Closely related receptor complexes differ in their ABA selectivity and sensitivity


For correspondence (fax +49 8161 715471; e-mail


The recent discovery of a variety of receptors has led to new models for hormone perception in plants. In the case of the hormone abscisic acid (ABA), which regulates plant responses to abiotic stress, perception seems to occur both at the plasma membrane and in the cytosol. The cytosolic receptors for ABA have recently been identified as complexes between protein phosphatases 2C (PP2C) and regulatory components (RCAR/PYR/PYL) that bind ABA. Binding of ABA to the receptor complexes inactivates the PP2Cs, thereby activating the large variety of physiological processes regulated by ABA. The Arabidopsis genome encodes 13 homologues of RCAR1 and approximately 80 PP2Cs, of which six in clade A have been identified as negative regulators of ABA responses. In this study we characterize a novel member of the RCAR family, RCAR3. RCAR3 was identified in a screen for interactors of the PP2Cs ABI1 and ABI2, which are key regulators of ABA responses. RCAR3 was shown to repress ABI1 and ABI2 in vitro, and to stimulate ABA signalling in protoplast cells. RCAR3 conferred greater ABA sensitivity to the PP2C regulation than RCAR1, whereas stereo-selectivity for (S)-ABA was less stringent with RCAR3 as compared with RCAR1. In addition, regulation of the protein phosphatase activity by RCAR1 and RCAR3 was more sensitive to ABA for ABI1 than for ABI2. Based on the differences we have observed in transcriptional regulation and biochemical properties, we propose a model whereby differential expression of the co-receptors and combinatorial assembly of the receptor complexes act in concert to modulate and fine-tune ABA responses.


Higher plants are sessile organisms that have evolved a high plasticity for adaptation to environmental challenges. Pathogens and abiotic stress such as drought and salt stress severely impact plant performance and productivity. The phytohormone ABA serves as an endogenous messenger in biotic and abiotic stress responses (Christmann et al., 2006; Melotto et al., 2006; Adie et al., 2007; Hirayama and Shinozaki, 2007). Abiotic stress such as drought and high salinity results in strong increases of ABA levels accompanied by a major change in gene expression and in adaptive physiological responses (Seki et al., 2002; Rabbani et al., 2003; Priest et al., 2006; Christmann et al., 2007). ABA is also required to fine-tune growth and development under non-stress conditions and regulates a large number of physiological processes such as seed dormancy, growth and stomatal adjustment. Stomatal aperture is regulated by an ABA-triggered alteration of ion fluxes in guard cells (Levchenko et al., 2005; Vahisalu et al., 2008). ABA regulates physiological responses in concert with other phytohormones such as gibberellic acid, auxin and ethylene (De Smet et al., 2003; LeNoble et al., 2004; Zhang et al., 2009). In the integration of hormone signalling, auxin, ethylene and ABA are thought to converge on the DELLA proteins, which act as a common crosstalk node (Santner and Estelle, 2009). The DELLA proteins are transcriptional repressors that interact with the gibberellin receptor in a gibberellin-dependent manner. ABA appears to inhibit the gibberellic acid-mediated degradation of the DELLA proteins (Achard et al., 2006).

ABA is able to redirect the expression of approximately one-tenth of the Arabidopsis genome (Hoth et al., 2002; Nemhauser et al., 2006). A large number of ABA signalling components have been identified (Christmann et al., 2006; Hirayama and Shinozaki, 2007). Major players in ABA signalling are a subclass of Mg2+- and Mn2+-dependent serine/threonine phosphatases type 2C (PP2Cs). There are approximately 80 PP2Cs in Arabidopsis (Schweighofer et al., 2004) and six of the nine PP2Cs in clade A have been identified as negative regulators of ABA responses (Merlot et al., 2001; Kuhn et al., 2006; Robert et al., 2006; Saez et al., 2006; Yoshida et al., 2006b; Nishimura et al., 2007). Prototypes of these enzymes are ABI1 and its close homologue ABI2, which globally repress ABA responses and which have emerged as a hub in the network of ABA signal transduction (Yang et al., 2006; Moes et al., 2008).

ABA receptors have until recently remained either elusive or contested. However, a recent publication reported that two GTPases, GTG1 and GTG2, act as ABA receptors at the plasma membrane (Pandey et al., 2009). In addition, we and others have recently identified a different class of ABA receptor complexes that reside in the cytosol. These ABA receptors consist of a complex between PP2C protein phosphatases and an abscisic acid (ABA) binding protein. PP2Cs that have been shown to assemble into such complexes are ABI1, ABI2 and their close homologue HAB1. These three PP2Cs physically interact with the ABA-binding proteins RCARs/Pyrobactin Resistant 1 (PYR1)/PYR1-like (PYLs), which inhibit protein phosphatase activity in the presence of ABA (Ma et al., 2009; Park et al., 2009). RCARs/PYR/PYLs belong to a protein family with 14 members within the Bet v 1 superfamily of Arabidopsis. The Bet v 1 superfamily includes the birch pollen allergen Bet v 1a, class 10 pathogen-related proteins, polyketide cyclases, and lipid transfer proteins with a START domain (Radauer et al., 2008). Additionally, ABI1 physically interacts with a number of ABA signalling components and targets, such as the protein kinase OST1 (Yoshida et al., 2006a), calcineurin B-like sensor-associated protein kinases (Ohta et al., 2003), the redoxtransducer glutathione peroxidase (Miao et al., 2006) and the transcription factor AtHB6 (Himmelbach et al., 2002). The ABA-dependent inactivation of ABI1 through RCARs/PYR/PYLs promotes the ABA response through the action of SnRK2-type kinases, which are positive regulators of ABA responses and act downstream in the signal transduction pathway (Fujii and Zhu, 2009; Nakashima et al., 2009).

