Gibberellins accumulate in the elongating endodermal cells of Arabidopsis root

Abstract

Plant hormones are small-molecule signaling compounds that are collectively involved in all aspects of plant growth and development. Unlike animals, plants actively regulate the spatial distribution of several of their hormones. For example, auxin transport results in the formation of auxin maxima that have a key role in developmental patterning. However, the spatial distribution of the other plant hormones, including gibberellic acid (GA), is largely unknown. To address this, we generated two bioactive fluorescent GA compounds and studied their distribution in Arabidopsis thaliana roots. The labeled GAs specifically accumulated in the endodermal cells of the root elongation zone. Pharmacological studies, along with examination of mutants affected in endodermal specification, indicate that GA accumulation is an active and highly regulated process. Our results strongly suggest the presence of an active GA transport mechanism that would represent an additional level of GA regulation.

Adaptive growth of plants is regulated by small-molecule regulators called plant hormones (1). Plants regulate hormone response pathways at multiple levels including hormone biosynthesis, metabolism, perception, and signaling. In the case of auxin, elegant studies have shown that the regulation of auxin distribution through the action of specific transporters also has a central role in plant development (2, 3). The recent isolation of transporters for other hormones (4–6) suggests that the spatial distribution of these compounds may also be regulated.

Gibberellins (GAs) are a class of tetracyclic diterpenoid hormones that regulate many developmental processes such as seed germination, root and shoot elongation, flowering and fruit patterning (7–9). Over the years, more than 130 GAs have been identified, of which only a few, such as GA₁, GA₃, and GA₄, are bioactive (8). Much progress has been made in understanding how plants control GA response through regulation of biosynthesis, metabolism, and signaling (10–14). Although experiments with radiolabeled GAs as well as grafting studies have established that GAs move through the plant (15–18), little is known about either the mechanisms of transport or the distribution of GA.

To address these questions, we generated fluorescently tagged GA. The labeled GAs retained much of their bioactivity and thus were used as fluorescent GA surrogates to study the distribution of GA in the Arabidopsis root system.

Results

Fluorescently Labeled GAs Are Bioactive.

Four derivatives of fluorescein (Fl)-labeled GA₃ were synthesized (SI Appendix, Figs. S1–S4), varying primarily in the length of the linker between Fl and GA₃ (Fig. 1A and SI Appendix, Fig. S5A). Conjugation to GA₃ through amide formation on C6 was based on previous reports demonstrating its stability in vitro and in vivo (19). The four molecules were compared for their GA bioactivity. GAs are essential for seed germination in many plants including Arabidopsis (20, 21). The GA biosynthesis mutant ga1 germinates poorly, whereas exogenous application of GA₃ fully restores germination levels (22, 23). Whereas molecules 1 and 4 had a very small effect on ga1 germination (4% and 8%, respectively), treatment with molecules 2 and 3 resulted in 33% and 53% germination. Application of Fl to ga1 plants had no effect on germination (Fig. 1B and SI Appendix, Fig. S5B). Similarly, for wild-type (WT) plants treated with the GA-biosynthesis inhibitor paclobutrazol (Paclo), molecules 2 and 3 showed the highest rate of germination (38% and 27%, respectively), whereas 1 and 4 had a smaller effect (13% and 16%, respectively) and Fl had no effect at all (SI Appendix, Fig. S5C). GAs are also key regulators of hypocotyl and root elongation (24, 25). In 4-d-old WT seedlings treated with Paclo, molecules 1, 4, and Fl had little effect on Paclo-treated hypocotyls whereas 2 and 3 partially restored elongation (73% and 69%, respectively) (Fig. 1 C and D and SI Appendix, Fig. S5 D and E). Strikingly, in WT plants treated with Paclo, application of molecules 2 and 3 fully restored root elongation, compared with GA₃, whereas 1 and 4 had only a modest effect (Fig. 1E and SI Appendix, Fig. S5F). In this assay, the half-maximal effective concentration (EC₅₀) of the best performing derivative molecule 3 (termed “GA₃-Fl” from here on) was 130% that of GA₃ (1.13 ± 0.02 μM and 0.87 ± 0.11 μM, respectively) (Fig. 1F). Together, these results demonstrate that GA₃ can be labeled with fluorescein and retain biological activity. Structurally, there is a positive correlation between linker length and bioactivity. Reintroduction of the carboxylic acid in close proximity to its original location on GA’s C6 (molecule 4) did not have a positive effect on bioactivity. To further evaluate tagged GA function across species, we tested GA₃-Fl bioactivity in Solanum lycopersicum (tomato) compound leaf development. Tomato plants treated with GA₃-Fl presented simpler leaves with smooth margins, mimicking GA’s effect on tomato leaf shape (SI Appendix, Fig. S6).

