The PHYB–FOF2–VOZ2 module functions to fine-tune flowering in response to changes in light quality by modulating FLC expression in Arabidopsis

Abstract

Proper timing of flowering under different environmental conditions is critical for plant propagation. Light quality is a pivotal environmental cue that plays a critical role in flowering regulation. Plants tend to flower late under light with a high red (R)/far-red (FR) light ratio but early under light with a low R/FR light ratio. However, how plants fine-tune flowering in response to changes in light quality is not well understood. Here, we demonstrate that F-box of Flowering 2 (FOF2), an autonomous pathway–related regulator, physically interacts with VASCULAR PLANT ONE-ZINC FINGER 1 and 2 (VOZ1 and VOZ2), which are direct downstream factors of the R/FR light receptor phytochrome B (PHYB). We show that PHYB physically interacts with FOF2, mediates stabilization of the FOF2 protein under FR light and end-of-day FR light, and enhances FOF2 binding to VOZ2, which leads to degradation of VOZ2 by SCF^FOF2 E3 ligase. By contrast, PHYB mediates degradation of FOF2 protein under R light and end-of-day R light. Genetic interaction studies demonstrated that FOF2 functions downstream of PHYB to promote FLC expression and inhibit flowering under both high R/FR light and simulated shade conditions, processes that are partially dependent on VOZ proteins. Taken together, our findings suggest a novel mechanism whereby plants fine-tune flowering time through a PHYB–FOF2–VOZ2 module that modulates FLC expression in response to changes in light quality.

Light quality is a pivotal environmental cue for flowering regulation, but how plants fine-tune flowering in response to changes in light quality is not well understood. This study reveals that the F-box protein FOF2 acts downstream of PHYB to regulate VOZ2 stability and thus modulate FLC expression, ensuring proper flowering in response to changes in light quality.

Introduction

Flowering is one of the most important processes in plant growth and development, and it determines the reproductive and genetic ability of higher plants. Flowering time is affected by various external and internal factors such as day length, ambient temperature, plant hormones, and plant age (Li et al., 2016, 2022; Perrella et al., 2020). Plants integrate environmental cues into their developmental programs to achieve proper flowering and reproduction. Arabidopsis thaliana is a facultative long-day (LD) plant in which six major flowering time pathways—the photoperiod, temperature, vernalization, gibberellin (GA) biosynthesis, autonomous, and aging pathways—have been identified. They converge to regulate the floral integrators FLOWERING LOCUS T (FT), TWIN SISTER OF FT (TSF), and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1) (Fornara et al., 2010). Light quality can also contribute to floral induction in plants, and plants tend to flower late under light with a high red (R)/far-red (FR) light ratio and early under light with a low R/FR light ratio (Halliday et al., 1994; Lin, 2000; Cerdan and Chory, 2003; Verhage et al., 2014). However, how plants fine-tune flowering in response to changes in light quality is not well understood.

Photoreceptor phytochromes mainly absorb and trigger responses associated with R and FR light and detect changes in the R/FR ratio of incident light (Franklin and Quail, 2010; Legris et al., 2019). The phytochrome family has five members (PHYA, PHYB, PHYC, PHYD, and PHYE). PHYB regulates flowering in response to changes in R/FR ratio and plays a major role in shade-avoidance syndrome (SAS), although other related phytochromes, such as PHYD and PHYE, act redundantly (Cerdan and Chory, 2003; Franklin, 2008; Casal, 2013; Martinez-Garcia et al., 2014; Pierik and de Wit, 2014). SAS reactions can be induced under simulated shade by addition of FR light to white light (WL) (low R/FR) or by a short FR light treatment at the end of the photoperiod (EOD-FR) (Goto et al., 1991; Nagatani et al., 1991; Halliday et al., 1994; Cerdan and Chory, 2003). The phyB mutant exhibits a constitutive SAS response and early flowering, even under high R/FR conditions, and shows attenuated responses to low R/FR or EOD-FR treatment, indicating that PHYB negatively regulates shade-induced flowering (Goto et al., 1991; Nagatani et al., 1991; Robson et al., 1993; Halliday et al., 1994; Devlin et al., 1996). CO is a B-box transcription factor that regulates flowering time mainly by activating FLOWERING LOCUS T (FT) transcription (Putterill et al., 1995) and plays an important role in flowering-time regulation downstream of phytochromes. Both CO mRNA levels and CO protein abundance increase upon exposure to low R/FR (Kim et al., 2008; Wollenberg et al., 2008). However, reports have also shown that co mutation does not affect responsiveness to EOD-FR treatments in either wild-type or phyB mutant backgrounds (Devlin et al., 1996, 1998). Inigo et al. showed that PHYTOCHROME AND FLOWERING TIME 1 (PFT1) acts downstream of PHYB to promote flowering through CO-dependent and -independent pathways by activating CO transcription and FT transcription, respectively (Inigo et al., 2012). Recently, PHYTOCHROME-INTERACTING FACTORs 4, 5, and 7 (PIF4, PIF5, and PIF7) were reported to mediate shade-induced flowering and act directly downstream of PHYB to promote expression of FT and TSF (Galvao et al., 2019; Zhang et al., 2019). However, whether and how PHYB regulates flowering time through other regulators in response to changes in light quality remains largely unknown.

The VASCULAR ONE-ZINC FINGER 1 (VOZ1) and VOZ2 transcription factors have been identified as PHYB-interacting factors (Yasui et al., 2012). VOZ1/2 downregulate FLOWERING LOCUS C (FLC) expression and function downstream of PHYB in regulating flowering time (Yasui et al., 2012; Celesnik et al., 2013; Yasui and Kohchi, 2014). FLC encodes a MADS domain–containing protein that is a key repressor in the autonomous and vernalization pathways and inhibits flowering by directly repressing FT transcription (Michaels and Amasino, 1999; Helliwell et al., 2006). Mutation(s) of autonomous pathway genes such as FCA, FVE, FPA, FY, LUMINIDEPENDENS (LD), FLOWERING LOCUS K HOMOLOGY DOMAIN (FLK), and FLOWERING LOCUS D (FLD) cause elevated FLC expression and consequent late flowering (Simpson, 2004; Amasino, 2010). The phyB mutation can largely rescue the late flowering of mutants in several autonomous-pathway genes such as FCA and FVE (Halliday et al., 1994). Whether and how the PHYB–VOZ2 module regulates flowering time and FLC expression in response to changes in light quality are not clear.

We previously identified F-box of Flowering 2 (FOF2), which encodes an F-box protein that is a subunit of the SKP1–CUL1–F-box (SCF) E3 complex, as an autonomous pathway–related regulator downstream of FCA (Gagne et al., 2002; He et al., 2017). FOF2 functions redundantly with its homolog FOF2-LIKE 1 (FOL1) to negatively regulate flowering time by promoting FLC expression, and FOF2 itself is regulated by both the autonomous pathway gene FCA and light signals (He et al., 2017), suggesting that FOF2 is a possible linker between light conditions and the autonomous flowering pathway. In this study, we showed that FOF2 interacts with both VOZ proteins and PHYB and participates in PHYB-mediated regulation of flowering time under both high R/FR light and simulated shade conditions by regulating VOZ2 stability to modulate FLC expression, thereby ensuring that flowering occurs at an appropriate time in response to changes in light quality.

Results

FOF2 interacts with VOZ proteins

To investigate the molecular mechanism by which FOF2 regulates FLC expression, we performed a yeast two-hybrid (Y2H) screen to identify proteins that interact with FOF2. VOZ1, a transcription factor specific to higher plants (Mitsuda et al., 2004), was isolated (Supplemental Table 1). Arabidopsis has two members of the VOZ transcription factor family, VOZ1 and VOZ2, which interact with PHYB and redundantly promote flowering time and inhibit FLC expression (Yasui et al., 2012; Yasui and Kohchi, 2014). Therefore, we performed a detailed investigation of the FOF2–VOZ1 interaction using a targeted Y2H assay. As expected, FOF2 interacted with VOZ1, and we found that FOF2 also interacted with VOZ2 in yeast (Figure 1A–1D; Supplemental Figure 1A). To confirm the interaction between FOF2 and VOZ1/2 in plants, we performed a bimolecular fluorescence complementation (BiFC) assay using tobacco leaves. Clear YFP fluorescence was observed in the nucleus upon co-expression of cCFP-FOF2 and VOZ1-nYFP or VOZ2-nYFP (Figure 1E) but not in control cells (Supplemental Figure 2), suggesting that FOF2 interacts with VOZ1/2 in the nucleus. Co-immunoprecipitation (coIP) assays showed that FOF2 interacted with VOZ1/2 in plant cells (Figure 1F). Both semi-in vivo pull-down and in vitro pull-down assays further confirmed their direct interactions (Supplemental Figure 1B and 1C). Taken together, these results suggest that FOF2 physically interacts with VOZ1/2 in vitro and in vivo.

Figure 1.

FOF2 interacts with VOZ proteins.

(A) Schematic diagram of various constructs used in the Y2H assays. Numbers indicate amino-acid positions in FOF2.

(B) Y2H assays showing interactions of FOF2 and the LRR and FBD domains of FOF2 with VOZ1 and VOZ2. FOF2 and FOF2ΔF were fused with the DNA-binding domain (BD) of GAL4, and VOZ1 and VOZ2 were fused with the transcriptional activation domain (AD) of GAL4. Yeast cells co-expressing the indicated combinations of proteins were grown on non-selective (SD−TL) and selective (SD−TLH and SD−TLH + 1 mM 3-AT) media. FOF2ΔF, F-box deletion of FOF2 (containing the LRR and FBD domains).

(C) Schematic diagram of various constructs used in the Y2H assays. Numbers indicate amino-acid positions in VOZ2.

(D) Y2H assays showing interaction of FOF2 with the B domain of VOZ2. The FOF2 protein was fused with the BD of GAL4, and different constructs of VOZ2 were fused with the AD of GAL4. Yeast cells co-expressing the indicated combinations of proteins were grown on non-selective (SD−TL) and selective (SD−TLH + 3-AT) media. VOZ2a, A-domain deletion of VOZ2 (containing the B domain); VOZ2b, B-domain deletion of VOZ2 (containing the A domain).

