Endoplasmic reticulum-localized CCX2 is required for osmotolerance by regulating ER and cytosolic Ca²⁺ dynamics in Arabidopsis

Significance

Given the lack of a nervous system in plants, signaling molecules, such as Ca²⁺, are of vital importance for perceiving environmental stimuli. However, little information on the genes that control this network is available. Our demonstration that Arabidopsis CCX2, localized to the endoplasmic reticulum, modulates osmotic stress responses through the regulation of cytosolic and endoplasmic reticulum Ca²⁺ concentrations is one more step toward identifying all the players involved in plant stress responses. The elucidation of these pathways will help to meet the necessary requirements to face global challenges, such as high salinity on the world’s surface, including arable land. This work sheds light on an as yet uncharacterized and key player that tunes intracellular Ca²⁺ homeostasis, with particular significance for plants under salt stress.

Abstract

Ca²⁺ signals in plant cells are important for adaptive responses to environmental stresses. Here, we report that the Arabidopsis CATION/Ca²⁺ EXCHANGER2 (CCX2), encoding a putative cation/Ca²⁺ exchanger that localizes to the endoplasmic reticulum (ER), is strongly induced by salt and osmotic stresses. Compared with the WT, AtCCX2 loss-of-function mutant was less tolerant to osmotic stress and displayed the most noteworthy phenotypes (less root/shoot growth) during salt stress. Conversely, AtCCX2 gain-of-function mutants were more tolerant to osmotic stress. In addition, AtCCX2 partially suppresses the Ca²⁺ sensitivity of K667 yeast triple mutant, characterized by Ca²⁺ uptake deficiency. Remarkably, Cameleon Ca²⁺ sensors revealed that the absence of AtCCX2 activity results in decreased cytosolic and increased ER Ca²⁺ concentrations in comparison with both WT and the gain-of-function mutants. This was observed in both salt and nonsalt osmotic stress conditions. It appears that AtCCX2 is directly involved in the control of Ca²⁺ fluxes between the ER and the cytosol, which plays a key role in the ability of plants to cope with osmotic stresses. To our knowledge, Atccx2 is unique as a plant mutant to show a measured alteration in ER Ca²⁺ concentrations. In this study, we identified the ER-localized AtCCX2 as a pivotal player in the regulation of ER Ca²⁺ dynamics that heavily influence plant growth upon salt and osmotic stress.

Plant tolerance to abiotic stress relies on a wide range of physiological and molecular mechanisms involving a fine regulation of cytosolic ion homeostasis, including calcium (Ca²⁺) (1). In eukaryotes Ca²⁺ can act as a ubiquitous second messenger in triggered signal transduction cascades (2). In plants, several stresses trigger Ca²⁺ influx in the cytosol that can involve the activation of membrane-localized Ca²⁺-permeable channels (3). The homeostasis of [Ca²⁺]_cyt depends on a fine regulation of Ca²⁺ influxes and effluxes that occur at the plasma membrane (PM) and membranes of various subcellular compartments (4–7). Downstream of [Ca²⁺]_cyt elevation is the activity of various Ca²⁺ sensors (2) and reactive oxygen species (ROS)-producing enzymes that are synergistically activated by Ca²⁺ binding and phosphorylation by Ca²⁺-dependent protein kinases (3, 8, 9).

Cameleon Ca²⁺ sensors have recently been developed to study Ca²⁺ fluxes and dynamics in various organelles (4, 6, 10–13). In plants, despite the development of this technology, there is still a lack of molecular information regarding the regulation of [Ca²⁺]_cyt and the interplay between cytosol and cellular compartments, particularly the endoplasmic reticulum (ER), in the regulation of Ca²⁺ fluxes. Indeed, despite the important role of the ER in the regulation of Ca²⁺ dynamics in mammals (14) and in stress adaptation (15), little is known about its function in the control of Ca²⁺ fluxes in plants (4, 7, 16). Only two ER-localized Ca²⁺ pumps have been partially characterized, namely ECA1 and ACA2. Although their predicted function is to load cytosolic Ca²⁺ into the ER lumen, their role in establishing ER-cytosol Ca²⁺ dynamics is so far not clear (17, 18).

