Full length cDNA of Bradi1g13980 (BdPIL1; 445 aa) and Bradi2g11100 (BdPIL3; 549 aa) was cloned into pGADT7 vector and fused with the GAL4 activation domain at the C-terminus, and either the N-terminal domain (BdAN, 1∼612 aa; BdBN, 1∼660 aa) or C-terminal domain (BdAC, 568∼1131 aa; BdBC, 653∼1181 aa) of BdphyA (Bradi1g10520) or BdphyB (Bradi1g64360) were cloned into the pGBKT7 vector and fused with the GAL4 DNA-binding domain at the C-terminus. The primers used for cloning are listed in Supplementary Table 1. The vector constructs were then transformed into the yeast strain Y2H Gold (for pGBKT7 constructs) or Y187 (for pGADT7 constructs) using the lithium acetate transformation method (Gietz and Schiestl, 2007) and plated on synthetic dextrose without tryptophan (SD-Trp) or leucine (SD-Leu), respectively. After mating, the yeast cells were selected on SD-Trp-Leu (DDO) plates, and the selected cells were plated on SD-His-Trp-Leu-Ade/X-gal/Aureobasidin A (QDO/X/A) to analyze protein–protein interactions. For quantitative analysis, β-galactosidase assay was performed using the mated yeast cells, according to the Yeast Protocol Handbook (Clontech).

To determine whether the interaction was Pfr-specific, full-length cDNA of BdphyA and BdphyB were cloned into the pGBKT7 vector and used for Y2H assays in the presence of phycocyanobilin (PCB) as chromophore. The mated yeast cells were plated on non-selective (DDO) or selective (QDO/X/A) media containing 20 μM PCB, and incubated for 3 days in the dark or under continuous R light (3 μmol⋅m^–2⋅s^–1), representing the Pr and Pfr forms, respectively.

Full-length recombinant BdphyA, oat phytochrome A (AsphyA), and Arabidopsis thaliana phyB (AtphyB), with a 10-amino acid streptavidin affinity-tag (strep-tag; SAWRHPQFGG) at the C-terminus, were prepared using the Pichia pastoris protein expression system, as previously reported (Shin et al., 2016; Han et al., 2019). PCB was added (final concentration: 20 μM) before purification using streptavidin affinity chromatography under dim green light, and zinc fluorescence assay was performed to verify the ligation of PCB to the phytochromes. The purified phytochrome (as the Pr form) was exposed to R light to generate the Pfr form, which was verified using a diode array UV-Visible spectrophotometer (Cary; Varian Inc).

To prepare recombinant BdPIL1 and BdPIL3, full-length cDNA was cloned into the pGEX 4T-1 (GE Healthcare) vector with the strep-tag, named pStrep vector (Choi et al., 2021), using the primers listed in Supplementary Table 1. Glutathione S-transferase and streptavidin (GST/strep) tags were fused to the N- and C-termini of the recombinant proteins, respectively. The E. coli strain BL21-CodonPlus^TM (Agilent Technology) was used for protein expression and the GST/strep-tagged recombinant BdPIL1 and BdPIL3 proteins were purified using streptavidin affinity chromatography.

Pull-down experiments were performed to examine the in vitro protein–protein interaction between phytochromes and BdPILs, as previously described (Jeong et al., 2016; Shin et al., 2016). To 1 mL of pull-down buffer (100 mM Tris–HCl, pH 7.8, 1 mM EDTA, 150 mM NaCl, and 100 μg mL^–1 BSA), 2 μg of phytochrome (Pr or Pfr form) and 2 μg of GST/strep-tagged BdPIL1 or BdPIL3 were mixed and incubated for 60 min at 4°C with gentle rotation. Next, 50 μL of glutathione resin was added, and the solution was incubated for 30 min. After washing and pelleting glutathione bead-bound proteins, AsphyA/BdphyA, AtphyB, and GST/strep-fused BdPIL1/BdPIL3 proteins were detected using 1:5,000 AsphyA-specific monoclonal antibody (oat25), 1:2,000 AtphyB polyclonal antibody (aN-20; Santa Cruz Biotechnology), and 1:2,000 GST-specific monoclonal antibody (sc-138; Santa Cruz Biotechnology), respectively.

