Hd3a Protein Is a Mobile Flowering Signal in Rice

See allHide authors and affiliations

Science  18 May 2007:
Vol. 316, Issue 5827, pp. 1033-1036
DOI: 10.1126/science.1141753


Florigen, the mobile signal that moves from an induced leaf to the shoot apex and causes flowering, has eluded identification since it was first proposed 70 years ago. Understanding the nature of the mobile flowering signal would provide a key insight into the molecular mechanism of floral induction. Recent studies suggest that the Arabidopsis FLOWERING LOCUS T (FT) gene is a candidate for encoding florigen. We show that the protein encoded by Hd3a, a rice ortholog of FT, moves from the leaf to the shoot apical meristem and induces flowering in rice. These results suggest that the Hd3a protein may be the rice florigen.

The flowering time of plants is determined by a number of environmental factors (13), among which day length (photoperiod) is a major factor (4). On the basis of the day length, which promotes flowering, plants are grouped into two major classes: long-day (LD) and short-day (SD) plants. Arabidopsis is a LD plant and rice is a SD plant. FT is a major floral activator (5, 6), which is expressed in the vascular tissue of leaves (7, 8). FT protein interacts with a transcription factor FD, which is expressed only in the shoot apical meristem (SAM) (9, 10). The difference in expression site implies that FT protein must move to the SAM to interact with FD for flower induction. Therefore, FT is a primary candidate for encoding florigen (11), a mobile flowering signal.

A tomato ortholog of FT, SFT, induced early flowering, and grafting sft mutant shoots to 35S::SFT donors induced normal flowering in the sft shoots (12). However, SFT mRNA was not detected in the SAM of the grafted tomato plants (12), suggesting that SFT mRNA does not move through graft junctions in tomato. Furthermore, a previous study suggesting that florigen was an RNA molecule has been retracted (13). Therefore, although FT is a candidate for encoding florigen, the exact nature of florigen remains to be determined.

Previous studies indicate that Hd3a is the major activator of flowering in rice, a SD plant, under SD conditions, and that Hd3a complements Arabidopsis ft mutants (1417). Therefore, we examined Hd3a transcript levels in several tissues by real-time polymerase chain reaction (PCR) under inductive conditions for flowering (Fig. 1A). Hd3a mRNA accumulates in leaf blade tissue, but is present at very low abundance in leaf sheath (Fig. 1A). Quantitative comparisons of Hd3a mRNA in leaves and the shoot apex indicate that its accumulation in the shoot apex is on the order of 10–4 of that in leaf blade, indicating that Hd3a mRNA is virtually absent from the shoot apex of rice plants when flowering is induced under SD conditions. Therefore, it is unlikely that Hd3a mRNA moves from leaf to the SAM in any appreciable amount.

Fig. 1.

Expression of Hd3a mRNA in rice under SD conditions. (A) Real-time quantitative RT-PCR of Hd3a mRNA accumulation in rice tissue. Samples of plants were harvested at ZT 0 to 4. Hd3a mRNA was quantified relative to Ubiquitin (Ubq) mRNA. (B to D) GUS staining of Hd3a::GUS. (B) Leaf blade ofthe Hd3a::GUS transgenic rice plant at ZT4 on day 35 under SD conditions. (C) Transverse section of a leaf blade in (B). (D) Longitudinal section of the SAM (arrow) of the same transgenic plant as in (B) and (C). Scale bars: 1 mm (B), 20 μm (C), 50 μm (D).

To determine the tissue and cell specificity of Hd3a mRNA expression, we analyzed the activity of an Hd3a::GUS transgene in leaf blades and SAMs of transgenic rice. The promoter activity of Hd3a was detected in phloem and xylem parenchyma cells of leaf blade (Fig. 1, B and C), and no GUS activity was detected in the SAM (Fig. 1D). This was consistent with the quantitative reverse transcription–polymerase chain reaction (RT-PCR) results (Fig. 1A) and similar to the tissue specificity of FT expression in Arabidopsis (7, 18). Hd3a expression is thus restricted to the vascular tissues of rice leaves under inductive SD conditions.

