PhyB is a key component in light signalling for SAT

Light signalling is closely linked to plant growth and stress tolerance. To investigate the effects of light on SAT in rice, we compared the salt damage rates in rice roots under full light (with shoots and roots fully exposed to continuous light) and full dark conditions using a seedling growth assay in 0.5 × MS medium containing 100 mm NaCl or 80 mm NaHCO₃. The results indicated that rice grown under full light had lower salt damage rates than those grown in the full dark (Figure 1a,b), except for the saline-alkaline treatment, where the opposite trend was observed, with lower inhibition rates in the full dark (Figure 1b). As the key regulators of plant growth under various light conditions, phytochromes were examined for their role in regulating rice SAT. Following saline and alkaline treatments, the salt damage rate of phyB mutants was significantly lower than that of wild-type Dongjin (DJ) (Figure 1c–e), whereas no significant differences were observed for phyA and phyC mutants (Figure 1c–e). Previous studies have suggested that NH₄⁺ is a crucial factor for light-regulated root growth (Hirano et al., 2008). To further explore this, we treated rice with methylammonium (MeA), a toxic ammonium analog, under different light conditions (Figure 1f). Root growth was significantly reduced under full light, with or without MeA, compared to the full dark conditions, where MeA treatment significantly inhibited root growth (Figure 1f,g). However, MeA showed no inhibitory effect under full light conditions (Figure 1f,g). These findings suggest that different light conditions affect NH₄⁺ uptake in rice roots, with higher uptake occurring in the full dark, indicating that light signalling could play distinct roles in salt and saline-alkaline tolerance mechanisms in rice.

To further confirm the role of PhyB in regulating rice SAT, we evaluated the salt damage rates under stress conditions using DJ, phyB, AMT1 RNA interference (RNAi), and phyB/AMT1 RNAi. The results demonstrated that the SAT of phyB was completely inhibited by AMT1 RNAi (Figure 1h), as indicated by the similar salt damage rates in phyB/AMT1 RNAi and AMT1 RNAi, both of which were significantly higher than that of phyB and DJ (Figure 1i,j). These findings suggest that PhyB is a key regulator of rice SAT mediated by phytochrome-regulated light signalling.

PhyB interacts with CIPK31

Recent findings have revealed that PhyB inhibits NH₄⁺ uptake by inhibiting the transcription factors IDD10 and BZR1, which directly activate AMT1;2 in rice (Jung et al., 2023). Data have shown that phyB/idd10 mutants accumulate more NH₄⁺ than the idd10 mutants alone (Jung et al., 2023). Additionally, it has been previously demonstrated that brassinosteroid (BR) signalling promotes NH₄⁺ uptake via BZR1 in rice (Xuan et al., 2017), with BZR1 being a key transcription factor in BR signalling (Wang et al., 2002). To explore this further, phyB/idd10 and idd10 mutants were treated with the BR biosynthesis inhibitor propiconazole (PCZ) (Hartwig et al., 2012), and NH₄⁺ levels were measured. These results indicate that phyB/idd10 mutants accumulate more NH₄⁺ than idd10 mutants after PCZ treatment (Figure 2a), suggesting that PhyB may regulate NH₄⁺ uptake through additional pathways beyond the IDD10-BZR1 pathway.

