For general growth and seed harvesting, Arabidopsis thaliana plants were grown in a greenhouse with 16 h light/8 h dark cycles at 22°C. All A. thaliana plants used in this study were in the Col-0 ecotype background. The full-length coding sequences of BBX18, BBX18ΔBox1, and BBX18ΔBox2 were cloned into the gateway-compatible vector pX-YFP (35S promoter-gateway cassette-YFP) to generate transgenic plants BBX18-OX, BBX18ΔBox1-OX, and BBX18ΔBox2-OX. The bbx18-cr1 and bbx18-cr2 mutants were generated by transforming the CRISPR-Cas9 construct pHEE401-UBQ10-BBX18 – targeting the BBX18 locus – into the wild-type plants (Wang Z. P. et al., 2015). Gene-specific primers used for vector construction are listed in the Supplementary Table 1. The PRR5-OX seeds were kindly provided by Nakamichi et al. (2012). The bbx18-4 (SALK_061956) seeds were kindly provided by Yuan et al. (2021).

Sterilized seeds by 70% (v/v) ethanol and 0.01% (v/v) Triton X-100 were plated on Murashige and Skoog (MS) medium (PhytoTechnology Laboratories) and supplemented with 0.75% phytoagar. After 3 days at 4°C, seeds were treated with white light for 6 h to stimulate germination, and then incubated under specific light and temperature conditions. Seedlings were photographed, and hypocotyl lengths were measured using ImageJ.¹

Plants were harvested and ground in liquid nitrogen. Proteins were extracted using protein extraction buffer (100 mM Tris–HCl of pH 6.8, 25% glycerol, 2% SDS, 0.01% bromophenol blue, and 10% beta-mercaptoethanol). BBX18-YFP protein levels were determined by western blotting using anti-GFP antibody (1:5000 dilution, Takara). PRR5-FLAG protein levels were determined by western blotting using the anti-FLAG antibody (1:2000 dilution, Sigma-Aldrich).

For co-immunoprecipitation (co-IP) assays using Arabidopsis mesophyll protoplasts, 2 × 10⁴ isolated mesophyll protoplasts were transfected with a total of 20 μg of DNA (35S::PRR5-Myc and 35S::BBX18-GFP). The IP buffer was used to lyse the transfected protoplasts (50 mM Tris–Cl pH 7.5, 1 mM EDTA, 75 mM NaCl, 0.1% Triton X-100, 5% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor). After centrifugation at 20,000 g for 10 min, the supernatant was incubated for 1 h with anti-GFP antibody (A11122, Thermo Fisher Scientific) immobilized on magnetic beads (Dynabeads™ Protein G, Thermo Fisher Scientific). After three washes with IP buffer, the immunoprecipitated proteins were eluted and immunoblotted with anti-Myc (1:5000 dilution, Cell Signaling Technology) and anti-GFP (1:5000 dilution, Takara) antibodies.

To identify interactions between BBX18 and PRRs, different fragments of BBX18 or PRR complementary DNA were subcloned into the gateway compatible pGADT7 or pGBKT7 vectors (Clontech). The resulting yeast constructs containing various fragments of BBX18 or PRRs cDNA were co-transformed into yeast AH109 cells (Clontech). The yeast clones were grown on synthetic dropout media deficient in Leu and Trp (-LT) and in Leu, Trp, and His (-LTH) but containing various concentrations of 3-amino-1,2,4-Triazole (3-AT).

Total RNA was extracted from plants using the MiniBEST Plant RNA Extraction Kit (Takara) according to the manufacturer’s instructions. For cDNA synthesis, M-MLV reverse transcriptase (Thermo Fisher Scientific) was utilized. We performed quantitative real-time PCR (qRT-PCR) using the Bio-Rad CFX96 Real-Time PCR detection system and the EvaGreen master mix (Solgent). The expression of each gene was normalized to that of PP2A. The gene-specific primers are listed in the Supplementary Table 1.

