A cell surface receptor mediates extracellular Ca2+ sensing in guard cells

To study the putative CAS further, we attempted to clone CAS from Arabidopsis using a functional screening approach, as described for the capsaicin receptor⁹. Messenger RNA from leaves was used to construct a cDNA library, which was subdivided into pools. Each pool was transiently transfected into human embryonic kidney (HEK293) cells. Transfected cells were screened for CICI using ratiometric imaging of the fluorescent Ca²⁺-sensitive dye Fura-2. Ca²⁺_o at 2.5 mM caused [Ca²⁺]_i increases in a few cells transfected with empty vector (Fig. 1, top). A positive pool, pool-58, was identified, which resulted in more Ca²⁺_o-sensitive cells (Fig. 1, middle). This pool was divided into ten sub-pools and re-screened. After subdividing the pool several times, a single 1.4-kb cDNA was isolated that conferred Ca²⁺_o sensitivity to transfected HEK293 cells (Fig. 1, bottom), and was thus named CAS.

Figure 1 Calcium-imaging-based expression cloning of a Ca2+o-sensing receptor from an Arabidopsis cDNA library.   Full legend   High resolution image and legend (117k)

The CAS cDNA encodes a protein of 387 amino acids with a calculated molecular mass of 41,268 Da (Fig. 2a). CAS is a single-copy gene in the Arabidopsis genome. Database searches identified several homologues in other plant species but not in animals. Hydrophobicity analysis showed that CAS has a single transmembrane domain (Fig. 2a, b). The carboxy terminus is possibly intracellular, and has sequence similarity to a rhodanese domain, which is likely to have a role in protein–protein interactions. The putative extracellular amino terminus has no significant homology to any known functional domains. It should be noted that in contrast to cytosolic Ca²⁺ sensors, such as calmodulin (CaM)¹¹, CAS is unlikely to contain high-affinity Ca²⁺-binding sites, such as EF-hands, as physiological Ca²⁺_o is in the submillimolar range^{2, 14}. Nevertheless, acidic amino acids have been reported to act as low-affinity Ca²⁺-binding sites in animal cells¹⁵. The N terminus contains several acidic amino acids and its isoelectric point is 5.3 (Fig. 2a).

Figure 2 CAS encodes a low-affinity/high-capacity Ca2+-binding protein.   Full legend   High resolution image and legend (109k)

To test whether CAS binds to Ca²⁺, we prepared recombinant proteins for both the N terminus and the C terminus (Fig. 2c, left), and assayed their abilities to bind Ca²⁺. When blotting with ⁴⁵Ca only, the N terminus and CaM bound to ⁴⁵Ca, but not the C terminus (Fig. 2c, middle). When blotting with ⁴⁵Ca and cold CaCl₂, only the N terminus bound to ⁴⁵Ca (Fig. 2c, right), showing qualitatively that the N terminus is indeed a Ca²⁺-binding domain, and contains low-affinity/high-capacity Ca²⁺-binding sites compared with CaM. To analyse the Ca²⁺-binding quantitatively, we conducted equilibrium dialysis using the purified N-terminal protein. Figure 2d shows that the N terminus bound 10 to 12 Ca²⁺ per CAS molecule with a dissociation constant K_d 1.2 mM, confirming the ⁴⁵Ca-binding data. Furthermore, we analysed the kinetics of activation and inactivation of CAS in HEK293 cells, which correlate with the association and dissociation between CAS and Ca²⁺_o, respectively. The activation time course could be described by a simple sigmoid equation¹⁶, with an activation time-constant of 28.6 s (Fig. 2e). A triple exponential decay function could be fitted¹⁶ to the inactivation of CAS (Fig. 3e) with an inactivation time-constant of 18.2 s. A Hill curve could be fitted¹⁶ to the data of CAS activation (Fig. 2f) with a K_d of 1.5 mM and a Hill coefficient of 2, consistent with the K_d and multiple Ca²⁺_o-binding sites observed using equilibrium dialysis. Taken together, these data demonstrate that CAS contains low-affinity/high-capacity Ca²⁺-binding sites on the N terminus.