In plants, three classes of membrane transporters are predicted to mediate Ca²⁺ fluxes: Ca²⁺-permeable channels, Ca²⁺-ATPases, and Ca²⁺/cation antiporters (CaCAs). The CaCA superfamily comprises five families, including cation/Ca²⁺ exchangers (CCX) and Ca²⁺/proton exchangers (CAX) (5, 18, 19). Although Arabidopsis CAX members are well characterized (8, 20–23), CCX’s function in plants remains largely unknown. A recent characterization of Arabidopsis thaliana CCX1 supports its role in leaf senescence, as well as a higher sensitivity of the ccx1-1 mutant to Ca²⁺ deprivation (24). Functional characterization based on heterologous expression in yeast suggests that AtCCX3 and AtCCX4 have affinity for Mn²⁺, K⁺, and Na⁺ (25), while AtCCX5 appears to mediate K⁺ uptake and Na⁺ transport (26). Phylogenetic analysis showed that CCX5 and, to a lesser extent CCX3 and CCX4, genes are limitedly close to a mammalian K⁺-dependent Na⁺/Ca²⁺ exchanger (NCKX6) (27, 28). CCXs are indeed annotated as cation/Ca²⁺ exchangers, even if CCX3/4/5 are characterized as H⁺/K⁺ antiporters and do not exhibit affinity for Ca²⁺ (25, 26). The rice vacuolar cation/Ca²⁺ transporter OsCCX2 was recently characterized. This gene enhanced tolerance of Saccharomyces cerevisiae cells in the presence of excess Na⁺, Li⁺, Fe²⁺, Zn²⁺, and Co²⁺ (29).

In this study, we describe the role of AtCCX2 (At5g17850), which encodes a putative cation/Ca²⁺ exchanger. We show that AtCCX2 is an ER-localized protein whose activity impacts both cytosolic and ER [Ca²⁺] regulation under both standard and NaCl stress conditions and plant growth during salt and osmotic stresses.

Results

AtCCX2 Is Induced by a Wide Range of Osmotic Stresses in A. thaliana.

The expression of the five CCXs genes was retrieved from the electronic Fuorescent Pictograph (eFP) (30) database (Fig. S1 A and B). These data indicate that salt and mannitol treatments caused the highest induction of CCX1, -2, and -3 gene expression in roots and shoots. Interestingly, CCX2 showed the earliest changes (after 30 min of treatment) and the strongest induction upon salt and mannitol treatments in roots and shoots. These data suggest the possible involvement of CCX2 in salt and osmotic-stress responses, which were further investigated.

AtCCX2 Mutants Showed Contrasting Responses upon Salt and Osmotic Stress.

To test whether CCX2 is involved in stress tolerance, both loss- and gain-of-function approaches were followed. CCX2 knockout (ccx2), two ccx2 + p35S::AtCCX2-GFP complemented lines (ccx2-C), and four independent lines of p35S::AtCCX2 [CCX2-overexpressed (OE)] Col-0 Arabidopsis plants (Fig. S1 C–G) were exposed to different stress conditions. While no significant differences in growth parameters were noticed in the control condition compared with the WT (Fig. S1 E–G), clearly sensitive and resistant phenotypes were observed for ccx2 and CCX2-OE plants, respectively, when exposed to both salt and osmotic stresses (Fig. 1 and Fig. S2). ccx2 root growth was strongly impaired at all tested NaCl concentrations, showing a reduction in the length of lateral roots compared with WT, while ccx2-C lines showed root growth similar to the WT upon NaCl stress (Fig. 1 A and B). Conversely, the root length of the CCX2-OE lines was longer upon NaCl treatment, compared with the WT. In particular, lateral roots were on average two- to threefold longer in CCX2-OE lines with respect to the WT, depending on NaCl concentration (Fig. 1 B and D). The same trend was observed for shoot area in the ccx2, ccx2-C, and CCX2-OE plants (Fig. 1 C and D). ccx2 and CCX2-OE plants were also tested for their ability to grow in the presence of KCl, mannitol (Fig. S2 A–D, G, and H), and of EGTA Ca²⁺ chelator (Fig. S2 E, F, and I). The results pinpointed a lateral root growth reduction in ccx2 with respect to WT, while CCX2-OE showed longer roots and a larger shoot area compared with the WT. Together, these results support a role for CCX2 in salt and osmotic stress responses, with a possible function in regulating Ca²⁺ homeostasis.

Fig. 1.