To investigate the molecular mechanism of PhyB-mediated NH₄⁺ uptake, PhyB-interacting proteins were isolated using a yeast two-hybrid (Y2H) assay. Among the identified interactors, the C-terminus of CIPK31, an NH₄⁺-induced gene (Xuan et al., 2019), was identified to interact with PhyB-C (Figure 2b,c). The BiFC assay confirmed that CIPK31 interacted with PhyB in both the cytosol and the nucleus (Figure 2d). Additionally, a co-immunoprecipitation (co-IP) assay in tobacco leaves, where PhyB-GFP/CIPK31-FLAG or GFP/CIPK31-FLAG were co-expressed, revealed that PhyB-GFP, but not GFP alone, interacted with CIPK31-FLAG (Figure 2e), confirming the in vivo interaction. Upon photoactivation, PhyB converts from its Pr to Pfr form, where it interacts with transcription factors to inhibit their activity, either by degrading them or inhibiting their DNA binding (Dong et al., 2020; Leivar and Monte, 2014; Yasui et al., 2012). Co-expression of PhyB-GFP/CIPK31-FLAG or GFP/CIPK31-FLAG in tobacco leaves, followed by the western blot analysis, showed that CIPK31-FLAG levels remained similar with or without PhyB-GFP coexpression (Figure 2f), indicating that the interaction between PhyB and CIPK31 does not result in CIPK31 degradation. Additionally, CPK6 and CPK12 interact with and phosphorylate PhyB to regulate its nuclear translocation in Arabidopsis (Zhao et al., 2023). CIPK31-C, rather than its N-terminal kinase domain (CIPK31-N), interacted with PhyB (Figure 2b–d). To assess the impact of this interaction on PhyB localisation, we co-expressed CIPK31-FLAG/PhyB-GFP and FLAG/PhyB-GFP in tobacco leaves and analysed PhyB accumulation in the cytoplasm and nucleus. Western blot results demonstrated that co-expression of CIPK31-FLAG did not alter cytosolic or nuclear PhyB-GFP levels (Figure 2g). These findings suggest that although PhyB interacts with CIPK31, this interaction does not affect PhyB localisation or CIPK31 degradation.

CIPK31 promotes NH₄⁺ uptake by activation of AMT1 expression

CIPK31 expression is sensitive to NH₄⁺ supply (Xuan et al., 2019). Therefore, NH₄⁺ content was measured in Dongjin (DJ), cipk31 mutants, Zhonghua11 (ZH11), and CIPK31 OX plant roots. The results indicated that cipk31 mutants accumulated less NH₄⁺, whereas CIPK31 OXs accumulated more NH₄⁺ than their respective wild-types (Figure 2h). Subsequently, AMT1 expression levels were examined in CIPK31 plant roots. The RT-qPCR results revealed that the expressions of AMT1 genes (AMT1;1, AMT1;2, and AMT1;3) were significantly lower in cipk31 than in DJ plants, whereas their expressions in CIPK31 OXs were higher than that in ZH11 plants (Figure 2j–k). Previous studies have indicated that CIPK23 inhibits AMT1 by phosphorylating its C-terminal, reducing its NH₄⁺ transport function (Straub et al., 2017). However, CIPK31 did not interact with AMT1 in tobacco leaves (Figure S1), suggesting that CIPK31 promotes NH₄⁺ uptake via a mechanism different from that of CIPK23.

A previous study found that cipk31 mutants exhibited hypersensitivity to abiotic stress during the early seedling stage (Piao et al., 2010). Given the higher NH₄⁺ content in CIPK31 OX plants (Figure 2h) and the known link between NH₄⁺ uptake and SAT (Jung et al., 2023), the role of CIPK31 in rice SAT was examined using a seedling growth assay in 0.5 × MS medium containing 100 mm NaCl or 80 mm NaHCO₃. After the saline and alkaline treatments, the results revealed that the CIPK31 OX plants exhibited higher SAT (Figure 2l–n), consistent with the trend in AMT1;2 expression in CIPK31 plants (Figure 2j–k), while the cipk31 mutants were more sensitive to the salt-alkaline treatment (Figure 2o–q). Collectively, these findings indicated that CIPK31 promoted NH₄⁺ uptake and SAT by activating AMT1 expression.

PhyB disrupts the interaction between CIPK31 and TCP19 to repress AMT1;2

CIPK31 is a protein kinase, raising the question of how it regulates NH₄⁺-dependent AMT1 expression. Interestingly, TCP19, a key regulator of NUE in rice (Liu et al., 2021), was identified to interact with CIPK31 (Figure 3a). A BiFC assay confirmed that CIPK31 and TCP19 interacted in the nucleus (Figure 3b), and a co-IP assay further validated the interaction between TCP19-MYC and CIPK31-GFP but not GFP alone (Figure 3c), indicating that CIPK31 interacted with TCP19 in vivo. As shown in Figure 2b, CIPK31 contained a kinase domain at the N-terminus and a regulatory domain at the C-terminus. Further analysis revealed that the CIPK31-C terminal, rather than the N-terminus, of CIPK31 interacted with TCP19 (Figure 3d). To test whether phosphorylation affected the interaction between CIPK31 and TCP19, we co-expressed both proteins in tobacco leaves and used a pSer/Thr antibody to detect immunoprecipitated TCP19. However, the western blot analysis did not detect phosphorylated TCP19 (Figure 3e).

