CNGC2 Is a Ca²⁺ Influx Channel That Prevents Accumulation of Apoplastic Ca²⁺ in the Leaf

Arabidopsis CNGC2 is a Ca²⁺-permeable channel that mediates Ca²⁺ influx into mesophyll cells near minor veins but has no direct role in the hypersensitive response or leaf senescence.

Abstract

Ca²⁺ is absorbed by roots and transported upward through the xylem to the apoplastic space of the leaf, after which it is deposited into the leaf cell. In Arabidopsis (Arabidopsis thaliana), the tonoplast-localized Ca²⁺/H⁺ transporters CATION EXCHANGER1 (CAX1) and CAX3 sequester Ca²⁺ from the cytosol into the vacuole, but it is not known what transporter mediates the initial Ca²⁺ influx from the apoplast to the cytosol. Here, we report that Arabidopsis CYCLIC NUCLEOTIDE-GATED CHANNEL2 (CNGC2) encodes a protein with Ca²⁺ influx channel activity and is expressed in the leaf areas surrounding the free endings of minor veins, which is the primary site for Ca²⁺ unloading from the vasculature and influx into leaf cells. Under hydroponic growth conditions, with 0.1 mm Ca²⁺, both Arabidopsis cngc2 and cax1cax3 loss-of-function mutants grew normally. Increasing the Ca²⁺ concentration to 10 mm induced H₂O₂ accumulation, cell death, and leaf senescence and partially suppressed the hypersensitive response to avirulent pathogens in the mutants but not in the wild type. In vivo apoplastic Ca²⁺ overaccumulation was found in the leaves of cngc2 and cax1cax3 but not the wild type under the 10 mm Ca²⁺ condition, as monitored by Oregon Green BAPTA 488 5N, a low-affinity and membrane-impermeable Ca²⁺ probe. Our results indicate that CNGC2 likely has no direct roles in leaf development or the hypersensitive response but, instead, that CNGC2 could mediate Ca²⁺ influx into leaf cells. Finally, the in vivo extracellular Ca²⁺ imaging method developed in this study provides a new tool for investigating Ca²⁺ dynamics in plant cells.

Deciphering the molecular mechanisms underlying plant mineral nutrition will provide a foundation for improving the productivity of modern crops and forests (Karley and White, 2009; Liu et al., 2015; Pilbeam, 2015). Although significant progress has been made in elucidating the cellular and genetic pathways of nitrogen, phosphorus, and potassium uptake and distribution in plants in the past decade (Baker et al., 2015; Krapp, 2015; Wang and Wu, 2015), the corresponding pathways remain largely unknown for Ca (Gilliham et al., 2011; Kumar et al., 2015), an essential mineral that stabilizes cellular structures and acts as an important intracellular messenger (Dodd et al., 2010; Kudla et al., 2010; Reddy et al., 2011; Spalding and Harper, 2011; Yang et al., 2012). Ecological evidence indicates that Ca concentrations have decreased in forest soils and fresh water worldwide over the past century (Johnson et al., 2008; Bedison and Johnson, 2010; Talhelm et al., 2012; Leys et al., 2016). This Ca decrease has increased the susceptibility of forest and crop plants to stressful conditions (Hong-Bo et al., 2008; Genc et al., 2010; Halman et al., 2011) and reduced the diversity of plant species (Närhi et al., 2011). Understanding the genetic mechanisms by which Ca is absorbed by the roots and distributed in the leaves is important for addressing this environmental challenge and for breeding crops with better Ca nutrition (Robertson, 2013).

Ca²⁺ is absorbed by root cells, uploaded into the xylem, and further transported to the aerial parts of the plant, where it is unloaded from the xylem and redistributed into all types of leaf cells (White, 2001; White and Broadley, 2003; Gilliham et al., 2011). Whereas Ca²⁺ absorption and distribution have been studied at the physiological level (Kumar et al., 2015), the underlying molecular mechanisms remain largely unknown (Kudla et al., 2010; Spalding and Harper, 2011). Suberin deposits in the roots form a barrier that limits the accumulation of mineral nutrients, including Ca²⁺, in the shoots (Baxter et al., 2009). Tonoplast-localized Ca²⁺/H⁺ transporter CATION EXCHANGER1 (CAX1) and CAX3 transport Ca²⁺ from the cytosol to the vacuole of Arabidopsis (Arabidopsis thaliana) leaf mesophyll cells (Hirschi et al., 1996; Conn et al., 2011; Manohar et al., 2011; Kumar et al., 2015; Pittman and Hirschi, 2016). A cax1cax3 loss-of-function double mutant exhibits leaf growth defects and early senescence in the leaves when grown in soil and increased sensitivity to high Ca²⁺ concentrations when grown on agar medium (Cheng et al., 2005). This mutant accumulates higher levels of Ca in the extracellular space than does Columbia-0 (Col-0), as revealed by x-ray analysis and field emission scanning electron microscopy-based mineral element microanalysis of cross sections of frozen-hydrated leaves (Conn et al., 2011).

The identity of the protein responsible for initial extracellular Ca²⁺ influx into the leaf cells after Ca²⁺ unloading from the xylem is unknown. In Arabidopsis, CNGC2 is predicted to be a cyclic nucleotide-gated channel and was initially identified as a key component of the hypersensitive response (HR), a rapid, local cell death response evoked by avirulent pathogens (Yu et al., 1998; Clough et al., 2000). The HR is an important self-protection mechanism whereby plants preserve whole leaves by undergoing localized cell death at the infection site. Although the cngc2 mutant, which harbors a disruption in CNGC2, displays no localized cell death when grown in soil, its leaves maintain defense to avirulent pathogen infection (Yu et al., 1998; Clough et al., 2000). This observation has led to the notion that the HR is not necessary for plant cells to withstand infection by avirulent pathogens. Thus, the cngc2 mutant also was named defense no death (dnd1; Yu et al., 1998). However, the soil-grown mutant displays a leaf growth inhibition phenotype, constitutively accumulates high levels of salicylic acid (SA) and reactive oxygen species (Yu et al., 1998; Clough et al., 2000; McDowell et al., 2013), and exhibits conditional early senescence symptoms, reduced Ca concentrations in the leaf (Clough et al., 2000), and altered Ca and magnesium (Mg) contents in the seeds (McDowell et al., 2013). Despite this, and similar to the cax1cax3 double mutant, cngc2 grows as well as the wild-type Col-0 on agar-based medium in sterile conditions, unless the Ca²⁺ concentration is increased to 10 mm (Chan et al., 2003). To date, no clear mechanism has been proposed to explain all of the phenotypes of the cngc2 mutant. In this study, we show that all of the reported leaf phenotypes of soil-grown cngc2 could be indirect effects of its sensitivity to the Ca²⁺ supply, suggesting that the primary function of CNGC2 is to mediate the influx of apoplastic Ca²⁺ into the leaf cells. In addition, we report a new in vivo method to monitor extracellular Ca²⁺ dynamics in the leaf.