Fig. 1.

Characterization of labeled GAs. (A) Molecular structure of GA₃-Fl. (B) Germination of GA biosynthesis mutant ga1 treated with GA₃-Fl, Fl, or GA₃ (100 μM). (C and D) Hypocotyl elongation of seedlings treated with Paclo (2 μM) plus GA₃-Fl, Fl, or GA₃ (10 μM). Shown are averages ± SE (n = 10). (D) Graphical representation of the experiment in C. (Scale bar, 1 mm.) (E) Root elongation of seedlings treated with Paclo and GA₃-Fl, Fl, or GA₃ (10 μM). Shown are averages ± SE (n = 10). (F) Response of Paclo-treated seedlings to increasing concentrations of GA₃ and GA₃-Fl (molecule 3). EC₅₀ is for GA₃ and GA₃-Fl. Shown are averages ± SE (n = 10). (G) Molecular structure of GA₄ labeled with Fl_. (H) Effects of Paclo (2 μM) and GA₄-Fl or GA₄ (10 μM) on germination of WT seeds. Shown are averages ± SE (n = 70). (I and J) Fluorescence HPLC chromatograms of root extracts from GA₃-Fl–treated (I) or GA₄-Fl–treated (J) plants and reference compounds. Green, reference molecule; blue, 1-d root extract; yellow, 2-d root extract; red, 3-d root extract. (I) Peaks 1 and 2 elute similarly to the GA₃-Fl isomer reference and were further identified by HRMS. (J) Peak 1 elutes similarly to GA₄-Fl reference and was further identified by HRMS. Peak 2 is an unidentified GA₄-Fl adduct. Arrows point to retention times expected for GA₃-Fl and GA₄-Fl cleavage products: blue: C6 amide, black: 5- and 6-carboxyfluorescein (SI Appendix, Fig. S10E). *Significantly different relative to respective Paclo treatment at P ≤ 0.001 by Student t test.

The minor structural differences between GAs prompted us to test whether the strategy used for GA₃ labeling is also effective for other bioactive GAs. Thus, GA₄ was labeled with fluorescein similarly to GA₃-Fl (Fig. 1G) and tested for its activity. The fluorescent conjugate (GA₄-Fl) had comparable activity to GA₄ with respect to germination (Fig. 1H) and root and hypocotyl elongation (SI Appendix, Fig. S7). This suggests that the GA-labeling strategy described here could be applied to other GA derivatives. GA₃-Fl and GA₄-Fl have spectroscopic properties characteristic of fluorescein (λ_ex = 496 nm, λ_em = 523 nm, Φ_f = 0.81 and λ_ex = 495 nm, λ_em = 525 nm, Φ_f = 0.75, respectively) (SI Appendix, Fig. S8), making them suitable for detection by fluorescence imaging.

GA-Fls Interact with the GA Receptors and Are Not Significantly Metabolized in Vivo.