ABA levels can vary by two orders of magnitude between non-stress and stress conditions, ranging from low nm to low μm levels (Harris et al., 1988; Christmann et al., 2005). The high dynamic range of ABA levels and the distinct responses evoked imply the existence of mechanisms that differentially regulate distinct facets of ABA signalling in response to varying cytosolic phytohormone concentrations. As both components of the cytosolic ABA receptor complex belong to gene families in Arabidopsis, there are a large number of possible combinations for the assembly of potential receptor complexes. Our working model is that these homologous yet distinct receptor complexes differ with respect to their sensitivity to different concentrations of ABA. In order to test this model, we have identified a novel RCAR family member, RCAR3. In this study, we characterize RCAR3 and compare it with RCAR1. The receptor complexes between RCAR1 or RCAR3 and the PP2Cs ABI1 or ABI2 differ considerably with respect to their sensitivity to ABA. Selectivity for (S)-ABA compared with (R)-ABA and trans-ABA was less stringent for RCAR3 as compared with RCAR1. The results are compatible with a combinatorial assembly of ABA receptor complexes, which can differ both in their ABA stereo-selectivity and ABA sensitivity.



Physical interaction between RCAR3 and ABI1 or ABI2

As key regulators of ABA responses, the ABI1 and ABI2 protein phosphatases are promising starting points for the elucidation of the integrative network of ABA signalling. We used the yeast two-hybrid system to screen for proteins interacting with ABI2 and thereby identified six clones that showed lacZ activation and histidine autotrophy in dependence on the expression of the cDNA fusion protein (Yang et al., 2006). Two of the positive cDNA clones encoded the co-receptor RCAR1/PYL9 (Ma et al., 2009) and the structurally related RCAR3/PYL8 (At5g53160). RCAR1 and RCAR3 belong to clade I of the 14-member protein family, and share 70% amino acid identities and 82% similarity in their primary structure. The interaction between RCAR3 and ABI2 (Figure 1a) was more than 80% reduced by the single amino acid exchange present in abi2 (ABI2G168D). The abi1 (ABI1G180D) and abi2 mutations impair Mg2+ binding, which is required for the phosphatase activity of PP2Cs, and negatively affect protein interactions. The abi1-1 and abi2-1 mutants are ABA-Insensitive and both alleles are dominant. The full-length ABI1 protein was not examined in yeast because of the autoactivation of the reporter system (Himmelbach et al., 2002). Instead, we analysed N-terminally truncated versions of the wild-type and mutant proteins, ABI1121–434 and abi1121–434. The single point mutation present in abi1 and in the catalytically non-active ABI1 (NAPD177A) also impaired RCAR3 binding. The partial inhibition of the protein interaction suggests that the structure of the phosphatase domain may affect RCAR3-PP2C complex formation. Indeed, a truncated ABI11–180 devoid of the phosphatase domain was incapable of binding to RCAR3.

Figure 1.

 Physical interaction between RCAR3 and ABI1 and ABI2.
(a) The specificity of the ABI1 and ABI2 interaction with RCAR3 was examined in the yeast two-hybrid system. (Left) The analysis included different ABI1 and ABI2 variants (wild-type versus mutant) fused to the GAL4 DNA binding domain (BD fusions). (Right) Binding of RCAR3 to the PP2Cs is indicated by transactivation of the β-gal reporter activity above basal levels. Reporter activity is given in Miller units and was calculated as the mean value of three independent experiments (±SD).
(b) Bimolecular fluorescence complementation assays in Arabidopsis protoplasts indicated the interaction of RCAR3-YFPN with ABI1-YFPC and ABI2-YFPC (top and bottom, respectively). YFP signal (left part), YFP and chloroplast autofluorescence (middle part) and bright-field images of the analysed protoplast (right part). The arrows depict fluorescence of the nucleus. Scale bar: 10 μm.

The protein interactions between RCAR3 and ABI1 and ABI2 were confirmed in plant cells by bimolecular fluorescence complementation (Walter et al., 2004). Co-expression of RCAR3-YFPN and PP2C-YFPC in Arabidopsis protoplasts yielded YFP signals both in the cytosol and in the nucleus (Figure 1b). The presence of ABA did not detectably affect the interaction. The data support the yeast two-hybrid analyses and provide evidence for a physical interaction between RCAR3 and PP2C proteins.