To determine whether the expression of BdPIL1 and BdPIL3 was light-inducible, the lemma of Brachypodium inbred line Bd21-3 (wild-type) seeds were removed and surface-sterilized using 70% (v/v) ethanol and 2% (w/v) sodium hypochlorite for 5 min each with gentle shaking. After washing with sterile distilled water, the seeds were placed on half-strength Murashige and Skoog (MS) medium plates containing 0.8% phytoagar (pH 5.8). After 7 days of cold and dark treatment, the seeds were exposed to white light (WL) for 12 h at 22°C to induce germination. Five days after growth in the dark (D), the seedlings were exposed to WL (100 μmol⋅m^–2⋅s^–1) for 1 or 4 h before harvesting. Moreover, to investigate light-inducible expression under different light conditions, the dark-grown seedlings were exposed to R (λ_max = 654 nm; bandwidth = 25 nm; intensity = 20 μmol⋅m^–2⋅s^–1), FR (λ_max = 738 nm and bandwidth = 42 nm; intensity = 20 μmol⋅m^–2⋅s^–1), and blue (B; λ_max = 450 nm and bandwidth = 20 nm; intensity = 20 μmol⋅m^–2⋅s^–1) light for 4 h in an LED growth chamber (Vision Science Co., Korea). To investigate tissue-specific expression, Bd21-3 plants were grown under long day conditions (18-h light/6-h dark cycle) at 22°C. Root, stem, and leaf tissues were collected from 4-week-old plants, and flower tissues were collected from 6-week-old plants.

For gene expression analysis, seedling and tissue samples were frozen in liquid nitrogen immediately after collection, total RNA was isolated using RNAiso Plus (Takara Bio), and cDNA was synthesized using RNA to cDNA EcoDry Premix kit (Takara Bio). To determine the expression levels of BdPIL1 and BdPIL3, qRT-PCR was performed using Stratagene Mx3005P with Brilliant III Ultra-Fast SYBR Green Q-PCR Master Mix (Agilent Technologies) and corresponding primers (Supplementary Table 1). The expression of BdUBC18, a housekeeping gene, was used for data normalization, and the relative expression levels were estimated by setting the transcript level in the dark-grown or root samples as 1.

To determine the role of BdPIL1 and BdPIL3, we investigated the phenotype of transgenic Bd21-3 plants transformed with RNAi constructs. For this, we cloned partial sequences of BdPIL1 (68∼550 bp) and BdPIL3 (152∼685 bp), based on a previous report (Kerschen et al., 2004), into the pFGC5941 vector using the corresponding primers listed in Supplementary Table 1. The interference fragments were cloned into the RNAi vector in the opposite orientation, separated by the ChsA intron. The binary vector constructs were then transformed into Agrobacterium tumefaciens strain AGL1, and Brachypodium transformation was performed using embryogenic calli induced from immature embryos, according to a previously described method (Alves et al., 2009). The transformed plantlets with well-developed roots were then transferred to the soil, grown under long day conditions for 2 weeks, and sprayed with 0.8% (v/v) BASTA^® to select putative transgenic plants. Herbicide resistance was determined after 7 days, and herbicide-resistant plants were further analyzed using PCR with total genomic DNA isolated from the leaves of mature plants. The coding regions of the 35S promoter (P_35S)-BdPIL1 or P_35S-BdPIL3 and BAR were PCR-amplified using the primers listed in Supplementary Table 1. BdUBC18 was also PCR-amplified using the same template to serve as loading controls.

To obtain homozygous lines with a single transgene integration, the herbicide-resistant plants exhibiting 3:1 segregation in the T2 generation were selected (Supplementary Tables 2, 3), and the plants from the T3 or T4 generations were used for subsequent analyses. To verify RNAi-suppression of BdPIL1 and BdPIL3 in the homozygous lines, 5-day-old dark-grown seedlings were transferred into the growth chamber, exposed to WL for 4 h, and harvested for qRT-PCR analysis. BdUBC18 expression levels were used for data normalization, and relative expression levels were estimated by setting the transcript level in Bd21-3 as 1.