To study the function and localization of Hd3a protein in rice, we fused the 1.7-kb Hd3a promoter used for GUS analysis to green fluorescent protein and introduced the resulting construct (Hd3a:GFP) into rice plants by Agrobacterium-mediated transformation. The leaf diurnal expression pattern of transgenic plants was similar to that of the endogenous Hd3a gene (Fig. 2L), but varied among transgenic plants. Transgenic rice plants flowered (headed) significantly earlier than wild-type plants (Table 1 and fig. S1A), suggesting that expression of Hd3a:GFP causes early flowering, because expression of endogenous Hd3a mRNA in transgenic rice plants did not change relative to that in wild-type plants.

Fig. 2.

Confocal microscopy of Hd3a::Hd3a-GFP transgenic rice. (A to J) Confocal images of Hd3a:: Hd3a-GFP transgenic plants. (A to H) Longitudinal sections through the SAM. (I and J) Longitudinal section through vascular bundles indicated by the red squares in (G) and (H). (A), (C), (E), (G), and (I) are composite images of the fluorescein isothiocyanate (FITC) and transmission channels. (B), (D), (F), (H), and (J) show the spectrally unmixed images. Hd3a-GFP fluorescence is shown in green, and plant autofluorescence in red. Scale bars, 50 μm. Arrows indicate a SAM. (K) Diagram of the SAM and the upper part of the rice stem. V, vascular bundles; SAM, shoot apical meristem. (L) Real-time quantitative RT-PCR of Hd3a-GFP and endogenous Hd3a mRNAs under SD conditions in Hd3a::Hd3a-GFP transgenic rice plants. White and black bars at the bottom represent light and dark periods, respectively.

Table 1.

Flowering (Heading) times of transgenic plants under SD conditions.

GenotypeDays to flowering (days ± SE) n
Wild type 50.4 ± 7.6 5
Hd3a::Hd3a:GFP 32.8 ± 11.2 6
RPP16::Hd3a:GFP 14.8 ± 3.3 5
RPP16::GFP 64 2
rolC::Hd3a:GFPView inline 19.5 ± 13.6 11
rolC::GFPView inline 88.6 ± 11.3 5
  • View inline* Indicates significant difference from control by Student's t test (P = 0.0000007).

  • To examine tissue localization of the Hd3a protein in Hd3a:GFP transgenic plants, we analyzed GFP fluorescence in the SAM, the upper part of the stem, and in the leaf blade by confocal laser scanning microscopy. GFP fluorescence was limited to the inner conelike region of the SAM in transgenic rice (Fig. 2, A to D, G and H). The GFP signal was detected in the SAM (Fig. 2, C and D) and stem vascular tissue (Fig. 2, I and J). GFP signal was also detected in the vascular tissue of the upper part of the stem and in the region just beneath the meristem where nodes are present (Fig. 2, E and F), suggesting that Hd3a:GFP protein moves from the end of the vascular bundles through the basal cells andintothe SAM.

    Hd3a:GFP protein is thus found in the inner region of the SAM and in stem and leaf blade vascular tissues, suggesting that it is produced in the vascular tissue of the leaf blade, transported through stem phloem tissue, unloaded at the upper end of the vascular tissue, and translocated to the SAM, probably through the region just beneath the SAM. These results suggest that the Hd3a protein, but not Hd3a mRNA, is a candidate for the florigen in rice.

    We expressed the Hd3a:GFP gene in phloem tissue by fusing it with two phloem-specific promoters, the Agrobacterium rhizogenes rolC promoter (8, 18) and Rice Phloem Protein 16 (RPP16) promoter (19). The rolC promoter is specifically active in the phloem (18), and rolC::CO is known to induce extremely early flowering in Arabidopsis (8). The RPP16 gene encodes a phloem-specific protein in rice (19). Rice plants expressing RPP16::Hd3a:GFP and rolC::Hd3a:GFP flowered very early compared to the wild-type plant (Table 1 and fig. S1, B and C), indicating that the vascular-specific expression of the Hd3a:GFP gene induced early flowering in rice. GFP signals were detected in the vascular tissues of leaf blades and in the stems of rolC::Hd3a:GFP and RPP16::Hd3a:GFP transgenic plants (Fig. 3, B, D, J, and L). In transverse sections of the leaf blade, GFP signals were detected in cells near the phloem (Fig. 3, A, B, I, and J). The intact Hd3a:GFP protein was detected by immunoblotting with antibody to GFP in the leaf extract (fig. S2). Fluorescence was detected in the SAMs of both transgenic lines (Fig. 3, E, F, M, and N), and in leaves adjacent to SAMs (Fig. 3, E, F, M, and N). Because the free GFP protein diffused in many tissues in rice, the Kaede reporter protein (20, 21) was used to localize promoter activity. The Kaede protein forms a monotetrameric complex of 116 kD and is retained in cytoplasm (20). Kaede fluorescence was not detected in the SAM (Fig. 3, G, H, O, and P) and was detected only in the vascular tissues of rolC::Kaede and RPP16::Kaede transgenic plants (fig. S3), demonstrating that the rolC and RPP16 promoters are not active in the SAM. This result confirms that Hd3a protein is translocated from stem vascular tissue to the SAM.