In vitro growth test conducted on Col-0 WT, ccx2, ccx2-Complemented (ccx2-C) and CCX2-OE (OE) lines subjected to NaCl treatment. Relative primary (PR, A) and lateral (LR, B) root length (L) and shoot area (SA, C) were recorded after 7 d of treatment. Histograms represent relative growth values with respect to WT. Error bars represent the SD (n = 25–45 for root; n = 15–35 for shoot). Asterisks indicate statistically significant differences (P = 0.05) between WT and mutants by ANOVA Duncan’s new multiple-range test. (D) WT, ccx2, ccx2-C and CCX2-OE plants 7 d after 100 mM NaCl treatment.

The most noteworthy phenotype for CCX2 mutants was observed in response to NaCl. We focused on this stress and monitored the expression of CCX1, -2, and -3 genes (Fig. S2J), in addition to others reported to play a role in salt stress and Ca²⁺ homeostasis (Fig. S2K). WT, ccx2, and CCX2-OE plants were transferred to the control and 300 mM NaCl solutions and sampled at 2 and 6 h after stress initiation (HSI). CCX2 showed a 10-fold increase in expression in NaCl-treated WT plants at 6 HSI compared with control plants. Interestingly, CCX2 transcript level was still increased, even in NaCl-treated CCX2-OE seedlings, with respect to the control condition as early as 2 HSI. CCX1 and CCX3 genes were also induced upon NaCl stress in WT plants, respectively, at 2 and 6 HSI, but to a lesser extent compared with CCX2 (Fig. S2J). RD29B.1 and SOS1 showed no significant differences in expression between WT and the CCX2 mutants (Fig. S2K). Conversely, at 6 HSI, the SOS3 calcium sensor (31) gene was induced by NaCl treatment in CCX2-OE plants but not ccx2 plants. Finally, ACA2, which encodes an ER-localized Ca²⁺ pump, was more highly expressed at 6 HSI in CCX2-OE mutants, with respect to WT and ccx2.

Subcellular Localization of AtCCX2.

To determine the subcellular localization of CCX2, we analyzed the localization of the C-terminal GFP-tagged CCX2 fusion protein in the ccx2 complemented line (Fig. 1). Hence, Fig. 2 A–F provides representative confocal microscope images of ccx2-C Arabidopsis seedlings, demonstrating that the chimeric protein was expressed in leaf (Fig. 2A and Movie S1), hypocotyl (Fig. 2B), and root cells (Fig. 2C). The GFP signal decorated structures resembling the typical ER morphology and the associated nuclear envelope (32). To confirm CCX2 localization, the same construct was transiently expressed in agro-infiltrated tobacco leaf (Fig. 2 G–J). In this case, the AtCCX2-GFP signal largely colocalized with the mCherry-HDEL ER-marker (Fig. 2 I and J and Movie S2). Moreover, coexpression of AtCCX2-GFP with untagged RFP (cytosolic and nuclear localization) demonstrated that the GFP signal only surrounded the nucleus and was absent from nucleoplasm (Fig. 2 K–M), further confirming its membrane association. However, in this latter case the pixel scatter-plot analysis still revealed a partial colocalization of GFP and free RFP, but this was assumed to be the presence of the large vacuole that pushes both the ER and cytoplasm against the PM (Fig. 2N). Overall, our results clearly indicate that CCX2 is an ER-localized protein.

Fig. 2.

AtCCX2-GFP subcellular localization in the ccx2 Arabidopsis knockout background and in agro-infiltrated leaves of Nicotiana benthamiana. (A–C) Representative fluorescence images of AtCCX2-GFP in a selected focal plane of a ccx2/AtCCX2-GFP Arabidopsis plantlet in different tissues/cell types: cotyledon epidermal cells (A); root mature zone cells (B); and root tip cells (C). (D–F) Bright-field images corresponding to the fluorescence images shown in A–C. (G) Fluorescence of AtCCX2-GFP in a representative tobacco epidermal cell coexpressed with mCherry-HDEL used as an ER⁺ marker. (H) Fluorescence of mCherry-HDEL of the cell shown in G. (I) Merged fluorescent signals of GFP and mCherry. (J) Two-dimensional scatter plot visualizing the correlation of the GFP and mCherry pixel intensities over all pixels in the images in G and H. (K) Fluorescence of AtCCX2-GFP in a representative tobacco epidermal cell coexpressed with free RFP used as cytosolic and nuclear-positive marker. (L) Fluorescence of RFP of the cell shown in K. (M) Merged fluorescent signals of GFP and RFP. (N) Two-dimensional scatter plot visualizing the correlation of the GFP and RFP pixel intensities over all pixels in the images in K and L. NE, nuclear envelope. (Scale bars, 25 µm.)