RESULTS

CNGC2 Is Required for Plant Ca Nutrition But Not for Leaf Development and the HR

To test for a role of CNGC2 in Arabidopsis Ca nutrition, we used a hydroponic system to control the amount of Ca²⁺ available to the root (Fig. 1). Col-0 and cngc2-1 plants were initially grown in medium containing 0.1 mm Ca²⁺ for 3 weeks and then transferred to medium containing 0.1, 1, 5, or 10 mm Ca²⁺ for 8 d. When grown in 0.1 mm Ca²⁺, cngc2-1 grew as well as Col-0 (Fig. 1A); however, when subjected to 1, 5, or 10 mm Ca²⁺, the cngc2-1 plants but not Col-0 showed [Ca²⁺]-dependent growth inhibition and developed leaf senescence symptoms (Fig. 1, A and B). The root length (Supplemental Fig. S1A) and fresh weight (Supplemental Fig. S1B) were comparable between the mutant and Col-0 under these conditions. In addition, the Ca²⁺-sensitive phenotype of cngc2-1 could be complemented by expressing wild-type CNGC2 driven by its native promoter (COM, which denotes complemented with CNGC2; Fig. 1C). Extending the treatment to 30 d strongly inhibited the stem elongation and inflorescence development of cngc2 but not of the Col-0 or COM plants (Fig. 1D). These results demonstrated that CNGC2 has an essential role in Ca²⁺ nutrition.

Figure 1.

Arabidopsis hydroponics-grown cngc2, cngc4, and cax1cax3 mutants are sensitive to the increasing [Ca²⁺]. Arabidopsis plants were grown in hydroponic medium supplemented with 0.1 mm Ca²⁺ for 3 weeks and then transferred to medium containing the indicated concentrations of Ca²⁺. A, Col-0 and cngc2-1 were treated with 0.1, 1, 5, or 10 mm [Ca²⁺] for 8 d. B, Leaf fresh weight of plants from A. Asterisks indicate significant differences between Col-0 and cngc2-1: **, P < 0.01. Values represent means ± se; n = 3 biological trials with 40 plants per trial. C, Representative plants of Col-0, cax1cax3, cngc2-1, and cngc2-1 complemented with CNGC2 (COM) treated with 0.1 or 10 mm Ca²⁺ for 8 d. D, Representative plants of Col-0, cax1cax3 (1/3), cngc2-1 (2-1), and COM treated with 0.1 or 10 mm Ca²⁺ for 30 d. E, Representative plants of Col-0, cngc2-3, cngc4-5, and cngc2-3cngc4-5 after 8 d of 10 mm Ca²⁺ treatment.

The Arabidopsis genome contains 20 CNGC genes (Ward, 2001), 15 of which we tested for their involvement in the Ca²⁺ growth phenotype, using the corresponding T-DNA insertion mutants. Besides cngc2, the only mutant that showed a leaf growth inhibition phenotype when grown on 10 mm Ca²⁺ was cngc4 (Supplemental Fig. S2). CNGC4 is the closest homolog of CNGC2 in the Arabidopsis genome, and soil-grown cngc4 has the same leaf-growth inhibition phenotype as cngc2 (Jurkowski et al., 2004). To further explore this result, we grew two cngc2 and cngc4 mutant alleles as well as the cngc2cngc4 double mutant and found that all of them showed similar [Ca²⁺]-sensitive symptoms as the cngc2-1 mutants (Fig. 1E). These results further demonstrated the vital role of CNGC2 in Ca²⁺ nutrition.

Leaves of soil-grown cngc2-1 plants have spontaneous lesions, occasional cell death spots, and constitutively elevated SA contents and expression of PR1 (Yu et al., 1998; Clough et al., 2000; McDowell et al., 2013). In this study, under 0.1 mm Ca²⁺ conditions, the leaves of cngc2-1 and Col-0 had similar low levels of staining for cell death (Fig. 2A) and H₂O₂ (Fig. 2B) as well as basal levels of SA (Fig. 2C) and PR1 gene expression (Fig. 2D). Treatment with 10 mm Ca²⁺ for 8 d significantly increased cell death (Fig. 2A), H₂O₂ (Fig. 2B), and SA accumulation (Fig. 2C) as well as PR1 gene expression (Fig. 2D) in the leaves of cngc2-1 but not of Col-0 and the COM plants. These findings demonstrated that the Ca²⁺ sensitivity of cngc2-1 accounts for the spontaneous high SA, H₂O₂ production, and cell death spots in cngc2-1 leaves.

Figure 2.

Increasing [Ca²⁺] promotes cell death and H₂O₂ and SA production in the leaves of cax1cax3 and cngc2 plants. A, Col-0, cax1cax3, cngc2-1, and cngc2-1 complemented with CNGC2 (COM) were initially grown on 0.1 mm Ca²⁺ and then treated with 0.1 or 10 mm Ca²⁺ for 8 d. Leaves with typical symptoms were examined with a dissecting microscope and stained with Trypan Blue for cell death, shown by dark blue spots. Images were taken before and after the staining. B, Plants were treated as in A and stained with 3,3-diaminobenzidine (DAB) for H₂O₂ accumulation, shown by brown spots. C, Free SA contents in the leaves of Col-0 and cngc2-1 grown in the hydroponics system after 3 d of treatment with 0.1 or 10 mm Ca²⁺. D, Plants were treated as in C, and PR1 gene expression was examined and normalized to the internal standard and the PR1 expression in Col-0 under 0.1 mm Ca²⁺ conditions. Values in C and D represent means ± se; n = 3 biological trials. **P < 0.01 between the 0.1 and 10 mm Ca²⁺ treatments.

