4 under the control of the maize ubiquitin promoter were generated in ZH11 background. Of the OsTCP4-overexpressing lines, 2 independent T₀-generation lines OsTCP4-OE-7 and OsTCP4-OE-27 that constitutively express enhanced levels of OsTCP4 transcripts were selected from more than 20 T₀-generation seedlings for further study (Fig. 3B). T-DNA insertion assay further confirmed the existence of OsTCP4 sequence fused with 3*Myc tag sequence in OsTCP4-OE-7 and OsTCP4-OE-27 overexpression lines (Fig. S10).

For three consecutive years (2023, 2024, and 2025), the morphological characteristics of the ostcp4 (homozygous and Cas9-free) and WT at the tillering stage and the maturity stage were compared in four paddy fields (Chenggong, Chengjiang, Yuanjiang, and Yuanmou) of Yunnan province, China (Figs. 3C–F, S3, S8, S11, and S12). Plant height of the ostcp4 lines were markedly higher than that of the WT control under both ZH11 and Nip background at the tillering stage (Fig. 3C and E) and the maturity stage (Fig. 3D and F, S3). For two consecutive years (2024 and 2025), the morphological characteristics of the OsTCP4-OE and WT at the tillering stage and the maturity stage were compared in three paddy fields (Chenggong, Chengjiang, and Yuanmou) of Yunnan province, China, in which the plant height of the OsTCP4-OE-7 and OsTCP4-OE-27 plants were significantly shorter than that of the WT plant at the two developmental stages (Fig. 3C–F, S3, and S11).

Therefore, OsTCP4 negatively regulates the plant height at the tillering stage and the maturity stage.

3.4. OsTCP4 regulates plant height through modulating cell length of internode

Rice plant height in this paper refers to the height from the base of stem to the top of mature panicle, which approximately equals to the sum of the panicle length and the length of all internodes (Fig. S13).

As rice height is mainly determined by the length and number of internodes, we found that the number of internodes was the same among ZH11 WT plants, ostcp4 mutant plants (ZH11) and OsTCP4-overexpressing plants (ZH11), which are all four internodes (Fig. 4A and E). However, the 1st, the 2nd, the 3rd, and the 4th internodes of ostcp4 (ZH11) plants were all significantly longer than those of ZH11, while the 1st and the 2nd internodes of OsTCP4-OE (ZH11) plants were significantly shorter in length than those of ZH11 (Fig. 4A, B, E, and F), indicating a repression role for OsTCP4 in the modulation of internode elongation.

The longitudinal image of the internode cells was examined by fluorescence microscope (DM4B; Leica®, Wetzlar, Germany) and the cell length was measured by ImageJ software. We observed the cell length of the 1st intenode was increased in ostcp4 (ZH11) plants significantly while reduced in OsTCP4-OE (ZH11) plants with strong significance (Fig. 4C, D, G, and H), indicating a repressing role for OsTCP4 in the regulation of cell elongation in the 1st internode. Moreover, the measurement significance of internode length and the measurement significance of cell length basically coincide with each other (Fig. 4B, D, F, and H), indicating a positive correlation between the internode length and the cell length of internode.

3.5. OsTCP4-SD1 module regulates rice plant height

It is known that SD1 regulates rice plant height, and mutations in this locus cause varying degrees of dwarfism in rice (Sasaki et al., 2002). As the subject of this study, OsTCP4 was derived from a yeast one-hybrid screen targeting the 200 bp fragment of SD1pro (Figs. 1A and S1), which was further verified by EMSA (Fig. 1B). Therefore, we speculated that OsTCP4 protein regulates the expression of SD1, thereby modifying the plant height of rice.

To examine the transcriptional activity of OsTCP4 protein at SD1 promoter, a dual-luciferase reporter assay was conducted in rice protoplasts. Protoplasts were co-transformed with effector plasmid that containing YFP or OsTCP4 in the downstream of CaMV35S promoter, and reporter plasmid that containing REN in the downstream of CaMV35S promoter and LUC in the downstream of SD1 promoter (Fig. 5A). The result showed that compared with the YFP control, OsTCP4 significantly repressed the transcription activity of the SD1 promoter (Fig. 5A). Furthermore, we analyzed the expression of SD1 in TCP4-transgenic rice plants by RT-qPCR. As shown in Fig. 5B, the expression of SD1 was markedly reduced in OsTCP4-OE lines as compared to ZH11, but increased significantly in ostcp4 lines as compared to ZH11. Considered together, these results indicate that OsTCP4 protein directly suppresses the transcription of SD1.