For the photoresponse analysis, seeds without lemma were surface-sterilized, stratified at 4°C for 7 days in the dark, and the embryos were placed vertically on 0.8% phytoagar plates containing half-strength MS salts and vitamins. The seedlings were then exposed to WL for 12 h to promote germination, returned to darkness at 22°C for 1 day, and grown further for 5 days in the dark (D), under continuous R light (cR, 20 μmol⋅m^–2⋅s^–1) or continuous FR light (cFR, 20 μmol⋅m^–2⋅s^–1). Subsequently, the lengths of the coleoptile and first and second leaves of the seedlings were measured using ImageJ.

For the phenotypic analysis, Brachypodium plants were grown at 22°C in a culture room under long day conditions (18-h light/6-h dark cycle), and flowering time, plant height, and chlorophyll content were measured. The flowering time was estimated from germination to the day when the spike emerged (i.e., days to heading), and plant height was measured from the base to the highest point of the plant. Total chlorophyll content was determined using the second and third leaves of the primary tillers of 8-week-old plants. 100 mg of leaf sample was ground in liquid nitrogen, and incubated in 1 mL of 80% acetone, with overnight-shaking in the dark. Absorbances were measured using the UV-Visible spectrophotometer (Cary), and the total chlorophyll content was estimated using the equation: chlorophyll_a+b = 7.15 × A₆₆₀ + 18.71 × A₆₄₇. In addition, stem internode numbers and lengths were also measured. The internode number was counted using the highest tiller, in which the internode lengths were measured after removing the leaf sheath.

Considering the increased height exhibited in the BdPIL1/RNAi and BdPIL3/RNAi plants, histochemical analysis was performed to determine the number and size of cells using the cross-sections of the first internode of the highest tiller. The edge of the first internode was excised using scalpel blade No. 11 (Sigma-Aldrich), subjected to 0.02% toluidine blue staining for 30 s, and rinsed with distilled water. Stem diameter, the number of vascular bundles and pitch cells, and pitch cavity were then estimated, as previously described (Matos et al., 2013; Sakai et al., 2021).

Bd21-3 and transgenic plants harboring the RNAi constructs of BdPIL1 and BdPIL3 (BdPIL1/RNAi and BdPIL3/RNAi, respectively) were grown for 4 weeks under long day conditions, and leaf samples were collected for RNA-seq analysis at 4 h after the start of light cycle. RNA concentration and purity were determined using NanoDrop 2000 (Thermo Fisher Scientific), and TruSeq RNA Sample Preparation Kit V2 (Illumina) was used for library construction. Sequencing was performed using a HiSeq4000 platform (Illumina) with 150-nt paired end sequencing at Macrogen (South Korea). FastQC v.0.11.4 was used for the quality examination of the paired-end reads. Cutadapt v.1.15 and Sickle v.1.33 were used to filter low-quality reads and adaptors. After trimming, reads were aligned to the Phytozome 9.0 B. distachyon reference genome (Bdistachyon_192_hardmasked.fa.gz) using TopHat2 version 2.0.10 with default parameters. Cufflinks version 2.2.1 was used to calculate FPKM (Fragments Per Kilobase of transcripts per Million mapped reads) values. The cuffdiff was carried out for the selection of DEGs (fold change ≥ 2). Gene Ontology and KEGG pathway enrichment analyses were performed using DAVID ver. 6.8 and CluGO ver. 2.5.5 in cytoscape ver. 3.7.1. Heat maps, Venn diagram, and hierarchical clustering were performed with R scripts.

To validate RNA-seq results, we selected several genes involved in the regulation of elongated growth and cell number, including Bradi1g28120 (LOC100838644), Bradi1g51490 (LOC100828255), Bradi2g24980 (LOC100830094), Bradi2g57027 (LOC100823873), Bradi3g46930 (LOC100840545), and Bradi4g10290 (LOC104584443), and performed qRT-PCR using RNA extracted from 4-week-old or 8-week-old plant leaves and the corresponding primers listed in Supplementary Table 1.