The cngc2 is also named dnd1, since the mutant maintains defense to avirulent pathogens without displaying HR (Yu et al., 1998; Clough et al., 2000). To our knowledge, all of the reported cngc2-1 (dnd1) HR phenotype characterizations were performed with leaves of soil-grown mutant plants showing early senescence or growth inhibition phenotypes (Yu et al., 1998; Clough et al., 2000; Ahn, 2007; Ali et al., 2007; Genger et al., 2008; Chin et al., 2013) similar to those in plants given 10 mm Ca²⁺ treatment using the hydroponics system in this study (Fig. 1). We reasoned that it was possible that the dnd1 phenotype of cngc2-1 is also Ca²⁺ supply dependent. To test this hypothesis, we infiltrated low (10⁷ colony-forming units [cfu] mL⁻¹) and high (10⁸ cfu mL⁻¹) densities of the avirulent pathogen Pseudomonas syringae pv tomato DC3000 expressing avrRpm1 into leaves of Col-0, cngc2-1, cngc2-3, and COM plants. Under 0.1 mm Ca²⁺ growth conditions, all four genotypes displayed a typical HR (Fig. 3). A 3-d treatment with 10 mm Ca²⁺ did not cause obvious leaf senescence in the mutants and Col-0 but inhibited development of the HR in the leaves of cngc2-1 but not of Col-0 and COM, in response to the low-density pathogen (Fig. 3). However, with the 10 mm Ca²⁺ treatment, all three genotypes displayed a normal HR in response to the high-density pathogen (10⁸ cfu mL⁻¹; Fig. 3).

Figure 3.

HR of Arabidopsis Col-0, cngc2, and cax1cax3 mutants. Arabidopsis Col-0, cngc2-1, cngc2-1 complemented with CNGC2 (COM), and cax1cax3 plants were first grown in hydroponic solution supplemented with 0.1 mm Ca²⁺ for 3 weeks and then subjected to 3-d treatments with 0.1 or 10 mm Ca²⁺. The leaves were infiltrated with either water or a suspension of P. syringae pv tomato DC3000 avrRpm1 in water at a density of 1 × 10⁷ or 1 × 10⁸ cfu mL⁻¹. Leaves with typical symptoms were photographed 24 h after infection.

These data demonstrated that CNGC2 plays an essential role in plant Ca nutrition and that the previously observed leaf phenotypes of the soil-grown cngc2 mutant could be an indirect result of CNGC2 function in Ca nutrition.

The cngc2 Mutant and cax1cax3 Double Mutants Have Similar Ca-Dependent Phenotypes

To dissect the underlying mechanism of CNGC2-mediated Ca nutrition, we measured total Ca contents in the leaves and roots of Col-0 and cngc2-1 plants that we initially cultured on 0.1 mm Ca²⁺ and then treated for 3 d with 0.1, 1, or 10 mm Ca²⁺. The 3-d treatments had no noticeable effects on the leaf growth and development of Col-0 and cngc2-1. Col-0 and cngc2-1 had similar Ca contents in their leaves at 0.1 and 1 mm Ca²⁺, and cngc2-1 had significantly less Ca than Col-0 at 10 mm Ca²⁺ (Fig. 4A), which is similar to results with soil-grown plants (Clough et al., 2000; Ma et al., 2010). The total Ca content in cngc2-1 roots was comparable to that of Col-0 (Fig. 4B). The Mg contents showed no difference between Col-0 and the mutant in both leaf and root (Fig. 4, C and D). These data imply that cngc2-1 primarily was defective in Ca²⁺ distribution in the leaf but not in Ca²⁺ transport from the root to the leaf. This hypothesis was further supported by the observation that the detached leaves of cngc2-1 developed similar senescence symptoms to those of intact plants under 10 mm [Ca²⁺] (Fig. 4E). These findings suggested that cngc2-1 was defective primarily in Ca²⁺ distribution in the leaf but not in Ca²⁺ transport from the root to the leaf.

Figure 4.

Ca and Mg contents of Col-0 and cngc2-1 plants and their detached leaf responses to [Ca²⁺]. A to D, Col-0 and cngc2-1 initially were grown under hydroponic conditions with 0.1 mm Ca²⁺ for 3 weeks, then given 0.1, 1, or 10 mm Ca²⁺ for 3 d. Ca contents in the leaves (A) and roots (B), and Mg contents in the leaves (C) and roots (D), were measured. Values represent means of three biological replicates ± se. **, Significant difference at P < 0.01 with Student’s t test between Col-0 and cngc2-1 under the same treatment. E, Col-0 and cngc2-1 initially were grown under hydroponic conditions with 0.1 mm Ca²⁺ for 3 weeks. Fully expanded leaves were detached and inserted into a 200-µL tube with full nutrition medium containing either 0.1 or 10 mm Ca²⁺ for 8 d. Leaves with typical symptoms are shown.

The two most important genes known to be involved in Arabidopsis leaf Ca²⁺ nutrition are CAX1 and CAX3, which encode Ca²⁺/H⁺ antiporters localized in the tonoplast of leaf mesophyll cells (Hirschi et al., 1996; Manohar et al., 2011; Kumar et al., 2015). The Arabidopsis cax1cax3 double mutant has a very similar phenotype to that of cngc2-1: the soil-grown cax1cax3 plants have growth defects and early senescence phenotypes; moreover, cax1cax3 plants grown on agar-based medium are sensitive to 20 mm Ca²⁺ (Cheng et al., 2005). Disruption of CAX1 and CAX3 blocks cytosolic Ca²⁺ storage in the vacuole and causes Ca²⁺ to overaccumulate in the apoplast (Conn et al., 2011). By analogy, we wondered whether CNGC2 mediates Ca²⁺ influx into the mesophyll cells and if disrupting CNGC2 causes Ca²⁺ accumulation in the apoplast, which would account for the mutant’s sensitivity to increasing Ca²⁺ supply to its roots. If CNGC2, CAX1, and CAX3 indeed function in same pathway mediating extracellular Ca²⁺ deposit to leaf cells, the cax1cax3 and cngc2 mutants might have similar Ca-dependent phenotypes. To test this hypothesis, we first grew Col-0, cngc2-1, COM, and cax1cax3 plants on the agarose-based medium supplied with different [Ca²⁺] levels. Similar to cngc2-1, the growth of cax1cax3 plants was no different compared with Col-0 at 0.1 mm Ca²⁺, but its leaf growth (Supplemental Fig. S3, A and B) and root elongation (Supplemental Fig. S3C) were suppressed significantly by increasing [Ca²⁺]. In the hydroponic condition with 0.1 mm Ca²⁺, the cngc2-1 and cax1cax3 mutants grew as well as Col-0 and COM (Fig. 1, C and D); however, the 8-d (Fig. 1C) or 30-d (Fig. 1D) 10 mm Ca²⁺ treatment inhibited cax1cax3 leaf growth as much as that of cngc2-1, promoted senescence and cell death (Fig. 2A), induced the accumulation of H₂O₂ (Fig. 2B) and SA (Fig. 2C), and suppressed the HR induced by infection with avirulent pathogen at low density (Fig. 3). These results implied that CNGC2 and CAX1/CAX3 have similar functions in leaf Ca nutrition.