(E) Bimolecular fluorescence complementation (BiFC) assay showing interactions of FOF2 with VOZ1 and VOZ2. cCFP-FOF2 and VOZ1-nYFP or VOZ2-nYFP were co-transformed into tobacco leaves. Bright, bright field. Merged, overlay of the YFP and bright-field images. Scale bars correspond to 50 μm.

(F) CoIP assay showing FOF2–VOZ1 and FOF2–VOZ2 interactions in Arabidopsis. VOZ1-FLAG, VOZ2-FLAG, VOZ1-FLAG/Myc-FOF2, and VOZ2-FLAG/Myc-FOF2 seedlings were grown on ¹/₂ Murashige & Skoog (MS) solid medium under continuous WL (cW) for 8 days. Total proteins (input) or the IP product of anti-Myc (IP (Myc)) antibody in immunoblots were probed with anti-Myc and anti-FLAG antibodies.

(G and H) CoIP assay showing interaction of FOF2 and the LRR and FBD domains of FOF2 with VOZ1 (G) and VOZ2 (H). VOZ1-GFP and VOZ2-GFP alone or together with FOF2-Myc or FOF2ΔF-Myc were transiently expressed in tobacco leaves. The leaves were infiltrated with 50 μM MG132 for 5 h before sampling. Immunoprecipitates against anti-Myc antibody (IP (Myc)) or crude extracts (input) were analyzed via immunoblots using anti-Myc antibody and anti-GFP antibody. FOF2ΔF, F-box deletion of FOF2 (containing the LRR and FBD domains).

(I) CoIP assay showing interaction of FOF2 with VOZ2 and the B domain of VOZ2. VOZ2-GFP, VOZ2a-GFP, and VOZ2b-GFP alone or together with FOF2-Myc were transiently expressed in tobacco leaves. The leaves were infiltrated with 50 μM MG132 for 5 h before sampling. Immunoprecipitates against anti-Myc antibody (IP (Myc)) or crude extracts (input) were analyzed via immunoblots using anti-Myc antibody and anti-GFP antibody. VOZ2a, A-domain deletion of VOZ2 (containing the B domain); VOZ2b, B-domain deletion of VOZ2 (containing the A domain).

(J–L) Genetic interaction between FOF2 and VOZ proteins in flowering. Images of representative flowering phenotypes (J) of plants grown under long days (LD, 16-h light/8-h dark) in WL (high R/FR light ratio) conditions for 50 days. The days to flowering (K) and the number of rosette leaves (L) on the day floral buds became visible were measured. Standard deviations (n ≥ 20) are shown. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(M) mRNA levels of FLC and FT in fof2fol1, voz1voz2, and fof2fol1/voz1voz2 mutants. Seedlings were grown on ¹/₂ MS solid medium under LD in high R/FR light conditions for 12 days and harvested 4 h after being exposed to light. Gene expression was normalized to that of ACTIN2 (ACT2). Bars represent the standard deviations of three independent experiments. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

To better understand the interaction between FOF2 and VOZ proteins, we selected VOZ2 for analysis of their domains of interaction because both VOZ1 and VOZ2 are homologs and interact with FOF2. FOF2 has an F-box domain for interacting with the ASK14 protein (He et al., 2017) and leucine rich repeat (LRR) and FBD domains for substrate recognition (Xu et al., 2009; Hua and Vierstra, 2011) (Figure 1A). VOZ2 has two conserved domains, domains A and B, and domain B contains a zinc-coordinating motif and a basic region (Mitsuda et al., 2004) (Figure 1C). Our Y2H and coIP analyses demonstrated that an F-box deletion of FOF2 (FOF2ΔF, containing the LRR and FBD domains) interacted with VOZ1/2 (Figure 1B, 1G, and 1H; Supplemental Figure 1A). In addition, domain B, but not domain A, of VOZ2 interacted with FOF2 (Figure 1D and 1I). These observations suggest that the LRR and FBD domains of FOF2 and domain B of VOZ2 are the domains for FOF2 and VOZ2 protein interaction.

To further investigate the genetic relationship between FOF2 and VOZ in flowering, we first used the voz1voz2 (voz1-1voz2-1) mutant, which exhibited a late-flowering phenotype (Yasui et al., 2012; Supplemental Figure 3), for genetic analysis. Accordingly, we crossed voz1voz2 with the fof2fol1 (CR-fof2fol1-m1) mutant, which exhibited an early-flowering phenotype (He et al., 2017), and analyzed the flowering phenotype of the fof2fol1/voz1voz2 quadruple mutant under LD in WL conditions (Supplemental Figure 4A and 4B; high R/FR [R/FR∼4.5]). voz1voz2 completely suppressed the early-flowering phenotype of fof2fol1 (Figure 1J), and the flowering time and leaf number of fof2fol1/voz1voz2 at bolting were the same as those of the voz1voz2 mutant (Figure 1K and 1L). Consistent with the flowering phenotype, FLC expression was upregulated in the voz1voz2 mutant but downregulated in the fof2fol1 mutant (Yasui et al., 2012; Yasui and Kohchi, 2014; He et al., 2017; Figure 1M and Supplemental Figure 5). In agreement with the notion that FT is regulated by the floral repressor FLC via direct binding (Helliwell et al., 2006), FT expression was downregulated in the voz1voz2 mutant and upregulated in the fof2fol1 mutant (Figure 1M and Supplemental Figure 5). The expression levels of FLC and FT in the fof2fol1/voz1voz2 mutant were similar to those in the voz1voz2 mutant (Figure 1M and Supplemental Figure 5). These results demonstrate that FOF2 acts upstream of VOZ proteins to regulate FLC and FT expression and flowering time under high R/FR conditions.

FOF2 is involved in VOZ2 protein degradation via the 26S proteasome pathway under FR light

We previously showed that FOF2 is a component of SCF E3 ligase and interacts with the ASK14 protein (He et al., 2017). VOZ2 has been reported to be degraded in response to FR light via the 26S proteasome pathway (Yasui et al., 2012). Therefore, we wondered whether FOF2 might be involved in VOZ2 protein degradation. We first examined the light regulation of VOZ2 protein stability using VOZ2-FLAG transgenic lines constitutively expressing VOZ2-FLAG (Luo et al., 2020). Intriguingly, the VOZ2-FLAG protein was degraded in response to FR light or in the dark but was stable in response to R light or blue light (Supplemental Figure 6A and 6B), although the transcript abundance of the VOZ2-FLAG transgene was not significantly affected by light (Supplemental Figure 7). By pre-treating the seedlings with MG132, a 26S proteasome–specific inhibitor, we reconfirmed that VOZ2 degradation induced by FR light was dependent on the 26S proteasome (Yasui et al., 2012; Supplemental Figure 6C and 6D). We then examined the role of FOF2 in the FR light–induced degradation of the VOZ2 protein using a cell-free system as described previously (Wang et al., 2009; Yan et al., 2020). Compared with wild-type Col-0, the fof2fol1 mutant showed slower degradation of GST-VOZ2, whereas FOF2 overexpression resulted in faster GST-VOZ2 degradation (Figure 2A and 2C; Supplemental Figure 8). Moreover, GST-VOZ2 degradation was efficiently blocked by MG132 (Figure 2B and 2D; Supplemental Figure 8). We further examined the degradation of the VOZ2 protein in vivo. Seedlings were grown in continuous WL (Supplemental Figure 4A and 4B) for 10 d, then exposed to FR light for the indicated times. VOZ2-FLAG protein degraded more slowly in the fof2fol1 mutant but more quickly in the Myc-FOF2-overexpressing plants compared with the wild-type background (Figure 2E and 2G; Supplemental Figure 8). Moreover, VOZ2-FLAG degradation in Myc-FOF2-overexpressing plants was effectively inhibited by MG132 (Figure 2F and 2H; Supplemental Figure 8). In addition, to exclude the effect of reduced total light density in the light changes from continuous WL to FR light on protein stability, we examined VOZ2-FLAG protein stability using WL to simulated shade conditions (continuous WL [cW] + far-red light) (the same photosynthetically active radiation [PAR] with WL) (Supplemental Figure 4C and 4D). Similarly, VOZ2-FLAG was degraded in response to simulated shade, and FOF2 was involved in this degradation (Supplemental Figure 9A and 9C). We then examined the effect of FOF2 on VOZ1 protein levels under FR light conditions. VOZ1-FLAG protein levels decreased markedly in Myc-FOF2-overexpressing plants, but no obvious change was observed in the wild-type background (Supplemental Figure 10A and 10C). Furthermore, the FOF2-triggered reduction in VOZ1-FLAG protein could be inhibited by MG132 treatment (Supplemental Figure 10B and 10D). Collectively, these data indicate that FOF2 is involved in the degradation of VOZ1 and VOZ2 via the 26S proteasome pathway in response to FR light.

Figure 2.

FOF2 participates in FR light–induced degradation of VOZ2 protein via the 26S proteasome, partially dependent on PHYB.

(A and B) Representative immunoblots showing the degradation of GST-VOZ2 in cell-free degradation assays. Wild-type Col-0, fof2fol1 mutant, and Myc-FOF2-overexpressing seedlings were grown on ¹/₂ MS solid medium under cW for 10 days, then exposed to FR light (30 μmol m⁻² s⁻¹) for 24 h and sampled. E. coli–purified recombinant GST-VOZ2 was incubated with equal amounts of total protein extracts from seedlings for the indicated times in the absence (A) or presence (B) of 50 μM MG132. Protein levels were analyzed by immunoblotting using anti-GST antibody. The anti-actin antibody was used as the loading control. Each assay was biologically replicated three times.

(C and D) Relative protein levels of GST-VOZ2 in (A) and (B). GST-VOZ2 protein levels were normalized to actin levels. The values of starting points were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (Student’s t-test).