To account for the delayed flowering in BdPIL1/RNAi and BdPIL3/RNAi plants, we analyzed the expression of Brachypodium genes that induce flowering, FLOWERING LOCUS T (BdFT1) and CONSTANS (BdCO1) (Lv et al., 2014; Feng et al., 2017; Qin et al., 2019). As the expression of these genes is regulated by diurnal rhythm, we cultivated Brachypodium plants under long day conditions for 8 weeks, and leaf samples were harvested every 2 h in a day. After RNA extraction and RT-PCR, the expression levels of these genes were quantified from DNA gel images with normalization using BdUBC18 expression levels. Then, qRT-PCR was used to analyze the transcript levels in the samples harvested at ZT8 for BdFT1 and ZT20 for BdCO1. The primers used for these analyses are listed in Supplementary Table 1.

To compare chlorophyll biosynthesis between Bd21-3 and the RNAi plants, we analyzed the expression levels of two important genes in chlorophyll biosynthesis, GLU-tRNA REDUCTASE (BdHEMA1; Bradi3g30160) and PROTOCHLOROPHYLLIDE OXIDOREDUCTASE (BdPOR; Bradi5g26230), which were identified using BLASTP with corresponding Arabidopsis proteins as the query sequences. qRT-PCR was performed using RNA samples extracted from 5-day-old dark-grown seedlings and the corresponding primers for BdHEMA1 and BdPOR (Supplementary Table 1).

To determine the DNA-binding ability of BdPIL1 and BdPIL3, EMSA was performed as previously reported (Moon et al., 2008). The sequences of 3 kb upstream region of three genes (Bradi1g51490/BdSAUR50, Bradi5g26230/BdPOR, and BdMIR156H/LOC104794734) were analyzed using PlantPAN 3.0, and 60-bp promoter fragments containing a G-box (CACGTG), E-box (CANNTG), or N-box (CACG(A/C)G) sequence that was scored over 0.98 for the binding to BdPIL1 and BdPIL3 were selected as probes (Supplementary Table 1). After labeling the probes with ³²P-ATP using DNA 5′-End-Labeling System (Promega), 1 pmol of the ³²P-labeled probe was incubated with 2 μg of BdPIL1 or BdPIL3 in a 20 μL reaction mixture (50 mM Tris–HCl, pH 7.5, 50 mM NaCl, 200 mM KCl, 5 mM MgCl₂, 5 mM EDTA, 5 mM DTT, and 250 mM BSA) for 30 min. Cold competitor probes were generated from dimerized oligos without labeling. The reaction mixtures were resolved on 5% native polyacrylamide gels and dried under vacuum before autoradiography.

To examine the effect of phytochrome interaction on the DNA-binding ability of BdPIL1 or BdPIL3, different amounts (0.1, 0.5, 1, 1.5, and 2 μg) of BdphyA or AtphyB (in the Pfr form) were added to the reaction mixture containing 2 μg of BdPIL1 or BdPIL3 before EMSA. For negative control, 2 μg of the phytochrome protein was incubated with 1 pmol of the ³²P-labeled probe only. In addition, to quantify the DNA-binding ability of BdPIL1 and BdPIL3 in the presence of phytochromes, the DNA probe was stained using Electrophoretic Mobility-Shift Assay Kit with SYBR^TM Green and SYPRO^TM Ruby EMSA stains (Thermo Fisher Scientific). The relative DNA binding was estimated by setting the signal from BdPIL1-DNA or BdPIL3-DNA complex in the sample without phytochromes as 1.

Analysis of variance and Duncan’s multiple range test were performed to determine the significant differences in multiple comparisons using IBM SPSS Statistics 20 software. The significant differences in mean values were compared using LSD at P < 0.05 (labeled ‘^∗’) or P < 0.01 (labeled ‘^∗∗’).