In Vivo Imaging of Extracellular Ca²⁺ in cngc2 and cax1cax3 Leaves

Disruption of the pathway from the cytosol to the vacuole for Ca²⁺ storage causes the apoplastic (extracellular) space of cax1cax3 mutants to overaccumulate Ca, which was originally measured using x-ray microanalysis in fixed leaf cross sections (Conn et al., 2011). Although several luminescence- and fluorescence-based methods have been developed to measure intracellular Ca²⁺ distribution and dynamics in living plant cells (Knight et al., 1991; Behera et al., 2013), there has not been an in vivo method reported to date to visualize extracellular Ca²⁺ distribution in plant cells. In animal and human cells, Oregon Green BAPTA 488 5N (OGB-5N), a well-established membrane-impermeable low-affinity Ca²⁺ fluorescent dye (White and McGeown, 2002; Jiménez-Moreno et al., 2008), has been used to monitor extracellular Ca²⁺ dynamics (Rusakov and Fine, 2003), endocytosis-mediated extracellular Ca²⁺ uptake (Gerasimenko et al., 1998; Sherwood et al., 2007), and intracellular Ca²⁺ dynamics in giant cells like oocytes (Callamaras and Parker, 2000) and muscle cells (Hollingworth et al., 2009) through microinjection or the pipette during patch-clamp whole-cell recordings. Here, we developed a method to use OGB-5N in plants to track extracellular Ca²⁺ transport and distribution in the leaf.

We first grew Col-0 and cngc2-1 in the hydroponic system with 0.1 mm Ca²⁺ and then applied a 3-d 0.1 or 10 mm Ca²⁺ treatment. Afterward, we placed the petioles of detached mature leaves from the hydroponics-cultured plants into the corresponding growth solutions containing OGB-5N. At time 0, no fluorescence was observed in all tested leaves (Fig. 5A). After 20 min, clear green fluorescence with similar patterns and intensity appeared in the vascular systems of Col-0 and cngc2-1 in the 0.1 and 10 mm Ca²⁺ conditions (Fig. 5A). After 4 h of staining, leaves of both Col-0 and cngc2-1 treated with 0.1 mm Ca²⁺ showed extracellular [Ca²⁺] ([Ca²⁺]_ext)-dependent green fluorescence primarily in the third and minor veins, with no obvious difference between genotypes in terms of pattern and intensity (Fig. 5A). No detectable fluorescence was present in the nonvascular tissues (e.g. veinlet, the leaf areas divided by the veins; Fig. 5A). Thus, it appears that Ca²⁺ unloaded from the vascular system was actively redistributed into other leaf cells and that very little free Ca²⁺ was extracellular. In the leaves treated with 10 mm Ca²⁺, the [Ca²⁺]_ext-dependent green fluorescence was still restricted to the vascular tissues in Col-0 but it spread to the nonvascular tissues (veinlet) in cngc2-1 leaves (Fig. 5A). This difference implied that disrupting CNGC2 causes extracellular Ca²⁺ accumulation around the leaf cells. Furthermore, in a separate set of experiments, extracellular Ca²⁺ accumulation around the leaf cells was observed with two alleles of cngc2 and cax1cax3 and was rescued back to the Col-0 phenotype in the complementation line COM (Fig. 5B).

Figure 5.

Fluorescence microscopy of extracellular Ca²⁺ in leaves of Col-0, cngc2, and cax1cax3. Col-0, cngc2-1, cax1cax3, and cngc2-1 complemented with CNGC2 (COM) were hydroponically grown at 0.1 mm Ca²⁺ for 3 weeks and then given 3-d treatments of 0.1 or 10 mm Ca²⁺. Fully expanded leaves were detached and inserted into 200-μL tubes with their corresponding growth medium in the presence of 20 μm Ca²⁺-dependent fluorescent dye OGB-5N. A, Bright-field and fluorescence images of these leaves were taken at different time points of the OGB-5N treatment with either 2× or 10× objective lenses. The areas defined by the yellow squares in the 4-h (2×) images were magnified and are shown in the 4-h (10×) images. B, Treatment was the same as in A, and images were taken at 4 h.

To visualize the extracellular Ca²⁺ distribution in the leaf with higher spatial resolution, we used confocal laser scanning microscopy. The cell wall boundaries of the leaf cells were defined using the red fluorescent dye propidium iodide. At 0.1 mm Ca²⁺, extracellular Ca²⁺ was detected exclusively in the vascular system (veins), with very little if any green fluorescence detected in the surrounded leaf areas (veinlet) of Col-0, cax1cax3, cngc2-1, and COM plants (Fig. 6). At 10 mm Ca²⁺, the [Ca²⁺]_ext-dependent green fluorescence was still almost exclusively present in the vascular tissues (veins) of Col-0 and COM, especially the minor veins, with very limited signal in the nonvascular leaf areas (veinlet; Fig. 6). However, a significant amount of extracellular Ca²⁺-dependent green fluorescence was detected around veinlet areas in cngc2 and cax1cax3 leaves (Fig. 6).

Figure 6.