It is known that the SD1 genomic DNA sequence of ZH11 (from Rice RC database) (Qin et al., 2021) is almost the same as that of Nip (from RGAP database) (Kawahara et al., 2013; Hamilton et al., 2025), which includes 3300 bp promoter, 1170 bp exon, and 1000 bp 3' UTR sequence, except for 2 bp divergence at the 2nd intron (see the magnified part of Fig. S14). To determine whether the higher plant height phenotype of ostcp4 is caused by activated SD1 expression, we constructed ossd1 single mutant, ostcp4 single mutant, and ossd1ostcp4 double mutant in Nip background (Fig. S15), whose plant height were examined at the heading stage and the maturity stage. Compared with the WT, the plant height of ostcp4 mutants was significantly higher, whereas the plant height of ossd1 mutant was markedly shorter. The ossd1 ostcp4 double mutants demonstrated dramatically shorter plant height compared with the WT, which exhibited the same trend as the plant height comparison between ossd1 and the WT (Fig. 5C–E). Overall, these results suggested that SD1 acts downstream of OsTCP4 in the gene regulatory network and the OsTCP4-SD1 module mediates the regulation of plant height in rice.

3.6. Distinction of plant height in ostcp4 lines and OsTCP4-OE lines is caused by different endogenous concentrations of GA₁

Elongation or shortening in internode length commonly results from impairments in the biosynthetic pathways or signal transduction of phytohormones gibberellin (GA) or brassinosteroid (BR) (Sakamoto et al., 2004; Tong et al., 2014; Yamaguchi, 2008). SD1 is known to encode OsGA20ox2, an enzyme that governs the production of biologically active GAs and thus serves as a critical regulator of GA biosynthesis (Monna et al., 2002; Spielmeyer et al., 2002; Sasaki et al., 2002). As shown in Fig. 4D and H, only the cell length of 1st-internode in ostcp4-2 (ZH11) and OsTCP4-OE-7 (ZH11) show significant difference simultaneously compared with WT. Therefore, we speculated that one or more biologically active GAs may exhibit differential accumulation in the 1st internode between ostcp4 lines and OsTCP4-OE lines. In order to test which biologically active GA(s) correlate(s) with the disparity of 1st internode length between ostcp4 lines and OsTCP4-OE lines, the internode that includes the first node was sampled to measure the concentrations of four kinds of biologically active GAs (i.e., GA₁, GA₃, GA₄, and GA₇). Compared with WT, we found that among GA₁, GA₃, GA₄, and GA₇, only GA₁ exhibits significantly higher concentrations in two ostcp4 lines and significantly lower concentrations in two OsTCP4-OE lines, which is the same variation pattern as the main panicle length between WT and transgenic lines used for GAs measurements (Figs. 6A and S16). These results imply that the distinction of plant height in ostcp4 lines and OsTCP4-OE lines is caused by different endogenous concentrations of GA₁.

3.7. Genetic model of TCP4-driven rice plant height regulation

According to previous literatures, TCP TFs tend to form dimers because TCP family TFs require dimerization to bind DNA, specifically, they dimerize via amino acids outside the TCP domain and then bind DNA via the TCP domain (Kosugi and Ohashi, 1997; Kosugi and Ohashi, 2002; Trémousaygue et al., 2003; Andriankaja et al., 2014). Using Y2H system, we observed that OsTCP4 can interact with itself (Fig. 6B), indicating that OsTCP4 might form homodimers.

Therefore, based on the existing experimental results, we proposed the following model. In WT plants, OsTCP4 protein forms a homodimer and bind to the 'GTGGG' and 'CCCAC' motifs of SD1 promoter to restrict its transcription. When OsTCP4 is overexpressed, more OsTCP4 homodimers form and their repression on SD1 transcription reduces plant height significantly. In ostcp4 mutant plants, function of OsTCP4 protein is disrupted, abolishing its transcription repression on SD1 promoter, so the expression of SD1 is elevated and rice plant height is promoted (Fig. 6C).