The seedlings were cross-linked for 20 min with 1% formaldehyde under vacuum. After cross-linking, seedlings were ground in liquid nitrogen. The cross-linked chromatin complex was resuspended in nuclear lysis buffer (50 mM HEPES at pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% Na deoxycholate, and 0.1% SDS) and sheared by sonication to reduce the average DNA fragment size to approximately 0.3–0.5 kb. The fragmented chromatin complex was then immunoprecipitated using the anti-FLAG antibody (Sigma-Aldrich). The beads were washed with low-salt buffer (50 mM Tris–HCl at pH 8.0, 2 mM EDTA, 150 mM NaCl, and 0.5% Triton X-100), high-salt buffer (50 mM Tris–HCl at pH 8.0, 2 mM EDTA, 500 mM NaCl, and 0.5% Triton X-100), LiCl buffer (10 mM Tris–HCl at pH 8.0, 1 mM EDTA, 0.25 M LiCl, 0.5% NP-40, and 0.5% deoxycholate), and TE buffer (10 mM Tris–HCl at pH 8.0, and 1 mM EDTA), and eluted with the elution buffer (1% SDS and 0.1 M NaHCO₃). After reverse cross-linking, the eluted DNA fragments were purified using a PCR purification kit (Macherey-Nagel) and quantified by real-time PCR using specific primers (Supplementary Table 1).

BBX18 was previously shown to promote hypocotyl growth in a temperature-dependent manner (Ding et al., 2018). To confirm this, we generated transgenic plants overexpressing BBX18 fused with YFP at the C-terminus (BBX18-OX) and measured the hypocotyl growth of these plants at two temperatures 20 and 28°C. Consistent with the previous report, the hypocotyls of BBX18-OX plants were longer than those of the wild-type at 28°C (Figure 1A). However, in contrast to the previous study, we noticed that all three BBX18-OX plants had longer hypocotyls than the wild-type at 20°C (Figures 1A,B). Leaf hyponasty was also strongly promoted in the BBX18-OX plants at 20°C (Supplementary Figure 1). These results indicate that overexpression of BBX18 causes constitutive thermomorphogenic responses at normal temperatures. Next, we examined whether BBX18 protein levels are affected by high temperatures. The levels of BBX18-YFP in BBX18-OX plants under the two growth temperatures were not significantly different (Figure 1C). These results show that the protein stability of BBX18 is not affected by the ambient temperature changes.

It was previously shown that hypocotyl growth is hyposensitive to high temperatures in the bbx18 loss-of-function mutants (Ding et al., 2018). To confirm this result, we generated two bbx18 loss-of-function mutants (bbx18-cr1 and bbx18-cr2) using CRISPR/Cas9 against BBX18. bbx18-cr1 mutants contained a 37-nucleotide deletion in the first exon, and bbx18-cr2 mutants had a single-nucleotide insertion in the first exon (Figure 2A). These indel mutations lead to a frameshift and resulted in the generation of premature stop codons (Figure 2B); therefore, both bbx18-cr1 and bbx18-cr2 mutants are likely to be null mutants. However, the hypocotyl growth response to high temperatures was not significantly affected in both bbx18-cr1 and bbx18-cr2 mutants (Figure 2C). We also examined the thermoresponsive hypocotyl growth of bbx18-4 mutants with a T-DNA insertion in the 3′-UTR of the BBX18 gene. The hypocotyl growth in bbx18-4 mutants at high temperature was comparable to that of the wild-type at both the temperatures (Supplementary Figure 2). In line with these results, the expression of PIF4 and PIF5 was not significantly altered in response to the introduction of bbx18 mutations (Figures 2D,E). Expression of PIF4 target genes (YUC8, IAA19, and IAA29) was similarly induced by high temperature in wild-type and bbx18-cr1 plants (Figures 2F–H). Together, these results suggest that genes with functions similar to those of BBX18 might be present in the Arabidopsis genome.

A previous study proposed that BBX18 promotes hypocotyl growth by reducing the levels of ELF3, which negatively regulates the expression of PIF4 and its homolog PIF5 (Ding et al., 2018). However, it was not shown whether BBX18 influences the expression of these PIFs (Ding et al., 2018). Therefore, we determined the expression of PIF4 and PIF5 in the wild-type and BBX18-OX plants at two zeitgeber times (ZT0 and ZT12). PIF4 expression was higher in BBX18-OX plants than in the wild-type plants at ZT0 and ZT12 (Figure 3A). PIF5 expression was similar in the wild-type and BBX18-OX at ZT0, but was higher in the BBX18-OX than in the wild-type at ZT12 (Figure 3B). Given that ELF3 represses the expression of PIF4 and PIF5 in the evening, these results corroborate the hypothesis that BBX18-mediated ELF3 degradation stimulates hypocotyl growth through PIF4 and PIF5. To further confirm this hypothesis, we examined if PIFs are necessary for the promotion of hypocotyl growth by BBX18. Overexpression of BBX18 was not able to promote hypocotyl growth in the pif quadruple (pifq) mutants lacking PIF1, PIF3, PIF4, and PIF5 at normal temperature (20°C) (Figure 3C). BBX18 could not promote hypocotyl growth even at high temperature (28°C) in the pifq mutants (Figure 3C). These results indicate that the promotion of hypocotyl growth by BBX18 is dependent on the PIF activity, irrespective of the temperature.