Confocal microscopy of extracellular Ca²⁺ in leaves of Col-0, cngc2, and cax1cax3 plants. A, The plant growth and treatment procedures were the same as in Figure 7. The extracellular Ca²⁺ (Ext Ca²⁺) and the cell wall boundary of the epidermal cells were visualized using the Ca²⁺-dependent green fluorescent dye OGB-5N and red fluorescent dye propidium iodide, respectively, and observed by confocal laser scanning microscopy. Ext Ca²⁺ + Cell Wall indicates the merged images of Ext Ca²⁺ and Cell Wall. COM denotes the cngc2-1 line complemented with CNGC2. B, Relative fluorescence intensity between the propidium iodide-dependent red and OGB-5N-dependent green signals was compared between Col-0 and mutants under 0.1 or 10 mm Ca²⁺ conditions in the vein and nonvein (veinlet) areas. Values represent means of three biological trials ± se, with five leaves per trial and five areas per leaf. **P < 0.01 between the 0.1 and 10 mm Ca²⁺ treatments.

Ca in plants can be in a water-soluble ionic form or chelated by cell wall components and other acidic compounds like oxalate (Gilliham et al., 2011). It has been suggested that Ca²⁺ overaccumulation in the extracellular space of the cax1cax3 leaf can increase Ca²⁺ deposition into the cell wall and decrease Ca²⁺ in the cytoplasm (Conn et al., 2011). To investigate this hypothesis, Ca in the leaves of Col-0, cngc2-1, and cax1cax3 with 3 d of exposure to either 0.1 or 10 mm Ca²⁺ was sequentially extracted with water, acetic acid, and HCl to isolate water-soluble Ca²⁺, the Ca²⁺ bound to the cell wall, and Ca²⁺ in complexes with acid compounds, respectively. After water and acetic acid extractions, we did not detect Ca in the HCl extract. With the 0.1 mm Ca²⁺ treatment, there was no significant difference between Col-0 and cngc2-1 or cax1cax3 in either water-soluble (Fig. 7, A and C) or acetic acid-soluble (Fig. 7, B and D) Ca²⁺ contents. However, at 10 mm Ca²⁺, cngc2-1 and cax1cax3 plants had significantly less water-soluble Ca²⁺ (Fig. 7, A and C) but more acetic acid-soluble Ca²⁺ than Col-0 (Fig. 7, B and D). This result was consistent with the hypothesis that increasing the accumulation of Ca²⁺ in the extracellular space promotes Ca²⁺ deposition into the cell wall of cngc2 and cax1cax3 plants under high-Ca²⁺ growth conditions.

Figure 7.

Water- and acetic acid-soluble Ca in the leaves of Col-0, cngc2-1, and cax1cax3. Col-0, cngc2-1, and cax1cax3 plants were grown hydroponically with 0.1 mm Ca²⁺ for 3 weeks and then given a 3-d treatment with 0.1 or 10 mm Ca²⁺. Fully expanded leaves were detached and extracted with water (A and C) for Ca content measurement. After the water extraction, the concentrated tissues were sequentially extracted with acetic acid (AA) for Ca content measurement (B and D). Values represent means ± se; n = 3 biological trials, five replicates per trial, and leaves from four plants per replicate. **, P < 0.01.

CNGC2 Is Expressed around Minor Veins and Encodes a Ca²⁺ Influx Channel

The Ca²⁺ overaccumulation in the extracellular space of cngc2-1 leaf cells appeared preferentially in the areas surrounding the minor veins (Figs. 5 and 6), which might reflect cell-specific expression of CNGC2. For the 12-d-old seedlings grown on the agarose-based medium, CNGC2 promoter activity was detected primarily in the leaves (Supplemental Fig. S4A) and weakly in the vascular tissue of the roots (Supplemental Fig. S4, B and C). For the mature plants grown on the hydroponics system, the CNGC2 promoter activity was weakly detected in the stem (Supplemental Fig. S4D), flowers (Supplemental Fig. S4E), developing siliques (Supplemental Fig. S4F), and not in the roots (Supplemental Fig. S4G). In the leaf, besides the activity found in the cells surrounding trichomes (Supplemental Fig. S4H), CNGC2 was expressed specifically in the vascular tissues and the leaf cells proximate to the vascular system, especially in the leaf areas around the free endings of minor veins, under both 0.1 and 10 mm Ca²⁺ conditions (Fig. 8A).

Figure 8.

CNGC2 minor vein area-specific expression and inward Ca²⁺ channel activity in HEK293T cells. A, GUS reporter gene-based CNGC2 gene expression pattern (blue areas) in the leaf at 0.1 mm Ca²⁺. The red circles indicate the positions of the minor vein free-ending areas. The pattern in the leaf at 10 mm Ca²⁺ was similar to that at 0.1 mm. B, CNGC2 was transiently expressed in HEK293T cells. Patch-clamp whole-cell recording was used to monitor current across the plasma membrane under continuous voltage change from +20 to –180 mV. Representative recordings are shown under the indicated conditions. C, Average current-voltage curves are shown for the cell transformed with empty vector and recorded with 200 μm 8-bromoadenosine cAMP (8-Bromo-cAMP) and 10 mm Ca²⁺ in the bath solution (n = 7) and the cells transformed with CNGC2 in bath solution containing 200 μm 8-Bromo-cAMP and 10 mm Ca²⁺ (n = 6), 10 mm Ca²⁺ + 1 mm GdCl₃ (n = 5), or 30 mm Ca²⁺ (n = 9). Values represent means ± se.

Free-ending areas of these minor veins are the primary places at which mineral ions, including Ca²⁺, are unloaded from the vascular system and redistributed into the surrounding leaf cells (Ahn, 2007). Thus, it seemed that CNGC2 might function as a Ca²⁺ influx transporter mediating Ca²⁺ influx into the leaf cells immediately after Ca²⁺ unloading from the vascular tissues. Indirect evidence has suggested that CNGC2 is a Ca²⁺ influx channel: the plasma membrane of cngc2 guard cells lacks the cAMP-activated Ca²⁺ influx current (Ali et al., 2007), and CNGC2 mediates a cAMP-induced increase in cytosolic Ca²⁺ when it is expressed in mammalian HEK293 cells (Leng et al., 1999). Recently, CNGC18, one of 20 CNGCs in Arabidopsis, was demonstrated to have Ca²⁺ influx channel activity when expressed in HEK293T cells, as measured by a patch-clamp whole-cell configuration (Gao et al., 2014, 2016) with an experimental setup similar to that used for previous recordings of the CRACM1 Ca²⁺ channel (Vig et al., 2006). To test directly for CNGC2 channel activity, we used the same protocol to express and study CNGC2 in HEK293T cells.