PP2C phosphatase regulation by RCAR3

Analysis of RCAR1, 3, 8, 11 and 12 revealed an ABA-dependent inactivation of ABI1, ABI2 and the homologue of ABI1 (HAB1) (Ma et al., 2009; Park et al., 2009). We wished to establish whether differences between RCARs exist in terms of ABA sensitivity and ligand selectivity. To this effect, we directly compared four different receptor complexes generated by different combinations of RCAR1/RCAR3 and ABI1/ABI2. In the presence of RCAR3, purified ABI2 was efficiently blocked in its phosphatase activity by 1 μm (S)-ABA (Figure S1a) with half-maximal inhibition 30 sec after ABA administration. Titration of the physiologically active (S)-ABA to ABI1 and RCAR3 revealed an IC50 value for phosphatase inhibition of approximately 18 nm (Figure 2a). Under comparable experimental conditions, ABI1 and RCAR1 yielded an IC50 of 35 nm (S)-ABA. (R)-ABA and trans-ABA were effective in inhibiting ABI1 to approximately 80%, albeit at much higher ligand levels (Figure 2b). Indeed, compared with the 18 nmIC50 value for (S)-ABA, more than 1 μm (R)-ABA or 3 μmtrans-ABA were required to accomplish half-maximal inhibition. Similar results were recorded for ABI2 with a half-maximal inhibition at 30 nm (S)-ABA for RCAR3 and at 60 nm for RCAR1 (Figure 2c). Regulation of the PP2Cs by ABA and RCAR3 was clearly impaired by the amino acid exchange present in abi1 and abi2, which were inhibited less than 40 and 50% by 1 μm (S)-ABA, respectively (Figure S1b). The inhibition level under these conditions for the corresponding wild-type proteins was close to saturation (approximately 95%). The less efficient ABA-mediated inhibition of mutant as opposed to wild-type PP2Cs can be explained by their less efficient interaction with RCAR3 (Figure 1a). In contrast to RCAR3, RCAR1 complexes with either ABI1 or ABI2 were only inhibited to 30% at 30 μm (R)-ABA and trans-ABA (Figure 2b,d). The full extent of ABI2 inhibition was recovered by supplementation of the ABA analogues with 1 μm (S)-ABA. ABI2 inactivation was found to be heat labile (Figure S2a). To ensure that the ABA-mediated inhibition of PP2C by RCAR3 is not limited by the artificial substrate umbelliferylphosphate, we also tested a phosphopeptide substrate and found that this alternative yielded comparable results (Figure S2b). Taken together, RCAR1 conferred almost absolute stereo-selectivity to the receptor complex, whereas complexes with RCAR3 responded to the (R)-ABA and trans-ABA stereoisomers, albeit with more than 10-fold lesser sensitivity than to (S)-ABA. We conclude that the different receptors differ in terms of stereo-selectivity for and sensitivity to the ABA ligand.

Figure 2.

 Comparison of four receptor complexes.
(a) Inhibition of ABI1 by increasing concentrations of (S)-ABA. Half-maximal inhibition of ABI1 occurred at ∼35 nm and ∼18 nm of physiologically active (S)-ABA in the presence of RCAR1 (○) or RCAR3 (•), respectively.
(b) Inhibition of ABI1 by (S)-ABA (•), (R)-ABA (▪, bsl00066), and trans (R,S)-ABA (□, Δ) for RCAR1 (dotted line) and RCAR3 (solid line).
(c, d) Corresponding analysis of ABI2 as shown in (a,b). Half-maximal inhibition of ABI2 occurred at ∼60 nm and ∼30 nm of (S)-ABA with RCAR1 (○) or RCAR3 (•), respectively. The analysis was performed at a constant molar ratio of PP2C:RCAR of approximately 1:4 and with the PP2C level at 0.05 μm.

The PP2C inhibition imposed by ABA and RCAR3 is independent of substrate concentration and relies on a non-competitive inactivation of the enzyme (Figure 3a,b). The Michaelis–Menten constant of the ABI2-catalysed reaction was not affected by increasing RCAR3/enzyme ratios, whereas Vmax was reduced (Figure 3b). In the presence of saturating ABA levels, serial dilutions of up to 5 nm ABI2 in an RCAR3-containing solution maintained a rather constant inhibition level (Figure 3c). The stability of the inhibition level was only observed at saturating ABA levels, consistent with an ABA-mediated stabilisation of the RCAR3-ABI2 complex in the low nanomolar range.

Figure 3.

 Regulation of PP2Cs catalytic activities by RCAR3.
(a) Substrate dependence of ABI2 activity in the absence (•) or presence of RCAR3 at a molar ratio 0.2, 0.4, 0.6, 0.7 and 1 (○, ▪, □, bsl00066 and Δ, respectively). The analysis was performed in the presence of 1 mm (S)-ABA (SD < 4%).
(b) Lineweaver–Burk plot of data from (a).
(c) Regulation of ABI2 activity at a constant molar ratio of RCAR3 and ABI2 (approximately 1:1.8, respectively). The protein phosphatase activity was analysed at different levels of ABI2 in the presence (•) or absence (○) of 1 mm (S)-ABA. Activity is presented in relative fluorescence units (RFU) and the ABI2 inhibition level is given as a percentage (bsl00066).