(E) Representative immunoblots showing protein levels of VOZ2-FLAG in VOZ2-FLAG/Col-0, VOZ2-FLAG/fof2fol1, and VOZ2-FLAG/Myc-FOF2 seedlings. Seedlings were grown on ¹/₂ MS solid medium under cW for 10 days, exposed to FR light (FR, 30 μmol m⁻² s⁻¹) for the indicated times, and sampled. Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(F) Representative immunoblots showing protein levels of VOZ2-FLAG in VOZ2-FLAG/Myc-FOF2 seedlings treated with the proteasome inhibitor MG132. Seedlings were grown on ¹/₂ MS solid medium under cW for 10 days, treated with 50 μM MG132 (+MG132) or DMSO (–MG132) for 5 h, exposed to FR light (30 μmol m⁻² s⁻¹) for the indicated time, and sampled. Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(G and H) Relative protein levels of VOZ2-FLAG in (E) and (F). VOZ2-FLAG protein levels were normalized to HSP82 levels. The values of starting points were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗∗p < 0.01, ∗∗∗p < 0.001 (Student’s t-test).

(I) Ubiquitination analysis of VOZ2 protein in tobacco. VOZ2-Myc alone or together with FOF2-FLAG was transiently expressed in tobacco leaves. The leaves were infiltrated with 50 μM MG132 for 12 h before sampling. Immunoprecipitates against anti-Myc antibody (IP) or crude extracts (input) were analyzed via immunoblotting using anti-Myc antibody, anti-FLAG antibody, and anti-ubiquitin antibody (Enzo Life Sciences; BML P-W8810-0100). Ponceau staining was used as a loading control.

(J) The ubiquitination of VOZ2-FLAG in VOZ2-FLAG and VOZ2-FLAG/Myc-FOF2 seedlings in response to FR light or R light. Seedlings were grown in cW for 10 days, treated with 50 μM MG132 for 5 h, and then exposed to FR light (30 μmol m⁻² s⁻¹) or R light (20 μmol m⁻² s⁻¹) for 24 h. Total proteins were extracted and incubated with anti-FLAG magnetic beads at 4°C overnight. The immunoprecipitates were analyzed via immunoblotting with anti-Myc antibody, anti-FLAG antibody, and anti-ubiquitin antibody (CST, 3936S). The anti-actin antibody was used as the loading control.

(K–M) Representative immunoblots showing protein levels of VOZ2-FLAG in VOZ2-FLAG/Col-0 (K), VOZ2-FLAG/phyB(L), and VOZ2-FLAG/Myc-FOF2/phyB(M) plants in response to FR light. Seedlings were grown on ¹/₂ MS solid medium under cW for 10 days and exposed to FR light (FR, 30 μmol m⁻² s⁻¹) for the indicated times. Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(N–P) Relative protein levels of VOZ2-FLAG in (K), (L), and (M). VOZ2-FLAG protein levels were normalized to HSP82 levels. The values of starting points were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (Student’s t-test).

Because both VOZ1 and VOZ2 were degraded by FOF2, we next used VOZ2 for immunoprecipitation (IP) assays in tobacco leaves to explore whether VOZ proteins are ubiquitinated by FOF2. As shown in Figure 2I, there were stronger ubiquitinated smear ladders of VOZ2-Myc protein in the samples co-expressing VOZ2-Myc and FOF2-FLAG proteins than in the samples expressing VOZ2-Myc alone (Figure 2I), indicating that FOF2 promoted VOZ2 ubiquitination. VOZ2 ubiquitination was also detected in the samples without FOF2 (Figure 2I), presumably owing to other E3 ligases such as BRUTUS (BTS) (Selote et al., 2018) and AVRPIZ-T INTERACTING PROTEIN 10 (APIP10) (Wang et al., 2021) or to some as-yet-unidentified E3 ligase(s) that mediate ubiquitination of VOZ2-Myc in tobacco.

To determine whether the effects of FOF2 on VOZ2 ubiquitination show FR light specificity, we performed IP assays using VOZ2-FLAG and VOZ2-FLAG/Myc-FOF2 seedlings. Intriguingly, unique VOZ2-FLAG polyubiquitinations were markedly promoted in VOZ2-FLAG/Myc-FOF2 seedlings exposed to FR light, whereas polyubiquitination signals in the VOZ2-FLAG/Myc-FOF2 seedlings in response to R light showed almost no change compared with those in VOZ2-FLAG seedlings (Figure 2J). These results demonstrate that FOF2 is specifically involved in the FR light–induced polyubiquitination of VOZ2.

The photoreceptor PHYB is required for FOF2-mediated degradation of the VOZ2 protein under FR light

VOZ2 has been reported to degrade under FR light in a phytochrome-dependent manner (Yasui et al., 2012). To investigate the role of PHYB in the FOF2-mediated degradation of VOZ2 under FR light, we introduced VOZ2-FLAG and VOZ2-FLAG/Myc-FOF2 transgenes into the phyB background via genetic crossing. Consistent with a previous study (Yasui et al., 2012), the VOZ2-FLAG fusion protein showed a much slower rate of degradation in the phyB mutant than in the wild-type background after 24 h of FR light irradiation (Figure 2K, 2L, 2N, and 2O; Supplemental Figure 8). Notably, Myc-FOF2 overexpression was barely capable of promoting the degradation of VOZ2-FLAG in the phyB background upon FR light exposure (Figure 2M and 2P; Supplemental Figure 8), indicating that PHYB is essential for FOF2-mediated degradation of the VOZ2 protein. Similarly, VOZ2 degradation by FOF2 in response to simulated shade was partially dependent on PHYB (Supplemental Figure 9B and 9D). Nevertheless, the protein levels of VOZ2-FLAG in both VOZ2-FLAG/phyB and VOZ2-FLAG/Myc-FOF2/phyB plants declined gradually after 24 h of FR light exposure (Figure 2L, 2M, 2O, and 2P; Supplemental Figure 8), suggesting that other phytochromes may also perform redundant functions. Together, these data indicate that FOF2-mediated degradation of VOZ2 under FR light requires PHYB.

PHYB interacts with FOF2 under both FR light and R light

Because the FOF2-mediated degradation of VOZ2 requires PHYB, we next assessed whether PHYB physically interacts with FOF2 using a BiFC assay in tobacco leaves. Obvious YFP fluorescence was observed in the nucleus of leaves co-expressing PHYB-nYFP and cCFP-FOF2 but not in those co-expressing PHYB-nYFP and cCFP or nYFP and cCFP-FOF2 (Figure 3A), indicating that PHYB interacts with FOF2 in plant cells. To examine whether the FOF2–PHYB interaction was light dependent, we first performed a semi-in vivo pull-down assay. Surprisingly, purified recombinant His-TF-FOF2 pulled down PHYB-GFP from PHYB-GFP transgenic seedlings both in the dark and under FR light or R light (Figure 3B). CoIP assays using Myc-FOF2/PHYB-GFP double-transgenic seedlings further demonstrated that Myc-FOF2 was co-immunoprecipitated by PHYB-GFP from seedlings irradiated with FR light or R light (Figure 3C). However, FOF2 did not interact with PHYA under either R light or FR light (Supplemental Figure 11). We next detected the localization of PHYB-GFP protein and the PHYB–FOF2 interaction when WL-grown seedlings were exposed to R light or FR light. Immunoblots showed that PHYB-GFP was located in both the nucleus and cytoplasm in response to R light or FR light, although more nuclear PHYB-GFP protein was observed in R light-treated seedlings than in FR light-treated seedlings (Supplemental Figure 12). CoIP analysis showed that PHYB interacts with FOF2 in the nuclear fraction under both R light and FR light (Figure 3D). Similar evidence has been reported showing that the PHYB–SPA1 interaction occurs in the nucleus under both R light and FR light (Zheng et al., 2013). Taken together, these results indicate that PHYB is physically associated with FOF2 in the nucleus under both FR light and R light.

Figure 3.

PHYB interacts with FOF2 in response to FR light and R light.

(A) BiFC assay showing the interaction between PHYB and FOF2. cCFP-FOF2 and PHYB-nYFP were co-transformed into tobacco leaves. Bright, bright field. Merge, overlay of the YFP and bright-field images. Scale bars correspond to 50 μm.

(B) Semi-in vivo pull-down assay showing the interaction between PHYB and FOF2 in response to FR light and R light. PHYB-GFP seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, transferred to dark conditions for 2 days, kept in darkness (D), or exposed to FR light or R light for 24 h. E. coli–purified His-TF-FOF2 proteins were mixed with total proteins extracted from PHYB-GFP seedlings following incubation with Ni-NTA agarose. The eluted proteins were separated on an SDS–PAGE gel and probed with anti-His and anti-GFP antibodies.

(C) CoIP assay showing interaction between PHYB and FOF2 in response to FR light and R light. Myc-FOF2 and Myc-FOF2/PHYB-GFP seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, treated with 50 μM MG132 for 5 h, and exposed to FR light or R light for 24 h. Total proteins (input) or IP products of anti-GFP (IP (GFP)) antibody in immunoblots were probed with anti-Myc and anti-GFP antibodies.

(D) CoIP assay showing that PHYB interacts with FOF2 in the nuclear fraction under both R light and FR light. Myc-FOF2 and Myc-FOF2/PHYB-GFP seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, treated with 50 μM MG132 for 5 h, and exposed to FR light or R light for 12 h. Crude extracts (input) or IP products of anti-GFP (IP (GFP)) antibody in immunoblots were probed with anti-GFP and anti-Myc antibodies. Histone H3 was used as a nuclear marker and UGPase (UDP-glucose pyrophosphorylase) as a cytoplasmic marker.

(E) Semi-in vivo pull-down assay showing the interaction of PHYB-NT with FOF2. Myc-PHYB-NT, Myc-PHYB-CT, and Myc-GUS seedlings were grown on ¹/₂ MS solid medium under cW for 10 days and sampled. E. coli–purified His-TF-FOF2 proteins were mixed with total proteins extracted from seedlings following incubation with Ni-NTA agarose. The eluted proteins were separated on an SDS–PAGE gel, stained with Coomassie brilliant blue (the input of His-TF-FOF2), or probed with anti-His (the IP of His-TF-FOF2) and anti-Myc antibodies.