With 10 mm Ca²⁺ in the bath, an inward current was observed in HEK293T cells expressing CNGC2 but not in the cells with empty vector (Supplemental Fig. S5). The CNGC2-dependent current showed a moderate response to the membrane-permeable lipophilic cAMP analog 8-bromoadenosine cAMP (Supplemental Fig. S5). There are five significant ions in the recording solutions: Ca²⁺, Na⁺, Cs⁺, Mg²⁺, and Cl⁻ (Vig et al., 2006; Gao et al., 2014, 2016). At room temperature (23°C), with the assumption that only a single ion has the ability to cross the membrane, the theoretical equilibrium potentials for the individual ions are 140 mV (Ca²⁺), 69 mV (Na⁺), −63 mV (Cs⁺), −13 mV (Mg²⁺), and −48 mV (Cl⁻). The measured reverse membrane potential for the CNGC2-dependent current was close to 0 mV (Supplemental Fig. S5), which was not close enough to any of the theoretical equilibrium potentials that the ionic composition of the current could be determined. However, additional observations suggested that the CNGC2-dependent current had a Ca²⁺ influx component. First, this was strongly inhibited by adding Gd³⁺, an inward Ca²⁺ channel inhibitor, into the bath solution (Fig. 8, B and C). Second, the theoretical equilibrium potential for Ca²⁺, with 175 nm free Ca²⁺ inside the cell, is predicted to increase from 140 to 154 mV when the bath [Ca²⁺] is increased from 10 to 30 mm; experimentally, the reverse potential of the CNGC2-dependent current showed a similarly positive shift from 0 to around 17 mV (Fig. 8, B and C). In addition, the measured reverse membrane potential for the CNGC2-dependent current (Fig. 8, B and C) was similar to those reported for CNGC18 (Gao et al., 2014, 2016) and CRACM1 (Vig et al., 2006) Ca channels under the same ionic recording conditions. These data support the conclusion that CNGC2 has Ca²⁺ influx channel activity.

CNGC2 shows inward K⁺ channel activity when expressed in oocytes (Leng et al., 1999). By contrast, CNGC18 does not have inward K⁺ channel activity in HEK293T cells (Gao et al., 2014). Here, we did not detect inward K⁺ channel activity in the CNGC2-expressing HEK293T cells (Supplemental Fig. S5), although the KAT1 positive control showed typical inward K⁺ channel activity (Supplemental Fig. S6). In summary, we conclude that CNGC2, similar to CNGC18 (Gao et al., 2014), can mediate Ca²⁺ but not K⁺ influx across the plasma membrane in HEK293T cells.

DISCUSSION

Plant growth and development depend on mineral nutrients absorbed by the roots and redistributed into leaves through the apoplast (extracellular space) and symplast (intracellular compartments) pathways, facilitated by transpiration and membrane-located mineral ion transporters, respectively (Kumar et al., 2015). Ca is an essential plant macronutrient with key roles in cellular signaling and structural integrity (Dodd et al., 2010; Kudla et al., 2010; Reddy et al., 2011; Yang et al., 2012). The physiological pathways mediating Ca absorption and distribution in plants are well studied (White and Broadley, 2003). However, the corresponding genetic mechanisms are largely unknown (Kudla et al., 2010; Spalding and Harper, 2011). In this study, we have provided several lines of evidence to support a key role of CNGC2 in mediating Ca²⁺ influx into Arabidopsis leaf cells after unloading from the vascular tissues (Fig. 9). First, CNGC2 shows Ca²⁺ influx channel activity when expressed in HEK293T cells (Fig. 8, B and C). Second, CNGC2 is expressed predominantly in the leaf cells surrounding the free endings of the minor veins, which are the primary sites for Ca²⁺ unloading from the vascular tissues into the leaf cells (Fig. 8A). Third, disrupting CNGC2 caused a significant amount of Ca²⁺ to accumulate in the extracellular space of leaf epidermal cells, similar to cax1cax3 double mutants, which are known to be defective in Ca²⁺ transport from the extracellular space to the intracellular compartments (Figs. 5 and 6). Finally, at 0.1 mm Ca²⁺, the cngc2 mutant grew as well as the wild type; however, increasing the [Ca²⁺] induced cell death, H₂O₂ and SA accumulation (Fig. 2), as well as growth inhibition in the leaves of cngc2-1 but not of the wild type (Fig. 1), which indicates the importance for plant cells to maintain low extracellular Ca²⁺ through CNGC2-mediated deposition of Ca²⁺ into leaf cells.

Figure 9.

Model of CNGC2, CAX1, and CAX3 function in the transport of Ca²⁺ from the apoplast to the vacuole in Arabidopsis leaves. Ca²⁺ is absorbed by roots and transported upward to the apoplastic space of the leaf through the xylem driven by the leaf transpiration stream. After the Ca²⁺ in the xylem unloads into the apoplastic space of the leaf cells, it subsequently is transported into the cytosol through the plasma membrane-localized CNGC2 channel. The Ca²⁺ in the cytosol is sequestered into the vacuole via the tonoplast-localized Ca²⁺/H⁺ antiporters CAX1 and CAX3.

Since the discovery of the first CNGC2 loss-of-function mutant dnd1 (cngc2-1) in 1998, it has been noted that the mutant leaves on soil-grown plants are smaller and senesce earlier than those of wild-type plants (Yu et al., 1998). In addition, the mutant leaves have constitutively higher accumulation of H₂O₂ and SA than the wild-type leaves when grown in soil (Yu et al., 1998; Clough et al., 2000). However, the mutant grows normally, like the wild type, when grown on agar medium in petri dishes unless the [Ca²⁺] increases to 10 mm (Chan et al., 2003). In this study, by controlling the Ca²⁺ supply to the roots through a hydroponic growth system (Fig. 1), we demonstrated that it was the increasing [Ca²⁺]_ext that was responsible for the observed leaf phenotypes of the mutant. It is possible that disrupting CNGC2 decreases the Ca²⁺ uptake capacity of the leaf cells near the minor veins. With low Ca²⁺ supply, the remaining Ca²⁺ uptake capacity of the cngc2 leaf cells still would efficiently absorb the Ca²⁺ in the extracellular space that was unloaded from the vascular tissues, to maintain a low [Ca²⁺]_ext. However, when the Ca²⁺ supply increased, the amount of Ca²⁺ absorbed by the mutant leaf cells would not be able to keep pace with the amount of Ca²⁺ unloading from the vascular tissue, which in turn would lead to Ca²⁺ overaccumulation in the extracellular space of the mutant cells (Fig. 9).