RCAR3 binding to (S)-ABA was examined by isothermal titration calorimetry. The analysis revealed binding of (S)-ABA to RCAR3, with an apparentKd of approximately 970 ± 150 nm ABA (Figure S3a). The stoichiometry of the RCAR3/PP2C complex was analysed by titration of ABI2 with increasing levels of the binding protein in the presence of saturating ABA concentrations [1 mm (S)-ABA] (Figure S3b). Half-maximal inhibition occurred at an RCAR3 to ABI2 ratio of approximately 0.5. This value ranged between 0.3 and 0.7 for different protein preparations. Combined, the data are in support of a one-to-one ratio of the heteromeric protein complex as reported for RCAR1 (Ma et al., 2009). Circular dichroism analysis of RCAR3 (Figure 4a) is indicative of α-helical and β-sheet structures (Unneberg et al., 2001). The presence of ABA (100 μm) did not significantly alter the secondary structure that was, however, affected at temperatures of 40°C, consistent with high thermal sensitivity of RCAR3 (Figure 4b).

Figure 4.

 Circular dichroism (CD) analysis of RCAR3.
(a) CD spectrum of RCAR3 in the absence (black line) and presence (red line) of 100 μm (S)-ABA in the UV range from 195 to 260 nm. Spectra were plotted to the scale given in mean molar ellipticity (θ) units (deg × cm2 × dmol−1).
(b) Temperature-induced changes of RCAR3 secondary structures in the absence (black line) and presence (red line) of 100 μm (S)-ABA were monitored by CD measurements at 222 nm, indicative of α-helical signatures. Data were plotted as θ222 (deg × cm2 × dmol−1) versus temperature (°C).

Differential regulation of RCAR and PP2C expression throughout development and in response to abiotic stress

We mined the Genevestigator database (Zimmermann et al., 2005) to assess whether transcript levels of the diverse members of the cytosolic ABA receptor complexes may vary in vivo at different developmental stages, in different tissues or under different stress conditions. We found an interesting differential regulation of the expression of RCAR1, RCAR3, ABI1 and ABI2 under the conditions examined.

Whereas RCAR1 is upregulated in the seed coat, RCAR3 exhibits peak transcript levels in the xylem (Figure S4a). ABI1 is upregulated in radicles, in senescent leaves, in leaf primordia and in the root hair zone (Figure S4a). ABI2 shows a considerably lower level of expression than ABI1, and is upregulated predominantly in senescent leaves (Figure S4a) and in the endodermis ( RCAR1 and RCAR3 are expressed throughout development, with RCAR1 showing maximal expression levels in flowers and siliques (Figure S4b). By contrast, ABI1 and to a lesser extent ABI2 are upregulated late in development, in mature siliques (Figure S4b).

Transcriptional profiling (Zimmermann et al., 2005) showed that the RCAR1/3 and ABI1/2 genes are differentially regulated by light quality, duration and intensity, by a broad range of chemical and hormone treatments, and by stress conditions such as heat and cold. We list a subset of conditions that impact gene expression of the RCAR genes in Table 1. ABA treatment strongly upregulates ABI1 and ABI2 but downregulates RCAR3, whereas RCAR1 levels are either constant or upregulated. In contrast to ABA treatments, ethylene treatment upregulates RCAR3 and ABI1 without affecting RCAR1 or ABI2 levels (Table 1). Osmotic stress, salt stress and drought strongly upregulate ABI1 and ABI2 but downregulate RCAR3. Under these conditions, RCAR1 levels are either upregulated or constant. We extended this analysis to 10 RCAR (listed in Table S1) and six PP2C family members linked to ABA responses, and found that whereas the PP2Cs are uniformly upregulated by ABA treatment, salt or osmotic stress and drought, different RCAR genes vary in their responses, being either unaffected, up- or downregulated under these stress conditions (Table 1). Indeed, RCAR3 and RCAR10 were consistently downregulated under the conditions we examined, whereas RCAR13 was slightly upregulated (Table 1). With few exceptions, it can be stated that exogenous ABA as well as conditions that increase endogenous ABA upregulate PP2C levels by up to 75-fold, leave RCAR1 levels either constant or slightly increased but downregulate RCAR3 and RCAR10 by a factor of up to 25 (Table 1). We conclude that ABA-related stress conditions or treatments alter the relative levels of RCAR family members and increase the PP2C:RCAR ratio.

Table 1.   Transcriptional profiling upon hormone treatment and stress conditions, based on mining the Genevestigator database https://www.genevestigator.comThumbnail image of

Altering the PP2C:RCAR ratio influences the sensitivity of PP2C regulation

As transcriptional profiling showed that exogenous ABA or abiotic stress conditions differentially regulated expression levels of RCARs and PP2Cs, we attempted to assess the effect of varying the PP2C:RCAR protein ratios. We kept the ABI1 and ABI2 PP2C levels constant while altering those of RCAR1 or RCAR3 in in vitro experiments. IC50 values correspond to the ABA concentration required to achieve a 50% inhibition of the phosphatase activity and, as such, these are a good measure of the sensitivity of the PP2C regulation by ABA. Figure 5 shows that the PP2C:RCAR ratio has a large impact on IC50 values. As RCAR levels increase, IC50 values decrease, indicating a more ABA-sensitive regulation of phosphatase activity. Conversely, the PP2C regulation with RCAR1 and RCAR3 becomes less sensitive to ABA, when PP2C:RCAR ratio increases, which possibly simulates what is seen for transcript levels under stress conditions (Table 1).