As TCP19 is a key transcription factor in NUE (Liu et al., 2021), its regulation of AMT1 gene expression was examined. Using the CRISPR/Cas9 system, two in-dependent homozygous TCP19 knockout lines, tcp19-1 and tcp19-2, with 1-bp insertions in the exon, were generated (Figure S2a). Additionally, TCP19-GFP overexpressors (TCP19 OXs) were produced. RT-qPCR results confirmed significantly higher TCP19 expression in TCP19 OXs than in wild-type ZH11 (Figure S2b). Transcriptomic analysis revealed that most AMTs were downregulated in TCP19 OXs (Liu et al., 2021) (Figure S2c). RT-qPCR further validated these findings, showing higher AMT1 expression in tcp19 mutants and lower expression in TCP19 OXs than in ZH11 (Figure 3f–h). To determine whether TCP19 directly regulates AMT1 gene expression, promoter sequences of AMT1 were analysed, and the JASPAR 2024 database (Rauluseviciute et al., 2024) predicted the cis-element GGNCCCAC as the TCP19-binding motif (Figure 3i). Promoter sequence analysis identified the putative TCP19-binding site GGGCCCAC in the AMT1;1 and AMT1;2 promoter regions (Figures S2d and 3k), but not in the AMT1;3 promoter. The yeast-one hybrid (Y1H) assay confirmed that TCP19 directly activated the AMT1;2 promoter but did not bind to the AMT1;1 or AMT1;3 promoter (Figures 3j and S2e, f). To examine the in vivo binding of TCP19 to the GGNCCCAC motif, a chromatin immunoprecipitation assay was conducted, which showed significant enrichment of AMT1;2 promoter fragments (P2) containing the motif, while fragments P1 and P3 did not show enrichment (Figure 3k). This interaction was further validated by electrophoretic mobility shift assay (EMSA). A specific shift band was generated when TCP19 protein was added, while this shift band was competed by unlabeled P2 probe (Figure 3l). To verify the regulation of AMT1;2 by TCP19, a transient transactivation assay was performed on tobacco leaves using a dual-luciferase reporter (DLR) system. The relative LUC/REN ratio was lower in leaves co-expressing TCP19-GFP with AMT1;2_pro:LUC, confirming that TCP19-GFP, but not GFP, directly represses AMT1;2 (Figure 3m–o). These results suggested that TCP19 directly repressed AMT1;2 expression via promoter binding. The C-terminal regulatory domain of CIPK31, rather than its kinase domain, interacts with TCP19 (Figure 3d), prompting an analysis of CIPK31's role in the TCP19-mediated repression of AMT1;2. Transient assay results demonstrated that CIPK31 did not activate or repress the AMT1;2 promoter on its own. However, the co-expression of CIPK31 inhibited the TCP19-mediated repression of AMT1;2 (Figure 3m–o), suggesting that CIPK31 may inhibit TCP19 repression to AMT1;2. promoter.

Since PhyB interacted with CIPK31 and CIPK31 interacted with TCP19, the potential interaction between PhyB and TCP19 was tested. The Y2H results demonstrated that PhyB did not interact with TCP19 (Figure S2), suggesting that PhyB may inhibit CIPK31-TCP19 interaction through competitive binding with CIPK31. To test this hypothesis, a yeast three-hybrid (Y3H) assay was conducted using constructs expressing CIPK31 (bait), TCP19 (prey), and PhyB-C (bridge protein) (Figure 3p). The Y3H results showed that PhyB-C coexpression inhibited the yeast growth, indicating that PhyB-C inhibited CIPK31-TCP19 interaction in yeast cells (Figure 3q). Additionally, a competition immunoprecipitation assay was used to further test the effect of PhyB on CIPK31-TCP19 interaction. We transiently infiltrated the tobacco leaves with equal proportions of TCP19-MYC, CIPK31-FLAG, and PhyB-GFP. As increasing amounts of PhyB-GFP were added, the interaction affinity between CIPK31-FLAG and TCP19-MYC decreased (Figure 3r). Immunoblotting analysis confirmed that PhyB-GFP inhibited the interaction between CIPK31-FLAG and TCP19-MYC (Figure 3r). Similar results were observed in the luciferase complementation imaging (LCI) assay (Figure 3s,t), where the interaction affinity between CIPK31-CLuc and TCP19-NLuc was weakened by coexpression of PhyB-GFP. These findings suggest that PhyB inhibits the CIPK31-TCP19 interaction, thereby affecting the formation of the CIPK31-TCP19 complex.

