Environmental Nitrate Stimulates Abscisic Acid Accumulation in Arabidopsis Root Tips by Releasing It from Inactive Stores

Nitrate stimulates abscisic acid accumulation in root tips by inducing expression of the gene encoding the ABA-glucose ester-deconjugating enzyme β-GLUCOSIDASE1, thereby regulating root growth.

Abstract

Abscisic acid (ABA) signaling plays a major role in root system development, regulating growth and root architecture. However, the precise localization of ABA remains undetermined. Here, we present a mechanism in which nitrate signaling stimulates the release of bioactive ABA from the inactive storage form, ABA-glucose ester (ABA-GE). We found that ABA accumulated in the endodermis and quiescent center of Arabidopsis thaliana root tips, mimicking the pattern of SCARECROW expression, and (to lower levels) in the vascular cylinder. Nitrate treatment increased ABA levels in root tips; this stimulation requires the activity of the endoplasmic reticulum-localized, ABA-GE-deconjugating enzyme β-GLUCOSIDASE1, but not de novo ABA biosynthesis. Immunogold labeling demonstrated that ABA is associated with cytoplasmic structures near, but not within, the endoplasmic reticulum. These findings demonstrate a mechanism for nitrate-regulated root growth via regulation of ABA accumulation in the root tip, providing insight into the environmental regulation of root growth.

INTRODUCTION

Abscisic acid (ABA) mediates plant responses to environmental inputs, such as temperature, salinity, nitrate (NO₃⁻), water stress, and the presence of some pathogens (Lee and Luan, 2012; Hong et al., 2013). These environmental inputs shape the growth of the plant, most of which occurs postembryonically. In particular, changes in these environmental factors can have profound effects on the architecture of the root system, promoting root growth in one area and inhibiting it in another (Zhang and Forde, 1998; Walch-Liu et al., 2006; Harris, 2015). Depending on the concentration, ABA can inhibit or stimulate the functioning of the root meristem (Cheng et al., 2002), thereby modulating root growth.

NO₃⁻ is an essential nutrient required for many metabolic processes. As such, plants are continuously sensing its presence and using this information to regulate development. Elaboration of root architecture is strongly regulated by the level of NO₃⁻ in the environment, and the distribution of NO₃⁻ in the root environment is as important as its concentration. If the environment surrounding the root system is uniformly high in NO₃⁻, lateral root growth will generally be inhibited (Zhang and Forde, 2000; Walch-Liu et al., 2006). When present only in a patch, NO₃⁻ locally stimulates lateral root elongation and, in some species, initiation (Hackett, 1972; Drew, 1973; Zhang and Forde, 1998). In Arabidopsis thaliana, this local stimulation of lateral root elongation by patches of NO₃⁻ requires ABA signaling (Signora et al., 2001).

ABA levels in the plant are regulated by a combination of biosynthesis, degradation, transport, conjugation, and deconjugation (Finkelstein, 2013). In response to drying soil or a change in salinity, ABA levels in the stem xylem rise rapidly (Jeschke et al., 1997). Movement of ABA from the portion of the root system that senses the change in soil salinity or water availability to the leaves is important in the coordination of the plant response (Dodd et al., 2008, 2010; Puértolas et al., 2015). In response, patterns of root growth are altered, with some root tips inhibited and others stimulated, and stomata in the leaf close (Daszkowska-Golec and Szarejko, 2013; Harris, 2015). The rapid increase in xylem sap ABA in response to salinity is thought to be due to multiple sources: increased biosynthesis, phloem import of ABA from the leaves to the roots, and release of ABA from the conjugated form, ABA-glucose ester (ABA-GE) (Jiang and Hartung, 2008).

Unlike jasmonic acid, which is most active when conjugated to isoleucine (Staswick and Tiryaki, 2004), ABA conjugates are thought to be inactive forms, which are stored in the vacuole and transported in the xylem (Jiang and Hartung, 2008). In response to salt stress, ABA can be rapidly released from ABA-GE by β-glucosidases in the endoplasm reticulum (ER) or the cell wall space (Dietz et al., 2000; Lee et al., 2006) In Arabidopsis, β-GLUCOSIDASE1 (BG1) has been shown to function in the ER to release ABA from ABA-GE stores in response to salt stress (Lee et al., 2006). Similarly, the vacuolar β-glucosidase, BG2, plays a role in ABA release during dehydration stress (Xu et al., 2012). Thus, release of ABA from ABA-GE pools is an important mechanism for regulating ABA levels both locally and within the plant as a whole.

ABA, like auxin, is a weak acid and thus can exist either in a protonated, uncharged form (ABAH) that is predicted to passively diffuse across membranes or in a charged form (ABA⁻) that is unable to cross membranes without the aid of a transporter (Wilkinson and Davies, 2002; Schachtman and Goodger, 2008). The charged form is favored at more basic pH levels, as found in the mitochondrial matrix or chloroplast stroma, whereas the uncharged form is likely to predominate in areas with a more acidic pH, such as the cell wall. Thus, ABA⁻ is predicted to be trapped in more basic compartments of the cell (the anion trap hypothesis) until it is released via a transporter (Slovik et al., 1995). Over the past few years, both efflux and influx ABA transporters have been identified in Arabidopsis. The ABC subfamily G proteins, ABCG40 and ABCG25, mediate ABA influx and efflux, respectively (Kuromori et al., 2010). The ABA-IMPORTING TRANSPORTER1 (AIT1), also known as NRT1.2 or At-NPF4.6 (Tsay et al., 2007; Kanno et al., 2012; Léran et al., 2014), mediates influx as well, demonstrating the variety of transporter types that are able to mediate ABA transport. Recently, the ABC subfamily C (ABCC) transporters ABCC1 and ABCC2 have been shown to transport ABA-GE into the mesophyll vacuole (Burla et al., 2013).

Despite the importance of ABA signaling to root growth and development, we know very little about the biosynthesis or mobilization of ABA in root tips in response to environmental signals. Analysis of this problem has been hampered by the lack of an efficient method to directly visualize the location of ABA in plant tissues. Until now, approaches have largely involved indirect methods, such as monitoring the expression of ABA-induced genes, which reflects both the responsiveness of the tissue as well as the presence of ABA (Ishitani et al., 1997; Xiong et al., 1999; Christmann et al., 2005; Duan et al., 2013), or the expression of ABA biosynthetic genes, which indicates the possibility of ABA biosynthesis (Frey et al., 2012). Direct approaches include ABA biosensors (Jones et al., 2014; Waadt et al., 2014) and ABA immunolocalization approaches (Zhang et al., 1996; Schraut et al., 2004; Peng et al., 2006).

Here, we used an optimized immunocytochemistry technique to directly visualize ABA in plant tissues and cells. We found that ABA accumulates in the Arabidopsis root tip in a characteristic pattern, with a peak in the endodermis. ABA biosynthesis is not required for this pattern, which can be recreated when ABA biosynthesis is blocked and exogenous ABA added. We found that NO₃⁻ strongly stimulates root tip ABA accumulation by stimulating release of ABA from the storage form, ABA-GE, via the ER-localized BG1 β-glucosidase. Finally, we used immunogold labeling to demonstrate that ABA accumulates in the cytoplasm near the ER. Together, these data provide a mechanism for NO₃⁻-regulated root growth via the regulation of ABA accumulation in the root tip, providing insight into the environmental regulation of root growth.