Figure 5.

 Dependence of ABA-mediated inhibition on receptor complex composition and co-receptor ratios.
(a, b) ABI1 (a) and ABI2 (b) were titrated with increasing levels of RCAR1 (•) and RCAR3 (○). The half-maximal inhibitory ABA concentration was determined at different molar ratios of PP2C:RCAR, and a constant PP2C level at 0.05 μm.

The efficiency of ABA-mediated phosphatase inhibition was higher with ABI1 than with ABI2 and higher with RCAR3 than with RCAR1. The IC50 values of PP2C inhibition were approximately twofold lower with RCAR3 versus RCAR1 at a 2:1 RCAR:PP2C ratio. Under these experimental conditions, ABI1 was approximately twofold more sensitive to ABA regulation than ABI2. Half-maximal inhibition of RCAR3/ABI1 was observed at 23 nm ABA, whereas RCAR1/ABI2 revealed a more than fourfold higher IC50 value of 95 nm ABA (Figure 5a,b). The finding reflects major differences in the heteromeric receptor complexes with respect to ABA-mediated inhibition. PP2C inhibition requires RCAR binding to the PP2C, and increasing the RCAR:PP2Cs ratio shifts the equilibrium towards complex formation. The differences between IC50 values of ABI1/RCAR3 and ABI2/RCAR1 were reduced with increasing RCAR levels, and were almost abolished at high excess levels of RCAR (PP2C:RCAR value of 0.1). Interestingly, the IC50 values were more responsive to changes in RCAR1 than RCAR3 levels by a factor of 2.5 and 5.4 for ABI1 and ABI2, respectively. The data imply a higher affinity of RCAR3 for PP2C interaction compared with RCAR1. Thus, both the PP2C and RCAR components modulate the sensitivity of ABA-mediated PP2C inactivation. Inactivation of the PP2Cs is required to overcome their negative regulation of the ABA signal pathway and to allow the activation of the ABA response via SnRKs (Fujii and Zhu, 2009; Nakashima et al., 2009), with RCAR3 and ABI1 providing greater ABA-sensitive regulation.

ABA responses

The ABA response was quantitatively assayed in protoplasts by using reporter constructs consisting of the ABA-upregulated promoters pRab18 (Figure 6a,b) and pRD29B (Figure 6c,d) driving luciferase expression (Moes et al., 2008). Transient expression of RCAR3 in the Arabidopsis cells resulted in an enhanced ABA response. The upregulation of gene expression was also observed in the absence of exogenous ABA (Figure 6a–d). To examine whether this RCAR3-mediated activation might result from endogenous ABA levels, protoplasts derived from the ABA-deficient Arabidopsis mutant aba2-1 were analysed (Figure 6a,c). The mutant cells had an approximately twofold lower level of reporter expression compared with the wild type in the absence of exogenous ABA. Ectopic expression of RCAR3 stimulated luciferase activity of wild-type cells by a factor of 8.0 ± 2.2 (Figure 6a) and 18.0 ± 1.9 (Figure 6c), with the ABA-responsive reporter constructspRAB18::LUC and pRD29B::LUC, respectively. The RCAR3-mediated activation in aba2-1 protoplasts was limited to an increase of 3.9 ± 0.9 (pRAB18::LUC) and 8.3 ± 4.3 (pRD29B::LUC), respectively. Administration of 3 μm (S)-ABA to the transfected protoplasts resulted in a 14- and 38-fold increase of reporter expression in the wild type, and in a 68- and 140-fold enhancement in aba2-1, in the presence of AtRab18 and AtRD29B promoters, respectively, compared with cells not expressing the effector protein and not exposed to ABA. In this respect, aba2-1 mutant protoplasts display a hypersensitivity to ABA that we have observed previously (Ma et al., 2009). The findings of reduced reporter expression in the ABA-deficient protoplasts at endogenous ABA levels, irrespective of RCAR3 expression, and the observed recovery of the ABA response in the presence of exogenous ABA, are supportive of an RCAR3-controlled stimulation of ABA signalling.

Figure 6.

 RCAR3-activated and PP2C-antagonized ABA responses.
(a–f) The ABA-induced upregulation of gene expression was monitored with the ABA-responsive reporter constructs pRAB18::LUC (a, b) and pRD29B::LUC (c–f) in Arabidopsis protoplasts and was measured as relative light units (RLU/RFU). Each data point represents the mean value of three independent transfections.
(a, c) Regulation of gene expression by the effector RCAR3 (3 μg) in the absence (white bars) and presence of 3 μm (S)-ABA (black bars). The analysis was performed with ABA-deficient aba2-1 protoplasts (left panel) and wild-type protoplasts (right panel).
(b, d) The RCAR3- and ABA-stimulated reporter expression is inhibited by concomitant expression of various PP2Cs (1 μg) in the absence (white bars) and presence of 3 μm (S)-ABA (black bars).
(e) The RCAR3-mediated stimulation of the ABA response is antagonized by co-expression of ABI2. The levels of RCAR3 effector constructs were 0, 0.3 and 3 μg plasmid (○, ▪ and bsl00066, respectively).
(f) The activation of the ABA response by RCAR3 is partially antagonized by co-expression of an RNAi construct (1 μg) targeting RCAR3 in the absence (white bars) and presence of 3 μm (S)-ABA (black bars). cRNAi: control RNAi. The -fold change of induction was compared with the wild-type sample not exposed to (S)-ABA and not targeting the RCAR3 effector construct.