Accession numbers to the referenced genes are in Supplementary Table 4. The RNA-seq data included in this study were deposited into Korean Bioinformation Center (KOBIC)¹ under the accession codes of KBRS20191011_0000024 to KBRS20191011_0000032.

Among 146 bHLH transcriptional factors in the genome of Brachypodium (Niu et al., 2017), we identified the following five B. distachyon PIF-like (BdPIL) genes containing conserved active phytochrome binding (APB) and bHLH motifs using BLASTP analysis with Arabidopsis PIFs as query sequences (Supplementary Figures 1A,B): Bradi1g13980, Bradi2g11100, Bradi1g58230, Bradi1g06670, and Bradi5g33170. They belong to the sixth subfamily of Brachypodium bHLH transcriptional factors in the 24 phylogenetic groups classified previously (Niu et al., 2017). As PIF1 and PIF3 are the most studied Arabidopsis PIFs (Pham et al., 2018), we selected two Brachypodium PIFs that are closely related to PIF1 and PIF3 from the phylogenetic analysis (Supplementary Figure 1C), i.e., Bradi1g13980 (BdPIL1) and Bradi2g11100 (BdPIL3) for the present study.

To determine whether BdPIL1 and BdPIL3 are genuine PIFs, we first verified their interaction with phytochromes by yeast two-hybrid (Y2H) assays. Using N- and C-terminal domains of B. distachyon phytochromes A (BdphyA) and B (BdphyB) as baits, we observed the interaction of both BdPIL1 and BdPIL3 with C-domains of both phytochromes (Figures 1A,B). It was also noted that the interactions of BdPILs with the C-domain of BdphyB was stronger than that with the C-domain of BdphyA. In addition, we performed Y2H assays using full-length BdphyA and BdphyB in the presence of phycocyanobilin (PCB) as chromophore, and observed that BdPIL1 and BdPIL3 interacted with both BdphyA and BdphyB under red light condition, but not in the dark (Supplementary Figure 2). These results suggest that both BdPIL1 and BdPIL3 interact with Brachypodium phytochromes in a Pfr-specific manner.

To verify the Pfr-specific interaction further, we performed in vitro pull-down assays using BdPIL1 and BdPIL3 proteins with phytochromes. For this, we initially expressed three phytochromes found in Brachypodium [BdphyA (Bradi1g10520), BdphyB (Bradi1g64360), and BdphyC (Bradi1g08400)] using the Pichia pastoris protein expression system as reported previously (Han et al., 2019). However, BdphyB and BdphyC were not expressed in this system, and only BdphyA could be expressed and purified in a sufficient amount for spectroscopic analysis (Supplementary Figures 3A,B). Thus, we used A. thaliana phyB (AtphyB), as well as Avena sativa phyA (AsphyA), for the pull-down assays. In the case of BdPIL1 and BdPIL3, recombinant proteins with the fusion of glutathione S-transferase (GST) were prepared using an E. coli protein expression system (Supplementary Figure 3C). The pull-down assays revealed the Pfr-specific interaction of BdPIL1 and BdPIL3 with all three phytochromes (Figure 1C). Overall, these results suggest that the conserved APB motifs in BdPIL1 and BdPIL3 function similarly to Arabidopsis PIF1 and PIF3 for the interaction with both phyA and phyB in an active form (Pfr)-specific manner.

To characterize BdPIL1 and BdPIL3 in Brachypodium, we determined whether their gene expression was regulated by light. Both BdPIL1 and BdPIL3 were upregulated under light conditions, compared with the transcript level in the dark (Figure 1D). It was also noted that the expression level of light-induced BdPIL3 was higher than that of BdPIL1. We then analyzed which wavelengths of light were effective for the expression of BdPIL1 and BdPIL3, and found that the expression of both genes increased under R and white light, but not by blue light (Supplementary Figure 4). Furthermore, the expression of BdPIL1 and BdPIL3 was also induced by FR light, but lesser than that induced by R light.