Next, to examine whether BBX18 regulates hypocotyl growth exclusively through the ELF3-dependent pathway, we determined the effects of BBX18 overexpression on hypocotyl growth in an elf3-1 mutant background (BBX18-OX;elf3-1). Interestingly, overexpressed BBX18 significantly promoted hypocotyl growth in the absence of functional ELF3 (Figure 3D). This finding indicates that BBX18 promotes hypocotyl growth via ELF3-dependent and -independent pathways.

Given the inability of BBX18 to enhance hypocotyl growth in the pifq mutants (Figure 3C), it is likely that the ELF3-independent pathway also increases PIF abundance and/or activity to induce hypocotyl growth promotion. In line with this hypothesis, the expression levels of PIF4 target genes (YUC8, IAA19, and IAA29) were significantly higher in BBX18-OX;elf3-1 plants than those in elf3-1 plants (Figures 3E–G). In contrast, PIF4 expression was not significantly increased in response to the overexpression of BBX18 in the elf3-1 mutant, implying that BBX18 potentiates PIF4 activity (Figure 3H). To identify the molecular mechanism by which BBX18 promotes hypocotyl growth independent of ELF3, we tested the interaction of BBX18 with proteins known to regulate PIF activities using yeast two-hybrid assays (Supplementary Figures 3A,B). This screening identified PRR5 as a BBX18-interacting protein (Figure 4A). PRR5 was previously shown to inhibit the transactivation activity and transcript-level expression of PIF4 (Nakamichi et al., 2012; Zhu et al., 2016). We confirmed the interaction between BBX18 and PRR5 in vivo using co-IP assays (Figure 4B). BBX18 contains two tandem repeats of B-Box domains (B-Box1 and B-Box2). To identify the domains of BBX18 which are required for interaction with PRR5, we performed the yeast two-hybrid assays with the truncated versions of BBX18 proteins (Figures 4C,D). The assays revealed that the B-Box2 domain is necessary for the interaction of BBX18 with PRR5 (Figure 4D). PRR5 has a pseudo-receiver (PR) domain at the N-terminus and CONSTANTS, CONSTANS-LIKE, and TOC1 (CCT) domains at the C-terminus (Figure 4E). Yeast two-hybrid assays revealed that the PRR5 truncation containing the PR domain, and not the CCT domain, interacted with BBX18 (Figure 4F).

We next examined whether PRR5 interacts with BBX19, a homolog of BBX18, which promotes hypocotyl growth through PIFs (Wang C. Q. et al., 2015). Figure 4G shows that PRR5 interacts with BBX19 as well as BBX18. Additional yeast two-hybrid assays revealed that BBX18 interacts with PRR7 but not PRR9, and BBX19 interacts with PRR9. Taken together, these results indicate that BBX18 and BBX19 interact with multiple PRR proteins.

In contrast to BBX18, PRR5 inhibits hypocotyl growth by inhibiting the expression of PIF4 and repressing the transactivation activity of PIF4 (Nakamichi et al., 2012; Zhu et al., 2016). Therefore, it is likely that the BBX18-PRR5 interaction interferes with PRR5 to promote hypocotyl growth. To test this hypothesis, we first determined whether the B-Box2 domain is required by BBX18 to promote hypocotyl growth because the B-Box2 domain mediates the BBX18-PRR5 interaction. We generated transgenic plants overexpressing BBX18 with the deletion of B-Box2 (BBX18ΔBox2-OX). In contrast to BBX18-OX plants, the hypocotyls of BBX18ΔBox2-OX plants were not significantly longer than those of wild-type plants (Figure 5A), although the levels of BBX18ΔBox2-YFP were higher than those of BBX18-YFP (Figure 5B). Consistent with this hypocotyl growth pattern, the overexpression of BBX18ΔBox2 only marginally increased PIF4 expression and had no effect on the expression of PIF5 (Figures 5C,D). Given that the B-Box2 domain is required for the interaction of BBX18 with PRR5 (Figure 4D), BBX18 may promote hypocotyl growth, at least partially, through the inhibition of PRR5. Additionally, we generated transgenic plants overexpressing BBX18 with the deletion of the B-Box1 domain (BBX18ΔBox1-OX) and measured their hypocotyl growth. The overexpression of BBX18ΔBox1 did not affect hypocotyl growth (Supplementary Figure 4). The B-Box1 domain of BBX19 was previously shown to be essential for its interaction with ELF3 (Wang C. Q. et al., 2015). Therefore, the deletion of the B-Box1 domain is likely to abolish the interaction between BBX18 and ELF3, which might be partially responsible for the inability of BBX18ΔBox1 to promote hypocotyl growth.