Figure 3 Expression pattern and subcellular location of CAS.   Full legend   High resolution image and legend (71k)

Subsequently, we determined the expression and subcellular localization of CAS in Arabidopsis. RNA blot analysis shows that CAS is predominantly expressed in shoots (Fig. 3a). A CAS promoter::GUS fusion shows that CAS is expressed in guard cells (Fig. 3b). We expressed 35S::CAS–GFP and 35S::GFP in onion epidermal cells, and found that CAS–GFP (green fluorescent protein) was exclusively localized to the cell surface, whereas GFP alone was localized throughout the cells (Fig. 3c). After the epidermal cells were incubated in 0.8 M mannitol, the fusion proteins and the plasma membrane retracted from the cell wall due to plasmolysis (not shown). We also found CAS in the membrane fraction but not in the soluble fraction of proteins from leaves (Fig. 3d). Finally, we confirmed the guard-cell surface localization of CAS using immunolocalization¹⁷ (Fig. 3e).

To assess the function of CAS in plants, first we examined whether CAS is required for guard-cell Ca²⁺_o signalling. As we were unable to find CAS-null lines, we introduced a 35S::CAS antisense construct into Arabidopsis. CAS expression was dramatically reduced at the mRNA and protein levels in all six antisense lines tested (AS1 to AS6) with three of them shown (Fig. 4a, b). To analyse CICI in guard cells, we introduced a cameleon construct into AS1 and AS2 plants, and found that the CICI was disrupted significantly in these lines compared with the wild type (WT; Fig. 4c, d). Ca²⁺_o-induced stomatal closure was also abolished in these lines (Fig. 4e). However, ABA-induced stomatal closure was virtually unaffected (not shown), suggesting that CAS antisense specifically knocked out Ca²⁺_o signalling, and that CAS has an essential role in guard-cell Ca²⁺_o signalling. Note that, apart from Ca²⁺_o, stomatal movements are regulated by various abiotic and biotic factors^{5, 8, 18}. Apparently, guard cells must integrate several factors to control stomatal apertures appropriately^{8, 18}; CAS, several Ca²⁺ influx and release mechanisms, and Ca²⁺ pumps must act in concert to regulate cytosolic free Ca²⁺ status^{3, 8}.

Figure 4 CAS is required for guard-cell Ca2+o signalling and plant development.   Full legend   High resolution image and legend (53k)

Calcium is an abundant and essential divalent cation nutrient in higher plants, and has a central role in numerous physiological and developmental processes^{2, 3, 10, 14}. To examine whether CAS also functions in processes other than stomatal closure, we analysed plant growth and development in response to limited Ca²⁺ supplies. Figure 4f shows that AS1 plants showed a severe defect in bolting in response to Ca²⁺ deficit compared with the wild type. AS1 plants bolted eventually at reduced Ca²⁺ but much later than the wild type (Fig. 4g). In contrast to wild-type plants, AS1 plants could not flower and displayed severe symptoms of Ca²⁺ deficiency (not shown). The Ca²⁺ content of these plants was further analysed and no significant differences between AS1 and wild-type plants were found under the imposed growth conditions (100% humidity, not shown). Similar phenotypes were observed in AS2 to AS6 plants. Bolting represents an important developmental switch in flowering plants¹⁹. It is possible that a threshold of Ca²⁺_o has to be reached and sensed by CAS to signal the switch to bolting. In the antisense plants, although the Ca²⁺_o level is not changed, the malfunction of Ca²⁺_o sensing cannot sensitively detect the level of Ca²⁺_o and signals insufficient Ca²⁺_o for bolting, suggesting that CAS has a key role in Arabidopsis development.