RESULTS

Immunofluorescence Reveals a Radial Pattern of ABA Localization in the Growing Root Tip

ABA regulates root growth in response to environmental signals such as drought, salt stress, and nitrate, affecting meristem size and cell differentiation and triggering the induction of ABA-responsive gene expression (Signora et al., 2001; Sharp et al., 2004; Zhang et al., 2010; Duan et al., 2013; Geng et al., 2013). However, the spatial localization of ABA within the growing root tip is unknown. We developed an optimized immunofluorescence technique using a commercial rabbit anti-abscisic acid antibody (anti-ABA) that recognizes ABA and ABA conjugates (Hradecká et al., 2007; Turecková et al., 2009). By using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), which activates carboxyl groups to conjugate to nearby amino groups, we immobilized ABA within the tissue (Sotta et al., 1985; Sossountzov et al., 1986; Peng et al., 2006). ABA conjugates such as ABA-GE are formed by the addition of the conjugating molecule to the ABA carboxyl group, thus blocking the carboxyl group, thereby rendering the ABA unavailable for EDC-mediated cross-linking, which leads to its loss during subsequent tissue processing steps. Using confocal microscopy, we found that ABA localizes to the growing Arabidopsis root tip, where it is found in the central portion of the root, with a maximum level in a single cell layer surrounding the root stele (Figure 1; Supplemental Figures 1 and 2). The outer cell layers of the roots displayed almost no immunoreactivity, indicating that ABA is localized within the central root tissues (Figure 1; Supplemental Figures 1 and 2). Addition of exogenous ABA increased the immunofluorescence (Figure 1B; Supplemental Figure 1) and treatment with fluridone (FLU), an inhibitor of ABA biosynthesis (Bartels and Watson, 1978; Han et al., 2010; Talboys et al., 2011), significantly reduced the immunoreactivity of ABA in the growing root tip (Figure 1C; Supplemental Figure 1G). Similarly, aba2-1 mutants, which are defective in ABA biosynthesis (Léon-Kloosterziel et al., 1996; Cheng et al., 2002), had almost undetectable levels of ABA immunofluorescence (Supplemental Figure 3). Adding exogenous ABA to FLU-treated roots or to aba2-1 mutants restored ABA immunofluorescence (Figure 1D; Supplemental Figure 3), indicating that this technique is specific for ABA. Quantification of fluorescence intensity along a single transverse or vertical line through the root tip confirmed the presence of an ABA peak in a single cell layer surrounding the stele, with lower levels in the stele and the virtual absence of ABA immunoreactivity in the epidermis, cortex, and root cap (Figures 1M to 1P).

Figure 1.

ABA Localization in Situ with Immunofluorescence in Growing Root Tips.

(A) to (P) Confocal micrographs and corresponding line scan graphs showing ABA localization in situ with immunofluorescence in 7-d-old primary root tips in Col-0. Position of transverse line scan (green) is indicated by a dashed green double arrow in (A). Position of vertical line scans (yellow) is indicated by the dashed yellow double arrow in (A). White arrowheads indicate the endodermis, and green arrowheads point to the highest peak of fluorescence intensity. Brackets indicate the absence of ABA staining in the root cap.

(A) to (D) ABA localization using anti-ABA/Alexa Fluor 488.

(E) to (H) Bright-field micrographs.

(I) to (L) Merged images of anti-ABA/Alexa Fluor 488 and bright-field.

(A), (E), and (I) Control roots (0.5× MS).

(B), (F), and (J) ABA-treated roots (2 d).

(C), (G), and (K) FLU-treated roots (2 d).

(D), (H), and (L) Roots treated with FLU for 1 d followed by ABA for 1 d.

(M) to (P) Plots of transverse and vertical line scans of ABA fluorescence intensity of images in (A) to (D), respectively.

(Q) to (V) Confocal micrographs and corresponding line scan graphs showing colocalization of ABA accumulation (magenta) and ProSCR:erGFP expression (green) in the endodermis of ProSCR:erGFP lines.

(Q) ProSCR:erGFP root stained with anti-ABA/Alexa Fluor 555 to show ABA localization (magenta arrowheads).

(R) ProSCR:erGFP expression localized to the endodermal layer (green arrowheads).

(S) Merged image of ABA/Alexa Fluor 555 and ProSCR:erGFP expression indicating colocalization.

(T) and (U) Plots of transverse line scans showing ABA/Alexa Fluor 555 and ProSCR:erGFP fluorescence intensities (dashed magenta and green double arrows in [Q] and [R], respectively). White arrowheads indicate peaks of highest fluorescence intensity of both ABA and ProSCR:erGFP observed in the endodermal layer.

(V) Overlay of line scans from (T) and (U) showing a strong correlation between the spatial localization of ABA and ProSCR:erGFP expression.

RI, relative intensity. Treatment concentrations: 10 µM ABA and 5 µM FLU. Bars = 50 µm in (A) to (L) and (Q) to (S).

The Root Tip ABA Maximum Is in the Endodermis

We consistently observed an ABA maximum in a cell layer surrounding the stele. To determine the identity of this layer, we transformed plants with the transcriptional Pro-SCARECROW-erGFP (ProSCR-erGFP) fusion construct (Supplemental Figure 4) to label the endodermis and quiescent center with GFP (Wysocka-Diller et al., 2000). We examined the ABA immunofluorescence in the proSCR-erGFP roots and found that the ABA peak colocalized with ProSCR:erGFP expression in the root tip (Figures 1Q to 1S). Horizontal line scans confirmed that the GFP and ABA peaks were in the same position radially along the root (Figures 1T to 1V), indicating that the ABA maximum is in a continuous band of cells. Higher magnification revealed strong ABA immunofluorescence in the cortical/endodermal initial and the endodermal daughter cell and a weaker signal in the quiescent center (Supplemental Figure 2).

The SCR transcription factor is involved in radial patterning of the root meristem and is required for the fate of the endodermis and quiescent center (Di Laurenzio et al., 1996). Since the peak of ABA staining matches the pattern of SCR expression, we wondered whether ABA might also regulate SCR expression. We found that both ABA and nitrate decreased the fluorescence of the ProSCR-erGFP plants and that cotreatment with both ABA and NO₃⁻ strongly decreased the fluorescence in both the endodermis and the quiescent center (Supplemental Figure 4). Interestingly, ProSCR-erGFP plants treated with both ABA and NO₃⁻ formed root hairs nearer to the root tip, suggesting that meristem function may be compromised (Supplemental Figure 4).

ABA Biosynthesis Does Not Play a Major Role in Establishing the Root Tip ABA Pattern

We observed a consistent ABA immunofluorescence pattern in the root tip (Figure 1; Supplemental Figures 1 and 2) and found that even following a 2-d ABA treatment, the pattern was maintained, merely increasing in intensity (Figures 1B and 1N; Supplemental Figure 1G). Despite the continuous exposure of the epidermis to ABA in the medium, peak immunofluorescence remained in the endodermis, with the next highest level in the stele itself and only a weak signal in the outer epidermal and cortical layers (Figures 1B, 1J, and 1N). Conversely, in FLU-treated roots, the pattern remained the same, only much fainter (Figure 1C, 1K, and 1O). These observations suggest that the pattern of ABA in the root tip is actively maintained by biosynthesis, degradation, conjugation, or transport.

To test the role of ABA biosynthesis in establishing and maintaining the root tip ABA accumulation pattern, we added ABA to plants in which biosynthesis had been inhibited either genetically, with the aba2-1 mutation, or pharmacologically, with FLU treatment. We observed the regeneration of the typical ABA root tip pattern, with a maximum in the endodermis, strong immunofluorescence in the stele, and weak to no immunoreactivity in outer cell layers (Figures 1D and 1P; Supplemental Figures 1G and 3B). This observation suggests that ABA biosynthesis is not a key factor in establishing or maintaining the root tip ABA pattern.

NO₃⁻ Increases ABA Accumulation in the Growing Root Tip

Nitrogen is essential for plant growth, and the presence of NO₃⁻ in the medium outside the root has profound effects on root architecture, including the inhibition of lateral root elongation when present globally and the stimulation of lateral root elongation when present locally (Zhang and Forde, 1998). ABA signaling is required for the response of root branching to NO₃⁻ (Signora et al., 2001). We wondered whether increasing the NO₃⁻ concentration of the medium would alter ABA accumulation and localization. To test this, we transferred 5-d-old seedlings from half-strength Murashige and Skoog (MS) medium, which contains 20 mM NO₃⁻, to control medium (0.5× MS) or to 0.5× MS supplemented with either 10 mM KCl or an additional 10 mM KNO₃, resulting in a final concentration of 30 mM NO₃⁻. This concentration causes neither NO₃⁻ privation nor NO₃⁻ excess and falls within the range of standard growth conditions used in many Arabidopsis studies (i.e., still 10 mM less NO₃⁻ than in full-strength MS medium and 5 mM more than in 1× B5). We harvested roots after 2 d and found that increasing the NO₃⁻ concentration in the medium increased the ABA signal ∼3-fold (Figures 2C and 3A; Supplemental Figure 5C). In fact, the overall ABA fluorescence signal was higher in NO₃⁻-treated roots than in those treated with ABA (compared with Figures 2B, 2C, and 3A). In each case, the ABA localization pattern was maintained, with a peak in the endodermis and moderate staining in the stele, but the level in each tissue was higher than in the control (Figures 2 and 3; Supplemental Figure 5). This change in ABA levels in response to increased NO₃⁻ was due to the added 10 mM NO₃⁻, not the potassium, as adding an equivalent amount of KCl resulted in an ABA immunofluorescence pattern indistinguishable from that of the control (Figures 2A and 3A; Supplemental Figure 5A).