The data presented in this study demonstrate that maintaining a low [Ca²⁺]_ext is essential for the physiological functions of leaf cells and that CNGC2 plays a key role in this process. The increase of the [Ca²⁺]_ext due to the inactivation of CNGC2 interrupts multiple aspects of leaf function through largely unknown cellular and genetic pathways. Increasing [Ca²⁺]_ext activates class III peroxidase in the apoplast, which further promotes the degradation of H₂O₂ (Hepler and Winship, 2010). The plasma membrane-localized superoxide-generating NADPH oxidase RBOHD can be activated by the cytosolic [Ca²⁺] either directly or indirectly, which is responsible for the H₂O₂ burst in the apoplast (Hepler and Winship, 2010). In addition, H₂O₂ can activate a Ca²⁺-permeable channel in the plasma membrane of the root and guard cells (Hepler and Winship, 2010). An increase in cytosolic [Ca²⁺] can activate SA biosynthesis through Ca²⁺-binding proteins (Seyfferth and Tsuda, 2014). However, it is unknown how the increase of [Ca²⁺]_ext is linked to H₂O₂ and SA production. Moreover, increases in [Ca²⁺]_ext can suppress water flow in leaf, which might interfere with photosynthesis (Gilliham et al., 2011).

It is well documented that the leaves of soil-grown dnd1 (cngc2-1) mutants lack an HR but maintain their defenses against avirulent pathogen infection (Yu et al., 1998; Ali et al., 2007). In this study, the mutant leaf had a perfect HR when its root was supplied with 0.1 mm Ca²⁺ (Fig. 3). Three-day treatment with high [Ca²⁺] (10 mm) could suppress the HR induced by infection with the pathogen at low density but not at high density (Fig. 3). These data demonstrated that CNGC2 is not essential for the HR. The mutant’s defective HR in response to the low-density avirulent pathogen in this study could be an indirect consequence of the overaccumulation of Ca²⁺ in the apoplast. We hypothesize that the increase of [Ca²⁺]_ext enhances the interaction of the Ca²⁺ with pectin and further strengthens the cell wall (Hepler and Winship, 2010), which could protect the leaf cell from death as a response to the avirulent pathogen (Choi et al., 2013; Mortimer et al., 2013; Johansson et al., 2014) but also could limit cell extensibility and decrease its growth rate. This possibility is supported by the findings that cax1cax3 mutant leaf cell wall rigidity is increased under high [Ca²⁺] conditions (Conn et al., 2011).

In conclusion, this study not only identifies CNGC2 as the key factor mediating Ca²⁺ influx into the leaf cells after Ca²⁺ is unloaded into the apoplast from the vascular tissues but also provides an explanation for the growth and HR-defective phenotypes of the soil-grown cngc2-1 (dnd1) mutant known since 1998 (Yu et al., 1998; Ali et al., 2007). Finally, the in vivo extracellular Ca²⁺ imaging method developed in this study provides a new tool for investigating Ca²⁺ dynamics in plant cells.

MATERIALS AND METHODS

Arabidopsis Growth Conditions and Pathogen Inoculation

To grow Arabidopsis (Arabidopsis thaliana) hydroponically, the seeds were sterilized with 75% and 100% ethanol for 15 and 5 min, respectively. The ethanol-treated seeds were dried on sterilized filter paper inside a hood. Twelve seeds were evenly distributed on an agarose-based control medium inside a petri dish, which contained the control solution with 0.1 mm CaCl₂ (Wang et al., 2015). After a 3-d treatment at 4°C, the petri dishes with seeds were put horizontally inside a growth chamber with 12/12-h light/dark period and 100 to 120 μmol m⁻² s⁻¹ light intensity. Three-week-old seedlings were transferred to a 5-mL tube hydroponics box containing the same nutrient solution as a control but without sugar and agar. The hydroponic solution was changed weekly. After another 3 weeks, the solution was changed to that with different concentrations of Ca²⁺ and other nutrients for the times indicated in the text. The cngc2-1 (dnd1), cngc2-3 (salk_066908), cngc4-1 (dnd2), cngc4-5 (salk_081369), and cax1cax3 mutants were reported previously (Clough et al., 2000; Cheng et al., 2005; Genger et al., 2008; Liu et al., 2011; Chin et al., 2013). The cngc2-3cngc4-5 double mutant was identified using PCR in the F2 population from F1 seeds created by cross-pollination of the two single mutants. Other cngc SALK mutants (Alonso et al., 2003) and primers used for homozygous identification are listed in Supplemental Table S1.

For the hypersensitive response assay, after 3-d treatments with either 0.1 or 10 mm Ca²⁺, we infiltrated the avirulent pathogen Pseudomonas syringae pv tomato DC3000 avrRpm into the abaxial side of the leaves with a 1-mL syringe as described previously (Li et al., 2013).

Ca and Mg Content Measurement by Atomic Absorption Spectrometry

The leaf or root tissues for Ca and Mg content measurements in Figure 4 were harvested and dried in an oven at 105°C for 5 to 6 d. The dried tissues were weighed and ground to powder, which was further extracted with 1.5 m HCl for 2 d. The samples were centrifuged at 12,000g for 30 min, and the supernatants were used for Ca and Mg measurement with an atomic absorption spectrometer (TAS-990; Beijing Purkinje General).

For the sequential Ca extraction in Figure 7, the leaves were immediately frozen in liquid nitrogen after harvest and ground into fine powder for weighing (Borer et al., 2004, 2012). The powder was transferred into a 50-mL tube, and 20 mL of ultrapure water from the ELGA Option-Q ultrapure water system was added. The basal Ca content in the water was below 10 μm as measured with the Amplite Rhod Red fluorescence-based Ca quantification kit (Biolite). The powder-containing tube was shaken at 180 rpm, at room temperature, overnight and subsequently centrifuged at 8,000g for 30 min. The supernatant was saved as the water-soluble Ca for Ca measurement. The pellet was washed once with the ultrapure water and put into a 50°C oven overnight. The dried pellet was dissolved in 20 mL of acetic acid and shaken at 180 rpm, at room temperature, overnight. After centrifugation at 8,000g for 30 min, the supernatant was saved as the acetic acid-soluble Ca for Ca measurement. The pellet was subjected to further extraction with HCl.