We then asked whether co-expression of the PP2Cs would influence the RCAR3-mediated response (Figure 6b,d). Expression of ABI1 reduced the RCAR3- and ABA-stimulated reporter expression by a factor of 4.6 and 16 with the ABA-responsive reporter constructs pRAB18::LUC and pRD29B::LUC, respectively. The ABI2 expression diminished the RCAR3- and ABA-stimulated reporter expression by a factor of 10 (pRAB18::LUC) and 17 (pRD29B::LUC). The ABA response was almost fully blocked by PP2C expression in cells not transfected with the RCAR3 expression cassette. In these analyses, the abi1 and abi2 mutant proteins were much more effective in blocking the RCAR3-mediated stimulation of ABA signalling. Titration of the RCAR3-stimulatory effect by increasing quantities of co-transfected ABI2 effector resulted in complete abrogation of ABA signalling (Figure 6e). Co-expression of a construct targeting RCAR3 and related transcripts by RNA interference (RNAi) partially but significantly antagonized the RCAR-mediated activation of the ABA response (Figure 6f).


Plant adaptation requires the fine-tuning of a large number of responses to subtle or extreme changes in environmental conditions. Cytosolic levels of the phytohormone ABA can range from the nanomolar to the micromolar range depending on environmental challenge and/or developmental stage (Priest et al., 2006; Christmann et al., 2007). The high dynamic range of ABA levels and the distinct responses evoked imply the existence of mechanisms to fine-tune distinct ABA signalling pathways in response to varying ABA concentrations. Different heteromeric receptor complexes may provide a means of adjusting the sensitivity of ABA perception and signalling. In order to test this hypothesis, we compared four different receptor complexes formed by a combinatorial assembly of the co-receptors RCAR1/3 and of the PP2Cs ABI1/2.

A model for fine-tuning the ABA response

Several overlapping mechanisms may act to modulate ABA responses. First, the combinatorial assembly of receptor complexes showed that different complexes had different properties with respect to the sensitivity of the response. Second, transcriptional profiling showed that the different members of the RCAR and PP2C gene families were differentially expressed in different tissues and at different developmental stages. Third, ABA-related stress conditions alter the transcript levels of distinct co-receptors differentially, which in turn may alter the sensitivity and plasticity of the response. Fourth, different PP2Cs are known to target overlapping but distinct targets (Vranova et al., 2001; Himmelbach et al., 2002; Ohta et al., 2003; Miao et al., 2006; Yang et al., 2006; Yoshida et al., 2006b), which is likely to modulate the nature of ABA responses.

RCAR1 and RCAR3 in combination with ABI1 and ABI2 generate high-affinity receptors for ABA, which function within the nanomolar range of the phytohormone. Stress conditions, however, may lead to very high ABA levels, within the micromolar range, and these high levels may persist in the plant over long periods of time. Transcriptional profiling suggests that stress conditions such as drought, salt stress and osmotic stress upregulate the levels of PP2C phosphatases and downregulate some RCAR family members, such as RCAR3 and RCAR10, while leaving the levels of other RCARs, such as RCAR1 and RCAR2, relatively constant. The data are entirely consistent with the transcriptional profiling presented by Santiago et al. (2009). Our in vitro measurements suggest that higher PP2C:RCAR ratios and higher levels of RCAR1 over RCAR3 would lead to a desensitization of the ABA response. In addition, higher PP2C:RCAR ratios greatly enhance the difference between the RCAR1 and RCAR3 receptor complexes, yielding a greater plasticity of the response. Thus, although RCAR1 and RCAR3 may, under physiological conditions, function at low ABA concentrations, their differential regulation under stress conditions may provide a means for the plant to cope with sustained high levels of ABA. It is also important to note that we have looked at only four of a potential of at least 84 different receptor complexes. Although these four receptor complexes have shown considerable differences, all four are highly sensitive to ABA. An analysis of different combinations of RCARs and PP2Cs may further our understanding of the entire dynamic range of ABA levels in the cell.