Figure 2.

NO₃⁻ Stimulates ABA Accumulation in the Root Tip Even When ABA Biosynthesis Is Blocked.

Confocal micrographs and corresponding line scan graphs showing ABA localization in situ with immunofluorescence in 7-d-old wild-type primary roots.

(A) to (F) ABA localization in roots using anti-ABA/Alexa Fluor 488.

(G) to (L) Bright-field micrographs of roots.

(M) to (R) Merged images of anti-ABA/Alexa Fluor 488 and bright field.

(A), (G), and (M) Roots treated 2d with KCl.

(B), (H), and (N) Roots treated for 2 d with ABA.

(C), (I), and (O) Roots treated for 2 d with KNO₃.

(D), (J), and (P) Roots cotreated for 2 d with ABA and KNO₃.

Roots treated for 1 d with FLU, followed by 1 d of either KNO₃ treatment ([E], [K], and [Q]) or FLU plus KNO₃ ([F], [L], and [R]). Note the increased ABA immunoreactivity in the root tip after KNO₃ treatment, with or without FLU treatment ([D] to [F]).

(S) to (X) Plots of transverse and vertical line scans showing Alexa Fluor 488 fluorescence intensity of ABA labeling. Green plots represent transverse line scans measured as indicated by dashed green double arrow in (A). Yellow plots represent vertical line scans measured as indicated by dashed yellow double arrow in (A). White and green arrowheads indicate the highest peak of immunoreactivity of ABA in the endodermal layer. Brackets indicate the absence of ABA staining in the root cap. Asterisks indicate increased immunoreactivity of ABA in the root cap of KNO₃-treated roots.

(T) to (X) Vertical line scans (yellow) indicate a strong immunoreactivity of ABA observed in the root elongation zone of both ABA- and KNO₃-treated roots (yellow arrowheads).

RI, relative intensity. Treatment concentrations: 10 mM KCl, 10 mM KNO₃, 10 µM ABA, and 5 μM FLU. Bars = 50 µm in (A) to (R).

Figure 3.

Stimulation of ABA Accumulation by NO₃⁻ at the Growing Root Tip.

(A) ABA immunofluorescence intensity plotted as a percentage of intensity in control roots or roots treated for 2 d with either KCl, ABA, or KNO₃. Fluorescence intensity was measured as the sum of total fluorescence of the transverse or vertical line scans of treated and untreated roots (n = 10 roots for each column). Positions of line scans on the root are indicated by double arrows in Figure 2A. Different letters denote a statistically significant difference among means at P < 0.05 according to two-way ANOVA (Tukey's test). Experiments were done in triplicate. Error bars represent the mean ± se. (B) Competitive ABA ELISA immunoassay showing the amount of root tip ABA plotted as the least squares means (LSM) log₁₀ of pmoles ABA/g fresh weight from 1-cm root tips treated for 2 d with either KCl, ABA, or KNO₃ (n = ∼400 root tips pooled for each sample). Results represent the LSM log₁₀ ABA concentrations of three independent experiments. Error bars represent the se of the LSM log₁₀ ABA concentrations. Letters above the error bars indicate the levels of difference; LSM log₁₀ values not connected by the same letter are significantly different P < 0.0001 with two-way ANOVA followed by Tukey’s HSD test. Treatment concentrations: 10 mM KCl, 10 mM KNO₃, and 10 µM ABA.

When roots grown on control medium were cotreated with 10 µM ABA and 10 mM NO₃⁻, the effect was additive, and the ABA immunofluorescence was stronger than with either individual treatment, such that the pattern with a peak in the endodermis was less obvious (Figure 2D). However, a line scan across the root tip revealed the pattern, with a peak in the endodermis and reduced immunoreactivity in the stele (Figure 2V). ABA immunofluorescence was high in the cortex and epidermis of roots treated with both ABA and NO₃⁻ compared with untreated roots, but the overall level remained lower in the cortex than in the stele and highest in the endodermis (Supplemental Figure 5), which is consistent with the pattern in untreated root tips.

To confirm our ABA immunofluorescence results, we performed an ELISA assay based on a monoclonal antibody specific for ABA. We found that NO₃⁻ treatment increased ABA levels in the 1-cm root tip by 3 to 4 orders of magnitude (Figure 3B). These results are consistent with our quantification of fluorescence intensity of both transverse and vertical line scans, which show a strong increase in ABA fluorescence intensity in NO₃⁻-treated root tips (Figure 3A). The difference in amount of stimulation reported by the two techniques is likely due to the difference in sensitivity of ELISA versus immunofluorescence and the relative affinity of the different antibodies for their ligand.

NO₃⁻ Stimulates ABA Accumulation in the Growing Root Tip Even in the Absence of de Novo ABA Biosynthesis

We wondered whether the increased ABA accumulation in growing root tips in response to NO₃⁻ treatment was due to stimulation of ABA biosynthesis, altered ABA transport in and out of the root tip, inhibition of degradation, or a change in the conjugation/deconjugation of ABA. To test the role of ABA biosynthesis in the stimulation of root tip ABA levels in response to NO₃⁻, we used FLU to block biosynthesis (1 d treatment), thereby significantly reducing the ABA pool in the growing root tip (Figure 1C). Subsequent addition of nitrate for 1 d strongly stimulated root tip ABA levels (Figure 2E), even with continued FLU treatment (Figure 2F) or in aba2-1 mutants, which are defective in ABA biosynthesis (Léon-Kloosterziel et al., 1996) (Supplemental Figure 3C). This finding indicates that NO₃⁻ does not act by stimulating de novo ABA biosynthesis or by inhibiting degradation, but rather by regulating transport or deconjugation.

NO₃⁻ Regulates the Levels of Bioactive Root Tip ABA Pools by Regulating the Expression of BG1 to Stimulate Release of ABA from ABA-GE

Since NO₃⁻-induced ABA accumulation in the root tip does not require de novo ABA biosynthesis, we wondered whether the increase in root tip ABA levels is a consequence of altered ABA transport in and out of the root tip or by release of ABA from conjugated stores. To test this, we examined the ability of NO₃⁻ to stimulate root tip ABA accumulation in plants deficient in either ABA transport or ABA-GE deconjugation. First, we asked whether plants defective in ABA transporters AIT1 and ABCG25 could stimulate root tip ABA accumulation in response to NO₃⁻. We found that both the ait1 and abcg25 mutants responded to NO₃⁻ with an increase in ABA accumulation (Supplemental Figures 6D to 6W), ruling out a major role for either the AIT1 or ABCG25 transporters in NO₃⁻-stimulated ABA accumulation. However, the response was weaker in the ait1 mutants than in the wild-type, suggesting that AIT1-mediated ABA transport may make some contribution to NO₃⁻ stimulation of root tip ABA levels.