PRR5 directly binds to the promoter of PIF4 and represses its expression (Nakamichi et al., 2012). Because BBX18 augments PIF4 expression (Figure 3A), we examined whether BBX18 promotes PIF4 expression by preventing PRR5 from binding to the PIF4 promoter using chromatin immunoprecipitation (ChIP) assays in PRR5-OX and PRR5-OX;BBX18-OX plants. Consistent with the findings of a previous report (Nakamichi et al., 2012), ChIP assays showed that PRR5 binds to the PIF4 promoter (Supplementary Figure 5). The binding of PRR5 to the PIF4 promoter was not significantly affected by the overexpression of BBX18 (Supplementary Figure 5). These results show that the interaction between BBX18 and PRR5 does not affect the DNA-binding ability of PRR5.

The expression of PRR5 is regulated by the circadian clock, and peaks at ZT12. The circadian clock-dependent accumulation of PRR5 mediates the circadian gating of high-temperature-mediated PIF4 activation and thermoresponsive hypocotyl growth (Zhu et al., 2016). High-temperature-mediated PIF4 activation is enhanced around ZT0 when the PRR5 levels are low, whereas it is suppressed around ZT12 when the PRR5 levels are high (Zhu et al., 2016). To test whether the interaction between BBX18 and PRR5 alleviates PRR5-mediated suppression of high-temperature responses, we first determined the high-temperature responses of PIF4 and PIF4 target genes – in terms of expression – at different circadian times (ZT8–12 and ZT20–24) in wild-type and BBX18-OX plants. Although the expression of PIF4 was similarly induced by high-temperature during ZT8–12 and ZT20–24 (Figure 6A), the expression of several PIF4 target genes (YUC8, IAA19, and IAA29) was increased in response to the high temperature at ZT20–24, but not at ZT8–12, in the wild-type (Figures 6B–D). In contrast, the expression of these PIF4 target genes was induced by high temperature both during ZT8–12 and ZT20–24 in BBX18-OX plants (Figures 6B–D). These results show that overexpressed BBX18 restores the thermoresponsiveness of the PIF4 target genes in the evening (ZT12).

To examine if BBX18 interferes with PRR5-mediated inhibition of the high-temperature responses, we measured the hypocotyl growth of wild-type, PRR5-OX, and PRR5-OX;BBX18-OX plants grown under two different temperature conditions (20 and 28°C). As previously reported (Zhu et al., 2016), while the hypocotyls of the wild-type were significantly elongated at high temperature, those of PRR5-OX were mostly insensitive to the high temperatures (Figure 7A). The impaired thermoresponsive hypocotyl growth in PRR5-OX plants was partially restored in response to the overexpression of BBX18 (Figure 7A). The levels of PRR5 were not reduced but rather increased upon the overexpression of BBX18 (Figure 7B), which indicates that the restored thermoresponsive hypocotyl growth in PRR5-OX;BBX18-OX plants could not be attributed to the reduced levels of PRR5. The expression of PIF4 target genes (YUC8, IAA19, and IAA29) in the PRR5-OX plants did not increase at high temperatures, although the expression of PIF4 was significantly increased (Figures 7C–F), confirming that PRR5 inhibits PIF4-mediated activation of YUC8, IAA19, and IAA29. The thermoresponsive expression of these genes was restored in PRR5-OX;BBX18-OX plants (Figures 7D–F), consistent with the hypocotyl phenotypes. Therefore, these results provide evidence that BBX18 inhibits PRR5-mediated suppression of PIF4 activity, concomitantly enhancing the thermoresponsive hypocotyl growth.