(F–I) Genetic interaction between PHYB and FOF2 in flowering. Images of the representative flowering phenotypes of 37-day-old plants (F) and 28-day-old plants (G) grown under LD in high R/FR light conditions are shown. The days to flowering and the number of rosette leaves on the day floral buds became visible were recorded for phyB mutant and Myc-FOF2-overexpressing plant combinations (H) and phyB and fof2fol1 mutant combinations (I). Standard deviations (n ≥ 20) are shown. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(J and K) mRNA levels of FLC and FT in phyB mutant and Myc-FOF2-overexpressing plant combinations (J) and phyB and fof2fol1 mutant combinations (K). Seedlings were grown on ¹/₂ MS solid medium under LD in high R/FR light conditions for 12 days and harvested at 4 h after being exposed to light. Gene expression was normalized to that of ACT2. Bars represent the standard deviations of three independent experiments. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

The PHYB C-terminal domain (PHYB-CT) is responsible for nuclear import, and the N-terminal domain (PHYB-NT) is functional in the nucleus (Yamaguchi et al., 1999; Kircher et al., 2002; Li et al., 2011). To examine which domain is responsible for the PHYB–FOF2 interaction, we performed semi-in vivo pull-down assays involving Myc-PHYB-NT and Myc-PHYB-CT transgenic seedlings. FOF2 pulled down the PHYB-NT protein but not the PHYB-CT protein or the β-glucuronidase (GUS) control protein (Figure 3E). Therefore, we concluded that PHYB physically interacts with FOF2 through its N-terminal domain.

It has been reported that PHYB is degraded by the E3 ubiquitin ligase CONSTITUTIVELY PHOTOMORPHOGENIC1 (COP1) and Light-Response Bric-a-Brack/Tramtrack/Broad (BTB) (LRB) E3 ubiquitin ligases in the nucleus (Jang et al., 2010; Ni et al., 2014). Therefore, we used an anti-PHYB antibody to examine whether FOF2 could also be involved in the regulation of PHYB abundance. There were no obvious differences in PHYB levels among the wild-type Col-0, the fof2fol1 mutant, and the Myc-FOF2-overexpressing plants (Supplemental Figure 13), indicating that FOF2 did not participate in regulation of PHYB protein abundance.

To determine the functional relationships between PHYB and FOF2, we examined their genetic interactions. We crossed Myc-FOF2-overexpressing plants and fof2fol1 mutants with phyB mutants and analyzed flowering time of the resulting plants under LD with high R/FR light conditions. Consistent with a previous study, the phyB mutants had an early-flowering phenotype (Reed et al., 1993), whereas the phyB/Myc-FOF2 plants had a delayed flowering time compared with that of the phyB plants (Figure 3F and 3H). The phyB/fof2fol1 triple-mutant plants did not flower earlier than the phyB single mutant (Figure 3G and 3I). These results suggest that FOF2 plays a negative role in flowering downstream of PHYB. Consistent with the flowering phenotype, expression levels of the flowering genes FLC and FT in the phyB/Myc-FOF2 mutant were similar to those in the Myc-FOF2-overexpressing plants (Figure 3J and Supplemental Figure 14A and 14B), and no additive effect on their expression was observed in phyB/fof2fol1 (Figure 3K and Supplemental Figure 14C and 14D). Thus, FOF2 overexpression largely rescued the low expression levels of FLC and the high expression levels of FT in the phyB mutant. Taken together, these results indicate that PHYB and FOF2 act in the same genetic pathway and that FOF2, at least in part, functions downstream of PHYB to promote FLC expression and inhibit flowering time under high R/FR light.

PHYB mediates FR-light stabilization of the FOF2 protein but mediates R-light degradation of the FOF2 protein

Given that PHYB interacts directly with FOF2 under both FR light and R light (Figure 3B and 3C), we first evaluated whether and how FR or R light and PHYB regulate FOF2 protein levels. Interestingly, the Myc-FOF2 protein was stable within 24 h of FR-light exposure but was degraded under R light in the Myc-FOF2 seedlings (Figure 4A–4D; Supplemental Figure 15), which may partially explain why VOZ2 protein degraded under FR light but was stable and accumulated under R light (Supplemental Figure 6A and 6B). Notably, Myc-FOF2 protein levels in the phyB/Myc-FOF2 seedlings decreased substantially in response to FR light but barely changed under R light (Figure 4A–4D; Supplemental Figure 15), indicating that PHYB is required for stabilization of FOF2 under FR light and degradation of FOF2 under R light. To exclude the effect of reduced total light density in the light changes from continuous WL to FR light or R light on protein stability, we examined Myc-FOF2 protein levels using WL to simulated shade condition (cW+ far-red light) or cW + red light (the same PAR with WL) (Supplemental Figure 4C–4F). As shown in Supplemental Figure 16, Myc-FOF2 was also stable under cW + FR light conditions and was degraded under cW + red light conditions, partially dependent on PHYB (Supplemental Figure 16). Moreover, FOF2 degradation was blocked by MG132 treatment in both the phyB/Myc-FOF2 mutant under FR light and the wild-type background under R light (Figure 4E–4H; Supplemental Figure 15). Collectively, these results suggest that PHYB mediates FR-light stabilization of FOF2 protein but mediates R-light degradation of FOF2 protein via the 26S proteasome pathway.

Figure 4.

PHYB mediates FR light stabilization of FOF2 protein and promotion of the FOF2–VOZ2 interaction but mediates R light degradation of FOF2 protein.

(A and B) Representative immunoblots showing protein levels of Myc-FOF2 protein in the phyB mutant in response to FR light (A) and R light (B). Myc-FOF2 and phyB/Myc-FOF2 seedlings were grown on ¹/₂ MS solid medium under cW for 10 days and then exposed to FR light (30 μmol m⁻² s⁻¹) or R light (R, 20 μmol m⁻² s⁻¹) for the indicated times. Protein levels were analyzed by immunoblotting with anti-Myc antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(C and D) Relative protein levels of Myc-FOF2 in (A) and (B). Myc-FOF2 protein levels were normalized to HSP82 levels. The values of starting points were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗p < 0.05, ∗∗∗p < 0.001 (Student’s t-test).

(E) FOF2 protein was degraded in the phyB mutant background in response to FR light via the 26S proteasome. phyB/Myc-FOF2 seedlings were grown on ¹/₂ MS solid medium under cW for 10 days, treated with 50 μM MG132 (+MG132) or DMSO (–MG132) for 5 h, and then exposed to FR light (FR, 30 μmol m⁻² s⁻¹) for the indicated times. Protein levels were analyzed by immunoblotting with anti-Myc antibody. The anti-HSP82 antibody was used as a loading control. Each assay was biologically replicated three times.

(F) FOF2 protein was degraded in response to R light via the 26S proteasome. Myc-FOF2 seedlings were grown on ¹/₂ MS solid medium under cW for 10 days, treated with 50 μM MG132 (+MG132) or DMSO (–MG132) for 5 h, and then exposed to R light (R, 20 μmol m⁻² s⁻¹) for the indicated times. Protein levels were analyzed by immunoblotting with anti-Myc antibody. The anti-HSP82 antibody was used as a loading control. Each assay was biologically replicated three times.

(G and H) Relative protein levels of Myc-FOF2 in (E) and (F). Myc-FOF2 protein levels were normalized to HSP82 levels. The values of starting points were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗∗p < 0.01, ∗∗∗p < 0.001 (Student’s t-test).

(I) Semi-in vivo coIP assay showing FR light promotion of the FOF2–VOZ2 interaction. Myc-FOF2 transgenic seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, treated with 50 μM MG132 for 5 h, and then kept in WL (cW) or exposed to FR light or R light for 24 h. Total proteins extracted from Myc-FOF2 transgenic seedlings were mixed with E. coli–purified GST-VOZ2 or GST proteins following incubation with anti-Myc magnetic beads. The eluted proteins were separated on an SDS–PAGE gel and probed with anti-GST and anti-Myc antibodies.

(J) CoIP assay showing FR-light promotion of the FOF2–VOZ2 interaction. VOZ2-FLAG/Myc-FOF2 seedlings were grown on ¹/₂ MS solid medium under cW for 8 days. Seedlings were treated with 50 μM MG132 for 5 h, then exposed to FR light or R light for 24 h. Total proteins (input) or the IP product of anti-Myc (IP (Myc)) antibody in immunoblots were probed with anti-FLAG and anti-Myc antibodies.

(K) CoIP assay showing that PHYB, FOF2, and VOZ2 are present in the same complex in plants. Col-0, Myc-FOF2, and VOZ2-FLAG/Myc-FOF2 seedlings were grown on ¹/₂ MS plates under cW for 8 days. Seedlings were treated with 50 μM MG132 for 5 h and sampled for the coIP assay. The total proteins (input) or the IP product of anti-FLAG (IP (FLAG)) antibody in immunoblots were probed with anti-FLAG, anti-Myc, and anti-PHYB antibodies.

(L) Semi-in vivo pull-down assay showing PHYB promotion of the FOF2–VOZ2 interaction in response to FR light. PHYB-GFP transgenic seedlings and phyB mutant seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, transferred to dark conditions for 2 days, exposed to FR light or R light for 24 h, and sampled for total protein extraction. E. coli–purified His-TF-FOF2 protein and GST-VOZ2 protein were mixed with total proteins extracted from PHYB-GFP or phyB seedlings following incubation with Ni-NTA agarose. The eluted proteins were separated on an SDS–PAGE gel and probed with anti-GST, anti-His, and anti-GFP antibodies.

(M) CoIP assay showing that FR light promotion of the FOF2‒VOZ2 interaction is partially dependent on PHYB. VOZ2-FLAG/Myc-FOF2 and VOZ2-FLAG/Myc-FOF2/phyB seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, treated with 50 μM MG132 for 5 h, and then exposed to FR light for 24 h. Total proteins (input) or the IP product of anti-Myc (IP (Myc)) antibody in immunoblots were probed with anti-Myc and anti-FLAG antibodies.