We then examined ABA accumulation in plants mutant for BG1 (Lee et al., 2006), which encodes an ER-localized β-glucosidase that hydrolyzes biologically inactive glucose-conjugated ABA to produce bioactive ABA. The loss of BG1 activity blocks release of ABA from the ABA-GE-conjugated form, leading to a low level of bioactive ABA (Lee et al., 2006). We genotyped three bg1 T-DNA insertional lines, Salk_024344C (bg1-3), Salk_075731C (bg1-2), and Salk_122533 (bg1-1) (Lee et al., 2006; ,Supplemental Figure 6A), using PCR, RT-PCR and qRT-PCR (Figure 4C; Supplemental Figure 7), which together confirmed that both bg1-2 and bg1-1 are homozygous for the insertion and lack the BG1 transcript and that both insertions cause loss-of-function mutations. The bg1-3 line produced smaller bands with RT-PCR, suggesting that cryptic transcription yields a truncated mRNA (Supplemental Figure 7C). Both bg1-1 and bg1-2 display shorter roots compared with the wild type (Figures 4A and 4B) and are insensitive to either the short or long-term effects of NO₃⁻ on root architecture (Supplemental Figure 8) and were thus chosen for further analysis. We found that NO₃⁻ was unable to stimulate accumulation of ABA in either bg1-2 or bg1-1 mutant roots (Figures 4D to 4O). These findings indicate that NO₃⁻ stimulation of ABA accumulation requires the function of the BG1 β-glucosidase and, thus, that the source of the NO₃⁻-induced ABA accumulation must be inactive ABA-GE pools in root tip cells.

Figure 4.

NO₃⁻ Is Unable to Stimulate ABA Accumulation in bg1 Mutant Root Tips.

(A) Phenotypes of 7-d-old wild-type, bg1-2, and bg1-1 seedlings.

(B) Graph of average primary root length.

(C) BG1 RNA levels in 7-d-old seedlings were quantified using qRT-PCR.

(D) to (O) Confocal micrographs and corresponding line scan graphs showing the in situ localization of ABA stained with anti-ABA/Alexa Fluor 488 in wild-type, bg1-2, and bg1-1 roots grown under control conditions ([D], [G], and [J]) or treated for 2 d with KCl ([E], [H], and [K]) or KNO₃ ([F], [I], and [L]). Insets of (D) to (L) show transverse line scans of the relative intensity (y axis) of anti-ABA/Alexa Fluor 488 fluorescence plotted relative to position (distance in pixels, x axis). Position of line scans on the root is indicated by arrowheads in the corresponding image.

(M) to (O) Fluorescence intensity plotted as a percent of intensity in control roots. Fluorescence intensity was measured as the sum of total fluorescence of the transverse line scan shown in the inset (n = 5 roots for each column).

(P) Expression of genes encoding ABA metabolic enzymes BG2 and ABA2, ABA-responsive genes ABI5, RAB18, and RD29A, and NO₃⁻-responsive genes NIA1, NIA2, and NIR1 in 7-d-old roots, as revealed by qRT-PCR. Error bars represent the mean ± se of three biological replicates, and different letters denote significant differences among means at P < 0.05 according to one- or two-way ANOVA (Tukey's test).

Treatment concentrations: 10 mM KCl and 10 mM KNO₃. Bars = 0.5 cm in (A) and 50 µm in (D) to (L).

To determine the mechanism of NO₃⁻ regulation of BG1, we asked whether increased NO₃⁻ levels in the environment could stimulate the expression of BG1 in roots. We found that NO₃⁻ strongly induced BG1 expression at all time points tested, with the greatest stimulation at 1 d (Figures 5A, 5D, and 5G). In contrast, NO₃⁻ had no effect on the expression of the BG2 β-glucosidase gene or the ABA biosynthetic gene, ABA2, in roots (Figures 5A, 5D, and 5G), suggesting that the BG1 is regulated by NO₃⁻ at the transcriptional level.

Figure 5.

Cross-Regulation of Gene Expression by ABA and NO₃⁻ in Primary Roots.

(A) to (I) Expression levels of ABA metabolic ([A], [D], and [G]), ABA-responsive ([B], [E], and [H]), and NO₃⁻-responsive ([C], [F], and [I]) genes. Expression levels were measured using qRT-PCR on 7-d-old wild-type roots after 2 h ([A] to [C]), 1 d ([D] to [F]), and 2d treatment ([G] to [I]) with 0.5× MS (control) or 0.5× MS supplemented with KCl, KNO₃, FLU, or ABA.

(J) to (U) Confocal micrographs of untreated and treated transgenic roots carrying ProRAB18:GFP. GFP fluorescence is indicated in green and propidium iodide counterstaining in red. Roots were treated for 1 d with the indicated treatments except for (N) and (O), which are time-course images of ProRAB18:GFP activity in KNO₃-treated roots (1, 2, and 3 to 4 d). Error bars represent the mean ± se of three biological replicates, and different letters denote significant differences among means at P < 0.05 according to two-way ANOVA (Tukey's test).

Treatment concentrations: 10 mM KCl, 10 mM KNO₃, 10 μM ABA, and 5 μM FLU. Bars = 50 µm in (J) to (U).

Since BG1 activity is required for NO₃⁻-induced ABA release, we tested whether BG1 regulates both ABA-responsive and NO₃⁻-responsive gene expression. We measured the expression of a set of three ABA-responsive genes, ABA-INSENSITIVE5 (ABI5) (Finkelstein and Lynch, 2000), RESPONSIVE TO ABA18 (RAB18) (Kim et al., 2011), and RESPONSIVE TO DESSICATION29A (RD29A) (Yamaguchi-Shinozaki and Shinozaki, 1993); three NO₃⁻-inducible genes, NITRATE REDUCTASE1 (NIA1) (Parinov et al., 1999), NIA2 (Wilkinson and Crawford, 1991), and NITRITE REDUCTASE1 (NIR1) (Hwang et al., 1997); the ABA biosynthesis gene ABA DEFICIENT2 (ABA2) (Léon-Kloosterziel et al., 1996); and ABA deconjugation enzyme BG2 (Xu et al., 2012; Figure 4P) in wild-type and bg1-1 and bg1-2 mutant roots using qRT-PCR. We found that all ABA-responsive genes tested, ABI5, RAB18, and RD29A, exhibited significantly lower expression in both mutants (Figure 4P), which is consistent with the reduced root ABA immunofluorescence in both mutants (Figures 4D to 4O). ABA2 expression was significantly reduced in the bg1-1 mutant, but not in the bg1-2 mutant; otherwise, the expression changes in both mutants were similar. We also found that the expression of the NO₃⁻-responsive gene, NIA2, but not NIR1 or NIA1, was also reduced in both bg1 mutant lines (Figure 4P). Thus, our data indicate that BG1 function is required for wild-type expression levels of both ABA-responsive and NO₃⁻-responsive genes.

ABA Stimulates the Expression of NO₃⁻-Responsive Genes and Acts Synergistically with NO₃⁻ to Stimulate ProRAB18:GFP Expression in the Growing Root Tip

Since NO₃⁻ stimulates ABA accumulation in the root tip, we asked whether ABA might function as a component of the NO₃⁻ signaling pathway. To test this possibility, we used qRT-PCR to examine whether ABA treatment would stimulate the expression of three NO₃⁻-responsive genes, NIA1, NIA2, and NIR1, in the root. We found that ABA significantly stimulated all three genes as early as 2 h after treatment (Figure 5C), with stronger induction of gene expression by 1 and 2 d (Figures 5F and 5I). These same conditions also stimulated the expression of ABA-responsive genes, RAB18 and RD29A, in roots (Figures 5B, 5E, and 5H). Since NO₃⁻ stimulates ABA accumulation, we asked whether NO₃⁻ regulates expression of these ABA-responsive genes. We found that NO₃⁻ stimulated the expression of the RAB18 ABA-responsive gene only after 2 d of treatment, but not before (Figure 5H). These results are consistent with our immunofluorescence and ELISA findings that show a strong increase in ABA accumulation in NO₃⁻-treated roots after 2d treatment (Figures 2C, 3A, 3B, and 4F), the point at which NO₃⁻ stimulates the expression of the NO₃⁻-responsive genes NIA1 and NIR1.

To confirm our ABA immunofluorescence and gene expression results, we analyzed the activity of the ABA-responsive reporter gene fusion ProRAB18:GFP (Kim et al., 2011) in response to both ABA and NO₃⁻. We find that both NO₃⁻ and ABA stimulated ProRAB18:GFP expression in the meristem and elongation zone of primary roots and emerging lateral roots (Figures 5J to 5U; Supplemental Figure 9), but the timing of the response to NO₃⁻ and ABA differed: ProRAB18:GFP fluorescence was stimulated by ABA at 1 d (Figure 5R) and by NO₃⁻ at 2 d (Figure 5N). When plants were cotreated with both ABA and NO₃⁻, root tip expression was strongly stimulated after 1d (Figure 5T), indicating that ABA and NO₃⁻ function additively to regulate RAB18 expression.