GA metabolism in plants is highly complex, and most of the metabolites are biologically inactive (8). To test whether the observed bioactivity of the tagged GAs is due to the intact molecules and not degradation products or secondary metabolites, we first studied their activity in vitro. GA response is initiated by the binding of GAs to the GID1 receptor to promote its association with repressors of GA response called DELLA proteins. This association triggers degradation of the DELLAs via the SCF^SLY2/SNE ubiquitin E3 ligase and activation of GA responses (12, 26). To test for interaction between GA-Fl and the receptor, GST-GID1b was expressed in Escherichia coli and incubated with Myc-RGA [Repressor of GA, synthesized in TnT (Promega)] in the presence or absence of the tagged GAs. GA₃-Fl and GA₄-Fl promoted interaction between GID1b and the DELLA protein, whereas molecule 1, which was biologically inactive in vivo, did not promote this interaction (SI Appendix, Fig. S9A). Similarly, in the yeast two-hybrid system, GA₃-Fl and GA₄-Fl enhanced the interaction between GID1a and RGA (SI Appendix, Fig. S9B). In both pull-down and yeast two-hybrid assays, GA₄-Fl showed a higher activity compared with GA₃-Fl.

Next, we evaluated the stability of the tagged GAs in plants. For this, extracts from plants treated with either GA₃-Fl or GA₄-Fl for up to 3 d (same conditions as used in bioactivity assays) were prepared and analyzed by HPLC-HRMS (high resolution mass spectrometry). The fluorescence HPLC chromatograms of GA₃-Fl–treated roots (Fig. 1I) show two major fluorescent components in the extracts (peaks 1 and 2 in Fig. 1I) with no buildup of other fluorescent components over time. Peaks 1 and 2 eluted similarly to the two GA₃-Fl isomers (synthesized from 5- and 6-carboxyfluorescein, peaks 2 and 1, respectively) and were further identified by HRMS [calculated for GA₃-Fl (M+H⁺): 835.3073; found: 835.3071 and 835.3066, peaks 1 and 2, respectively (SI Appendix, Fig. S10A)]. In the GA₄-Fl extracts (GA₄-Fl was synthesized from the single 5-carboxyfluorescein isomer) (Fig. 1J), the major fluorescent component (peak 1) is intact GA₄-Fl [calculated for GA₄-Fl (M+H⁺): 821.3280; found: 821.3293], yet an additional fluorescent component that increased over time with a molecular mass of 1,308 Da was detected (peak 2) (SI Appendix, Fig. S10B). The higher molecular mass of this substance (+488 Da, compared with GA₄-Fl) suggests that it is an adduct of GA₄-Fl. However, its molecular identity was not determined. Interestingly, formation of the GA₄-Fl metabolite was observed in root but not in shoot extracts, whereas in GA₃-Fl extracts there was no difference in the fluorescent content of root and shoot (SI Appendix, Fig. S10 C and D). It is important to note that no cleavage products of the tagged GA’s C6 amide or any other degradation products were detected (SI Appendix, Fig. S10E). Altogether, these results establish that the bioactivity of GA₃-Fl and GA₄-Fl arises from the intact molecules and demonstrate their in vivo stability. Because GA₃-Fl proved more resistant to in vivo metabolism, it was preferred for use in further studies.

GA₃-Fl and GA₄-Fl Accumulate to High Levels in Endodermal Cells of the Root Elongation Zone.