To understand the functions of BdPIL1 and BdPIL3 in Brachypodium, we also analyzed the expression of BdPIL1 and BdPIL3 in different tissues such as the root, stem, leaf, and flower. Both BdPIL1 and BdPIL3 were upregulated in flowers compared with their expression levels in roots (Figure 1E), suggesting their roles in flowering regulation. In addition, the expression of BdPIL3 was increased in the stem and leaf tissues, indicating its possible role in the regulation of leaf and stem growth.

To investigate the roles of BdPIL1 and BdPIL3 in Brachypodium, we used the RNAi approach with the expectation of hypersensitive photoresponses, because PIFs are known as negative regulators of phytochrome signaling in Arabidopsis. For this, we used Bd21-3 inbred line of Brachypodium to generate transgenic plants harboring the RNAi constructs of BdPIL1 and BdPIL3 (Supplementary Figure 5 and Supplementary Tables 2, 3). Among the generated homozygous lines (designated as BdPIL1/RNAi and BdPIL3/RNAi), we selected two independent lines exhibiting the most decreased expression of BdPIL1 (1–5 and 9–20) and BdPIL3 (4–10 and 5–16) for further analyses (Supplementary Figures 6A,B). To verify specific RNAi-suppression, we also analyzed the expression of the five BdPIL genes in the plants of both RNAi lines and found the specific suppression of BdPIL1 and BdPIL3 in the BdPIL1/RNAi and BdPIL3/RNAi plants, respectively (Supplementary Figure 6C).

Then, we determined the photoresponses of the transgenic plants cultivated under continuous red (cR) or continuous far-red (cFR) light. In the dark, the seedlings of both RNAi lines exhibited decreased length of coleoptiles but increased length of the first leaves (Figures 2A,D). Compared with the control plant (Bd21-3), the differences in the lengths of the coleoptiles and first leaves were higher in the BdPIL3/RNAi plants than in the BdPIL1/RNAi plants. Similar results were observed under cR and cFR light conditions (Figures 2B,C,E,F). The longest first and second leaves were observed in the BdPIL3/RNAi plants grown under cR light (Figures 2B,E). In monocots, exposure of seedlings to light inhibits the elongated growth of coleoptiles and causes leaves emerged from coleoptiles (Takano et al., 2009). Thus, the shorter coleoptiles and longer leaves in both RNAi lines suggest that BdPIL1 and BdPIL3 play negative roles in seedling development in response to light.

To determine the physiological functions of BdPIL1 and BdPIL3, we further analyzed the phenotypes of the transgenic plants. Among the phenotypes observed in both RNAi lines, late flowering was the most prominent (Figures 3A,B). In addition, the plants of both RNAi lines exhibited a longer vegetative growth period than Bd21-3, resulting in the tall phenotype (Figure 3C and Supplementary Figure 7A). Under the long-day conditions (18-h light/6-h dark cycle), the maximum height was attained approximately 60 days after germination by Bd21-3 plants, whereas the RNAi plants maintained the growth even after 80 days. It is also notable that the BdPIL3/RNAi plants exhibited more delayed flowering and taller phenotype than the BdPIL1/RNAi plants. As another apparent phenotypes in the plants of RNAi lines, pale green leaves were observed (Supplementary Figure 7B). Thus, we analyzed chlorophyll content in the leaves and found approximately two-fold lower total chlorophyll content in the RNAi plants than in Bd21-3 (Figure 3D). These findings suggest that BdPIL1 and BdPIL3 play roles in the transition to flowering and chlorophyll biosynthesis.

Moreover, to account for the tall phenotype, we investigated stems and found that the plants of both RNAi lines exhibited increased number of internodes compared with Bd21-3 (Figures 4A,B). Moreover, the lengths of the three longest internodes in the main stem were all longer than those in Bd21-3 (Figure 4C). Furthermore, we analyzed the anatomical features of the stems and observed increased stem diameter, more vascular bundles, and more pitch cells in the BdPIL1/RNAi and BdPIL3/RNAi plants (Figures 4D,E). By contrast, pitch cavity was reduced in the plants of both RNAi lines owing to the increased cell number (Figure 4E). These results suggest that the increased height of the both RNAi plants was largely dependent on the increased internode lengths because of the increase in the cell number.