The Subcellular Localization of ABA Is Associated with Cytoplasmic Structures Adjacent to the ER, but Not within the ER Compartment

In order to determine the subcellular localization of ABA, we used immunogold labeling and electron microscopy of both high-pressure freeze substituted (Figures 6A to 6E) and chemically fixed (Figures 6F to 6I) untreated and NO₃⁻-treated wild-type roots. The two different techniques for preparing the samples allowed us to examine different factors affecting the localization of ABA. In the high-pressure freeze substituted samples, there is no cross-linking step; thus, subcellular localization of the epitope is not dependent on the proximity of molecules or structures to cross-link to. However, this technique will capture the localization of unconjugated as well as conjugated ABA. The chemical fixation technique (see Methods) includes a cross-linking step with EDC, thus retaining free ABA in the tissue but allowing much of the conjugated ABA to wash out. However, this technique requires molecules or structures to be near the ABA molecules so that they can be cross-linked and retained. When analyzed together, we found that ABA localizes to the cytoplasm (Figures 6A to 6H), mitochondria (Figure 6B), nucleus (Figure 6C), and plastids (Figure 6D). This localization is consistent with what has been reported for ABA localization in Arabidopsis nurse cells and Chenopodium polyspermum tissues (Sossountzov et al., 1986; Peng et al., 2006), with the exception that our immunoelectron labeling, via both fixation methods, revealed numerous gold particles organized in clusters near ER tubules (Figures 6A, 6B, and 6E to 6H). This cytoplasmic localization of ABA near the ER supports the hypothesis that one source of bioactive ABA is the release of unconjugated ABA from inactive stores in the ER. Localization of ABA in specific cellular compartments with a basic pH, such as plastids or mitochondria, is consistent with the anion trap hypothesis (Slovik et al., 1995). To test the effect of NO₃⁻ treatment on ABA accumulation at a cellular level, we quantitatively analyzed the number of gold particles in endodermal and stele cells of untreated and NO₃⁻-treated roots (Figures 6F to 6I). We found that the number of gold particles increased significantly in both the endodermis and stele after NO₃⁻ treatment, with a fold increase of 2.02 and 2.4, respectively (Figures 6G to 6I). This observation indicates that NO₃⁻ treatment stimulates ABA accumulation in both endodermal and stele cells, confirming the results obtained by immunofluorescence and ELISA (Figure 3).

Figure 6.

ABA Localizes to Plastids, Nuclei, Mitochondria, and to Cytoplasmic Structures Near the ER, but Not within the ER Compartment.

(A) to (H) Immunogold labeling of high-pressure frozen/freeze substituted ([A] to [E]) and chemically fixed ([F] to [I]) 7-d-old Col-0 root cells using anti-ABA antibodies visualized with gold nanoparticle-conjugated secondary antibodies. All electron micrographs are of the root tip. White arrowheads mark 15-nm gold particles. Endodermal cell (A). Note the clusters of gold particles near the ER tubules. ABA localizes to the cytoplasm and mitochondria (B), nucleus (C), plastids, and occasionally the cell wall (D). Note that gold particles are not seen in vacuoles (D) or Golgi (E).

(F) to (H) Low-magnification micrographs showing ABA localization in control (0.5× MS) and KNO₃-treated root cells. Control endodermal cell (F), KNO₃-treated endodermal cell (G), KNO₃-treated stele cell (H), and quantitative analysis (I) of immunogold labeling in KNO₃-treated and untreated cells. Gold particle density per µm² of both endodermal and stele cells is given. Error bars represent the mean ± se of three biological replicates (n = 10 cells), and different letters denote significant differences among means at P < 0.05 according to two-way ANOVA (Tukey's test).

EN, endodermis; CW, cell wall; C, cytoplasm; Mt, mitochondria; nCW, new cell wall deposition; G, Golgi; N, nucleus; Pl, plastid; RC, root cap cell; St, stele; V, vacuole. Bars = 0.5 µm in (A) to (H).

DISCUSSION

Although environmental regulation of root growth is a major determinant of root architecture (Slovik et al., 1995; Jovanovic et al., 2008; Gruber et al., 2013), we are only beginning to uncover the mechanism by which sensing of environmental inputs is translated into changes in growth and development. Nitrogen is an essential component of major macromolecules, and, as such, plant foraging for nitrogen drives changes in root architecture. Local changes in rhizosphere NO₃⁻ concentration are sensed by the NRT1.1 (NPF6.3) nitrate sensor, which also transports auxin, thus providing a direct link between NO₃⁻ sensing, auxin accumulation, and root architecture (Krouk et al., 2010). Genetic analysis clearly places ABA downstream of NO₃⁻ perception as well in the regulation of root architecture (Signora et al., 2001; De Smet et al., 2003), but the mechanistic link between NO₃⁻ perception and ABA regulation has not been determined. Our work provides evidence for an ABA root tip pattern that is actively maintained and whose levels change in response to environmental changes in NO₃⁻ concentration. Furthermore, these findings support a model in which the endodermis is the primary site of ABA function in the root, and they reveal a mechanistic link between NO₃⁻ sensing and ABA signaling that involves the release of bioactive ABA from inactive ABA-GE stores.

An Endodermal ABA Root Tip Pattern That Is Actively Maintained

In order to visualize ABA localization within plant tissues, we developed an optimized immunolocalization protocol using commercially available anti-ABA polyclonal antisera. The advantage of this approach is that ABA itself is directly visualized within plant tissues. The recently developed ABA biosensors also provide a direct approach to obtaining very sensitive, real-time readouts of ABA levels (Jones et al., 2014; Waadt et al., 2014). ABA biosensors are powerful tools, but their application also requires a transformation system, thus limiting their use to well-developed genetic systems (Jones et al., 2014; Waadt et al., 2014). The ABA immunolocalization approach, although used on fixed tissue, is not species specific and does not depend upon having a well-developed transformation system.

ABA immunolocalization revealed a root tip pattern with a peak in the endodermis, moderate levels in the stele, and low levels in the cortex, epidermis, and root cap (Figure 1). Our data indicate that this pattern is actively maintained and that ABA biosynthesis is not required to generate the pattern. When ABA biosynthesis is blocked, either by mutation (aba2-1) or the addition of FLU, and ABA levels are depleted, the addition of ABA to the medium is sufficient to recreate the pattern (Figure 1; Supplemental Figure 3). Even though when applied exogenously, ABA is absorbed by the outer tissues of the root, the ABA peak nonetheless remains in the endodermis with a moderate level in the stele and a lower level in the outer tissues, just as it is in untreated roots (Figure 1), implicating a radial transport system necessary to move ABA from the outer tissues of the root to the inner tissues of the endodermis and stele. Why does ABA accumulate with a peak in the endodermis? The endodermal accumulation of ABA may simply reflect the intersection of two ABA transport pathways: one from the outside of the root toward the center and the other from the stele outwards, both of which end at the endodermis. Alternatively, trapping ABA in the endodermis, or degradation of ABA in other tissues, could also be a factor. Either way, the root tip ABA accumulation pattern must be actively maintained because no matter how ABA levels are restored to ABA-depleted roots, whether by adding ABA exogenously or by treating roots with nitrate to stimulate internal release of ABA into root tissues, the overall ABA root tip accumulation pattern is recreated (Figures 1 and 2).

Our observation that this pattern of ABA accumulation revealed by immunolocalization (Figure 1) differs from the expression pattern of the ABA-responsive RAB18 promoter (strong expression throughout the root tip, including strong expression in the epidermis and cortex; Figures 5J to 5U) suggests that there may be cell-type-specific differences in responsiveness to ABA or other developmental or environmental factors that cause the difference between ABA localization and ABA-induced gene expression changes. Thus, these types of analyses can be complementary, and they provide an added dimension to our understanding of ABA signaling in the root.