AtCCX2 Shows Ca²⁺ Transport Properties.

Despite AtCCX2 being predicted to be a cation/Ca²⁺ exchanger, no data have been provided about its transport properties. Hence, to assess CCX2 Ca²⁺ transport capability in vivo, we expressed AtCCX2 in the low-affinity Ca²⁺ uptake-deficient triple S. cerevisiae yeast mutant K667, which lacks the vacuolar ATPase (PMC1), the vacuolar exchanger (VCX1), and the cytosolic regulatory subunit (CNB1). The K667 triple mutant has been reported to show a Ca²⁺ uptake deficiency and reduced growth at high external [Ca²⁺] (29). This evidence was indeed confirmed by us and, importantly, the K667 Ca²⁺-sensitivity was partially suppressed by AtCCX2 expression (Fig. S3), hence supporting its role in Ca²⁺ transport.

AtCCX2 Regulates Cytosolic and ER Ca²⁺ Fluxes.

The overall results suggest a role for CCX2 in Ca²⁺ transport and convinced us to analyze in vivo Ca²⁺ dynamics in the cytosol of WT, ccx2, and CCX2-OE plants transformed with the NES-YC3.6 Cameleon sensor (12, 33) subjected to NaCl (Fig. S4 A–L and V). Cells from the root maturation zone were considered for the analysis (Movies S3–S5). Interestingly, the Cameleon ratios, expressed as cpVenus/CFP values that report Ca²⁺ level (higher ratio = higher Ca²⁺ concentration), showed significant differences in the cells of the region of interest (ROI) among the three genotypes already under resting conditions (Fig. S4W). Particularly, the cytosolic resting Ca²⁺ level was significantly lower and higher in ccx2 and CCX2-OE roots, respectively, in comparison with WT. Moreover, to study NaCl-induced Ca²⁺ increase in the three genotypes, a custom perfusion set-up for in vivo fluorescence microscopy of Arabidopsis roots was employed (as in refs. 4 and 34). Seedlings were challenged with 100 and 200 mM NaCl stress (8 min of treatment) and a first cytosolic [Ca²⁺] peak was observed (only at 200 mM NaCl), followed by a second and higher peak when the NaCl solution was replaced with a control solution (Fig. S4V). Because the resting Ca²⁺ concentrations were different among the genotypes (Fig. 3 B, E, H, and K), the normalized ratio changes were compared (Fig. 3 C and I and Fig. S5 A–C), and this revealed that the first peak (NaCl administration) showed a milder slope in the FRET curve and a lower maximum normalized peak in ccx2 at 200 mM (Fig. 3 D–F). Despite CCX2-OE plants showing a higher basal Ca²⁺ level compared with the WT (Fig. 3B and Fig. S4W), the CCX2-OE normalized FRET ratios were similar to the WT in response to NaCl (Fig. 3C). NaCl removal caused lower and higher maximal free [Ca²⁺]_cyt (Fig. 3G) and maximum normalized peaks (Fig. 3 I and L) in ccx2 and CCX2-OE mutants, respectively, compared with the WT. The slope of normalized FRET curves suggests that Ca²⁺ accumulation kinetics were slower in ccx2 and faster in CCX2-OE, with respect to WT (Fig. 3J).

Fig. 3.

[Ca²⁺]_cyt measured in root elongation zone in Col-0 WT, ccx2, CCX2-OE2 and CCX2-OE3 plants subjected to NaCl treatment. (A and G) Representative trace of cytosolic Ca²⁺ kinetics recorded in plants subjected to NaCl administration (first peak, A–F) and removal (second peak, G–L). Small black arrow indicates analyzed peak. (B and H) Cytosolic Ca²⁺ FRET raw ratios. (C and I) Normalized cytosolic Ca²⁺ FRET ratios. (D and J) Slope of Ca²⁺ accumulation curves. (E and K) Representative root-ratio images under perfusion before (140 and 630 s) and after (220 and 710 s) NaCl administration (E) and removal (K). (F and L) Maximum normalized Ca²⁺ peak. A t test was conducted to assess differences between WT and mutants (*P < 0.05, **P < 0.01). Error bars represent the SD (n = 10–12).

