VASCULAR PLANT ONE-ZINC FINGER1 (VOZ1) and VOZ2 Interact with CONSTANS and Promote Photoperiodic Flowering Transition

Abstract

In plants, endogenous and environmental signals such as light control the timing of the transition to flowering. Two phytochrome B-interacting transcription factors, VASCULAR PLANT ONE-ZINC FINGER1 (VOZ1) and VOZ2, redundantly promote flowering in Arabidopsis (Arabidopsis thaliana). In the voz1 voz2 mutant, the expression of FLOWERING LOCUS C (FLC) was up-regulated and that of FLOWERING LOCUS T (FT) was down-regulated, which was proposed to be the cause of late flowering in voz1 voz2 However, the detailed mechanism by which the VOZ genes promote flowering is not well understood. Here, we show that neither the reduced FT expression nor the late-flowering phenotype of voz1 voz2 is suppressed in the voz1 voz2 flc triple mutant. Genetic interaction experiments between voz1 voz2 and constans-2 (co-2) mutants reveal that the VOZs and CO work in the same genetic pathway. Using in vitro pull-down, electrophoretic mobility shift, and bimolecular fluorescence complementation assays, we show that VOZ1 and VOZ2 interact with CO. The voz1 voz2 35S::CO:YFP plants show suppression of the early-flowering phenotype induced by CO overexpression, suggesting that CO requires VOZ for the induction of flowering. Determination of the VOZ consensus-binding site followed by genome-wide sequence analysis failed to identify any VOZ-binding sites near known flowering time genes. Together, these results indicate that the VOZ genes regulate flowering primarily through the photoperiod pathway, independent of FLC, and suggest that VOZs modulate CO function to promote flowering.

Figure 1.

VOZ1 and VOZ2 redundantly promote flowering only under LDs. A, Schematic showing the T-DNA (black triangles) insertions in the VOZ1 and VOZ2 loci. Allele names are shown above the respective insertions. White rectangles represent 5′ and 3′ untranslated regions (UTRs), gray rectangles represent exons, and black lines represent introns. B, Five-week-old plants of the indicated genotypes grown under LDs. Bar = 1 cm. C, Number of rosette leaves at bolting of the indicated genotypes grown under LDs. Data are shown as means ± sd (n ≥ 16). D, Days from germination to flowering of Col-0 and voz1 voz2 plants grown under LDs. Data are shown as means ± sd (n = 28). E, Number of rosette leaves at bolting of Col-0 and voz1 voz2 plants grown under SDs. Data are shown as means ± sd (n = 21). Asterisks in C and D indicate significant differences from Col-0 (P < 0.001, unpaired Student’s t test).

Figure 2.

VOZ1 and VOZ2 regulate flowering independently of FLC. A and B, Relative transcript levels of FT (A) and FLC (B) analyzed by RT-quantitative PCR (qPCR). Asterisks indicate significant differences from Col-0 (P < 0.001, unpaired Student’s t test). C, Five-week-old plants of the indicated genotypes under LDs. Bar = 1 cm. D, Number of rosette leaves at bolting of the indicated genotypes under LDs. Data are shown as means ± sd (n = 15). Letters shared between the genotypes indicate no significant difference (P < 0.001, one-way ANOVA, Tukey’s multiple comparison test). E and F, Relative transcript levels of FLC (E) and FT (F) analyzed by RT-qPCR. Letters shared between the genotypes indicate no significant difference (P < 0.05, one-way ANOVA, Tukey’s multiple comparison test). For RT-qPCR, total RNA was isolated at zeitgeber time (ZT)-16 (A and F) and ZT-8 (B and E) from 14-d-old seedlings grown under LDs. RNA extraction was performed three times independently. The transcript levels were normalized to ACTIN2 (ACT2). Data are shown as means ± sd (n = 3).

Figure 3.

VOZs and CO function together in the photoperiod pathway to promote flowering. A, Relative CO transcript levels as analyzed by RT-qPCR. B, Relative VOZ1 and VOZ2 transcript levels as analyzed by RT-qPCR. The asterisk indicates a significant difference from Ler (P < 0.05, unpaired Student’s t test). C, Relative VOZ1 and VOZ2 transcript levels as analyzed by RT-qPCR. D, Five-week-old plants of the indicated genotypes grown under LDs. Bar = 1 cm. E, Number of rosette leaves at bolting of the indicated genotypes under LDs. Data are shown as means ± sd (n ≥ 15; P < 0.01, one-way ANOVA, Tukey’s multiple comparison test). F, Relative transcript levels of FT as analyzed by RT-qPCR. Letters shared between the genotypes indicate no significant difference (P < 0.05, one-way ANOVA, Tukey’s multiple comparison test). For RT-qPCR, total RNA was isolated at zeitgeber time (ZT)-12 (A), ZT-8 (B and C), and ZT-16 (F) from 14-d-old seedlings (A–C) and 11-d-old seedlings (F) grown under LDs. RNA extraction was performed three times independently. The transcript levels were normalized to ACT2. Data are shown as means ± sd (n = 3). WT, Wild type.

Figure 4.

VOZ1 and VOZ2 interact with CO in vitro. A and B, GST pull-down assay showing the interaction of VOZ1 and VOZ2 with CO. His-VOZ1 (A) or His-VOZ2 (B) proteins were incubated with GST or GST-CO together with glutathione-Sepharose beads in a column. The bead-bound proteins were subjected to 10% (w/v) SDS-PAGE and detected by immunoblot (IB) analysis using anti-His antibody. For the input blots, 0.5% (v/v) input extracts were loaded to detect His-VOZ1 or His-VOZ2. C, Electrophoretic mobility shift assay (EMSA) showing the interaction of VOZ1 and VOZ2 with CO. [γ-³²P]ATP-labeled wild-type (WT) probe containing the VOZ2-binding site (GCGTGTGATACACGT) was incubated with His-VOZ2 alone or in combination with GST-CO or GST. As controls, GST-CO and GST alone were incubated separately with the wild-type probe. Additionally, His-VOZ2 was incubated with a probe containing a mutant VOZ2-binding site (tCGTGTGATACACGT).

Figure 5.

VOZ1 and VOZ2 interact with CO in vivo. BiFC analysis shows the interaction of VOZ1 (top) and VOZ2 (bottom) proteins with CO in N. benthamiana epidermal cells. Red arrows indicate YFP signal, and white arrows indicate the nuclei stained with 4′,6-diamidino-2-phenylindole (DAPI). BF, Bright field; Merge, merge of YFP, BF, and DAPI. Bars = 20 µm.

Figure 6.

A, 15 representative unique sequences obtained for VOZ1 SELEX. B, 18 representative unique sequences obtained for VOZ2 SELEX. C, Consensus logo for VOZ1 (top) obtained from the software tool MEME and the count matrix (bottom) showing the occurrence of each base. D, Consensus logo for VOZ2 (top) obtained from the software tool MEME and the count matrix (bottom) showing the occurrence of each base.

Figure 7.

Genetic interaction between 35S::CO:YFP and voz1 voz2. A, Five-week-old plants of the indicated genotypes under LDs. B, Number of rosette leaves at bolting of the indicated genotypes under LDs. Data are shown as means ± sd (n = 15). Letters shared between the genotypes indicate no significant difference (P < 0.001, one-way ANOVA, Tukey’s multiple comparison test).