    Fig. 3.

    Confocal microscopy of transgenic rice plants expressing a fusion of reporter protein with phloem-specific promoters. Confocal images of transgenic rice plants. (A), (C), (E), (G), (I), (K), (M), and (O) are composite images of FITC and transmission channels. (B), (D), (F), (H), (J), (L), (N), and (P) show the spectrally unmixed images. Hd3a-GFP and Kaede-green fluorescence are shown in green, and autofluorescence is in red. (A) and (B) Transverse sections through a leaf of rolC::Hd3a-GFP. (C) and (D) Longitudinal sections through the stem and SAM of rolC::Hd3a-GFP. (E) and (F) Longitudinal section through a SAM of rolC::Hd3a-GFP. (G) and (H) Longitudinal sections through a SAM of rolC::Kaede. (I) and (J) Transverse section through a leaf of RPP16::Hd3a-GFP. (K) and (L) Longitudinal sections through a stem, including the meristem of RPP16::Hd3a-GFP. (M) and (N) Longitudinal sections through a meristem of RPP16::Hd3a-GFP. (O) and (P) Longitudinal sections through the SAM of RPP16::Kaede. Scale bars: 25 μm [(A), (B), (M), and (N)]; 50 μm [(C to L), (O), and (P)]. Arrows indicate SAM. Arrowheads indicate GFP fluorescence.

    Hd3a protein fulfills the requirements for a florigen (11), but Hd3a mRNA cannot be completely ruled out as a florigen because Hd3a transcripts are present in the shoot apex in extremely low abundance. A recent proteomic study of phloem sap obtained from the inflorescence stem of Brassica napus identified FT protein (22) as a sap constituent. The presence of FT ortholog in the corresponding tissues of this distantly related plant supports our conclusion that it is the Hd3a protein that acts as the main florigen. Our results strongly suggest that the protein encoded by FT/Hd3a acts universally as aflorigen (2325).

    Because there is no vascular connection between the upper end of the vascular bundles and thebaseofthe SAM, there mustbesome mechanisms that regulate the movement of Hd3a protein into the SAM. There may be intercellular transport proteins which help Hd3a protein move toward the center of the stem just beneath the SAM. Once Hd3a protein enters the SAM, it may be localized in the nucleus. A recent report on the maize FD ortholog (26) shows that its mRNA is localized in the inner region of the SAM, similar to the region where GFP signal was detected in Hd3a:GFP transgenic rice. These results suggest that an FD-like nuclear protein may regulate intracellular localization of Hd3a protein in the SAM.

    The morphology of vegetative organs changes when there is a phase transition to flowering in some species. It has recently been shown that FT overexpression induces changes in leaf morphology and stem branching in tomato (12) and in Arabidopsis leaf morphology (27). In aspen trees, FT was shown to regulate growth cessation and bud dormancy (28). We found that transgenic rice plants expressing RPP16::Hd3a:GFP or rolC::Hd3a:GFP had alterations in multiple traits in vegetative organs such as elongation of internodes, which is known to occur after the transition to flowering and increased tillering. These alterations were induced by ectopic expression of the Hd3a protein in the vascular tissues. These results may suggest that many, if not all, of the changes associated with the transition from vegetative to reproductive growth and development induced by day length are induced by Hd3a protein. Therefore, Hd3a/FT protein may be a general mobile morphogen that regulates multiple phases of plant growth by photoperiod.

    Supporting Online Material

    Materials and Methods

    Figs. S1 to S3


    References and Notes

    Stay Connected to Science

    Navigate This Article