Phytohormone perception

The heteromeric ABA receptor complexes have a greater affinity for ABA than the ABA-binding regulatory components alone. The affinity of RCAR1 (Ma et al., 2009), RCAR3 and RCAR8/PYL5 (Santiago et al., 2009) for (S)-ABA did not considerably differ, with Kds of 0.7, 1.0 and 1.1 μm, respectively. By contrast, (S)-ABA binding to heteromeric receptor complexes revealed a pronounced shift in affinity. The low Kds of 64 nm for ABI2/RCAR1 and 38 nm for a truncated HAB1/RCAR8 combination (Ma et al., 2009; Santiago et al., 2009) are reflected by low IC50 values of the complexes with 60 and 35 nm (S)-ABA, respectively. Similarly, RCAR3 revealed half-maximal inhibition of ABI1 and ABI2 in the range of 15–40 nm, consistent with the generation of a high-affinity binding site for ABA by the heteromeric receptor complex. Varying IC50 values were also observed for other RCAR members and truncated HAB1, with RCAR11/PYR1 yielding the highest value of 390 nm ABA (Santiago et al., 2009).

The ABA ligand appears to promote or stabilize receptor complex formation. This conclusion is supported by enhanced interaction of some RCARs/HAB1 combinations in yeast in the presence of ABA (Park et al., 2009) and by stabilization of PP2C inhibition even in the low nanomolar range in the presence of high ABA levels, as observed for RCAR1 (Ma et al., 2009) and RCAR3 (this study). In protoplasts, the interaction of RCAR3 and ABI1/2 was not visibly enhanced by exogenous ABA, pointing to an efficient receptor complex formation under the in vivo conditions.

Plant hormones other than ABA also act to promote or stabilize protein interactions. Auxin, for example, acts as a ‘glue’ in the interaction between its receptor, the E3 Ubiquitin ligase TIR1, and the Aux/indole-3-acetic acid (IAA) transcriptional repressors that are targeted for degradation (Tan et al., 2007). Similarly, gibberellin binding to GID1 (GIBBERELIN-INSENSITIVE DWARF 1) stabilizes the interaction between an E3 Ubiquitin ligase SLY1 (SLEEPY1) and DELLA repressors, which are targeted for degradation (Murase et al., 2008; Shimada et al., 2008). A third example is that of brassinolides. Brassinolides are required for the assembly of a heterodimeric receptor complex that consists of two leucine-rich repeat receptor-like kinases (LRR RLKs), BRI1 (BRASSINOSTEROID INSENSITIVE 1) and BAK1 (BRI1-ASSOCIATED RECEPTOR KINASE 1) (Wang et al., 2008). These three examples are reminiscent of our findings for ABA, in which the hormone may act to promote or stabilize protein interactions between the components of the heteromeric co-receptors (Ma et al., 2009; Park et al., 2009).

Phytohormone-binding proteins and their targets or co-receptors are encoded by gene families in Arabidopsis, which provides a plethora of similar yet distinct possible combinations. The Arabidopsis genome encodes, for example, six TIR1 family members and 29 Aux/IAAs (Santner and Estelle, 2009), three GID1 genes and five DELLA proteins (Suzuki et al., 2009), and 14 RCAR genes and nine clade-A PP2Cs. In this study, we have shown that different combinations of RCARs and PP2Cs alter the sensitivity of the ABA receptor complex. The transcript levels of the different RCARs and PP2Cs varies throughout development and in response to environmental challenge in vivo. Furthermore, changes in the combinatorial assembly of the receptor or in the relative proportions of RCAR and PP2C proteins affect the sensitivity of the receptor complex in vitro, lending support for changing combinatorial receptor complexes. Given the complexity of the gene families involved, a combinatorial assembly and altered relative ratios of receptors, co-receptors and their targets may act to regulate not only ABA responses, but might provide a general model for the fine-tuning of hormone responses.

Experimental procedures


All chemicals were obtained from Sigma-Aldrich (, Fluka (now part of Sigma-Aldrich) and Merck ( Abscisic acid was purchased from Lomon Bio Technology [(S)-ABA;], Sigma-Aldrich, [(R)-ABA] and A.G. Scientific [trans (R,S)-ABA;]. The ABA enantiomers and trans (R,S)-ABA were purified by HPLC as previously described (Ma et al., 2009).

Plant materials

All the Arabidopsis lines used in this study were in the ecotype Columbia (Col) and Landsberg erecta (L-er). Plants used for protoplast preparation were grown for 4 weeks in a perlite–soil mixture in a controlled growth chamber at 23°C under long-day conditions with 16 h of light (250 μE m−2 sec−1) (Moes et al., 2008).

Plasmid constructs

The pRAB18::LUC and pRDB29::LUC reporter plasmids used in this study have been described previously (Moes et al., 2008; Ma et al., 2009). For heterologous expression, the cDNA of RCAR3 (At5g53160.2) was amplified with the primer pair 5′-ATTCTGGATCCGCATGCATGGAAGCTAACGGG-3′ and 5′-TGGGAGCTCCTTTAGACTCTCGATTCTGTC-3′. The PCR fragment was subsequently cloned via BamHI and SacI sites into the pQE30 vector (Qiagen, yielding pQE30-RCAR3 and verified by DNA sequence analysis. For transient expression, the plasmid pBI221-p35S::RCAR3 was generated by replacing the GUS-NOS terminator cassette in pBI221 with the RCAR3-NOS terminator cassette from pBI121-RCAR3 via BamHI and EcoRI.