ABA’s Role in the Endodermis

Our observation that ABA accumulates predominantly in the endodermis addresses a long-standing question about the role of ABA in roots and the function of the endodermis in ABA responses. The endodermis is a specialized tissue that regulates not only entry of water and solutes to the vasculature but also root growth under normal and salt-stressed conditions (Dinneny, 2014). Duan et al. (2013) have shown that ABA is required within the root endodermis to regulate root growth in response to NaCl. This observation could reflect either increased ABA levels within the endodermis, increased sensitivity of the endodermis to ABA, resulting in the activation of ABA responses at a lower threshold, or a requirement for ABA solely within the endodermis, since endodermal cell expansion is rate-limiting for root elongation (Ubeda-Tomás et al., 2008); thus, ABA signaling in other root tissues may be irrelevant for root growth. Our observation that ABA accumulates particularly within the endodermis supports the hypothesis that increased ABA levels within the endodermis mediate root growth in response to NaCl and underscores the importance of the endodermis in root physiology. Our data indicate that ABA accumulation within the endodermis is actively maintained and is likely to be responsible for some of its unique functions.

The transcription factor SCR is required for the determination of endodermal cell fate and has recently been shown to also regulate ABA signaling in the root by repressing expression of the ABI4 and ABI5 transcription factor genes (Scheres et al., 1995; Di Laurenzio et al., 1996; Cui et al., 2012). SCR is expressed in a sheath of cells in the root tip, comprising the endodermis, endodermal/cortical initials and the quiescent center (Di Laurenzio et al., 1996). This expression pattern precisely overlaps with the peak of ABA accumulation in the root (Figure 1), leading us to wonder whether ABA accumulation might also modulate SCR gene expression. We found that ABA and NO₃⁻ each individually reduced the expression of a ProSCR:erGFP fusion and, when applied together, strongly reduced expression of the SCR reporter gene (Supplemental Figure 4). This observation suggests that SCR not only acts upstream of ABA signaling but is itself regulated by ABA. In this way, ABA and NO₃⁻ could stimulate ABI4 and ABI5 expression by inhibiting an inhibitor, SCR. This observation also suggests that ABA might modulate not only the function of the endodermis, but perhaps also its developmental fate.

ABA as the Mediator of NO₃⁻ Responses in the Growing Root Tip

NO₃⁻ is one of the most studied metabolic and environmental signals in the root environment. NO₃⁻ modulates root architecture, the root:shoot ratio, flowering time, plant-microbe interactions, and root hair tip growth among other processes (Linkohr et al., 2002; Bloch et al., 2011; Castro Marín et al., 2011; Camañes et al., 2012; Vitor et al., 2013). Many of these physiological processes are also regulated by ABA, an important mediator of environmental signals. However, the relationship between NO₃⁻ and ABA has been opaque. Genetic analysis placed ABA signaling downstream of NO₃⁻ perception in the regulation of root architecture (Signora et al., 2001; Zhang et al., 2007), but upstream of nitric oxide production, in several physiological processes (Desikan et al., 2002; León et al., 2014). The link between environmental NO₃⁻ and the triggering of ABA signaling within the root, however, has just begun to be developed. Two NPF family members have recently been identified that transport both ABA and nitrate, thus potentially linking nitrate transport with ABA signaling. NPF6.8 transports both nitrate and ABA and mediates the Medicago truncatula nitrate response in the root modulating root length, implicating a function for ABA in potentially mediating the nitrate response, although the mechanism is unknown (Pellizzaro et al., 2014). The root-expressed NPF4.6/AIT1/NRT1.2 protein transports both ABA and nitrate in Arabidopsis, but nitrate has no effect on ABA transport, and the affinity of this transporter for ABA is several orders of magnitude higher than for nitrate, providing no obvious role for nitrate transport in this system (Huang et al., 1999; Kanno et al., 2012, 2013). Recently, the nitrate/auxin transporter NPF6.3 (NRT1.1) has been shown to interact with the ABA coreceptor, ABI2, such that ABA represses nitrate uptake by NRT1.1/AtNPF6.3 by inactivating ABI2, thus further strengthening the link between nitrate sensing and ABA signaling (Ma et al., 2009; Park et al., 2009; Léran et al., 2015).

Our data indicate that NO₃⁻ strongly stimulates ABA signaling by inducing ABA accumulation in the root tip, particularly in endodermal cells, by stimulating the release of ABA from the inactive storage form, ABA-GE. We find that NO₃⁻ stimulates expression of the ER-localized β-glucosidase, BG1, and that loss of BG1 function completely blocks NO₃⁻-induced ABA accumulation (Figure 4). The observation that a mutation in BG1 completely blocks the increase in root tip ABA accumulation by NO₃⁻ (Figure 4) indicates that a contribution by the vacuolar β-glucosidase, BG2, to the ABA pool must be insufficient to compensate for a loss of BG1. BG1 is localized to the ER, and the accumulation of ABA immunoreactivity in the cytoplasm near ER tubules (Figure 6) is consistent with the origin of this ABA being ER-localized ABA-GE (Lee et al., 2006). BG1 functions in response to both drought and salt stress (Lee et al., 2006), but its expression is not enriched in the root endodermis in response salt or iron stress (Dinneny et al., 2008), suggesting that the accumulation of ABA in the endodermis is due to other factors. Our data reveal an additional role for BG1 in the response to changes in external NO₃⁻. Not only does NO₃⁻ treatment strongly stimulate BG1 expression (Figure 5), it also subsequently stimulates expression of the ABA-responsive gene RAB18 at 2 d (Figure 5H), most likely via release of ABA stores. Most importantly, ABA also regulates the expression of NO₃⁻-responsive genes involved in NO₃⁻ metabolism, NIA1, NIA2, and NIR1 (Figure 5). This finding places ABA squarely in the middle of the NO₃⁻ signaling pathway, downstream of NO₃⁻ perception and upstream of NO₃⁻-responsive gene expression (Figure 7). These findings add a mechanistic dimension to our understanding of NO₃⁻ signaling and to the role of ABA in mediating environmental responses. ABA is required for meristem maintenance in Arabidopsis and thus exerts a positive effect on root growth in addition to the inhibitory effect seen with exogenous ABA treatment (Zhang et al., 2010; Talboys et al., 2011). Perhaps nitrate-induced ABA release confers a more fine-tuned regulation of ABA accumulation, due to slow, intracellular release.

Figure 7.

Proposed Model of NO₃⁻ Induction of ABA Accumulation via BG1-Mediated ABA-GE Deconjugation.

An increase in NO₃⁻ in the root environment initially stimulates expression of BG1, resulting in the slow release of bioactive ABA from the storage form ABA-GE, such that by day two, ABA has accumulated significantly in the root tip, and expression of the ABA-responsive RAB18 gene and the NO₃⁻-responsive NIA1 and NIR1 genes is induced. High levels of ABA continue to stimulate BG1 expression, maintaining root tip ABA accumulation.

Plants have evolved tremendous plasticity in root architecture in response to an inconsistent and changing environment. In addition, different plant taxa have evolved a variety of root architectures and strategies for success in response to different environmental changes. For example, legumes and non-legumes differ in root architecture responses to exogenous ABA (Liang and Harris, 2005). While the evolutionary origins of this difference are murky, we now have a tool to assess whether the differences in root architecture responses to ABA are due to changes in ABA localization or altered responsiveness to ABA and how this affects root responses to changes in environmental NO₃⁻ concentration. The immunolocalization approach we developed to visualize ABA spatially within plant tissues is platform independent, unlike genetic and genomic tools, which are tied to a specific genome and cannot be easily used for comparative analysis. Thus, our ABA localization technique can be used for comparative physiology to examine physiological responses of different plant taxa to changing environmental inputs. This approach may ultimately allow us to shed some light both on the control of root architecture plasticity and on the diversity of plant form.