To examine how BdPIL1 and BdPIL3 regulate the observed phenotypes, we investigated the transcriptional changes associated with the RNAi-suppression of BdPIL1 and BdPIL3 using RNA-seq analysis (see Supplementary Data 1, 2). FPKM analysis revealed significant differences between the RNAi lines and wild-type (Bd21-3), exhibiting downregulation of many genes (Figure 5A). Compared with Bd21-3 (using fold change ≥ 2), 444 genes were downregulated, whereas 166 genes were upregulated in the BdPIL1/RNAi plant (Figure 5B). Similarly, 487 genes were downregulated, whereas 144 genes were upregulated in the BdPIL3/RNAi plant. Furthermore, the correlation analyses of the FPKM values revealed that the expression of genes in both BdPIL1/RNAi (R² = 0.578) and BdPIL3/RNAi (R² = 0.569) plants was significantly different from those in Bd21-3 (Figure 5C). By contrast, the correlation between the expression of genes in the BdPIL1/RNAi and BdPIL3/RNAi plants was noticeably high (R² = 0.855), indicating largely overlapped differentially expressed genes (DEGs), i.e., 431 out of 610 and 631 DEGs in BdPIL1/RNAi and BdPIL3/RNAi, respectively (Figures 5C,D). Gene ontology (GO) analysis showed that the shared DEGs were involved in hormone signaling pathways such as auxin and gibberellin, cell wall and chlorophyll biosynthesis, and chloroplast development (Supplementary Figure 8 and Supplementary Data 3, 4). These data suggest that the functions of BdPIL1 and BdPIL3 overlap in Brachypodium, which are consistent with the similar phenotypes observed in the plants of both RNAi lines.

To validate the transcriptome data, we analyzed the expression of the following four upregulated and two downregulated genes using qRT-PCR: two small auxin upregulated RNA genes (BdSAUR; Bradi1g28120 and Bradi1g51490), two gibberellin 20 oxidase 2 genes (BdGA20ox2; Bradi2g57027 and Bradi2g24980), and two cell number regulator genes (BdCNR; Bradi3g46930 and Bradi4g10290). Similar to the RNA-seq results (Figure 5E), BdSAUR and BdGA20ox2 genes were upregulated and BdCNR genes were downregulated in the plants of both RNAi lines (Figure 5F). Considering the function of SAUR and GA20ox2 genes in Arabidopsis, such as the promotion of hypocotyl and internode elongation (Rieu et al., 2008; Dong et al., 2019), the upregulation of BdSAUR and BdGA20ox2 was correlated with the elongated growth observed in both RNAi lines. Moreover, a previous study suggested that, when the expression of maize CNR1 was suppressed, plant and organ size increased because of changes in cell number, but not cell size (Guo et al., 2010). In accordance, the present study found that the BdCNR genes were downregulated in the RNAi plants (Figure 5F), which might correlate with the increased cell number (Figure 4E). Collectively, these gene expression analyses verified the transcriptome data obtained in this study.