We have cloned and characterized a cell surface Ca²⁺_o-sensing receptor in Arabidopsis. We have also identified the essential role of CAS in Ca²⁺_o signalling in guard cells and bolting. Molecular genetic, cell biological, and physiological analyses lead us to propose that Ca²⁺_o activates CAS, which converts the Ca²⁺_o signal into cytosolic free Ca²⁺. It is possible that CAS may initiate the IP₃ pathway in Arabidopsis, as suggested by our preliminary observation (Supplementary Fig. S-1), consistent with IP₃-induced stomatal closure²⁰. In contrast, previous studies have shown that the CICI in Commelina guard cells is not affected by inhibition of phospholipase C, but by inhibition of Ca²⁺ channels^{6, 21}, suggesting that Ca²⁺_o triggers Ca²⁺ influx. Therefore, it remains to be determined which Ca²⁺-mobilizing machineries—such as Ca²⁺ releases triggered by IP₃, cyclic ADP-ribose or nicotinic acid adenine dinucleotide phosphate, or Ca²⁺ influx channels—are activated by CAS. Functional analysis of CAS-mediated CICI in heterologous expression systems and in plants, and identification of the downstream components of CAS, should help to clarify these issues.

The CAS-mediated Ca²⁺_o-induced stomatal closure is possibly physiologically relevant. The distribution of Ca²⁺ in plants is affected by the rate of transpiration, which is predominantly controlled by stomata^{5, 10}. Calcium moves in relatively large amounts to highly transpiring tissues, but much less moves to weakly transpiring tissues¹⁰. From the standpoint of Ca²⁺ supply, plants must downregulate transpiration when the Ca²⁺-unloading rate is high, and vice versa. It is likely that excessive water evaporation from stomatal pores results in an increase in the local concentration of Ca²⁺_o outside guard cells to submillimolar or millimolar. Consequently, the elevated Ca²⁺_o activates CAS, which mediates stomatal closure, preventing excessive Ca²⁺ unloading. Thus, Ca²⁺_o-induced stomatal closure could function as a feedback mechanism of Ca²⁺ supply. Interestingly enough, our preliminary data show that the Ca²⁺ content was 40% higher in antisense plants than in the wild type when they were grown in soil, supporting the feedback hypothesis. Precise information on the concentration range of Ca²⁺_o is lacking¹⁴, and data taken from the literature vary from 10 to 1,000 µM (refs 14, 22). It is especially difficult to precisely detect Ca²⁺_o outside guard cells, as water evaporation from stomatal pores may cause uneven distribution of apoplastic solutes, including Ca²⁺_o. Our finding that CAS is required for guard-cell Ca²⁺_o signalling and bolting suggests that CAS has a central role in diverse Ca²⁺_o-dependent processes. The cloning of CAS and manipulation of its activity now make it possible to assess the molecular function of Ca²⁺_o in plants.

Methods
Imaging of [Ca²⁺]_i in guard cells Leaf epidermal peels from soil-grown Arabidopsis thaliana (Columbia) were placed in a microwell chamber containing 50 µM CaCl₂, 5 mM KCl, 10 mM MES-Tris, pH 6.15. Cameleon-based Ca²⁺ imaging was performed as described7 using a fluorescence microscope (Axiovert 200; Zeiss) equipped with two filter wheels (Lambda 10-2; Sutter), and a cooled CCD camera (CoolSNAP_f; Roper). Excitation was provided at 440 nm, and emission ratiometric (F_535 nm/F₄₈₀ nm) images were collected using Imaging Workbench 4.0 software (Axon).

Stomatal aperture bioassay Rosette leaves were floated in solutions containing 50 µM CaCl₂, 10 mM KCl, 10 mM MES-Tris, pH 6.15, for 2 h. CaCl₂ at 2 mM was added to the solutions for 2 h to assay stomatal closure as described¹⁸.