TCP19 negatively regulates rice SAT by repression of AMT1

Our previous study demonstrated that NH₄⁺ accumulation, rather than NH₄⁺ metabolites, enhanced rice SAT (Jung et al., 2023). In rice, inhibition of AMT1 expression reduced NH₄⁺ uptake in the roots, and because TCP19 repressed AMT1;2 expression, the role of TCP19 in SAT was investigated. The seedling growth assay revealed that the tcp19 mutants were tolerant, whereas the TCP19 OX plants were sensitive to salt and saline-alkaline stress (Figure 4a). The salt damage rates were significantly lower in tcp19 plants than in wild-type ZH11, whereas TCP19 OX plants demonstrated higher damage rates (Figure 4b,c). Analysis of the natural variation in TCP19 revealed two major haplotypes (Haps), distinguished by a few single-nucleotide polymorphisms (SNPs) and deletions in the promoter and exon regions (Figure S4a). RT-qPCR analysis indicated that TCP19 expression was higher in the cultivars belonging to Hap_H (marked by the red line) than in Hap_L (Figure S4b,c). Salt stress responses in these haplotype plants indicated that cultivars with lower TCP19 expression exhibited better tolerance to salt stress (Figure S4b,c). These findings indicate that TCP19 negatively regulates salt tolerance in rice. Because AMT1 RNAi plants accumulate less NH₄⁺ and are more sensitive to saline-alkaline stress (Jung et al., 2023), tcp19-1 was crossed with AMT1 RNAi to produce double transgenic plants (tcp19-1/AMT1 RNAi). After saline and alkaline treatments, the salt damage rate of tcp19-1/AMT1 RNAi plants was higher than that of tcp19-1 (Figure 4d–f) and exhibited a similar damage rate to AMT1 RNAi (Figure 4d–f). This confirmed that AMT1 RNAi could reduce the saline-alkaline tolerance of tcp19 plants.

To determine whether the differences in SAT were due to reduced ammonium transporter activity and NH₄⁺ uptake in roots, we conducted sensitivity tests for MeA in TCP19 plants. In the absence of MeA, the root lengths of ZH11 and TCP19 OXs were similar but significantly shorter than those of the tcp19 mutants (Figure 4d–f). However, in the presence of MeA, tcp19 mutant roots were more sensitive to MeA, presenting a greater reduction in root length than ZH11 and TCP19 OXs (Figure 4g,h). Additionally, the tcp19-1/AMT1 RNAi plants exhibited a similar SAT to that of AMT1 RNAi (Figure 4d–f), and their MeA sensitivity results were consistent, showing similar hyposensitivity as tcp19-1 (Figure 4i,j). These findings suggest that AMT1 RNAi attenuates NH₄⁺ uptake in tcp19 plants. Collectively, these results indicated that TCP19 negatively regulated rice SAT and NH₄⁺ uptake by repressing AMT1 expression.