4. Discussion

Plant development is exquisitely regulated and coordinated via the interplay of several classes of plant hormones, among which GAs play a pivotal role in stimulating the growth of most organs through enhanced cell division and elongation (Colebrook et al., 2014). SD1, the pivotal green revolution gene, encodes OsGA20ox2, an enzyme that governs the production of biologically active GAs and thus serves as a critical regulator of GA biosynthesis (Monna et al., 2002; Sasaki et al., 2002; Spielmeyer et al., 2002).

Here, we describe a bHLH protein, OsTCP4, which is derived from Y1H screen (Fig. 1A) that has been verified by EMSA assay (Fig. 1B) and which shows homology to many other TCP proteins in other genera (Fig. 1C), as a new regulator of GA biosynthesis gene SD1 in rice. Expression analysis revealed that OsTCP4 is predominantly expressed in culm nodes and tiller base (Fig. 2A and B), sites of high GA accumulation and active cell elongation (Yamaguchi, 2008). This spatial expression pattern aligns with OsTCP4's role in fine-tuning GA biosynthesis to maintain GA abundance within a range that supports normal growth and development. SD1 is strongly expressed in the leaf blade, unopened flower, and stem (Sasaki et al., 2002). Coincidently, OsTCP4 gene is predominantly expressed in the node regions of stems (Fig. 2A and B), which adds evidence to the correlation between SD1 and OsTCP4 as both genes appear in the stem and might have the same or the opposite functions in stem development. The SD1 expression levels of ZH11, ostcp4-2 (ZH11), and OsTCP4-OE (ZH11) are proportional to their respective plant height (Fig. 3, Fig. 5B). For OsTCP4-OE plants, except for the reduced plant height (Fig. 3E and F), they also showed delayed heading time (Fig. S17) that belong to classic GA deficiency traits (Achard et al., 2007), while ostcp4 mutant lines displayed increased plant height (Fig. 3E and F) and advanced heading time (Fig. S18) that resembles classic GA sufficiency traits (Coles et al., 1999; Huang et al., 1998). EMSA (Fig. 1B) and dual luciferase reporter gene assay (Fig. 5A) substantiated this conjecture, OsTCP4 protein directly modulate SD1. This finding mirror a prior report where the Cys2/His2 zinc finger TF OsZFP207 was shown to negatively regulate plant height by negatively modulating SD1 (Duan et al., 2021). Likewise, a NAC TF, OsNAC2, has been shown to inhibit the GA pathway by directly promoting the expression of OsEATB and repressing the expression of OsKO2 (Chen et al., 2015).

Most of the sd1 mutant phenotypes were also observed in mutants with defects in GA biosynthesis or signaling (Li et al., 2011). Loss function of SD1 not only causes a semi-dwarf phenotype, but also results in decreases in yield per plant, filled grain number per panicle, and 1000-grain weight in rice (Wu et al., 2017). While the expression level of SD1 was significantly downregulated in OsTCP4-OE plants, their phenotypic characteristics are not totally the same as those of sd1 mutants. In contrast to sd1 mutants, OsTCP4-OE plants displayed not only dwarfism but also increased yield per plant, effective panicle number per plant, and filled grain number per panicle (Fig. S19). Moreover, according to previous report, SD1 can simultaneously affect both GA₁ pathway and GA₄ pathway (Su et al., 2021). But we only observed the accumulation difference of GA₁, instead of GA₄, aligned with the plant height divergence of ostcp4 lines and OsTCP4-OE lines (Figs. 6A and S16). SD1 (OsGA20ox2) encodes an oxidase enzyme that catalyses conversion of GA₅₃ and GA₁₂ into GA₂₀ and GA₉ by multi-step reactions; these GA intermediates are finally converted into functional GA₁ and GA₄, respectively (Itoh et al., 2001; Kuroha et al., 2018; Fig. S20). Moreover, the bioactivity of GA₄ is indeed much higher than that of GA₁ (Kuroha et al., 2018). However, GA₁ is predominant in rice vegetative tissues, while GA₄ levels peak in anthers during reproductive development of rice (Kuroha et al., 2018; Zhu et al., 2006). As the tissue we used to measure GAs' contents is the first node that include the connecting internode, which belongs to vegetative tissues. This may help to explain why it is GA₁ rather than GA₄ correlates with the plant height variations in this study (Figs. 6A and S16).

