FD, a bZIP Protein Mediating Signals from the Floral Pathway Integrator FT at the Shoot Apex

Flowering in Arabidopsis is regulated by several pathways that converge on the transcriptional regulation of the floral pathway integrators FT, SUPPRESSOR OF OVEREXPRESSION OF CO 1 (SOC1), and LEAFY (LFY) (1). FT is a direct target of CO, a key transcriptional regulator of the photoperiod pathway, and the role of FT as a potent promoter of flowering in response to photoperiods is conserved in Arabidopsis and rice (2–6). FT is expressed in the phloem tissues of cotyledons and leaves (7, 8) and encodes a 20-kD protein with homology to phosphatidylethanolamine binding protein or Raf kinase inhibitor protein (2, 3). However, the biochemical function of FT and downstream events leading to floral transition and floral morphogenesis at the shoot apex remain unknown.

bZIP protein FD is required for FT function. To understand how signals are mediated from FT to finally cause floral transition and floral morphogenesis, we searched for genes required for FT to promote flowering. Ectopic overexpression of FT by the 35S RNA promoter of cauliflower mosaic virus (35S::FT) causes a precocious-flowering phenotype (2, 3). We screened for suppressors of 35S::FT, because a similar approach with 35S::CO was successful in elucidating the downstream targets of CO, FT and SOC1 (4, 9). In addition to screening mutagenized populations of 35S::FT, we examined known late-flowering mutants (10) for their effect on the 35S::FT phenotype. Through the latter approach, we found that fd-1 is a strong suppressor of 35S::FT (Table 1). In contrast to a strong effect on 35S::FT, fd-1 had only a weak effect on a similar precocious-flowering phenotype of soc1-101D, an activation tagged allele of SOC1 (11) (Table 1), or 35S::LFY (12). These observations suggest that the FD activity is required specifically for the promotion of flowering by FT.

The FD gene was identified with AtbZIP14 (At4g35900) (13) by a map-based approach (figs. S1 and S2 and table S1). In seedlings, FD expression was observed mainly in the shoot apex (Fig. 1, A to E and J), did not show distinct circadian oscillation, and was not affected by photoperiods and CO activity (fig. S3). Under both short days (SDs) and long days (LDs), FD mRNA levels increased with time after germination (fig. S3). To examine the subcellular distribution of FD protein, we made a construct to express enhanced green fluorescent protein–FD (EGFP:FD) fusion protein by the FD promoter (FD::EGFP:FD) (14). That the fusion protein is functional was confirmed by the rescue of fd-1 (table S2). EGFP:FD was localized in the nucleus in cells of the shoot apex (Fig. 1, F and G). Similarly, nuclear localization of functional enhanced yellow fluorescent protein–FD (EYFP:FD) fusion protein (table S2) was observed in cells of the shoot apex and other tissues in 35S::EYFP:FD seedlings with various genetic backgrounds (fig. S4) (14), suggesting a constitutive nuclear localization.

• Download high-res image
• Open in new tab
• Download Powerpoint

To gain clues to the molecular basis of the requirement of FD for FT function, we investigated protein interactions. FT interacted with FD in yeast cells (Fig. 2A) and in vitro (Fig. 2C). In contrast, FD showed very weak interaction in yeast cells with TERMINAL FLOWER 1 (TFL1), the FT-related protein with an antagonistic role in the regulation of flowering (2, 3) (Fig. 2A and fig. S5). Because the subcellular distribution of the FT protein remains unknown, we examined the distribution of functional FT:EGFP fusion protein (table S2) expressed in the shoot apex of Arabidopsis by the FD promoter or in leaf epidermal cells of Nicotiana benthamiana by the 35S promoter (14). In both cases, FT:EGFP was observed in the nucleus and cytoplasm (Fig. 1, H and I, and fig. S4). That FT is able to function in the nucleus was supported by observations that FT protein fused to a glucocorticoid receptor (GR) expressed by the 35S promoter (35S::FT:GR) (14) promoted flowering on dexamethasone (Dex) treatment (Table 1). These findings suggest that FD and FT proteins coexist in the nucleus. We further analyzed the interaction of FT and FD proteins in plant cells using bimolecular fluorescent complementation (BiFC) (15). In tobacco leaf epidermal cells coexpressing the N-terminal half of EYFP fused to FD (YN-FD) and the C-terminal half of EYFP fused to FT (YC-FT) (14), YFP fluorescence was observed in the nucleus (Fig. 2D). These findings show that protein interaction is the basis of the dependence of FT on FD.

• Download high-res image
• Open in new tab
• Download Powerpoint

Floral meristem identity genes are regulatory targets of FT and FD. We next tried to identify the regulatory targets of FT and FD in the shoot apex that cause floral transition and morphogenesis. We obtained clues from analysis of ft; lfy and fd; lfy double mutants. A previous work showed that ft; lfy greatly reduced mRNA levels of AP1 and caused severe defects in floral meristem specification (16). These observations led to the suggestion that FT and LFY play redundant roles in up-regulation of AP1 and floral fate specification (1, 16). In agreement with a previous report, ft; lfy greatly reduced the amount and region of AP1 expression in the shoot apex and caused severe defects in floral development (Fig. 3, A to D, and fig. S6). fd; lfy plants had an inflorescence phenotype indistinguishable from that of ft; lfy and displayed a severely reduced amount and spatial extent of AP1 expression (Fig. 3, A to D, and fig. S6). These observations suggest that FT and FD together are involved in the regulation of AP1 redundantly with LFY.