PIL15 interacts with TCP19 and positively regulates rice SAT

Interestingly, PIL15, an Arabidopsis PIF homologue and a PhyB-interacting protein, also interacted with TCP19 (Figure 5a), as confirmed by BiFC and Co-IP assays (Figure 5b,c). To investigate whether PIL15 is involved in the regulation of AMT1;2 expression, PIL15 knockout lines and PIL15-GFP overexpressors (PIL15 OXs) were generated (Figure S5a,b). RT-qPCR results demonstrated that AMT1;2 expression levels were higher in PIL15 OXs and lower in pil15 mutants than in ZH11 (Figure S5c). As TCP19 directly repressed AMT1;2, we examined whether PIL15 could bind to AMT1;2. A previous study identified the PIL15 binding site (PBE-box) as CACATG (Figure 5e) (Li et al., 2022), and four PBE-boxes were observed in the AMT1;2 promoter region (Figure 5f). A subsequent Y1H assay confirmed that PIL15 interacted with the AMT1;2 promoter (Figure 5d). A ChIP assay further demonstrated that PIL15 binds to the AMT1;2 promoter in vivo, with significant enrichment of promoter fragments (P2) containing the CACATG motif, whereas P1 and P3 did not demonstrate enrichment (Figure 5f). Further experiment confirmed that PIL15 directly bound to the PBE-box element of AMT1;2 by EMSA (Figure 5g). As both PIL15 and TCP19 could bind to the AMT1;2 promoter, we analysed the role of the PIL15-TCP19 module in regulating AMT1;2 expression. A transient transactivation assay in tobacco leaves using the DLR system revealed that the co-expression of PIL15-GFP with AMT1;2_pro:LUC led to a higher LUC/REN ratio, confirming that PIL15-GFP directly activated AMT1;2 (Figure 5h–j). Interestingly, when PIL15 was co-expressed, it significantly inhibited TCP19-mediated repression of AMT1;2 and even activated its expression (Figure 5h–j).

Since PIL15 directly activated AMT1;2 expression, we investigated the response of PIL15 plants to SAT. PIL15 OX2 plants exhibited higher SAT (Figure 5k), whereas the pil15-2 mutants were more sensitive to salt-alkaline treatment (Figure 5k–m). Additionally, crossing PIL15 OX2 with AMT1 RNAi plants showed that after saline and alkaline treatments, the salt damage rate of PIL15 OX2/AMT1 RNAi plants was higher than that of PIL15 OX2 alone (Figure 5k–m). Sensitivity tests to MeA revealed that the PIL15 OX2 roots were more susceptible to MeA, with a greater reduction in root length compared to PIL15 OX2/AMT1 RNAi and pil15-2 plants (Figure 5n,o). These biochemical and genetic findings indicate that AMT1;2 is a direct target of PIL15 and that PIL15 positively regulates rice SAT and NH₄⁺ uptake by activating AMT1.

PIL15 interacts with IDD10 and BZR1 co-activate AMT1;2

We previously identified that PhyB inhibits IDD10 and BZR1, which directly activates AMT1;2 in rice (Jung et al., 2023). Given that PIL15 also activated AMT1;2, we investigated whether it interacted with IDD10 and BZR1. In Arabidopsis, BZR1 interacts with PIF to regulate light signalling and hypocotyl elongation (Oh et al., 2012). To explore this, we conducted a Y2H assay, which revealed that PIL15 interacted with both IDD10 and BZR1 (Figure 6a,b). These interactions were further confirmed in vivo using the BiFC and Co-IP assays (Figure 6c–f). A transient transactivation assay was conducted to assess the roles of PIL15, IDD10, and BZR1 in activating AMT1;2. The results indicated that PIL15, IDD10, and BZR1 individually activated AMT1;2, and the co-expression of PIL15 with either IDD10 or BZR1 resulted in an additive effect on AMT1;2 activation (Figure 6g–i). These findings suggest that PIL15 forms a transcriptional complex with IDD10 and BZR1 to collaboratively activate AMT1;2 expression and regulate NH₄⁺ uptake.