We then tested a milder stress by treating the seedlings with 100 mM NaCl. In this latter case we observed smaller differences among the genotypes, which were not statistically different. However, CCX2-OE plants still showed the tendency of a higher maximum Ca²⁺ peak compared with the WT, hence essentially confirming the behavior observed at 200 mM NaCl. Therefore, this phenotype was dependent upon the magnitude of the stimulus (Fig. S5 A–C). We argue that the absence of CCX2 or its overexpression leads, respectively, to loss- and gain-of-function phenotypes with respect to the maximum [Ca²⁺]_cyt accumulation in response to both NaCl administration and removal. Given the above data on [Ca²⁺]_cyt homeostasis (Fig. 3) and the fact that CCX2 is localized to the ER (Fig. 2), an additional step was to assess ER luminal [Ca²⁺] ([Ca²⁺]_ER) response to NaCl stress using the CRT-D4-ER Cameleon sensor (4) (Fig. S4 M–U and X). The CRT-D4-ER expression was lower compared with NES-YC3.6, which forced us to use a higher-magnification objective (20× instead of 4×) for ER luminal Ca²⁺ analyses. However, the same ROI was considered for ER and cytosol Ca²⁺ measurements (Fig. S4 V and X). Under resting conditions, WT and ccx2 genotypes did not show significant differences in the [Ca²⁺]_ER (Fig. S4Y) and NaCl addition did not induce any clear change in [Ca²⁺]_ER in WT or ccx2 plants (Fig. 4A). This result may depend on the sensitivity of the sensor used or the Ca²⁺ threshold required for its accumulation in the ER, similarly to that reported for mitochondria. Nevertheless, noteworthy results were observed for FRET ratios in CRT-D4-ER ccx2 mutants when NaCl was removed (Fig. 4 A and B). NaCl removal caused a gradual rise in [Ca²⁺]_ER (Movies S6 and S7), and notably, the FRET ratios significantly differed between WT and ccx2 seedlings, showing a higher maximal free [Ca²⁺]_ER in ccx2 (P = 0.009). Differences between WT and ccx2 in ER [Ca²⁺] were also in this case more pronounced at 200 mM than at 100 mM NaCl treatment (Fig. 4 B–E and Fig. S5 D–F).

Fig. 4.

[Ca²⁺]_ER measured in root elongation zone in Col-0 WT and ccx2 plants subjected to NaCl treatment. (A) Representative trace of cytosolic Ca²⁺ kinetics recorded in plants subjected to NaCl administration (first peak, not considered) and removal (second peak, black arrow). (B) Representative root-ratio images under perfusion before (680 s) and after (730 s) NaCl removal. (C) ER Ca²⁺ FRET raw ratios. (D) Normalized ER Ca²⁺ FRET ratios. (E) Maximum normalized Ca²⁺ peak. A t test was conducted to assess differences between WT and mutant (n.s. = not significant; P = P value). Error bars represent the SD (n = 12–13).

Finally, we tested cytosolic and ER Ca²⁺ dynamics in response to external ATP (eATP), that has no direct effect on the solution osmotic potential and induces a strong Ca²⁺ transient in the cytosol of root tip cells. No differences were observed for [Ca²⁺]_cyt between the WT and CCX2 loss- and gain-of-function mutants (Fig. S5 G–J and Movies S8–S10). Conversely, striking differences in the concentration, maximum peak, and accumulation kinetics of [Ca²⁺]_ER were observed between WT and ccx2, (Fig. S5 K–P and Movies S11 and S12), as already observed for NaCl treatment.

In conclusion, the ccx2 mutant showed lower cytosolic Ca²⁺ increase and higher ER Ca²⁺ accumulation, a phenotype that was particularly remarkable upon NaCl removal.

Growth and Ca²⁺ Accumulation Dynamics in ccx1 and ccx3 Mutants Give Insights into the Role of the AtCCX Family in Regulating NaCl Responses.

A step forward in the analysis was to shed light on the role of CCX1 and CCX3 genes, which are other members of the CCX family induced by osmotic and salt stresses (Fig. S1A). The ccx1 mutant displayed a reduced lateral root length compared with WT during 100 mM NaCl stress, while ccx3 showed no significant differences (Fig. S6 A and C). Similar results were observed after exposure to KCl and mannitol treatments (Fig. S6 D, F, G, and I).