• Download high-res image
• Open in new tab
• Download Powerpoint

Consistent with a role of FD in the regulation of AP1, ectopic expression of AP1 was observed in 35S::FD seedlings (Fig. 3E). Ectopic induction of AP1 expression by FD was abolished by ft mutation (Fig. 3F) or under SDs, which reduce FT expression (2, 3, 17, 18) (Fig. 3G). Upon a shift from SDs to LDs, which induces FT expression (8, 18), AP1 expression was induced (Fig. 3H). In young seedlings, FT is expressed mainly in the vasculature of cotyledons, and little is detected in mesophyll cells (7, 8). In 35S::FD seedlings, AP1 expression was detected in the vascular-rich fraction but not in the mesophyll-rich fraction from cotyledons (Fig. 3I). These observations suggest that activation of AP1 expression by FD requires the FT function. Finally, in the shoot apex around the stage of floral transition and in the young inflorescence apex, the region of AP1 expression was within the expression domain of FD (Fig. 1, J to M). Thus, AP1 seems to be a regulatory target of FD, which requires the FT activity through protein interaction. In support of this conclusion, several potential bZIP protein binding motifs were found in a 1.7-kb AP1 promoter (19) (fig. S7).

Because AP1 expression was observed only in a subset of the expression domain of FD, there should be factors that restrict AP1 expression to nascent lateral meristems. TFL1, which has a role antagonistic to FT (2, 3), is likely to be responsible for suppressing AP1 expression in the shoot apical meristem proper (20, 21). That the loss of AP1 alone does not affect the precocious-flowering phenotype of 35S::FT (3) suggests that other regulatory targets of FD contribute to the promotion of floral transition. FRUITFULL (FUL) and CAULIFLOWER (CAL), which act redundantly with AP1 to promote flowering (22), are candidates for such targets (Fig. 3, C and E to J, and fig. S8).

How FT regulates FD activity is another important question. Constitutive nuclear localization of FD and the presence of FT in the nucleus suggest the regulation of FD activity in the nucleus. Whether FD and FT form a stable transcriptional complex or interact only transiently remains to be investigated. FD protein has a potential phosphorylation site for Ca²⁺-dependent protein kinases (CDPKs) at the C terminus (Fig. 2B and fig. S2). Deletion or mutation of this site abolished interaction with FT (Fig. 2B), ectopic induction of AP1 expression (Fig. 3J), and the ability to complement fd-1 (fig. S9), although nuclear localization was not affected (fig. S4). These findings suggest the importance of phosphorylation of FD in the interaction with FT and in its functional regulation.

Mutual dependence and site of action of FT and FD. The mutual dependence of FT and FD, as shown above, is further supported by the observation that flowering of 35S::EYFP:FD plants was delayed under SDs, which reduce FT expression (table S3). Furthermore, the enhanced phenotype in 35S::FT; 35S::FD (tables S3 and S4) indicates that FT and FD are mutually limiting for the combined activity of FT and FD. These raise the question of the site(s) of action of FT and FD. In seedlings, FT is expressed in the vasculature of cotyledons, but not in the shoot apex (7) (fig. S10), whereas FD is expressed in the shoot apex but not in cotyledons and leaves (Fig. 1, A to E, and fig. S10). As expected, restoration of the FT function in the vasculature through expression by SULTR2; 1::FT (14) could rescue the late-flowering phenotype of ft (Table 1 and table S5). Restoration of the FT function in the shoot apex of ft, either in the whole region by FD::FT or in the outermost cell layer (L1) by PDF1::FT (14), also rescued the late-flowering phenotype (Table 1 and table S5). These findings agree well with those of a previous report that the late-flowering phenotype of co is suppressed by similar constructs for FT misexpression (23). We further observed that FD::FT, PDF1::FT, and SULTR2; 1::FT rescued delayed flowering, reduced AP1 expression, and severe floral defect in ft; lfy (Fig. 3, B and C, and fig. S6). By contrast, restoration of FT in root vasculature by IAA14::FT (14) failed to rescue ft (Table 1). These observations indicate that ectopically expressed FT in the shoot apex can exert an effect on flowering. FT expressed in the shoot apex requires the FD function, because fd-1 attenuated the rescued phenotype (Table 1 and table S5). In contrast, FD rescued the late-flowering phenotype of fd through expression in the whole shoot apex (by FD::FD), but not through ectopic expression in leaf vasculature (by SULTR2; 1::FD), where FT is expressed (Table 1). Therefore, FD acts in the shoot apex and seems to be required in all cell layers. These findings, together with observations that protein interaction is the basis for interdependence between FT and FD, suggest that the shoot apex is the site of the FT and FD action.

FT and the long-distance signal in flowering. It has long been believed that a long-distance signal, named florigen (24), is generated in leaves upon exposure to inductive photoperiods, is transported to the shoot apex, and acts there to promote flowering (25). However, the nature of the signal has remained elusive (25). Our present work supports an emerging hypothesis (7, 8, 23) that the FT products represent a part of the long-distance signal(s) generated in cotyledons and leaves (mainly in the phloem tissues) and act at the shoot apex to promote floral transition and to initiate floral development (fig. S11).