PhyB downstream TFs differ in ShB resistance

Given that NH₄⁺ uptake enhances resistance to ShB (Wu et al., 2022), which is one of the three major rice diseases, the role of PhyB and its interacting proteins in ShB resistance was investigated. Consistent with the trend in NH₄⁺ uptake, the phyB mutants showed increased resistance to ShB, whereas PhyB OXs were more susceptible (Figure 7a–d). Similar results were observed for CIPK31 and TCP19 plants. TCP19, which could inhibit NH₄⁺ uptake, negatively regulated ShB resistance (Figure 7e,f), whereas CIPK31, by inhibiting the TCP19-mediated AMT1;2 repression, positively regulated the ShB resistance (Figure 7g–j). Interestingly, in contrast to PIL15 OXs enhancing SAT (Figure 5k–m), pil15 mutants were less susceptible to ShB, whereas PIL15 OXs were more susceptible than wild-type ZH11, consistent with previous findings (Yuan et al., 2023) (Figure 7k,l). These results suggest distinct mechanisms in rice responses to SAT and ShB resistance involving PhyB-interacting transcription factors.

Plant growth condition and seedling growth assay

In the glass cultivation room at Nankai University, plants were grown under controlled conditions: temperatures between 24 and 30 °C, 70% relative humidity, and a 12-h light period. Nicotiana benthamiana plants were cultivated in growth rooms at 22–25 °C with a cycle of 16 h of light followed by 8-h darkness.

For the seedling growth assay, following the methods described (Leicher and Stegmann, 2024; Wang et al., 2021), the sterilised seeds were first germinated in ddH₂O and then transferred to 96-well hydroponic boxes (one seedling per well). Each box contained 1 L of liquid medium containing either 0.5× MS, 0.5× MS with 100 mm NaCl, or 0.5× MS with 80 mm NaHCO₃. Plants were grown under normal light/dark cycle conditions, for rice 12 h : 12 h, light : dark cycle and for tobacco 16 h : 8 h, light : dark cycle. Root length was measured after 5–10 days under these growth conditions. ImageJ software (version 1.5.3) (https://imagej.net/ij/) was employed for the root length analysis. Each experiment was conducted in triplicate.

Rhizoctonia solani inoculation assay

The R. solani strain AG1-IA was propagated on potato dextrose agar (PDA) at a constant temperature of 28 °C. The resistance of all rice plants to ShB was evaluated at the third-leaf stage following the methodology described by Cao (Cao et al., 2022). Lesion quantification was performed using ImageJ software. Each experiment was conducted in triplicate.

Methylammonium (MeA) sensitivity test

Sterilised seeds were germinated in ddH₂O and then transferred to a 96-well hydroponic box with one seedling per well. Each box contained 1 L of liquid medium, either 0.5× MS or 0.5× MS, supplemented with 3 mm MeA (Sigma-Aldrich). Root lengths were measured after 5–10 days under these growth conditions, and ImageJ software (version 1.5.3) (https://imagej.net/ij/) was used for root length analysis. Each experiment was conducted in triplicate.

Mutants and transgenic plants

The preparation of phyA, phyB, phyC, PhyB OXs, cipk31, CIPK31 OXs, AMT1 RNAi, idd10, idd10/phyB, and phyB/AMT1 RNAi plants has been described previously (Chen et al., 2023; Wu et al., 2022; Yuan et al., 2023). tcp19 and pil15 mutants were generated using CRISPR/Cas9 genome editing technology on the Zhonghua11 background (Ma et al., 2015). Genomic DNA was extracted from the rice transformants for mutation detection, and the PCR products were analysed using primers flanking the target site (Ma et al., 2015). Double mutants, including tcp19-1/AMT1 RNAi, PIL15 OX2/AMT1 RNAi, and tcp19-1/PIL15 RNAi, were obtained by crossing parental lines and genotyping double homozygous mutants in the F2 generation. To generate overexpression lines, the coding sequences of TCP19 and PIL15 were synthesized and cloned into the pCambia1381-Ubi vector. The constructs were introduced into Agrobacterium tumefaciens strain EHA105 and transferred into the rice cultivar Zhonghua11, following the method described (Hiei et al., 1994). The primers used for cloning are listed in Table S1.

Determination of NH₄⁺ concentration

Ammonium (NH₄⁺) concentration in the roots of 7-day-old rice seedlings was quantified using an F-kit (Roche) following the manufacturer's instructions (Oliveira et al., 2002).