Analysis of Ca²⁺ dynamics highlighted a significantly lower and higher steady state [Ca²⁺]_cyt in ccx1 and ccx3 mutants, respectively, with respect to the WT (Fig. S7B). Notably, ccx3 and ccx1 showed contrasting responses in the regulation of Ca²⁺ homeostasis upon NaCl, with a maximum peak after NaCl stress in the former and considerably lower levels in the latter (Fig. S7 C–M). Finally, ER Ca²⁺ measurements conducted on ccx1 plants (but not ccx3 because of silencing-related effects) showed similar responses to those observed in ccx2 (Fig. S7 N–S).

Discussion

Our results shed light on the role of AtCCX2, a putative cation/Ca²⁺ exchanger that is localized to the ER (Fig. 2) and give insights into the role of the AtCCX family in the regulation of [Ca²⁺]_cyt and [Ca²⁺]_ER dynamics upon salt stress. Salt stress is usually associated with a rise of [Ca²⁺] in the cytosol and the subcellular compartments of plant cells (35–37) and other responses involving PM depolarization and ROS homeostasis (38). In plants, the impact of ER on the regulation of Ca²⁺ fluxes and related responses have remained largely undecoded (7), and there is little evidence for a role of the ER as a regulator of [Ca²⁺]_cyt homeostasis upon stress in vivo (4). CCX2 characterization adds another piece to this puzzle and highlights the presence of a gene that impinges on the regulation of ER-cytosol intracellular Ca²⁺ dynamics upon abiotic stress.

AtCCX2 Loss-of-Function and Gain-of-Function Mutants Display Contrasting Responses to Salt Stress.

CCX2 loss- and gain-of-function mutants displayed impaired and enhanced growth compared with the WT upon NaCl stress and also upon osmotic stress treatment (Fig. 1 and Fig. S2). As already observed for other Arabidopsis mutants (39), the impaired root growth upon EGTA treatment may suggest that CCX2 is involved in NaCl-stress responses in a Ca²⁺-dependent way. However, the EGTA effect on intracellular Ca²⁺ is not fully understood, and long treatments can be misleading.

The link between CCX2 and salt stress was further strengthened by results obtained from CCX gene-expression analysis on NaCl-stressed seedlings (Fig. S2J). As already reported for other gain-of-function mutants (40), the CCX2 transcript steady-state level was significantly induced upon NaCl in CCX2-OE seedlings at 2 HSI, most likely indicating possible stress-dependent posttranscriptional regulation of expression. The unchanged expression of CCX1 and the slight induction of the CCX3 gene in the ccx2 background upon NaCl treatment suggest a lack of compensation for CCX2 expression during salt stress. These data support nonoverlapping functions for the CCX transporters. Expression analysis did not suggest a role for CCX2 in the regulation of RD29B.1 and SOS1 genes (Fig. S2K). As in our experiment, SOS1 was not induced, contrary to RD29B that is a well-known marker for osmotic stress, and the potential role of CCX2 on SOS1 regulation is difficult to deduce. In contrast, SOS3 (31) was less expressed in ccx2 plants, and ACA2 (17) was more expressed in CCX2-OE under NaCl treatment with respect to WT. Given that SOS3 and ACA2 are well-known players in Ca²⁺ regulation under NaCl stress, these results further suggest an important role for CCX2 in Ca²⁺ homeostasis. Interestingly, because ACA2-predicted function is to transport Ca²⁺ into the ER (17), its higher expression in CCX2-OE mutants, upon NaCl stress, might be related to the higher [Ca²⁺]_cyt response observed in these plants.

AtCCX2 Regulates Cytosolic and ER Ca²⁺ Dynamics upon NaCl.