(N) PHYB mediates FR light promotion of the FOF2–VOZ2 interaction in the nuclear fraction. VOZ2-FLAG/Myc-FOF2 and VOZ2-FLAG/Myc-FOF2/phyB seedlings were grown on ¹/₂ MS solid medium under cW for 8 days, treated with 50 μM MG132 for 5 h, and then exposed to FR light for 24 h and sampled for the coIP assay. Crude extracts (input) or the IP products of anti-Myc (IP (Myc)) antibody in immunoblots were probed with anti-Myc and anti-FLAG antibodies. Histone H3 was used as a nuclear marker and UDP-glucose pyrophosphorylase (UGPase) as a cytoplasmic marker.

PHYB mediates FR-light promotion of FOF2 binding to VOZ2

Light-induced degradation of VOZ2 by FOF2 can be achieved either by elevating FOF2 levels or by enhancing FOF2 activity toward VOZ. Because FR light exposure did not obviously alter FOF2 protein levels (Figure 4A and 4C), we investigated whether FR light affects FOF2 binding to VOZ2. We first evaluated this possibility using a semi-in vivo coIP assay. Total proteins extracted from Myc-FOF2 seedlings kept in WL or exposed to FR or R light were incubated with Escherichia coli–purified GST-VOZ2 protein. As shown in Figure 4I, substantially more VOZ2 protein was co-precipitated with FOF2 in the Myc-FOF2 plants treated with FR light, but this was not obvious for R-light-treated samples compared with samples kept in WL (Figure 4I). We also used VOZ2-FLAG/Myc-FOF2 double-transgenic lines to perform coIP assays, which showed that FOF2 co-precipitated more VOZ2 in the FR-light-treated seedlings than in the R light-treated seedlings (Figure 4J). These data indicate that FR light treatment enhances the association of FOF2 with VOZ2 in plants.

As PHYB interacts with both VOZ2 (Yasui et al., 2012; Supplemental Figure 17) and FOF2 (Figure 3A–3E), we investigated whether these proteins function in the same complex. For this, we used VOZ2-FLAG/Myc-FOF2 double-transgenic lines and blocked VOZ2 degradation by treatment with MG132. A coIP assay showed that VOZ2-FLAG, Myc-FOF2, and PHYB proteins could be immunoprecipitated by the anti-FLAG antibody at the same time (Figure 4K), suggesting the presence of a PHYB–FOF2–VOZ2 molecular complex in plants. Next, we performed a semi-in vivo pull-down assay to study whether PHYB affects FOF2 binding to VOZ2. Equal amounts of E. coli–purified GST-VOZ2 and His-TF-FOF2 proteins were incubated with cell extracts of PHYB-GFP seedlings or phyB mutant seedlings treated with FR or R light. FOF2 pulled down much more GST-VOZ2 in the presence of PHYB-GFP from the FR-light-treated seedlings than that from the R-light-treated seedlings (Figure 4L). However, there was no obvious difference in the abundance of GST-VOZ2 pulled down by FOF2 in the absence of PHYB from FR-light- and R-light-treated phyB seedlings (Figure 4L), indicating that PHYB plays a major role in the FR promotion of the FOF2–VOZ2 association. A coIP assay using VOZ2-FLAG/Myc-FOF2 and VOZ2-FLAG/Myc-FOF2/phyB seedlings confirmed the PHYB-dependent FR-light promotion of the FOF2–VOZ2 interaction in vivo (Figure 4M), which occurred in the nucleus (Figure 4N). Together, these data demonstrate that PHYB mediates FR-light promotion of FOF2 binding to VOZ2.

Regulation of flowering time and FLC expression by PHYB and FOF2 depends in part on VOZ proteins under high R/FR light

To determine whether the early-flowering phenotype of both the phyB and fof2fol1 mutants depends on VOZ genes under LD in high R/FR light, we generated the triple and quintuple mutants phyB/voz1voz2 and phyB/fof2fol1/voz1voz2 by crossing. This analysis confirmed that the triple mutant phyB/voz1voz2 exhibited a flowering time similar to that of voz1voz2 under LD in high R/FR light (Figure 5A and 5C), confirming epistasis of VOZ genes to PHYB (Yasui et al., 2012). Notably, like the fof2fol1/voz1voz2 (Figure 1J–1L) and phyB/voz1voz2 mutants (Figure 5A and 5C), the phyB/fof2fol1/voz1voz2 quintuple mutant showed a flowering phenotype similar to that of voz1voz2 (Figure 5B and 5D), although significant differences in leaf number and days to flowering were also observed between phyB/voz1voz2 or phyB/fof2fol1/voz1voz2 and voz1voz2 (Figure 5C and 5D). Consistent with their flowering phenotypes, the phyB/voz1voz2 and phyB/fof2fol1/voz1voz2 mutants showed higher FLC expression and lower FT expression than the wild-type, Col-0, and their expression was increased and was similar to that observed in voz1voz2 plants (Figure 5E and 5F and Supplemental Figure 18). However, a significant difference in FLC expression was still observed between phyB/voz1voz2 or phyB/fof2fol1/voz1voz2 and voz1voz2 (Figure 5E and 5F and Supplemental Figure 18). Collectively, these genetic results suggest that regulation of flowering and FLC expression by PHYB and FOF2 depends, at least in part, on VOZ proteins.

Figure 5.

PHYB and FOF2 regulate flowering and FLC expression in part via VOZ proteins under high R/FR light conditions.

(A and B) Images of 38-day-old phyB, voz1voz2, and phyB/voz1voz2(A) and 48-day-old phyB, fof2fol1, voz1voz2, and phyB/fof2fol1/voz1voz2(B) mutant plants grown under LD in high R/FR light conditions.

(C and D) Days to flowering and number of rosette leaves on the day floral buds became visible in phyB, voz1voz2, and phyB/voz1voz2(C) and phyB, fof2fol1, voz1voz2, and phyB/fof2fol1/voz1voz2(D) mutant plants grown under LD in high R/FR light conditions. Standard deviations (n ≥ 20) are shown. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(E and F) mRNA levels of FLC and FT in phyB, voz1voz2, and phyB/voz1voz2 mutant (E) and phyB, fof2fol1, voz1voz2, and phyB/fof2fol1/voz1voz2 mutant (F) plants. Seedlings were grown on ¹/₂ MS solid medium under LD in high R/FR light conditions for 12 days and harvested at 4 h after being exposed to light. Gene expression was normalized to that of ACT2. Bars represent the standard deviations of three independent experiments. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

FOF2 functions downstream of PHYB to regulate FLC expression and flowering time under simulated shade, partially dependent on VOZ proteins

Because PHYB is the principal photoreceptor for perception of the R/FR ratio signal and plays a major role in shade-induced flowering (Cerdan and Chory, 2003; Franklin, 2008; Casal, 2013; Martinez-Garcia et al., 2014; Pierik and de Wit, 2014), we wondered whether PHYB, FOF2, and VOZ proteins could act together in the regulation of flowering time in response to changes in light quality. We grew plants under LD in WL (high R/FR ratio) conditions supplemented with 15 μmol m⁻² s⁻¹ FR light for 30 min at the end of the photoperiod (EOD-FR) to mimic the effects of shade, as described previously (Xie et al., 2020). The flowering response to EOD-FR was determined by calculating the percentage reduction in leaf number (change in leaf number under EOD-FR/leaf number under WL) and days to flowering (change in days to flowering under EOD-FR/days to flowering under WL). This simulated shade treatment accelerated plant flowering. Compared with those of plants grown under WL conditions, the rosette leaf number and days to flowering of wild-type Col-0, Myc-FOF2, and voz1voz2 plants decreased significantly under EOD-FR (Figure 6A–6F; Supplemental Figure 19). Interestingly, Myc-FOF2 and voz1voz2 showed greater reduction in rosette leaf number (∼49.77% and 42.90%, respectively) than Col-0 (∼28.49%), but fof2fol1 showed less reduction (∼17.78%) (Figure 6E and 6F). These data indicate that the Myc-FOF2-overexpressing plants and the voz1voz2 mutant were more sensitive than the wild-type Col-0 plants to EOD-FR treatment, whereas the fof2fol1 mutant was less sensitive. Together, these results indicate that FOF2 and VOZ proteins play negative and positive roles, respectively, in flowering regulation under simulated shade.

Figure 6.

FOF2 functions downstream of PHYB to regulate flowering and FLC expression under simulated shade, dependent in part on VOZ proteins.

(A and B) Images of representative flowering phenotypes of phyB, Myc-FOF2, and phyB/Myc-FOF2(A) and phyB, fof2fol1, and phyB/fof2fol1(B) mutant plants grown under LD in WL (high R/FR ratio) conditions for 24 days or simulated shade (EOD-FR) conditions for 22 days. Plants were grown for 7 days under WL and either kept in WL or shifted to EOD-FR on day 8 until the onset of flowering.

(C and D) Images of representative flowering phenotypes of phyB, voz1voz2, and phyB/voz1voz2(C) and phyB, fof2fol1, voz1voz2, and phyB/fof2fol1/voz1voz2(D) mutant plants grown under LD in WL conditions or EOD-FR conditions for 28 days. Plants were grown for 7 days under WL and either kept in WL or shifted to EOD-FR on day 8 until the onset of flowering.