RNA extraction and qRT-PCR

Total RNA from three-week-old rice plants was extracted using TRIzol reagent (Invitrogen) following the manufacturer's instructions. The RNA was reverse-transcribed into cDNA using the HiScript III RT SuperMix for qPCR Kit (Vazyme), and qPCR was conducted with ChamQ Universal SYBR qPCR Master Mix reagent (Vazyme) as per the manufacturer's guidelines. RT-qPCR analysis was performed using an ABI-QS1 real-time PCR instrument (Thermo Fisher), with ubiquitin as the control (Xuan et al., 2013). The relative expression levels were calculated using the 2^−ΔΔCt method. The primers used for qRT-PCR are listed in Table S1.

Yeast-two hybrid (Y2H) assay

Y2H assays were conducted using the Matchmaker Gold Yeast Two-Hybrid System (Clontech). The coding sequences (CDS) of PhyB, CIPK31, PIL15, TCP19, IDD10, and BZR1 were cloned into pGAD424 and pGBT9 vectors to fuse with AD and BD, respectively, and co-transformed into the yeast strain Y2H Gold. Yeast transformants were selected on synthetic defined (SD) dropout medium. Colonies containing both vectors were first selected on SD medium without leucine and tryptophan (SD/−Leu-Trp) and then tested on a three-deficiency medium lacking leucine, tryptophan, and histidine (SD/−Leu-Trp-His). The plates were incubated at 30 °C for 3–7 days and then photographed. The primers used for Y2H are listed in Table S1.

Yeast-one hybrid (Y1H) assay and yeast-three hybrid (Y3H) assay

Y1H assays were performed using the Matchmaker Gold Yeast One-Hybrid System (Clontech). The 2 kb sequences upstream of AMT1;1, AMT1;2, and AMT1;3 start codon (ATG) were cloned into the pHisi-1 vector and transformed into yeast strain YM4271. Yeast transformants were selected on SD/-His medium. The CDS regions of TCP19 and PIL15 were cloned into the pGAD424 vector and transformed into reporter yeast cells containing pHisi-1-pAMT1. The interactions were analysed in SD/−Leu-His medium containing varying concentrations of 3-aminotriazole (3-AT). For the Y3H assay, the CDS regions of PhyB-C and CIPK31 were cloned into the pBridge vector and co-transformed with AD-TCP19 into the yeast strain AH109. Yeast transformants were selected on SD/−Leu-Trp medium, and the interactions were analysed on SD/−Leu-Trp-His or SD/−Leu-Trp-His-Met medium. The yeast cells were incubated at 30 °C for 3 days and then photographed. The primers used for the Y1H and Y3H assays are listed in Table S1.

Bimolecular fluorescence complementation (BiFC) assay

The coding sequences of PhyB, CIPK31, TCP19, PIL15, IDD10, BZR1, and AMT1 (without stop codons) were cloned into BiFC vectors pXNGW and pXCGW, respectively. All plasmids were transformed into the Agrobacterium tumefaciens strain GV3101 (Kim et al., 2009). To test for specific interactions, pairs of plasmids were co-transformed into the leaves of 5-week-old N. benthamiana plants by infiltration, as previously described. H₂B-mRFP1 was utilised as a nuclear marker, and fluorescence signals were imaged 36–48 h after transformation using an Olympus FV3000 laser scanning confocal microscope with 488 and 594 nm lasers to excite the fluorophores. The primers used for BiFC are listed in Table S1.

Co-immunoprecipitation (co-IP) and Western blot (WB) assay

Co-IP of N. benthamiana cells was performed as previously described. The coding sequences of PhyB, CIPK31, PIL15, TCP19, IDD10, and BZR1 (without stop codons) were cloned into pGD-3FLAG, -MYC, and -GFP vectors, respectively. The combinations to be tested were infiltrated into tobacco leaves by Agrobacterium strain GV3101. The samples were harvested at 48 hpi and extracted using an extraction buffer containing 75 mm NaCl, 50 mm Tris–HCl, 1 mm EDTA, 5% glycerol, 0.1% Triton X-100, 1 mm PMSF, and 1x protease inhibitor cocktail. Total protein was immunoprecipitated using anti-GFP magnetic beads (ABclonal Technology) or anti-FLAG magnetic beads (Sigma Aldrich), with 30 μL beads per sample. Protein samples were boiled with 2× loading buffer (1 M Tris–HCl, pH 6.8, 5% SDS, 0.25% Bromophenol Blue, and 25% glycerol) for 10 min and loaded onto 10% or 12.5% SDS-PAGE gels for immunoblot analysis. The WB blotting was performed using anti-GFP, anti-FLAG, and anti-MYC antibodies (Abmart), and signals were detected using a Tanon 4600 system (Tanon).