Because CCX2 showed its involvement in Ca²⁺ transport in yeast (Fig. S3) and is localized to the ER (Fig. 2), in vivo Ca²⁺ analysis was carried out on WT and CCX2 mutant roots expressing cytosolic and ER Cameleon sensors (4, 12). Differences between WT and CCX2 mutants were already recorded for cytosolic resting prestimulus Ca²⁺ levels, which were lower and higher in ccx2 and CCX2-OE, respectively, compared with the WT (Fig. S4W). Furthermore, the lower cytosolic Ca²⁺ peak (Fig. 3 F and L) and slower Ca²⁺ accumulation (Fig. 3 D and J) recorded upon 200 mM NaCl treatment in the ccx2 mutant compared with WT correlated well with the observed plant growth (Fig. 1). Stimulation of cytosolic and ER Ca²⁺ accumulation by NaCl were previously reported (4, 36), and a positive role for Ca²⁺ in the regulation of salt stress-related gene expression has already been established (38). Thus, differences in growth observed in NaCl-treated WT, ccx2, and CCX2-OE plants were most likely due to differential regulation of Ca²⁺ homeostasis that led to lower and higher [Ca²⁺]_cyt in the CCX2 loss- and gain-of-function mutants, respectively (Fig. 3). The significantly higher steady state of [Ca²⁺]_ER compared with [Ca²⁺]_cyt (considering different in vitro Ca²⁺ affinity for Ca²⁺ of the YC3.6 K_d = 250 nM and the D4 K_d = 195 µM), and higher Ca²⁺ peak observed in NaCl-treated ccx2 roots with respect to the WT (Fig. 4 C–E), allowed the identification of CCX2 as a pivotal player in the regulation of [Ca²⁺]_ER and [Ca²⁺]_cyt dynamics. Furthermore, ccx2 seedlings treated with eATP were characterized by a higher [Ca²⁺]_ER peak with respect to WT, suggesting an alteration of ER Ca²⁺ homeostasis in the mutant even though no differences were observed in the cytosol (Fig. S5). Although other Arabidopsis mutants showed different Ca²⁺ accumulation in subcellular compartments (e.g., MICU knockout mutants showed altered mitochondria Ca²⁺ levels compared with the WT), this did not lead to different cytosol Ca²⁺ dynamics (34). Indeed, in the case of CCX2, coupled cytosolic and ER Ca²⁺ changes upon NaCl removal were measured. We believe that CCX2 is involved in salt stress tolerance through its role in Ca²⁺ transport regulation across the ER and cytosol. The ER does not seem to be a main contributor for stimulus-induced increases in cytosolic Ca²⁺ concentration (4). However, luminal ER Ca²⁺ elevations typically follow cytosolic ones and when CCX2 is missing, this ER/Ca²⁺ dynamic is altered. The absence of AtCCX2 activity results in decreased cytosolic and increased ER Ca²⁺ concentrations in comparison with WT and the gain-of-function mutants in both salt and nonsalt osmotic stress conditions (Figs. 3 and 4 and Fig. S8), resulting in higher plant stress sensitivity. Finally, as observed for the rice vacuolar transporter OsCCX2 (29), AtCCX2 partially suppresses the Ca²⁺ sensitivity of a yeast mutant deficient in low-affinity Ca²⁺ uptake (Fig. S3), supporting its involvement in Ca²⁺ transport. The overall data suggest a direct involvement of CCX2 in the control of Ca²⁺ fluxes between the ER and the cytosol, which plays a key role in the ability of the plant to cope with osmotic stresses.

Toward Defining a Role of AtCCX Family.

In addition to the extended study of CCX2, CCX1 and CCX3 knockout mutants were analyzed. While CCX1 was proposed to be involved in modulating Ca²⁺ signaling (23), CCX3 was not previously associated with Ca²⁺ (24).

Growth tests (Fig. 1 and Fig. S6) and in vivo Ca²⁺ measurements (Fig. 3 and Fig. S7) of knockout mutants suggested that CCX2 and CCX1 genes are likely to play similar roles in plant stress tolerance, while CCX3 seems to play a different role from CCX1 and CCX2. In support of those results, a phylogenetic analysis of the CCX family (20, 28) placed Arabidopsis CCX1 and CCX2 closer in the tree than CCX3 and CCX4.

As already hypothesized for Arabidopsis, regulation of Ca²⁺ homeostasis is associated with different responses during salt stress (4, 36, 38), in which different members of the CCX family seem to play a role. Taken together, our results give insights into the unexplored role of the CCX gene family in planta and highlight a major role for CCX2 in the regulation of [Ca²⁺]_cyt and [Ca²⁺]_ER, with downstream effects on salt stress tolerance.

Methods

WT, ccx1, ccx2, ccx3, and CCX2-OE plants were in the Col-0 background. Growth conditions and tolerance tests, generation of transgenic lines, yeast complementation test, transcriptomic experiment, measurement of Ca²⁺ dynamics, statistical methods, protocols used for subcellular localization, and other imaging measurements are reported in SI Methods.