METHODS

Plant Materials, Growth Conditions, and Treatments

All Arabidopsis thaliana genotypes were in the Columbia (Col-0) accession, which was used as the wild type for all experiments. Transgenic lines T-DNA SALK insertion SALK_075731C and SALK_122533 for the BG1/BGL1 (At1g52400) gene were obtained from the ABRC. Seeds were surface-sterilized and grown on solid 0.5× MS basal salt mixture (Sigma-Aldrich) containing 1% (w/v) sucrose and 1% (w/v) Phytagel (Sigma-Aldrich) under a 16-h-light/8-h-dark photoperiod with a light intensity of 100 µmol m⁻² s⁻¹ at 20°C in a Conviron MTR30 growth chamber. At 5 d, the plants were transferred to fresh 0.5× MS medium supplemented with either ABA [(±)-ABA; Sigma-Aldrich], KCl, KNO₃, or FLU (Sigma-Aldrich) and grown for an additional 2 d. For FLU treatment followed by either ABA or KNO₃ treatment, seedlings were grown on 0.5× MS for 5 d, transferred to fresh plates containing 0.5× MS supplemented with 5 µM FLU, and grown for 1 d and then transferred again to fresh plates containing 0.5× MS with the second treatment and grown for another day. Plates were sealed with either Parafilm or surgical tape (Micropore) and placed vertically in the growth chamber. Concentrations of treatments were always as follows: 10 μM ABA, 10 mM KCl, 10 mM KNO₃, and 5 μM FLU. Note that the 0.5× MS already contains 20 mM NO₃⁻ (Murashige and Skoog, 1962), and an additional 10 mM NO₃⁻ results in a final concentration of 30 mM NO₃⁻. This modest change in NO₃⁻ concentration did not affect the growth of the root system; supplementation with exogenous ABA or FLU did result in a small but significant decrease in root length (Supplemental Figure 10).

Genotyping and Phenotyping Arabidopsis T-DNA SALK Insertion Lines

All Arabidopsis T-DNA mutant lines used for this experiment were grown on soil for 1 month, after which rosette leaves were harvested for genomic DNA (gDNA) isolation. One rosette leaf was harvested per genotype replicate for each condition tested. For PCR analysis, gDNA was isolated as previously described by Edwards et al. (1991). The gDNA quality and yield were assessed using a NanoDrop1000 spectrophotometer (Thermo Scientific). PCR was performed as previously described by Andème Ondzighi et al. (2008) and was optimized to produce a low-copy amplification of the gene product that was then analyzed by agarose gel electrophoresis. All of the T-DNA insertions were identified via PCR of Arabidopsis genomic DNA. The homozygous T-DNA lines were verified by PCR analysis of genomic DNA of self-fertilized plants using the left border primer and a gene-specific primer. For the RT-PCR, leaves were frozen over liquid nitrogen and ground to a powder using a micro-pestle. The RT-PCR was performed as previously described by Andème Ondzighi et al. (2008). The Ef-1α gene was used as an internal control. All gene-specific primers used in this study are listed in Supplemental Table 1. For phenotypic analysis, wild-type and bg1-1, bg1-2, and bg1-3 mutant plants were grown for 2 months and analyzed for morphological changes.

Primary Root Length Growth Analysis

Arabidopsis Col-0 seedlings were grown in a square Petri dish (10 × 10 × 1.5 cm³) containing 0.5× MS for 5 d, and then seedlings with roots of the same length were transferred onto 0.5× MS supplemented with the compounds of interest (10 µM ABA, 5 µM FLU, 10 mM KCl, 10 mM KNO₃, and both 10 µM ABA and 10 mM KNO₃) and grown for an additional 2 d. For FLU treatment followed by either ABA or KNO₃ treatment, seedlings were grown on 0.5× MS for 5 d, transferred to fresh plates containing 0.5× MS supplemented with 5 µM FLU, grown for 1 d, and then transferred again to fresh plates 0.5× MS containing either 10 µM ABA or 10 mM KNO₃ or both 10 mM KNO₃ and 5 µM FLU, and grown for another 1 d. Root length was recorded for each root from the top portion at the base of the hypocotyl to the root tip. Seven-day-old untreated and treated seedlings were scanned and the total length of root was determined using ImageJ. For the analysis of bg1 mutants, 7-d-old Col-0 wild-type and both bg1 mutants, bg1-1 and bg1-2, were grown together in the same square Petri dish containing 0.5× MS. To investigate whether nitrate has an effect on root growth or development of bg1 mutants, we subjected plants either to a transient, 2-d nitrate treatment or a continuous 7-d nitrate treatment. For the transient nitrate treatment, we grew Col-0 wild-type and both bg1-2 and bg1-1 seedlings on 0.5× MS control medium for 5 d and then transferred them to either 0.5× MS supplemented with either 10 mM KCl or 10 mM KNO₃ and grew them for an additional 2 d. For the continuous nitrate treatment, we grew seedlings continuously for 7 d on 0.5× MS supplemented or not with either 10 mM KCl or 10 mM KNO₃. Root length was recorded for each root by scanning seedlings and the total root length was determined using ImageJ. Statistical analysis was performed using a two-way ANOVA (Tukey’s test) with Graph Pad Prism software version GPW6 to determine significance of the final root length with 95% confidence. All experiments in this study were performed in triplicate.

Immunofluorescence and Confocal Microscopy

Immunofluorescence analysis of primary root growth in 7-d-old Arabidopsis primary roots was performed as previously described by Andème-Onzighi et al. (2002) with a few modifications. Whole 7-d-old Arabidopsis seedlings were vacuum-infiltrated for 2 h and then fixed overnight at 4°C in 50 mM sodium PBS, pH 7.2, containing 4% (v/v) paraformaldehyde (EMS), 0.2% (v/v) glutaraldehyde (Sigma-Aldrich), and 2% (v/v) EDC (Thermo Scientific), an essential step allowing the fixation of ABA in situ, via cross-linking by its carboxyl group (Sotta et al., 1985; Sossountzov et al., 1986; Peng et al., 2006). Fixed roots were washed three times with PBS and then were simultaneously blocked and permeabilized with 3% (w/v) nonfat dried milk solution in 0.01 M PBS (PBS-milk), pH 7.2, containing 0.1% cellulose, 0.1% pectinase, and 0.1% Triton for 1 h. Permeabilized root tips were washed three times with PBS and incubated with 0.25 µg anti-ABA polyclonal antibodies (rabbit anti-ABA; Hradecká et al., 2007; Turecková et al., 2009) (C1, AS09446; Agrisera) overnight at 4°C. The stained roots were washed and transferred to a 400-fold dilution of secondary antibody goat anti-rabbit IgG-conjugated to Alexa Fluor 488 (excitation, 488 nm; emission, 505 to 530 nm) (Invitrogen, Molecular Probes) or Alexa Fluor 555 (543-nm argon laser line and detected via a 560- to 615-nm band-pass filter [red]) for 2 h at room temperature. The stained roots were washed with PBS and mounted with Citifluor AF1 (Ted Pella) and then imaged by confocal microscopy using a Zeiss LSM 510 META Axiovert 200M inverted microscope with a Plan-Apochromat 20×/0.75 objective (Zeiss). GFP images were acquired using the same microscope and objective. GFP was excited with the 488-nm line of an argon laser and detected via a 505- to 530-nm band-pass filter (green). Propidium iodide was used to stain cell walls and was excited with the 543-nm line argon laser and detected via a 560-nm long-pass filter (red).

Quantification of Fluorescence Intensity

The relative fluorescence intensity was quantified using ImageJ software (http://rsb.info.nih.gov/ij/; Wang et al., 2012) on an original image of a root taken directly from the confocal microscope. Fluorescence intensity was measured transversely and vertically on each root, along a 10-pixel-wide line. A transverse line scan was done at a distance of 75 µm (100 pixels) from the quiescent region, and the vertical line was drawn from the root cap to the top of elongation zone of the root. Three repeated measurements were performed for each root and at least five roots from each treatment were used. The fluorescence intensity values were recorded and the significance was determined using a standard t test. P values reported are for two-tailed analyses.

Immunogold Labeling and Electron Microscopy

For immunogold labeling analysis, growing Arabidopsis root tips were preserved by either high-pressure freezing/freeze substitution techniques as described (Andème Ondzighi et al., 2008) or by chemical fixation. The chemical fixation method includes a cross-linking step with EDC as described above. In all control microscopy experiments, no fluorescence or gold particles were observed. These controls confirm the specificity of the primary and secondary antibodies for the root tip cells studied (Supplemental Figure 11).