Roots exhibit an apical-basal developmental gradient starting at the rootward end with the stem cell population in the root meristem. Upon emergence from the stem cell niche, cells undergo multiple rounds of cell division. As cells move farther shootward, they begin to mature and enter the elongation zone. About 6 h later, when they reach their final size, the cells become part of the differentiation zone (Fig. 2A and SI Appendix, Fig. S11) (27, 28). Cell-type–specific expression of a nondegradable form of the GA-signaling protein GAI revealed that inhibition of GA signaling in the endodermal layer had a particularly dramatic effect on root elongation (29). This suggests that these cells have a unique role in GA regulation of root growth. In addition, interaction between the transcriptional regulators SCL3 and the DELLAs plays a role in controlling GA response and biosynthesis in the endodermis (30, 31). To determine whether GA distribution might contribute to this specificity, we studied the distribution of GA₃-Fl in Arabidopsis roots. GA₃-Fl or Fl (5 and 0.5 μM, respectively) was applied to 6-d-old seedlings for 2 h by adding the compound to agar growth medium. Roots were comprehensively washed and immediately imaged. Interestingly, although Fl could be detected uniformly in all tissues (Fig. 2 C and D), GA₃-Fl was distributed in a highly specific pattern (Fig. 2 B–D), accumulating in the endodermal cells of the elongation zone (Fig. 2 B–E). Although GA₃-Fl was applied evenly to the whole root, very low levels could be detected in the meristematic or differentiation zones (Fig. 2 B and G). It is important to note that fluorescence intensity of Fl-treated roots was dramatically higher (∼22-fold) than for GA₃-Fl–treated roots (SI Appendix, Fig. S12). Remarkably, the boundaries of accumulation corresponded precisely to the transition zones from meristematic to elongating cells and from elongating to differentiating cells (SI Appendix, Fig. S13).

Fig. 2.

Distribution of GA₃-Fl and GA₄-Fl in the root. (A) Tissue organization and developmental zones of the Arabidopsis primary root. MZ, meristematic zone; EZ, elongation zone; DZ, differentiation zone. Green, epidermis; orange, cortex; purple, endodermis; red, pericycle. See SI Appendix, Fig. S11, for high-resolution image. (B) Confocal image of fluorescence distribution in elongating endodermal cells of roots treated with GA₃-Fl (green) (5 μM, 2 h). White arrows mark transition from MZ to EZ; pink arrow marks transition from EZ to DZ. Cell walls were stained with propidium iodide (red). (C) Fl and GA₃-Fl distribution at the transition between meristematic and elongation zones. White arrows mark transition from MZ to EZ (cortical cell is twice the size of shootward previous cell). (D) Fl and GA₃-Fl distribution in radial confocal sections of the elongation zone. GA₃-Fl accumulates in the endodermal cells. (E) Quantification of GA₃-Fl fluorescence intensity in different cell types of the root elongation zone. Shown are averages ± SE (3 root images, 10 cells/image, 2 sampling points/cell; n = 60). RFU, relative fluorescence units. (F) Quantification of GA₃-Fl fluorescence intensity in the vacuole and nucleus of cortical and endodermal cells of the elongation zone. Fluorescence intensity was normalized relative to nuclear marker 35S:H2B-RFP. Shown are averages ± SE (three root images, three cells/image, three sampling points/cell; n = 27) (SI Appendix, Fig. S16). (G) Time-lapse quantification of GA₃-Fl fluorescence intensity in endodermal cells at four different maturation points by confocal imaging. Point 1, meristematic; 2, rootward elongation; 3, shootward elongation; 4, rootward differentiation cells. Shown are averages ± SE (3 root images, 10 cells/image, 2 sampling points/cell; n = 60). (H) Distribution of GA₄-Fl (5 μM, 2 h) in the root. White arrow marks transition from MZ to EZ. Image on the right is a magnification of the elongation zone. (I) Distribution of GA₃-Fl or GA₄-Fl (5 μM) in the root after application for 16 h. (Scale bars, 50 μm.)

It is unlikely that the Casparian strip is responsible for the observed pattern as this structure is established shootward to the region of GA₃-Fl accumulation (32). Nevertheless, we examined GA₃-Fl distribution in the casp1-1 casp3-1 double mutant (33) and in WT plants treated with piperonylic acid, an inhibitor of monolingnol biosynthesis that blocks Casparian strip formation in newly forming cells (34). In both cases, the pattern of GA₃-Fl distribution was unaffected compared with WT plants (SI Appendix, Fig. S14), confirming that the Casparian strip does not determine GA₃-Fl accumulation in these cells.

Time-lapse confocal microscopy revealed a dynamic uptake process where GA₃-Fl begins to accumulate within 15 min in the endodermis (Fig. 2G). Interestingly, a strong reduction of GA₃-Fl signal could be detected 4 h after application (Fig. 2G and SI Appendix, Fig. S15).