Crop plant height stands as a pivotal agronomic trait, directly influencing yield and cultivation suitability. Our study revealed that OsTCP4, a gene of interest, exhibited robust expression in the culm nodes of rice (Fig. 2A and B), a tissue critical for internode development. Further phenotypic analysis demonstrated that both OsTCP4-OE and ostcp4 lines displayed altered internode lengths, with the first and the second internodes showing significant differences simultaneously in both OsTCP4-OE lines and ostcp4 lines (Fig. 4B and F). These observations suggest that OsTCP4 protein may broadly regulate internode elongation across the rice stem, especially the first and the second internodes. Previous work by Todaka et al. (2012) identified OsPIL1, a phytochrome-interacting factor-like protein, as a key regulator of internode elongation due to its high expression in nodes. Importantly, OsPIL1 is not involved in GA biosynthesis, implying that the regulatory pathways governing internode elongation for OsTCP4 protein and OsPIL1 protein may diverge. To resolve this, additional genetic and biochemical experiments are needed to clarify whether these two TFs interact or operate independently. A more recent study on PREMATURE INTERNODE ELONGATION 1 (PINE1), a gene encoding a single zinc finger protein with two EAR motifs, provided further insights for our research. PINE1 reduces stem elongation sensitivity to GAs, and its mutation leads to enhanced stem elongation (Gómez-Ariza et al., 2019), a phenotype analogous to that of ostcp4 lines in this study (Fig. 4A and B). GA biosynthesis and signaling pathways have been well-documented as essential for internode elongation in rice (Ayano et al., 2014). Building on these findings, our morphological and cellular analyses of wild-type, OsTCP4-OE, and ostcp4 plants revealed that OsTCP4 protein modulates internode elongation by restricting cell longitudinal expansion (Fig. 4A–H), a mechanism shared with GhWIP2, a WIP-type zinc finger protein in Gerbera hybrida (Ren et al., 2018). GhWIP2 acts as a transcriptional repressor to suppress cell expansion during organ growth, mirroring the function of OsTCP4 in rice. Several other TFs were also reported to participate in the regulation of GA biosynthesis in rice, like OsZFP207 (Duan et al., 2021), OsbZIP58 (Wu et al., 2014), GDD1 (Li et al., 2011), OsEATB (Qi et al., 2011), and OsNAC2 (Chen et al., 2015).

From the breeding perspective, the pursuit of an ideal plant architecture requires balancing lodging resistance (a benefit of shorter stems) and sufficient biomass production (a requirement for high yields). In this study, OsTCP4-OE-7 increased yield per plant by approximately 59.58% in 2024 Yuanmou paddy field (Fig. S19), while ostcp4-1 decreased yield per plant by about 20.51% in 2024 Yuanmou paddy field (Fig. S19) and decreased yield per plant by around 41.05% in 2024 Chenggong paddy field (Fig. S21). Given that OsTCP4 was somewhat expressed in the panicles (Fig. 2A and B), OsTCP4 may play a role in regulating the function of reproductive organ through precise manipulation of its GAs levels. Exploring whether genetic engineering of OsTCP4 expression can enhance rice yield in field practice is a particularly intriguing direction for future research. As such, precise modulation of plant height and crop production remains a central goal in crop improvement, with genes like OsTCP4 offering promising targets for molecular breeding.

Acknowledgements

We would like to thank Jiaojiao Zhang (ORCID: 0009-0007-9675-3873) of Yunnan Agricultural University for tremendous assistance with the rice field work and rice genotype evaluation work. We would like to thank Mrs Dong of Yunnan Hanfei Biotechnology Co., Ltd. for assistance with microscopic observations. This work was supported by the National Natural Science Foundation of China, China (32372030, 32560439), the Applied Basic Research Foundation of Yunnan Province, China (202401BC070005, 202501AT070184).

Credit authorship contribution statement

Chen Ruan: Experiment execution, Data collection and analysis, Validation, Writing − original draft − review & editing. Jun Yang: Screening. Sen Yang, Wulve Huang, Yang Zhou: Investigation. Diqiu Yu, Pengcheng Wang: Conceptualization, Funding acquisition, Investigation, Writing − review & editing, Supervision, Methodology, Validation.

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 papers.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2026.02.008.