For competitive co-IP assays, the components were mixed in equal volumes before Agrobacterium infiltration into tobacco leaves. Samples were harvested at 48 hpi, extracted, and incubated with the corresponding beads, followed by western blot analysis. Protein levels were quantified using the Bradford Protein Assay Kit, and ImageJ software was used for protein quantification. Primers used for co-IP are listed in Table S1.

Luciferase complementation imaging (LCI) assay

The coding sequence of CIPK31 was amplified and fused to the C-terminal fragment of luciferase (CLuc) to create CLuc-CIPK31, whereas that of TCP19 was amplified and fused to the N-terminal fragment of luciferase (NLuc) to create NLuc-TCP19. All plasmids were transformed into the Agrobacterium tumefaciens strain GV3101. To test specific interactions, the constructs were co-transformed into the leaves of 5-week-old tobacco plants via infiltration as previously described. After 36–48 h, tobacco leaves were treated with 1 mm luciferin (Promega) for 5 min, and luciferase activity was measured using an imaging apparatus (SH-523, SHbio). Primers used for LCI are listed in Table S1.

Chromatin immunoprecipitation (ChIP) assay

The chromatin immunoprecipitation (ChIP) assays were performed as previously described (Je et al., 2010). Briefly, 10-day-old PIL15-GFP and TCP19-GFP rice seedlings were used. The chromatin from these seedlings was cross-linked with 1% formaldehyde, which was sheared into 200–500 bp fragments by sonication, and used for immunoprecipitation with anti-GFP (ABclonal Technology). After reversing the cross-links, the immunoprecipitated DNA was analysed by qRT-PCR using primers specific to regions of the AMT1 gene. The primers used for target genes in the ChIP assay are listed in Table S1.

Electrophoretic mobility shift assay (EMSA)

The open reading frame (ORF) sequences of TCP19 and PIL15 were cloned into the pGEX-5X-1 vector and the GST-TCP19 or GST-PIL15 recombinant proteins expressed in the Escherichia coli BL21 (DE3) strain. The recombinant GST-TCP19 and GST-PIL15 fusion proteins were induced with addition of 0.5 mm isopropyl-β-D-thiogalactopyranoside (IPTG) for 6 h at 37 °C, followed by purification using GST purification columns (LABLEAD, China). Oligonucleotide probes were directly synthesized and were biotin end-labelled following annealed to double-stranded probe DNA. The assays were performed using the chemiluminescent EMSA kit (Thermo). Primers used for EMSA are listed in Table S1.

Transactivation assay

The pGreenII 0800-LUC cloning vector was used in the transient transactivation assay (Hellens et al., 2005). The promoter region of AMT1;2 (−1 to −1800 bp) was cloned into pGreenII 0800-LUC to drive the LUC gene and generate a reporter construct. Effector plasmids (p35S::TCP19, p35S:CIPK31, p35S::PIL15, p35S::IDD10, and p35S::BZR1) and a reporter plasmid (pAMT1;2) were co-transformed into tobacco leaves. After 36–48 h, the leaves were treated with 1 mm D-luciferin (Promega) for 5 min, and luciferase activity was measured using an imaging apparatus (SH-523, SHbio). The primers used for transactivation assays are listed in Table S1.

Statistical analysis

The statistical analysis was performed using GraphPad Prism software (version 9.4.1). All data were presented as the mean ± standard error (SE; n > 3). Student's t-test was applied to determine the statistically significant differences between two groups, while one-way analysis of variance (ANOVA) was used for comparisons between more than two groups. Significantly different means are indicated by different letters, as determined by two-way ANOVA with Tukey's post hoc test (P < 0.05). NS indicates no significant difference. All experiments were performed at least three times in the laboratory.