Subcellular imaging of the elongating endodermal cells showed that GA₃-Fl accumulates mainly in the vacuole and to a lesser extent in the nucleus (Fig. 2F and SI Appendix, Fig. S16). Importantly, the fluorescent signal in the nuclei of elongating cells was higher in endodermal cells than in other cell types. It should be noted that, in the acidic environment of vacuoles (pH ∼5.5), the fluorescence intensity of the Fl-labeled GAs would be decreased (fluorescein’s pK: ∼6.4); thus, the effective concentration difference between vacuole and nucleus is probably higher than that observed by fluorescence intensity ratio.

GA₄-Fl was distributed in the developing root similarly to GA₃-Fl with high levels of accumulation in the elongating endodermal cells, but was also detectable in the stele after 2 h of application (Fig. 2H). Examination of lateral roots showed that here, as well, GA₄-Fl accumulates mostly in the elongating cells of the endodermis (SI Appendix, Fig. S17). Additionally, GA₄-Fl accumulation was greatest in the endodermal cell of the main root opposite to the site of lateral root formation. Application of the tagged GAs for longer periods (16 h) resulted in additional accumulation in the transition zone from elongating to meristematic zones, and, for GA₄-Fl, very low levels were also detectable in the meristem’s cortex (Fig. 2I). The detection of GA₄-Fl in the ground tissues is in accordance with previous findings that GA signaling in ground tissue affects meristem size (35).

GA Marker GFP-RGA Shows Weaker Signal in the Elongating Endodermal Cells.

The specific distribution in the root observed for the tagged GAs prompted us to independently evaluate endogenous GA levels. To do this, we examined the expression of the DELLA reporter GFP-RGA in developing roots of pRGA:GFP-RGA transgenic plants. GFP-RGA is degraded in the presence of GA through the action of the E3 ligase SCF^SLY2/SNE (36, 37). Strikingly, GFP-RGA distribution is complementary to that of the tagged GA in both the meristematic and the elongation zones (Fig. 3A). In the presence of Paclo, when endogenous GA synthesis is inhibited, GFP-RGA signal was detected at high levels in the nuclei of all cells in the elongation zone (Fig. 3 B and C). In the absence of Paclo, although modest reduction in the GFP-RGA signal was observed in all cell types of the elongation zone (0.4-fold reduction in cortex cells), the most substantial drop was observed in the endodermis (17.2-fold reduction) (Fig. 3 B and D), indicating that endogenous GA level in the endodermis cells is higher than in other cell layers of the elongation zone.

Fig. 3.

GFP-RGA signal is reduced in elongating endodermal cells. (A) GFP-RGA localization in the root. MZ, meristematic zone; EZ, elongation zone. (B) Relative fluorescence intensity of GFP-RGA in the cortex and endodermal nuclei of the elongation zone with and without 2 M Paclo treatment. Shown are averages ± SE (three root images, three cells/image, three sampling points/cell; n = 27). (C and D) Images of the elongation zone showing GFP-RGA levels (C) with (D) without Paclo treatment. White arrows indicate the endodermal cell nuclei. *Significantly different relative to the respective cortex cell at P ≤ 0.001 by Student t test. (Scale bars, 50 μm.)

GA₃-Fl Accumulation in Roots Is Actively Regulated.