To generate the split-YFP constructs, RCAR3 cDNA with no stop codon was linked to the N-terminally truncated YFP genes via BamHI and SmaI sites in pSPYNE-35S (YFPN1–155, 1–155 aa) and pSPYCE-35S (YFPC156–239) vectors (Walter et al., 2004). The cDNA of RCAR3 with no stop codon was amplified with the primer pair 5′-ATCTTGGATCCATGGAAGCTAACGGG-3′ and 5′-AATACCCGGGGACTCTCGATTCTGTCG-3′ to generate p35S::SPYNE:RCAR3 and p35S::SPYCE:RCAR3. The RNAi construct used in the analysis was generated as described previously (Ma et al., 2009).

Yeast two-hybrid screen

The detection of interaction partners of ABI2 and the analysis of protein–protein interaction was performed as reported by Yang et al. (2006).

Expression and purification of RCARs and PP2Cs

All His-tagged proteins were expressed in Escherichia coli by using pQE30 and pQE70 clones (Ma et al., 2009). For protein expression, cells were grown overnight in 20 ml Luria Bertani (LB) broth and used for inoculation of 1 L of culture. The cells were grown at 37°C with vigorous shaking until an OD600 of 0.5–0.6 was reached. Protein expression was subsequently induced by administration of isopropyl-β-d-thiogalactopyranoside (IPTG; 0.5 mm final concentration) and cells were harvested by centrifugation at 4000 g for 20 min, 2 h (PP2Cs) or 4 h (RCARs) after induction. The cell pellet was used directly or stored overnight at −20°C prior to purification. After thawing, the pellet was lysed in 10 ml of lysis buffer (50 mm NaH2PO4, 300 mm NaCl, 5 mm imidazole, pH 8.0) and treated with lysozyme (1 mg ml−1) for 30 min. Cells were disrupted by sonication on ice (three times for 30 sec with 30-sec cooling intervals). A cleared protein lysate was obtained after centrifugation at 30 000 g for 30 min and loaded onto an Ni-TED 2000 column (Macherey-Nagel, In order to remove unspecifically bound proteins, 8 ml of washing buffer (50 mm NaH2PO4, 300 mm NaCl, 20 mm imidazole, pH 8.0) was applied to the column. Proteins of interest were eluted with 3 ml of elution buffer (50 mm NaH2PO4, 300 mm NaCl, 250 mm imidazole, pH 8.0) and dialysed three times against dialysis buffer (100 mm Tris–HCl, 100 mm NaCl, 2 mm dithiothreitol, pH 7.9).

Phosphatase assays

Phosphatase activity was measured using 4-methyl-umbelliferyl-phosphate (4-MUP) or phosphopeptide RRA(pT)VA as a substrate (Ma et al., 2009). Values are means ± SDs of three replicates.

Protoplast analysis

Preparation and analysis of Arabidopsis protoplasts was performed as described by Moes et al. (2008). In brief, Arabidopsis protoplasts were transfected with 10 μg DNA of the reporter plasmid (pRAB18::LUC, pRD29B::LUC), 0.1–10 μg DNA of effector plasmid and 2 μg of p35S::GUS plasmid as a control for internal normalisation of the expression. Protoplast suspensions were incubated in the presence or absence of ABA after transfection.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) measurements were performed on a MicroCal VP-ITC (MicroCal, The Origin software package (Origin Lab, was used for data analysis. Purified protein samples were dialysed overnight in buffer containing 100 mm Tris–Cl, pH 7.9, 100 mm NaCl and 2 mm dithiothreitol, and were degassed briefly before loading into the ITC cell. The syringe was filled with (S)-ABA in dialysis buffer. A typical experiment consisted of 5-μl injections (with 1-μl pre-injection) of (S)-ABA (in 5-min intervals) at 30°C under continuous stirring. The final concentration of RCAR3 in the cell was adjusted to 1 μm, whereas the injection syringe was filled with (S)-ABA solution of 400 μm.

Circular dichroism analyses

Circular dichroism (CD) spectra were performed in a J-715 with PTC343 peltier unit (Jasco, Far-UV spectra of RCAR3 (0.5 mg ml−1) in the absence or presence of (S)-ABA (100 μm) were registered in the range 195–260 nm using an optical-path cell of 0.1 cm. Samples were analysed in 15 mm sodium phosphate, pH 7.5, at a constant temperature of 20°C. Spectra were recorded with a 0.1-nm resolution at a scan speed of 20 nm min−1 and results were expressed as an average of 15 scans. Final spectra were baseline-corrected and ellipticities were calculated for a mean residue weight (Unneberg et al., 2001). To determine the thermal stability of the RCAR3 in the presence and absence of (S)-ABA (100 μm), the CD signal was monitored at 222 nm from 10 to 90°C at a protein concentration of 0.5 mg ml−1 and a heating rate of 20°C h−1.

Transcriptional profile analysis

Transcriptional profiling was carried out with the Genevestigator database (; Zimmermann et al., 2005).


The financial support of the Deutsche Forschungsgemeinschaft CH-182/5, EU Marie-Curie-Program MEST-CT-2005-020232 and the Fonds der Chemischen Industrie is gratefully acknowledged. We thank Johanna Berger, Christoph Heidersberger, Christian Kornbauer, and Knut Thiele for technical assistance.