(E) Quantitative analysis of rosette leaf number of the plants in (A) and (B). Light-gray and dark-gray bars indicate the rosette leaf number of plants grown under LD in WL conditions and EOD-FR conditions, respectively. Black bars indicate the percentage reduction in leaf number between WL and EOD-FR. Values shown are mean ± SD (n = 4). Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(F) Quantitative analysis of rosette leaf number of the plants in (C) and (D). Light-gray and dark-gray bars indicate the rosette leaf number of plants grown under LD in WL conditions and EOD-FR conditions, respectively. Black bars indicate the percentage reduction in leaf number between WL and EOD-FR. Values shown are mean ± SD (n = 4). Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(G) mRNA levels of FLC and FT in phyB, Myc-FOF2, fof2fol1, phyB/Myc-FOF2, and phyB/fof2fol1 mutant plants in response to EOD-FR treatment. Seedlings were grown on ¹/₂ MS solid medium under LD in WL conditions for 4 days and then shifted to EOD-FR for 7 days and sampled at the end of the EOD-FR treatment for RNA analysis. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

(H) mRNA levels of FLC and FT in phyB, voz1voz2, fof2fol1, phyB/voz1voz2, and phyB/fof2fol1/voz1voz2 mutants in response to EOD-FR treatment. Seedlings were grown on ¹/₂ MS solid medium under LD in WL conditions for 4 days and then shifted to EOD-FR for 7 days and sampled at the end of the EOD-FR treatment for RNA analysis. Letters indicate significant differences by one-way ANOVA with Duncan’s post hoc test (p < 0.05).

Consistent with previous reports (Goto et al., 1991; Nagatani et al., 1991; Devlin et al., 1996), seedlings of the phyB mutant flowered slightly earlier upon EOD-FR treatment, and the reduction in leaf number was only ∼11.58% (compared with ∼28.49% in Col-0) (Figure 6A–6F; Supplemental Figure 19). The sensitivity of the phyB/Myc-FOF2 (Figure 6A and 6E; Supplemental Figure 19A) and phyB/voz1voz2 (Figure 6C and 6F; Supplemental Figure 19B) mutants to the EOD-FR treatment was markedly increased compared with that of the phyB single mutant. Indeed, 35.54% and 45.10% reductions in leaf number were observed in phyB/Myc-FOF2 and phyB/voz1voz2 (compared with ∼11.58% in phyB) (Figure 6E and 6F), respectively, indicating that FOF2 and VOZ proteins function downstream of PHYB to negatively and positively regulate flowering, respectively, under EOD-FR. In addition, reductions in leaf number and days to flowering in the phyB/fof2fol1 mutant (16.83% and 23.26%, respectively) were similar to those in the fof2fol1 mutant (17.78% and 28.99%, respectively) (Figure 6E; Supplemental Figure 19A), further supporting the role of FOF2 in the PHYB pathway. Notably, the reduction in leaf number and days to flowering increased in phyB/fof2fol1/voz1voz2 and was similar to that in the voz1voz2 mutant (Figure 6F; Supplemental Figure 19B), suggesting that FOF2 functions downstream of PHYB to negatively regulate simulated shade-induced flowering, partially dependent on VOZ proteins. Consistent with these flowering phenotypes, the transcript levels of FLC and FT in phyB/Myc-FOF2 plants were similar to those in the Myc-FOF2 plants (Figure 6G). The expression levels of FLC and FT in phyB/fof2fol1 and phyB/fof2fol1/voz1voz2 mutants were similar to those in the fof2fol1 (Figure 6G) and voz1voz2 mutants (Figure 6H), respectively. Surprisingly, similar to but unlike a previous study in which low R/FR did not lower FLC expression in Arabidopsis (Wollenberg et al., 2008), our RT–qPCR analysis showed that the wild-type Col-0 plants grown under EOD-FR had significantly upregulated levels of FLC transcripts compared with their counterparts grown under WL conditions, although FT expression was significantly upregulated (Supplemental Figure 20). This discrepancy is likely due to differences in experimental conditions, including light conditions, incubators, and soil mixes. Our results also verified that seedlings grown under EOD-FR had significantly upregulated levels of FRUITFUL (FUL), LEAFY (LFY), and APETALA1 (AP1) transcripts (Xie et al., 2020) (Supplemental Figure 20). Overall, these results suggest that FOF2 functions downstream of PHYB to regulate FLC expression and flowering under simulated shade, a process that is partially dependent on VOZ proteins.

In addition to early flowering, the phyB mutant had shade-avoidance phenotypes such as longer petioles at the rosette stage; we therefore sought to determine whether FOF2 functions in the regulation of petiole extension. However, examination of petiole length revealed no obvious morphological differences between Myc-FOF2, fof2fol1, and wild-type Col-0 plants (Supplemental Figure 21). Moreover, the longer petioles of the phyB mutant were not affected by FOF2 (Supplemental Figure 21). The voz1 voz2 mutations have also been reported to not affect petiole elongation of the phyB mutant (Yasui et al., 2012). These results indicate that FOF2 and VOZ proteins likely affect shade avoidance in flowering time but not petiole extension.

PHYB mediates simulated shade stabilization of FOF2 and promotion of the FOF2–VOZ2 interaction and VOZ2 degradation by SCF^FOF2 E3 ligase

We evaluated whether and how EOD-FR treatment regulates FOF2 protein stability. Interestingly, as observed in response to FR light (Figure 4A and 4C), FOF2 protein accumulation did not show obvious changes, whereas the phyB mutation greatly decreased FOF2 protein abundance under EOD-FR (Figure 7A and 7C; Supplemental Figure 22). By contrast, FOF2 protein accumulation declined significantly after end-of-day red light (EOD-R) treatment, and the phyB mutation largely abolished EOD-R-triggered FOF2 degradation (Figure 7B and 7C; Supplemental Figure 22). EOD-FR-triggered degradation of FOF2 in the phyB mutant background and EOD-R-triggered degradation of FOF2 in the wild-type background were blocked by MG132 treatment (Supplemental Figures 23 and 24). Together, these data suggest that the FOF2 protein is stable under EOD-FR but unstable under EOD-R conditions, and both events partially require PHYB. We next assessed whether the EOD-FR treatment affected FOF2 binding to VOZ2 in plants using a semi-in vivo coIP assay. Much more GST-VOZ2 protein co-precipitated with FOF2 in the Myc-FOF2 seedlings upon EOD-FR treatment (Figure 7D). Moreover, PHYB-GFP from the EOD-FR-treated seedlings markedly enhanced the FOF2–VOZ2 interaction compared with that of the WL-treated seedlings, whereas the interaction of FOF2 with VOZ2 was attenuated in the absence of PHYB (Figure 7E). In accordance with these results, we observed that under EOD-FR, FOF2 overexpression enhanced, but fof2fol1 mutation reduced, the degradation of VOZ2 protein, which was blocked by MG132 treatment (Figure 7F–7K; Supplemental Figure 22). This result indicated that FOF2 is responsible for the 26S proteasomal degradation of VOZ2. We then examined the effect of FOF2 on VOZ1 protein levels under EOD-FR conditions. VOZ1-FLAG protein degraded more quickly in the Myc-FOF2-overexpressing background than in the wild-type background, and the FOF2-triggered degradation of VOZ1-FLAG was inhibited by MG132 treatment (Supplemental Figure 25), indicating that FOF2 also mediates the degradation of VOZ1 via the 26S proteasome. Importantly, the phyB mutation largely abolished the effect of FOF2 on VOZ2 degradation in response to EOD-FR treatment (Figure 7L and 7M; Supplemental Figure 22), suggesting that PHYB is necessary for FOF2-mediated degradation of VOZ2 protein under simulated shade. Collectively, these results demonstrate that simulated shade induces stabilization of the FOF2 protein, promoting FOF2 binding to the VOZ2 protein and VOZ2 degradation by SCF^FOF2 E3 ligase and that this process is at least partially dependent on PHYB.

Figure 7.

PHYB mediates EOD-FR stabilization of FOF2 and promotion of the FOF2–VOZ2 interaction and VOZ2 degradation by SCF^FOF2 E3 ligase.

(A and B) Representative immunoblots showing levels of Myc-FOF2 protein in the phyB mutant in response to EOD-FR (A) and EOD-R (B) treatment. Seven-day-old Myc-FOF2 and phyB/Myc-FOF2 seedlings grown under LD in WL (high R/FR ratio) conditions were treated with EOD-FR or EOD-R. Seedlings were harvested at the end of the EOD-FR or EOD-R treatment. Protein levels were analyzed by immunoblotting with anti-Myc antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(C) Relative protein levels of Myc-FOF2 in (A) and (B). Myc-FOF2 protein levels were normalized to HSP82 levels. The values of WL were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗∗∗p < 0.001 (Student’s t-test).

(D) Semi-in vivo coIP assay showing EOD-FR promotion of the FOF2–VOZ2 interaction. Seven-day-old seedlings grown under WL were treated with EOD-FR and sampled at the end of EOD-FR treatment. E. coli–purified GST-VOZ2 or GST proteins were mixed with total proteins extracted from Myc-FOF2 seedlings following incubation with anti-Myc magnetic beads. The eluted proteins were separated on an SDS–PAGE gel and probed with anti-GST and anti-Myc antibodies.

(E) Semi-in vivo coIP assay showing PHYB promotion of the FOF2–VOZ2 interaction in response to EOD-FR. Seven-day-old PHYB-GFP transgenic seedlings and phyB mutant seedlings grown under WL were treated with EOD-FR and sampled at the end of EOD-FR. E. coli–purified His-TF-FOF2 protein and GST-VOZ2 protein were mixed with total proteins extracted from PHYB-GFP or phyB seedlings following incubation with Ni-NTA agarose. The eluted proteins were separated on an SDS–PAGE gel and probed with anti-GST, anti-His, and anti-GFP antibodies.

(F) Representative immunoblots showing that FOF2 promotes VOZ2-FLAG protein degradation in response to EOD-FR. Seven-day-old seedlings grown under WL were treated with EOD-FR and sampled at the end of the EOD-FR treatment. Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(G and H) Representative immunoblots showing that FOF2 participates in VOZ2 degradation via the 26S proteasome in response to EOD-FR. Seven-day-old seedlings of VOZ2-FLAG/Col-0 (G) and VOZ2-FLAG/Myc-FOF2(H) were grown under LD in WL conditions, treated with 50 μM MG132 (+MG132) or DMSO (–MG132) for 5 h, and then exposed to FR light (15 μmol m⁻² s⁻¹) for 30 min at the end of the photoperiod (EOD-FR). Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(I–K) Relative protein levels of VOZ2-FLAG in (F), (G), and (H). VOZ2-FLAG protein levels were normalized to HSP82 levels. The values for WL (high R/FR ratio) were set to 1. Data are means ± SD (n = 3). Significant differences are indicated: ∗∗p < 0.01, ∗∗∗p < 0.001 (Student’s t-test).