The preferential accumulation of GA in elongating endodermal cells could be due to an active transport or a diffusive process. To address this question, we first evaluated the effect of unlabeled GA₃ on GA₃-Fl distribution. Cotreatment with GA₃-Fl and increasing concentrations of GA₃ led to a decreased fluorescence signal in the elongating endodermal cells and was completely blocked at 20-fold GA₃ excess (Fig. 4A and SI Appendix, Fig. S18). In contrast, no effect was observed on Fl signal intensity or distribution. We next examined the effect of temperature on GA₃-Fl distribution. In Fl-treated plants maintained at low temperature (4 °C), fluorescence intensity in the roots was relatively high (Fig. 4 D and F) and was reduced threefold compared with normal growth conditions (22 °C) (Fig. 4G and SI Appendix, Fig. S19). In contrast, exposure to lower temperature (4 °C) had a marked effect on GA₃-Fl distribution (40-fold reduction in fluorescence intensity compared with 22 °C), effectively inhibiting its accumulation (Fig. 4 D–G) and suggesting that the GA transport machinery is energy-dependent. To further test this possibility, GA₃-Fl and Fl accumulation were examined following a 2-h application of three mitochondrial ATP synthesis inhibitors: antimycin A, oligomycin A, and myxothiazol. Whereas Fl accumulation was unaffected by antimycin A and oligomycin A and elevated by myxothiazol, GA₃-Fl uptake was reduced approximately fivefold by all three ATP synthesis inhibitors (Fig. 4H and SI Appendix, Fig. S20).

Fig. 4.

GA₃-Fl is actively transported into the elongating endodermal cells. (A) Distribution of GA₃-Fl (5 μM) in the root elongation zone in the absence or presence of competing GA₃ (200 μM). At the far right is a radial confocal section of the elongation zone under competing conditions (compare with Fig 2D). (B) Quantification of the competition assay described in A. Fluorescence intensity was measured in the elongating endodermal cells. Shown are averages ± SE (3 root images, 10 cells/image, 2 sampling points/cell; n = 60) (SI Appendix, Fig. S18). (C) GA₃-Fl distribution in the elongation zone of the scr3-2 and shr-2 mutants. White arrows indicate the single cortex/endodermal layer. (D) Temperature dependence of Fl and GA₃-Fl accumulation in the elongation zone. Fl (0.5 μM) and GA₃-Fl (5 μM) were applied for 2 h at either 4 °C or 22 °C. (E) Fluorescence intensity in elongating endodermal cells in plants treated with GA₃-Fl (5 μM) for 2 h at the indicated temperature. Shown are averages ± SE (n = 60). (F) Fl and GA₃-Fl applied for 2 h at 4 °C. Fluorescent intensity was measured as described in B. (G) Fold change in fluorescence intensity of Fl (0.5 μM) and GA₃-Fl (5 μM) applied to seedlings exposed to 4 °C compared with 22 °C. Fluorescence intensity was measured as described in B. (H) ATP synthesis inhibitors antimycin A, oligomycin A, and myxothiazol disrupt GA₃-Fl but not Fl accumulation in the elongation zone. Seedlings were treated with 10 μM of the indicated inhibitor for 1.5 h and transferred to plates with inhibitor + GA₃-Fl or Fl (5 μM and 0.5 μM, respectively) for an additional 1.5 h. Fluorescence intensity was measured in the elongating endodermal cells. Shown are averages ± SE (n = 60). *Significantly different relative to the respective control at P ≤ 0.001 by Student t test. (Scale bars, 50 μm.)

To further understand the role of the endodermal cell type in GA accumulation, we tested GA₃-Fl distribution in scarecrow (scr) and short-root (shr) mutants. Both SCR and SHR are involved in specification of the ground tissue cell layers, the endodermis and cortex. In the scr-3 mutant, the ground tissue consists of a single layer of cells with attributes of both cortex and endodermis (38). The shr-2 mutant also has a single layer of ground tissue, but in this case the cells are cortical in nature (39, 40). GA₃-Fl application to the scr-3 mutant resulted in strong accumulation in the single layer of ground tissue (Fig. 4C). In contrast, no accumulation was observed in the ground tissue of shr-2 (Fig. 4C), indicating that the elongating endodermal cells have unique features that drive GA accumulation. Collectively, these results indicate that GA₃-Fl uptake is an active and saturable process that results in regulated accumulation in the endodermis of the elongation zone.

Ethylene Inhibits GA Accumulation in the Root.