Being late flowering and pale green leaves as apparent phenotypes of the BdPIL1/RNAi and BdPIL3/RNAi plants, we analyzed the expression of genes related to these phenotypes using RNA-seq data and qRT-PCR analysis. First, the transcriptome analysis revealed the decreased expression of genes that promote flowering, such as CONSTANS1 (BdCO1/Bradi1g43671), VERNALIZATION1 (BdVRN1/Bradi1g08340), and CO-like (BdCOL2/Bradi3g41500, BdCOL14/Bradi3g19011, and BdCOL16/Bradi3g57000) genes, and increased expression of genes that suppress flowering, such as BdCOL4 (Bradi3g15490) and BdCOL9 (Bradi1g43220) (Supplementary Figure 9A). Based on these results, we further analyzed the expression of BdCO1 and BdFT1 by qRT-PCR, because they play important roles in floral induction in Brachypodium (Lv et al., 2014; Feng et al., 2017; Qin et al., 2019). We compared the gene expression of BdCO1 and BdFT1 at the ZT20 and ZT8 stages, respectively, wherein the maximum expression was observed during the diurnal cycle (Supplementary Figure 10). The results showed that the expression of BdCO1 and BdFT1 was decreased in both RNAi lines (Figure 6A). Especially, BdFT1 expression was strongly decreased to similar levels in both RNAi lines, whereas the decrease in BdCO1 expression was more marked in the BdPIL3/RNAi plants than in the BdPIL1/RNAi plants. These findings were consistent with the extent of delay in flowering observed in the RNAi lines (Figures 3A,B). Thus, the RNAi-suppression of BdPIL1 and BdPIL3 repressed the expression of BdCO1 and BdFT1, resulting in delayed flowering. Therefore, our results suggest that both BdPIL1 and BdPIL3 positively regulate floral induction.

Next, transcriptome analysis also revealed the downregulation of genes involved in chlorophyll biosynthesis, such as BdChlH (Bradi1g19220), BdPBGD (Bradi3g05160), BdHEMA1 (Bradi3g30160), and BdChlI (Bradi1g49770) (Supplementary Figure 9B). Accordingly, we performed qRT-PCR analysis using dark-grown seedlings to investigate the expression of two important genes for chlorophyll biosynthesis, BdHEMA1 (Bradi3g30160) and BdPOR (Bradi5g26230). The results showed significant suppression of these genes in the plants of both RNAi lines (Figure 6B), indicating that both BdPIL1 and BdPIL3 positively regulate chlorophyll biosynthesis. Together, our gene expression analyses were consistent with the results that the RNAi-suppression delays flowering and pale green leaves in Brachypodium plants.

As PIFs are transcriptional factors with a bHLH motif, they bind to DNA. In Arabidopsis, it is known that phytochromes can inhibit the function of PIFs by preventing their DNA binding ability (Park et al., 2018; Favero, 2020). Therefore, we determined the DNA-binding ability of BdPIL1 and BdPIL3 using the promoter sequences of the putative target genes. As PIF proteins recognize G-, E-, and N-box motifs (Kim et al., 2016; Ji et al., 2019), we selected the following putative target genes containing G/E/N-box motifs in their promoter region: BdPOR (Bradi5g26230), BdMIR156H (LOC104794734), and BdSAUR50 (Bradi1g51490). It is notable that there is only one protochlorophyllide reductase (POR) gene in Brachypodium and its expression was light-repressible, whereas the expression of BdSAUR50 was induced under light conditions (Supplementary Figure 11). Using these promoter sequences containing one G/E/N box motif (Supplementary Table 1), we first verified the DNA-binding ability of both BdPIL1 and BdPIL3 by EMSA (Supplementary Figure 12A). Next, we further determined the effect of phytochrome interaction with BdPIL1 and BdPIL3 on their DNA-binding ability. For this, we first analyzed the DNA-binding ability of BdPIL1 in the presence of Pr and Pfr forms of BdphyA, and found that the DNA-binding was inhibited by the active Pfr form, but not by the inactive Pr form (Supplementary Figure 12B). Then, further analysis showed that both BdPIL1 and BdPIL3 bound to the promoter of BdPOR (Figures 7A,B), but the DNA binding was inhibited in the presence of BdphyA (Figures 7C,D). It is also notable that BdphyA inhibited the DNA binding of BdPIL3 more efficiently than that of BdPIL1. Moreover, BdphyA also prevented the binding of BdPIL1 and BdPIL3 to the promoter of BdSAUR50 (Supplementary Figure 13), and AtphyB also showed similar inhibitory effects on the DNA-binding ability of BdPIL1 and BdPIL3 (Figure 7E and Supplementary Figure 14). Overall, these findings elucidate a molecular mechanism underlying the regulation of BdPILs by phytochromes in Brachypodium, i.e., sequestration, wherein phytochromes interact with BdPILs and prevent their binding to the promoters of the target genes.