(L) Representative immunoblots showing protein levels of VOZ2-FLAG in VOZ2-FLAG/phyB and VOZ2-FLAG/Myc-FOF2/phyB seedlings in response to EOD-FR. Seven-day-old seedlings grown under LD in WL conditions were treated with EOD-FR and sampled at the end of the EOD-FR treatment. Protein levels were analyzed by immunoblotting with anti-FLAG antibody. The anti-HSP82 antibody was used as the loading control. Each assay was biologically replicated three times.

(M) Relative protein levels of VOZ2-FLAG in (L). VOZ2-FLAG protein levels were normalized to HSP82 levels. The values for WL were set to 1. Data are means ± SD (n = 3). The significant difference is indicated: ∗∗p < 0.01 (Student’s t-test).

(N) Proposed model describing how the PHYB–FOF2–VOZ2 module fine-tunes flowering in response to changes in light quality. Under WL conditions (high R/FR ratio), PHYB inhibits CO expression, resulting in deactivation of FT expression (Cerdan and Chory, 2003; Valverde et al., 2004; Inigo et al., 2012); however, PHYB destabilizes FOF2, which is downregulated by FCA (He et al., 2017), thereby attenuating FOF2-mediated VOZ2 degradation, which is required for CO function (Kumar et al., 2018). This in turn enhances CO activity, decreases FLC levels, and accordingly ensures sufficient FT expression and prevents late flowering. Under shade conditions (EOD-FR), PHYB-mediated inhibition of CO is lifted, resulting in activation of FT expression (Cerdan and Chory, 2003; Franklin, 2008; Kim et al., 2008; Wollenberg et al., 2008); however, PHYB stabilizes FOF2 protein and promotes binding of FOF2 to VOZ2, thereby enhancing VOZ2 degradation by FOF2. This in turn attenuates CO activity and increases FLC level, ensuring moderate FT expression and preventing early flowering .

Discussion

The R/FR light receptor PHYB regulates many light responses, from seed germination and shade avoidance to flowering (Smith, 2000). Only a few components involved in the PHYB-regulated flowering pathway have been identified by molecular and genetic investigations. Flowering regulators such as CO (Valverde et al., 2004), PFT1 (Cerdan and Chory, 2003), VOZ1, VOZ2 (Yasui et al., 2012), PHL (Endo et al., 2013), HOS1 (Lazaro et al., 2015), TZP (Kaiserli et al., 2015), PIF4, PIF5, and PIF7 (Galvao et al., 2019; Zhang et al., 2019) act downstream of PHYB to regulate flowering. Our group previously showed that FOF2 is an autonomous pathway–related regulator and inhibits flowering by promoting FLC expression (He et al., 2017). Here, we show that FOF2 is a novel component in PHYB signaling and regulates FLC expression and shade-accelerated flowering, in part via VOZ proteins. This notion is supported by several lines of evidence: (1) FOF2 is a new PHYB-interacting factor and acts downstream of PHYB to promote FLC expression and inhibit flowering under both high R/FR light and simulated shade conditions (Figures 3 and 6); (2) FOF2 directly interacts with VOZ proteins and functions upstream of VOZ proteins to promote FLC expression and inhibit flowering (Figure 1); (3) regulation of flowering and FLC expression by FOF2 and PHYB under both high R/FR light and simulated shade conditions are partially dependent on VOZ proteins (Figures 5 and 6); (4) in response to FR light and EOD-FR treatments, FOF2 was stable but VOZ2 degraded; in response to R light and EOD-R treatments, FOF2 was degraded but VOZ2 was stable, which at least in part required PHYB (Figures 2, 4, and 7). FOF2 therefore acts as a new regulator in PHYB signaling to modulate FLC expression and flowering time in response to changes in light quality.

PHYB mediates R-light regulation of the stability of key regulators such as AUX/IAAs (Xu et al., 2018), PIF3 (Xu et al., 2019), EIN3 (Shi et al., 2016), and CO (Lazaro et al., 2015), thereby regulating photomorphogenesis or flowering time. In addition, PHYB is involved in the degradation of VOZ2 protein induced by FR light (Yasui et al., 2012); however, the underlying mechanism is not clear. In this study, we showed that VOZ2 interacts with FOF2 (Figure 1B, 1E and 1F), a subunit of the SCF E3 complex (Gagne et al., 2002; He et al., 2017). Therefore, we reasoned that SCF^FOF2 E3 ligase might mediate VOZ2 protein degradation. As expected, we observed that the VOZ2 degradation induced by FR light was significantly reduced in the fof2fol1 mutant and enhanced in FOF2-overexpressing plants (Figure 2A, 2C, 2E, and 2G). Moreover, FOF2-mediated VOZ2 degradation was inhibited by the proteasome inhibitor MG132 (Figure 2B, 2D, 2F, and 2H). Intriguingly, FOF2 promoted VOZ2 protein polyubiquitination under FR light but not R light (Figure 2J). Similarly, FOF2 mediated 26S proteasomal degradation of VOZ2 under EOD-FR (Figure 7F–7K). Importantly, FOF2-mediated VOZ2 degradation under FR light and EOD-FR was partially dependent on PHYB (Figures 2K–2P and 7L and 7M). Therefore, we conclude that PHYB mediates degradation of VOZ2 under FR light and EOD-FR, in part via SCF^FOF2 E3 ligase.

One important question is how PHYB functions in the FOF2-mediated degradation of VOZ2 under both FR light and EOD-FR. Recently, PHYB was reported to enhance the binding of EIN3 to its E3 ligases EBF1/EBF2, resulting in EIN3 degradation mediated by SCF^EBF1/EBF2 under R light (Shi et al., 2016). Here, we found that PHYB enhanced FOF2 binding to the VOZ2 protein in response to FR light or EOD-FR treatment (Figures 4L–4N and 7E). Phytochromes have been reported to exist in two distinct forms in vivo, Pr and Pfr, and the Pr-to-Pfr conversion is reversible. Upon the absorption peak of Pfr in FR light, Pfr-to-Pr conversion dominates over Pr-to-Pfr reversion under FR light (Li et al., 2011). Pr PHYB was observed to reduce the COP1–SPA1 interaction in yeast cells (Sheerin et al., 2015; Li and Hiltbrunner, 2021). We therefore proposed that Pr PHYB might promote the FOF2–VOZ2 interaction under FR light or shade. Also, we cannot exclude the possibility that PHYA or other phytochromes might be involved in the enhanced interaction of FOF2–VOZ2 under FR light or shade, as enhancement of the FOF2–VOZ2 interaction induced by FR light or shade is partially dependent on PHYB (Figures 4L–4N and 7E). However, these hypotheses remain to be tested directly. These results expand the common understanding of the mechanisms of action of plant PHYB, in which light perception significantly changes the affinity of PHYB for its interaction partners. In addition, we showed that PHYB mediates stabilization of the FOF2 protein under FR light and EOD-FR (Figures 4A, 4C, 7A, and 7C). This may also partially contribute to the enhanced FOF2–VOZ2 interaction in response to FR light and EOD-FR treatment. Interestingly, PHYB mediates degradation of the FOF2 protein under R-light and EOD-R (Figures 4B, 4D, 7B, and 7C). This may partially explain why the VOZ2 protein was stable upon exposure to R light (Supplemental Figure 6A and 6B). Another important question is how PHYB regulates the stability of FOF2 in response to changes in light quality. The Pfr:Pr ratio of PHYB has been reported to change in response to environmental light changes (Li et al., 2011); hence, we proposed that stability of the FOF2 protein might be rapidly regulated by the shift in the Pfr:Pr ratio of PHYB. Importantly, we found that PHYB regulates the stability of FOF2 via the 26S proteasome pathway under both R and FR light (Figure 4A–4H). Therefore, we hypothesized that PHYB might regulate the activity of an E3 ligase under different light conditions. Under R light, Pfr PHYB might activate the E3 ligase and/or promote binding of FOF2 to the E3, which would then promote FOF2 degradation; under FR light, Pr PHYB might suppress activity of the E3 ligase and/or inhibit binding of FOF2 to the E3, thus stabilizing the FOF2 protein, leading to efficient modulation of VOZ protein abundance and ensuring signal dependence for light. However, these hypotheses remain to be tested directly.

PHYB represses flowering by antagonizing CO at the transcriptional and post-translational levels under R light or high R/FR conditions (Cerdan and Chory, 2003; Valverde et al., 2004; Inigo et al., 2012). When plants are exposed to EOD-FR or low R/FR conditions, the phytochrome Pfr form is photoconverted to the Pr form, thus decreasing Pfr PHYB and increasing active CO, which in turn activates FT expression (Cerdan and Chory, 2003; Franklin, 2008; Kim et al., 2008; Wollenberg et al., 2008). Recently, VOZ2 was reported to interact with CO, and CO requires VOZ2 for flowering induction (Kumar et al., 2018). VOZ proteins also negatively regulate FLC expression and promote flowering (Yasui et al., 2012). Here, we found that FOF2 functions downstream of PHYB to regulate flowering, at least in part, via VOZ proteins under both high R/FR and EOD-FR conditions (Figures 5 and 6). Moreover, our results, together with a previous report (Wollenberg et al., 2008), showed that low R/FR or EOD-FR did not downregulate FLC expression (Supplemental Figure 20). Hence, FT mRNA levels appear to be determined by a balance of CO and FLC activity in response to changes in light quality, thus ensuring proper flowering. On the basis of previous studies and the present study, we propose the following model to illustrate the mechanism by which the PHYB–FOF2–VOZ2 module fine-tunes flowering in response to changes in light quality (Figure 7N). Under WL (high R/FR ratio), PHYB inhibits CO expression at both the transcriptional and post-transcriptional levels to decrease FT expression (Cerdan and Chory, 2003; Valverde et al., 2004; Inigo et al., 2012); however, PHYB destabilizes FOF2, which is downregulated by the autonomous pathway gene FCA (He et al., 2017), causing a reduction in SCF^FOF2-mediated VOZ2 degradation, which is required for CO function (Kumar et al., 2018), and subsequently enhancing CO activity, decreasing FLC level, and thus ensuring sufficient FT expression to prevent late flowering. Under shade conditions (EOD-FR), PHYB-mediated inhibition of CO is relieved, promoting CO accumulation and activating FT expression (Cerdan and Chory, 2003; Franklin, 2008; Kim et al., 2008; Wollenberg et al., 2008); however, PHYB stabilizes FOF2 and enhances FOF2 binding to VOZ2, thus enhancing VOZ2 ubiquitination and degradation by SCF^FOF2, resulting in attenuated CO activity and increased FLC levels to override CO-mediated promotion of FT expression and prevent early flowering . This could be explained by the fact that the activation signal must be cautiously restricted to a certain extent to prevent overactivation. Together, our findings provide new insights into the mechanism by which plants fine-tune flowering time through the PHYB–FOF2–VOZ2 module that modulates FLC expression in response to changes in light quality.