Reactive Oxygen Species and Cell Death Assays in Arabidopsis Leaf

Arabidopsis leaf cell death was monitored with a Trypan Blue (Sigma-Aldrich) staining assay as described at http://commonweb.unifr.ch/biol/pub/mauchgroup/staining.html. The method for the 3,3-diaminobenzidine (Sigma-Aldrich)-based reactive oxygen species staining assay was as described previously (Zou et al., 2015).

SA Measurement and Real-Time Reverse Transcription-PCR

Arabidopsis plants were cultured in hydroponic conditions with 0.1 mm Ca²⁺ for 3 weeks and then given either 0.1 (as a control) or 10 mm Ca²⁺ treatment. Afterward, the well-expanded leaves were collected and ground into fine powder in liquid nitrogen, 50 to 100 mg of which was added into 1 mL of extraction buffer containing propanol:water:HCl (2:1:0.002, v/v/v) and 50 ng of ²H-labeled SA as an internal standard. The mixture was rotated at 100 rpm at 4°C for 1 h, after which 1 mL of dichloromethane was added and the sample was rotated at 100 rpm and 4°C for another 1 h. The mixture was then centrifuged at 12,000g at 4°C for 30 min. The tissue debris was found between the two layers of liquid phases. Solution (200 μL) from the bottom phase was taken for total SA quantification as described previously (Pan et al., 2010).

For RNA extraction, the leaf tissue fine powder was extracted with the TransZol Plant kit (Transgen Biotech). cDNA was synthesized with TransScript All-in-One First Strand cDNA Synthesis Supermix for qPCR with One-Step gDNA Removal (Transgen Biotech) following the manufacturer’s instructions. Primers for the internal standards TUBULIN β-SUBUNIT (At5g44340) and PR1 (At2g14610) were reported previously (Ma et al., 2010). Quantitative real-time PCR was performed using a SYBR Premix Ex Taq II (TaKaRa) in a qTOWER 2.2 real-time PCR system (Analytik). The level of gene transcript accumulation was first normalized to the internal standard TUBULIN β-SUBUNIT (At5g44340) and expressed relative to the PR1 gene value in Col-0 under the 0.1 mm Ca²⁺ condition.

GUS Reporter Gene-Based Promoter Activity Assay

A 2,186-bp DNA fragment was amplified from the genome region upstream of the CNGC2 start codon with Col-0 genomic DNA as a template. The primers used were 5′-CTAGAAGCTTCTCCAAGCCAAGCCTGGTTAAAATCGG-3′ and 5′-ATTGGATCCGATTGAAATAGAGGAACCACCATGGGAG-3′, the underlined portions of which are restriction enzyme sites for HindIII and BamHI, respectively. The amplified DNA fragment was integrated into the HindIII and BamHI cloning sites upstream of the GUS reporter gene inside the binary vector pORE R1. Agrobacterium tumefaciens strain GV3101 harboring the 2,186-bp DNA fragment-containing pORE R1 vector (Coutu et al., 2007) was used to transform Arabidopsis (Clough and Bent, 1998). Transgenic Arabidopsis screening and GUS assays were as described previously (Zou et al., 2015).

HEK293T Cell Culture, Transfection, and Patch-Clamp Whole-Cell Recording

The CNGC2 coding sequence was amplified with primers 5′-CGGAATTCATGCCCTCTCACCCCAACTTCAT-3′ and 5′-CCCCGGGTTATTCGAGATGATCATGCGGTCG-3′, the underlined portions of which are restriction enzyme sites for EcoRI and SmaI, respectively. The fragment was integrated into the EcoRI and SmaI sites of the pCI-neo vector. HEK293T cells were purchased from the Institute of Biotechnology and Cell Biology, Chinese Academy of Science. The procedures for HEK293T cell culture, transfection, and patch-clamp whole-cell recording were as described previously (Vig et al., 2006; Gao et al., 2014, 2016). For cAMP treatment, 200 μm 8-bromoadenosine cAMP (Sigma-Aldrich) was used. The theoretical equilibrium potentials were calculated with an online Nernst Potential Calculator at http://www.physiologyweb.com/calculators/nernst_potential_calculator.html.

Observation of Extracellular Ca²⁺ Distribution with Fluorescence and Confocal Microscopy

The low-affinity cell-impermeable Ca²⁺ fluorescent dye OGB-5N with excitation and emission at 494 and 521 nm, respectively, was purchased from Thermo Fisher Scientific (O-6812) and dissolved in DMSO to make a 5 mm stock stored at −20°C without exposure to light. For visualizing extracellular Ca²⁺ distribution, a fully expanded mature leaf was detached from a hydroponically grown Arabidopsis plant, and its petiole was put into a PCR tube containing 100 μL of growth solution including 20 μm OGB-5N and either 0.1 or 10 mm CaCl₂ for the indicated periods of time. Afterward, the leaf was stained with 20 μm propidium iodide for 10 min. Images presented in Figure 5 were acquired with a fluorescence microscope (Nikon Eclipse Ti). Images presented in Figure 6 were acquired with a confocal laser scanning microscope (Zeiss 710).

Accession Numbers

Sequence data for this article can be found in The Arabidopsis Information Resource at www.arabidopsis.org under the following accession numbers: AtCNGC2 (At5g15410), AtCNGC4 (At5g54250), AtCAX1 (At2g38170), and AtCAX3 (At3g51860).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Root responses of cngc2-1 to different Ca²⁺ concentrations in a hydroponic system.

• Supplemental Figure S2. Responses of hydroponically grown Arabidopsis cngc mutants to Ca²⁺.

• Supplemental Figure S3. Ca-dependent growth of cngc2-1 and cax1cax3 seedlings.

• Supplemental Figure S4. CNGC2 promoter-driven GUS reporter gene expression in Arabidopsis young seedlings and mature plants.

• Supplemental Figure S5. CNGC2 shows inward Ca²⁺ channel activity in HEK293T cells.

• Supplemental Figure S6. Inward K⁺ channel activity of CNGC2 and KAT1 in HEK293T cells.

• Supplemental Table S1. Primer list for the identification of Arabidopsis cngc homozygous mutants.

Glossary

Col-0

Columbia-0

HR

hypersensitive response

SA

salicylic acid

cfu

colony-forming units

OGB-5N

Oregon Green BAPTA 488 5N