The plant hormone ethylene has an important role in the regulation of root development (41–43). Application of ethylene to seedlings results in rapid inhibition of root elongation (44). Studies have shown that ethylene acts, in part, by stabilizing the DELLA proteins (45–47). The involvement of ethylene in GA signaling prompted us to explore its effect on GA distribution. Thus, seedlings were treated with the immediate precursor of ethylene, 1-aminocyclopropane-1-carboxylic (ACC) (48), for 2 h and transferred to medium with ACC and GA₃-Fl. Endodermal accumulation of GA₃-Fl was dramatically inhibited by ACC (Fig. 5 A and B). To determine if the canonical ethylene-signaling pathway regulates this response, we used constitutive triple response1-1 (ctr1-1), a mutation in a negative regulator of ethylene signaling, and the ethylene insensitive2-5 (ein2-5), a mutation in a key positive regulator of ethylene signaling (43). As expected, GA₃-Fl did not accumulate in the endodermis of ctr1-1 roots whereas accumulation was observed in ein2-5 roots in both the presence and the absence of ACC (Fig. 5 C and D). Coapplication of ACC and Paclo also inhibited GA₃-Fl uptake, implying that ACC does not affect GA₃-Fl accumulation by activating endogenous GA biosynthesis that leads to competition (SI Appendix, Fig. S21). Because roots exposed to ACC for 4 h and roots of the ctr1-1 mutant display developmental changes, including impaired elongation (44), the alteration in GA₃-Fl distribution could be a secondary effect. To clarify this, ACC and GA₃-Fl were applied simultaneously to seedlings roots and imaged after 1 h. Under these conditions, GA₃-Fl accumulation in the elongating endodermal cells was dramatically inhibited in the ACC-treated roots compared with the no-ACC treatment control (SI Appendix, Fig. S21), confirming that ethylene is responsible for the observed GA accumulation inhibition. Altogether, these results demonstrate that ethylene signaling affects GA distribution, suggesting that ethylene regulation of GA signaling in Arabidopsis roots may be mediated, at least part, through changes in GA transport.

Fig. 5.

Ethylene inhibits GA accumulation in the root. (A) Effect of ACC treatment on GA₃-Fl distribution in the elongation zone. Seedlings were treated with 1 μM ACC for 1.5 h, transferred to ACC + GA₃-Fl (1 μM + 5 μM, respectively), and imaged 1.5 h later. Center image is a magnification of the elongation zone. Right image is a control (GA₃-Fl distribution in the root elongation zone treated with GA₃-Fl for 1.5 h and imaged at the same conditions and settings). (B) Graphical representation of the experiment in A. Shown are averages ± SE (three root images, 10 cells/image, two sampling points/cell; n = 60) (C) GA₃-Fl distribution in ctr1-1 mutant. (D) GA₃-Fl distribution in ein2-5 mutant and WT (control) with and without 1 μM ACC. *Significantly different relative to control at P ≤ 0.001 by Student t test. (Scale bars, 50 μm.)

Discussion

Our studies demonstrate the value of fluorescent labeling of plant hormones to study dynamic processes involved in hormone localization and regulation. Recently, Irani et al. fluorescently labeled brassinosteroid (BR) to study BR receptor endocytosis during signaling (49).

Here, we use fluorescently labeled GAs to reveal the distribution of GA in Arabidopsis roots. We show that GA accumulates specifically in the endodermis of the root elongation zone. Our results suggest that the endodermis has a special role in GA regulation, consistent with previous studies that indicate that the endodermis is the major GA-responsive tissue in the root (29–31, 35).

Previous studies indicate that the expression of GA biosynthetic genes is particularly high in the meristem with lower levels in other cell types (50–53). Our observation that the labeled GAs accumulate in elongating endodermal cells suggests that GA synthesized in the meristem, as well as in the cortical and epidermal cell layers, moves to the endodermis where it regulates elongation growth. Cells enter the elongation zone for a short period where they increase their length by ∼10-fold over 5 h. This size increase would result in a corresponding rapid intracellular dilution of GA levels unless new GA is either synthesized in the endodermis or transported from surrounding tissues (52).