Materials and methods

Plant materials and growth conditions

The Arabidopsis used for all experiments was the Columbia (Col-0) ecotype. The mutants fof2fol1 (CR-fof2fol1-m1) (He et al., 2017), voz1voz2 (voz1-1voz2-1) (Yasui et al., 2012), and phyB (phyB-9) (Reed et al., 1993) and the transgenic lines overexpressing Myc-FOF2 (MycFOF2ox1) (He et al., 2017), VOZ2-FLAG (Luo et al., 2020), PHYB-GFP (Zheng et al., 2013), Myc-PHYB-NT, Myc-PHYB-CT, and Myc-GUS (Xu et al., 2019) have been described previously. The fof2fol1/voz1voz2, VOZ2-FLAG/fof2fol1, VOZ2-FLAG/phyB, and VOZ2-FLAG/Myc-FOF2 mutants were prepared by crossing fof2fol1 with voz1voz2 and crossing VOZ2-FLAG with fof2fol1. The phyB/fof2fol1, phyB/voz1voz2, phyB/fof2fol1/voz1voz2, phyB/Myc-FOF2, VOZ2-FLAG/Myc-FOF2/phyB, and Myc-FOF2/PHYB-GFP plants were prepared by crossing phyB with fof2fol1, voz1voz2, fof2fol1/voz1voz2, Myc-FOF2, or VOZ2-FLAG/Myc-FOF2 and crossing Myc-FOF2 with PHYB-GFP. Plants were grown on soil in a culture room at 23°C under LD (16-h light/8-h dark) in WL conditions (Supplemental Figure 4A and 4B; high R/FR ratio [R/FR ∼4.5]). The blue light, red light, and FR light sources used in this study were described in our previous study (Zhao et al., 2007), and cool white light-emitting diode (LED) lights were used as WL sources.

EOD-FR and EOD-R treatment

The EOD-FR treatment was performed as described previously (Xie et al., 2020). For RNA analysis, 4-day-old seedlings grown under LD in WL (high R/FR ratio) conditions were treated with 15 μmol m⁻² s⁻¹ FR light at the end of the light period (EOD-FR) for 30 min for 7 days and sampled for RNA assays. For protein analysis, 7-day-old seedlings were treated with EOD-FR for 30 min and sampled. For flowering-time determination, 7-day-old seedlings were treated with EOD-FR for 30 min every day until bolting. The EOD-R treatment was performed as described previously (Cerdan and Chory, 2003). Seven-day-old seedlings were treated with R light (20 μmol m⁻² s⁻¹) at the end of the light period (EOD-R) for 30 min and sampled for protein assays. All seedlings were grown under LD in WL (high R/FR ratio) conditions, with or without EOD-FR/R treatment.

Plasmid construction and plant transformation

Plasmid construction and plant transformation were performed as described previously (Luo et al., 2020; Yan et al., 2020), and details are provided in the supplemental methods. The VOZ1-FLAG plasmid was then transformed into wild-type Col-0 by the floral dip method (Clough and Bent, 1998). T₀ seeds were selected using 50 μg/ml hygromycin, and the hygromycin-resistant plants were confirmed by protein analysis. 35S:VOZ1-FLAG transgenic lines (termed VOZ1-FLAG) that exhibited VOZ1-FLAG protein expression (Supplemental Figure 26) were selected for further analysis. Primers are listed in Supplemental Table 2.

Y2H assay

Yeast transformation was performed as described in the Yeast Protocols Handbook (Clontech). The full-length cDNA of FOF2 or VOZ1/2, or the FOF2ΔF fragment, was cloned into both the pGADT7 and pGBKT7 vectors. The fragment of VOZ2a or VOZ2b was cloned into the pGADT7 vector. Combinations of AD-fusion and BD-fusion plasmids were co-transformed into the AH109 yeast cells. Yeast transformants were then grown on synthetic complete minimal medium without Trp and Leu (SD−TL) or without Trp, Leu, and His (SD−TLH) containing an appropriate amount of 3-amino-1,2,4-triazole (3-AT) for yeast selection.

mRNA and protein analysis

The mRNA and protein analyses were performed as described previously (Zhong et al., 2021). Details of the analyses are described in the supplemental methods.

BiFC assay

The BiFC assay was performed as described previously (Walter et al., 2004; Zhong et al., 2021). Details of the analysis are provided in the supplemental methods.

CoIP assay

For FOF2–VOZ1 and FOF2–VOZ2 interactions and the PHYB–FOF2–VOZ2 complex, 8-day-old seedlings of VOZ1-FLAG, VOZ2-FLAG, VOZ1-FLAG/Myc-FOF, VOZ2-FLAG/Myc-FOF2, Col-0, and Myc-FOF2 were treated with 50 μM MG132 (Selleck, S2619) for 5 h and sampled. For the FOF2–PHYB interaction and the effects of FR light on FOF2–VOZ2 interactions, 8-day-old seedlings of Myc-FOF2, Myc-FOF2/PHYB-GFP, and VOZ2-FLAG/Myc-FOF2 were treated with 50 μM MG132 for 5 h and then exposed to FR light (30 μmol m⁻² s⁻¹) or R light (20 μmol m⁻² s⁻¹) for 24 h and sampled. The coIP assay was performed as described previously (Yan et al., 2020). All seedlings were grown on ¹/₂ MS solid medium under continuous WL. In brief, equal amounts of total protein for each sample were incubated with anti-Myc magnetic beads (Bimake; lot 820023), anti-FLAG magnetic beads (Bimake; lot 580031), or anti-GFP magnetic beads (KT Health; # KTSM1334) at 4°C. The samples were analyzed by immunoblotting with anti-FLAG (Abmart), anti-GFP (Abmart), anti-Myc (Abmart), and anti-PHYB (PhytoAB) antibodies. The CoIP assay for the locations of the PHYB–FOF2 interaction is described in the supplemental methods.

Semi-in vivo coIP assay

To analyze the effect of FR light and R light on VOZ2 interactions by semi-in vivo coIP, Myc-FOF2 proteins were individually extracted from 8-day-old Myc-FOF2 seedlings that were treated with 50 μM MG132 for 5 h and then kept in WL or exposed to FR light (30 μmol m⁻² s⁻¹) or R light (20 μmol m⁻² s⁻¹) for 24 h. Equal amounts of GST-VOZ2 and GST protein were incubated with Myc-FOF2 protein extracts and immunoprecipitated with anti-Myc magnetic beads.

Semi-in vivo pull-down assay

The semi-in vivo pull-down assays for the FOF2–VOZ1, FOF2–VOZ2, and PHYB–FOF2 interactions and the effect of PHYB on the FOF2–VOZ2 interaction were performed as described in our previous study (Zhong et al., 2021), and details of the analysis are provided in the supplemental methods.

In vitro GST pull-down assay

The in vitro GST pull-down assay was performed for the FOF2–VOZ1 and FOF2–VOZ2 interactions. Details of the analysis are provided in the supplemental methods.

Cell-free protein degradation assay

The cell-free protein degradation analysis was performed as described previously (Wang et al., 2009; Yan et al., 2020). Details of the analysis are provided in the supplemental methods.

Ubiquitination assay

The ubiquitination assay using tobacco was performed as described previously (Liu et al., 2010; Yan et al., 2020). The VOZ2-Myc constructs were infiltrated alone or co-infiltrated with FOF2-FLAG into tobacco leaves via Agrobacterium. The leaves were treated with 50 μM MG132 for 12 h before sampling. The ubiquitination assay using Arabidopsis was performed as described previously (Xu et al., 2021). Total protein was extracted from 10-day-old seedlings of VOZ2-FLAG and VOZ2-FLAG/Myc-FOF2 that were grown under WL, treated with 50 μM MG132 for 5 h, and then exposed to FR light (30 μmol m⁻² s⁻¹) or R light (20 μmol m⁻² s⁻¹) for 24 h. The plant proteins were incubated with anti-Myc or anti-FLAG magnetic beads at 4°C overnight. The samples were analyzed by immunoblotting with anti-Myc, anti-FLAG, anti-ubiquitin (Enzo Life Sciences; BML-PW8810-0100), and anti-ubiquitin (CST, 3936S) antibodies. Ponceau staining or the anti-actin antibody was used as a loading control.

Funding

This work was supported by the National Natural Science Foundation of China (32170252, U20A2029), the Natural Science Foundation of Guangdong Province (2022A1515010968), the Natural Science Foundation of Hunan Province (2021JJ30097, 2022JJ30127), and the Natural Science Foundation of Changsha City (kq2202150).

Author contributions

L.Q., M.Z., F.D., X. Liu, and X.Z. designed the research. L.Q., M.Z., F.D., X. Li, J.Y., Q.Z., R.H., and D.T. performed the experiments. L.Q., M.Z., F.D., X. Liu, and X.Z. analyzed data. L.Q., M.Z., X. Liu, and X.Z. wrote the paper. All authors discussed the results and contributed to the article.

Published: April 14, 2024