Expression cloning mRNAs were isolated from Arabidopsis leaves. cDNAs were synthesized, and ligated into the SalI and NotI sites of pCMVSPORT6 vector (Invitrogen). Recombinant plasmids were introduced into Escherichia coli. The resulting 2 10⁶ clones were divided into 100 pools for transfection of HEK293 cells. HEK293 cells were grown and maintained in DMEM medium supplemented with 10% fetal bovine serum, 0.1% penicillin and streptomycin. For transfection, cells were seeded onto eight-well-chambered coverglasses (Nunc) for 24 h, and transfected with plasmid DNA from individual pools using LipofectAMINE 2000 reagent (Invitrogen). Cells were loaded with Fura-2-AM (5 µM; Molecular Probes), and a Ca²⁺-imaging assay was performed in the HEK293 cells as described⁹.

Ca²⁺-binding assays For ⁴⁵Ca blot analysis²³, the N terminus (1–180) and C terminus (211–386) were subcloned into pET-30a and expressed in E. coli. The purified N and C termini together with bovine CaM at 10 µg were separated by SDS–PAGE, and transferred to membranes (Millipore). The membranes were washed in Ca²⁺-free solution containing 5 mM MgCl₂, 60 mM KCl, 10 mM MES-KOH, pH 6.5, blotted with 2 µCi ml^-1 (3 µM) fresh ⁴⁵Ca (< 14 days; PerkinElmer, 370 MBq ml^-1) for 10 min, rinsed for 5 min in 50% ethanol, and exposed to X-ray films. Unlabelled Ca²⁺ at 5 mM was added as indicated. Equilibrium dialysis was performed using dialysis cells (100 µl) separated by dialysis membranes (Bel-Art) as described²⁴. Purified N terminus (66 µM) was dialysed in a solution containing 0.1 µCi ⁴⁵Ca (1.5 µM), 0–10 mM CaCl₂, 100 mM KCl, 10 mM MOPS, pH 7.5, for 24 h. Radioactivity was determined by liquid scintillation counting.

Expression and localization analysis Northern blot analysis, histochemical staining for GUS activity and immunolocalization were performed as described^{17, 25}. For northern blot, 15 µg RNA was loaded and blotted with a ³²P-labelled CAS probe. For GUS staining, a genomic fragment of 1.5 kb in the promoter region of CAS was amplified by PCR with primers: 5'-GCCCGCGGTGTTTGTTTGTTATTGTTTTGGGTA and 5'-GGCTCGAGGGTTTCTCTCTGCCACACTT, and cloned into a GUS binary vector pBI101. Transgenic plants were generated using Agrobacterium-mediated transformation and selected by kanamycin²⁶. Independent homozygous transformants carrying a single insertion in the T3 generation were analysed further. CAS antisense plants were generated similarly. For GFP-based analysis, 35S::CAS–GFP or 35S::GFP was transiently expressed in onion epidermal cells by bombardment²⁵. For immunolocalization, polyclonal antibodies were raised in rabbits using the purified N and C termini as the antigen. The leaf epidermal peels were placed on poly-L-lysine coated slides. Affinity-purified primary anti-CAS and fluorescein isothiocyanate-conjugated anti-rabbit secondary antibodies (Sigma) were diluted 1:10 and 1:200, respectively.

Western blot Total proteins (10 µg) from Arabidopsis leaves were separated on SDS–PAGE, and incubated with anti-CAS antibody followed by alkaline phosphatase-conjugated anti-rabbit antibody (Sigma), and developed using NBT/BCIP substrate. The leaf homogenate was centrifuged at 13,000g, and the supernatant was further centrifuged at 80,000g to separate the soluble proteins (supernatant) and membrane-associated proteins (pellet).

Ca²⁺-deficiency and Ca²⁺-content analysis Wild-type and antisense plants were grown in the MS-based media containing MS minor salts (Sigma), 1% sucrose (Sigma), 1% agar (Becton Dickinson) and MS major salts (Sigma) with Ca²⁺ supplemented to 0.1 mM or 1 mM CaCl₂. Each pot contained 50 ml media. For Ca²⁺-content analysis, shoots were digested with 70% nitric acid, and the total Ca²⁺ content was determined using an atomic absorption spectrometer. Data were analysed and presented as described^{16, 18}.