We observe a striking accumulation of GA in the vacuoles of endodermal cells. Vacuolar accumulation may act to buffer GA levels within these cells, thus controlling GA concentration in the cytoplasm and nucleus where the hormone is thought to act. The observation that both fluorescent and native GA are present at a higher level in the nuclei of elongating endodermal cells compared with other cell types supports this notion. In a similar way, the decrease in GA₃-Fl accumulation in elongating endodermis that was observed after 4 h (Fig. 2G) may reflect feedback inhibition of GA uptake. According to this model, new endodermal cells that form after several hours of treatment would accumulate less GA.

It is important to note that our results do not address the specificity of the suggested transport machinery. Thus, it is possible that this machinery transports additional plant hormones or other small molecules. Consistent with this possibility, a recent study demonstrated that AIT3, a member of the NRT1/PTR family, can transport abscisic acid (ABA) as well as GA₃ in yeast (6). At present, the mechanism of GA accumulation in the endodermis is unclear, but the fluorescent GAs can be used in combination with genetic screening to identify proteins involved in this process. In addition, introduction of a photo-crosslinking moiety into the fluorescent GA structure may facilitate the isolation of such proteins.

Physiological studies have revealed complex interactions between all of the major plant hormones during root development. The discovery that ethylene regulates GA distribution introduces an additional dimension to this complex picture. Based on our results, ethylene may regulate root growth in part by affecting the level of GA in the endodermis. Because ethylene was shown to regulate auxin transport and biosynthesis in the root elongation zone (54), it would be interesting to test whether ethylene’s effect on GA distribution is mediated by auxin. Dynamic regulation of GA levels by ethylene and perhaps other hormones may contribute to environmental regulation of root growth.

GAs regulate diverse processes throughout plant growth and development. It is likely that GA distribution, regulated by localized synthesis, transport, and inactivation, will be an important aspect of GA action in various plant tissues. For example, GA transport was recently shown to be involved in xylem expansion (6). We expect that the labeled GAs that we describe here will help to dissect this and other GA-regulated processes.

Unraveling the molecular details of plant hormone transport mechanisms will contribute to a more comprehensive understanding of hormone action. Our study emphasizes the significance of plant hormone mapping, demonstrating that fluorescent labeling is a compelling tool to understand spatial and temporal distribution of hormones and other small signaling molecules.

Methods Summary

Chemical Synthesis and Characterization.

Synthetic schemes, procedures, and compound characterization are reported in SI Appendix.

Bioactivity Assays.

For germination bioactivity assay with the GA biosynthesis mutant ga1-1, seeds were placed in liquid (100 μM) with the indicated treatments for 2 d at 4 °C and transferred to the dark at 22 °C for 2 more days. Subsequently, the seeds were plated on vertical MS plates, and germination was scored under a dissecting scope 3 d later. For germination assays with Paclo, seeds were placed on vertical plates with 2 μM Paclo and 10 μM of molecules 1–4, Fl, or GA₃. Germination was scored under a dissecting scope 5 d later. For root growth dose–response assays (Fig. 1F and SI Appendix, Fig. S7), 5-d-old seedlings were transferred onto fresh MS media with either GA₃ or GA₃-Fl for 7 additional days after which root length was measured. Root growth in Fig. 1E was performed in a similar way except the seedlings were transferred at day 4, and root growth was measured after a further 5 d. For hypocotyl assays, seedlings were transferred at day 4 and measured 5 d later. Ten-day-old tomato seedlings were treated three times a week for 2 wk (shoot apical meristem and small leaves) with liquid solution of 10 μM of the indicated compounds. The mature third leaf was imaged. Seed germination and root and hypocotyl growth were all imaged and measured using a Nikon SMZ1500 dissecting scope and ImageJ software (http://rsbweb.nih.gov/ij/index.html).

Complete methods are described in SI Appendix.