For the chemical fixation, 1 to 2 mm of growing untreated and NO₃⁻-treated root tips were cut and fixed as previously described (Andème-Onzighi et al., 2002) with a few modifications. Cut root tip apices were pre-vacuum infiltrated for 2 h and fixed with a 2% aqueous solution of water-soluble EDC (Sossountzov et al., 1986), 4% (v/v) glutaraldehyde (EMS), and 1% (v/v) paraformaldehyde (EMS) in 0.1 M sodium cacodylate, pH 7.2, overnight at 4°C. Samples were washed and then postfixed with 1% osmium tetroxide (OsO₄) for 1 h. OsO₄-treated root tips were washed three times with 0.1 M sodium cacodylate, followed by a water wash, and dehydrated in an acetone series, 10, 20, 40, 60, 80, and 90%, then four times in 100% acetone. Acetone-rinsed samples were embedded in Eponate resin (Ted Pella), resin/acetone 5, 10, 25, 50, and 75%, and three times 100% (24 h for each concentration). Eponate-infiltrated root tips were polymerized at 60°C overnight in flat bottom embedding capsules.

For immunogold labeling, 80- to 90-nm either high-pressure freezing/freeze substituted or chemically fixed Epon resin-embedded sections of root tips were placed on formvar-coated gold or nickel slot grids and were etched with 0.5 M sodium periodate (NaIO₄; Sigma-Aldrich) for 10 min, thoroughly washed, and then blocked for 30 min with 3% (w/v) nonfat dried milk solution in 0.01 M PBS, pH 7.2, containing 0.1% Tween 20. The sections were washed and then incubated with 0.50 µg anti-ABA primary polyclonal antibodies (Agrisera) for 2 h at room temperature. Sections were washed and transferred to a 25-fold dilution of secondary antibody goat anti-rabbit IgG-conjugated to 15-nm gold particles (Ted Pella) for 2 h at room temperature. After washing, samples were counterstained by uranyl acetate. All observations were performed using a JEOL 1400 transmission electron microscope. Quantitative analysis of anti-ABA was performed using Fiji (http://fiji.sc/Fiji). Two independent grids were analyzed, and from each grid five randomly chosen endodermis or stele cells were imaged. Labeling density of the anti-ABA antibodies was estimated on five cells. The density of gold particle labeling was calculated as the average number of particles per µm². To assess the specificity of the primary anti-ABA antibody, all immunofluorescence and immunogold-labeling controls involved omission of the primary antibody, substitution with the preimmune serum, and preincubation of the primary antibody with 10 µM (±)-ABA. All controls gave negative results (Supplemental Figure 11).

ABA Measurement

Arabidopsis seedlings were grown as described above for 5 d, after which 1-cm root tips were harvested for ABA extraction. Approximately 400 root tips were pooled for each treatment tested (control 0.5× MS, 10 mM KCl, 10 µM ABA, and 10 mM KNO₃) and then were weighed (0.10 to 0.20 g), frozen in liquid nitrogen, and immediately subjected to the ABA extraction following a procedure adapted from Walker-Simmons (1987) and Liang et al. (2007). Root tips were ground to a fine powder using a micropestle in 2-mL Eppendorf tubes with liquid nitrogen. Powdered root tips were suspended in methanol containing 2.5 mM citric acid monohydrate and 0.5 mM 2,6-di-ter-butyl-4-methly-phenol (B1378; Sigma-Aldrich) at a ratio of 10 mg per 1.5 mL of methanol solution. The extract was then stirred for 48 h in the dark at 4°C and centrifuged at 1500g for 10 min at 4°C. The supernatant was recovered and adjusted to 70% methanol and passed through a C18 Sep-Pak cartridge (WT023501; Waters). The eluates were dried with a speed vac, resuspended in TBS, and diluted as described (Walker-Simmons, 1987; Liang et al., 2007). ABA was quantified using the Phytodek ABA ELISA kit (Agdia), following the manufacturer’s instructions. All samples from the three biological replicates were run in parallel. Statistical analysis was done using JMP Pro 11.2.0 (64-bit). Effects of the four different treatments and replicates were analyzed by two-way ANOVA. The difference between each pair of treatments was analyzed by Tukey’s HSD test at α = 0.05.

RNA Extraction and Quantitative RT-PCR

Seedlings were grown as described above, after which either whole seedlings or roots were harvested for RNA isolation. Approximately 360 seedlings were pooled per sample. Each graph represents the average of three biological replicates. Total RNA was isolated using a Qiagen RNeasy Plant Mini Kit (Qiagen) and subsequently treated with Turbo DNase-free (Ambion) to remove genomic DNA contamination, then cleaned and concentrated with a Qiagen RNeasy MinElute Cleanup Kit (Qiagen). All RNA samples were checked on a 2100 Bioanalyzer for quality and integrity. First-strand cDNA was synthesized from 1 µg of total RNA using an iScript kit (Bio-Rad). The quality and yield of cDNA was assessed using a NanoDrop1000 spectrophotometer (Thermo Scientific). Quantitative PCR amplifications were conducted with the ABI StepOnePlus real-time PCR system (Applied Biosystems), using the SYBR Green reagent (VWR). Data were collected and analyzed according to the DDCT method and normalized to the geometric mean of the expression of the ACTIN2 (ACT2) gene with SDS 2.2 software (Applied Biosystems). Nucleotide sequences of primers used are listed in Supplemental Table 2. Statistical analysis was performed using a two-way ANOVA with Graph Pad Prism software version GPW6.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (GPW6) and JMP Pro 11.2.0 (64-bit). Data are plotted as mean ± se of three biological replicates. For analysis of gene expression data, significance of differences between the two treatment groups was assessed using a two-way ANOVA. A Tukey’s test was used to determine the significance of root length differences with 95% confidence. The P values reported are for two-tailed analyses. Significant differences are indicated in the figures by different letters, P < 0.05. All experiments in this study were performed in triplicate.

Accession Numbers

The TAIR accession numbers for the sequences used in this study are as follows: BG1/BGL1 (At1g52400), BG2 (At2g32860), AIT1 (At1g69850), ABA2 (At1g52340), ABCG25 (At1g71960), ABI5 (At2G36270), RAB18 (At5G66400), RD29A (At5G52310), NIA1 (At1g77760), NIA2 (At1g37130), and NIR1 (At2g15620).

Supplemental Data

• Supplemental Figure 1. In Situ Localization of ABA Accumulation in the Endodermis of the Growing Root Tip.

• Supplemental Figure 2. In Situ Localization of ABA in the Quiescent Center of the Growing Root Tip.

• Supplemental Figure 3. NO₃⁻ Is Able to Stimulate Active ABA Pools in the Growing Root Tip in the Arabidopsis aba2-1 Mutant.

• Supplemental Figure 4. Effect of Nitrate and ABA on the Pattern of ProSCR:erGFP Activity in the Growing Root Tip of Arabidopsis.

• Supplemental Figure 5. The Peak of ABA Staining Colocalizes with Root Tip SCR Expression in Nitrate-Treated Arabidopsis ProSCR:erGFP Roots.

• Supplemental Figure 6. ABA Transporter Gene Structure, Genotypes, and Immunofluorescence in Situ of ABA in the Growing Root Tip in Arabidopsis ABA Transporter Mutant Lines.

• Supplemental Figure 7. Arabidopsis β-GLUCOSIDASE1 (BG1) Gene Structure, Genotypes, and Phenotypes of bg1 Mutant Lines.

• Supplemental Figure 8. The ABA-Deficient bg1 Mutant Is Insensitive to the Effects of NO₃⁻ on Root Architecture.

• Supplemental Figure 9. NO₃⁻ Strongly Stimulates ProRAB18:GFP Activity in Primary and Lateral Roots.

• Supplemental Figure 10. Effect of NO₃⁻, ABA, and Fluridone on Root Growth in Arabidopsis Seedlings.

• Supplemental Figure 11. Immunological Controls for ABA Antibody Specificity in the Growing Root Tip at the Tissue and Cellular Level.

• Supplemental Table 1. Primer Sequences Used for PCR Genotyping.

• Supplemental Table 2. Primer Sequences Used for RT-qPCR Experiments.

Glossary

ABA

abscisic acid

ABA-GE

ABA-glucose ester

ER

endoplasm reticulum

EDC

1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride

MS

Murashige and Skoog

FLU

fluridone

gDNA

genomic DNA

LSM

least squares means