MIZ1 regulates ECA1 to generate a slow, long-distance phloem-transmitted Ca2+ signal essential for root water tracking in Arabidopsis.

Ever since Darwin postulated that the tip of the root is sensitive to moisture differences and that it "transmits an influence to the upper adjoining part, which bends towards the source of moisture" [Darwin C, Darwin F (1880) The Power of Movement in Plants, pp 572-574], the signal underlying this tropic response has remained elusive. Using the FRET-based Cameleon Ca2+ sensor in planta, we show that a water potential gradient applied across the root tip generates a slow, long-distance asymmetric cytosolic Ca2+ signal in the phloem, which peaks at the elongation zone, where it is dispersed laterally and asymmetrically to peripheral cells, where cell elongation occurs. In addition, the MIZ1 protein, whose biochemical function is unknown but is required for root curvature toward water, is indispensable for generating the slow, long-distance Ca2+ signal. Furthermore, biochemical and genetic manipulations that elevate cytosolic Ca2+ levels, including mutants of the endoplasmic reticulum (ER) Ca2+-ATPase isoform ECA1, enhance root curvature toward water. Finally, coimmunoprecipitation of plant proteins and functional complementation assays in yeast cells revealed that MIZ1 directly binds to ECA1 and inhibits its activity. We suggest that the inhibition of ECA1 by MIZ1 changes the balance between cytosolic Ca2+ influx and efflux and generates the cytosolic Ca2+ signal required for water tracking.

Plant adaptation to environmental changes requires continuous foraging for water to survive. Roots have evolved a yet-unexplained mechanism that directs their growth toward high water potential, a task that requires overcoming their default growth pattern along the gravity vector (gravitropism) (1–3). Landmark experiments demonstrated the importance of the root cap in sensing moisture and directing growth toward the water source (1, 4). If this is indeed the case, then numerous questions need to be addressed to elucidate the mechanism underlying hydrotropism: First, how is the water gradient detected; second, following the sensing, how does the detector transduce the sense to a signal, which is transmitted from the root cap to the elongation zone (EZ); third, which asymmetric cross-root signal underlies differential growth across the root, resulting in root bending toward the water source? Although it has been shown that the osmotic stress hormone abscisic acid (ABA) is required for hydrotropism (5, 6), there is no evidence for ABA signaling from the root cap to the EZ, nor is there evidence for the asymmetric distribution of ABA across the root in response to moisture gradients (6). The requirement of the hormone auxin, a regulator of some tropic responses (7–9), was revoked in relation to hydrotropism because it is not asymmetrically distributed following hydrostimulation (10, 11). Moreover, blocking of auxin polar transport, or TIR-dependent auxin signaling, enhanced hydrotropism (10, 11). Interestingly, reactive oxygen species (ROS) play an important role in tuning root tropic responses by acting positively in gravitropism and negatively in hydrostimulation (12, 13). In recent years, accumulating evidence has suggested that Ca2+ plays a key role in long-distance, systemic signaling in response to various stress stimuli (14, 15), for example, mediating the occlusion of phloem sieve tube elements in response to wounding (16), evoking electric signaling (15), and mediating rapid Ca2+ waves in roots responding to salt stress (14). Moreover, since both ABA and ROS signaling interact with Ca2+ signaling in plants, and since Ca2+ was suggested to be involved both in hydrotropic and gravitropic responses (17–20) and in cell elongation in different plant tissues (21, 22), we sought to assess the possible role of Ca2+ as a signal from the root cap to the EZ for root bending upon hydrostimulation.

Results

An MIZ1-Dependent Slow Shootward Cytosolic Ca2+ Signal Is Required for Root Hydrotropism.

To analyze cytosolic Ca2+ ([Ca2+]cyt) levels in the roots of wild-type (WT) Arabidopsis (Col-0), transgenic plants that express the cytosol-targeted, FRET-based Ca2+ sensor Cameleon (NES-YC3.6) (23) were studied by confocal microscopy. Confocal visualization of the NES-YC3.6 ratio intensity in Col-0 roots under control conditions revealed high levels of [Ca2+]cyt at the columella, meristematic zone, lateral root cap, and the EZ vasculature (Fig. 1). Strikingly, following 1 h of hydrostimulation in a split-agar/sorbitol system (), [Ca2+]cyt levels were elevated at the root tip (Fig. 1 and ) and at the vasculature of the meristem and elongation zones, with an apparent asymmetric distribution at the EZ, where higher [Ca2+]cyt levels were observed at the side that becomes convex upon bending (Fig. 1 and Movies S1 and S2). This result was also reproduced using a split-agar/mannitol system (), indicating that the [Ca2+]cyt elevation is not a specific response to sorbitol. To further assess the relationship between the long-distance Ca2+ signal and the cellular pathways mediating root curvature to moisture, we visualized [Ca2+]cyt in mutants of the Mizu-Kussey 1 (MIZ1) gene, which encodes an ER membrane-associated protein (24, 25), whose biochemical or cellular functions are unknown but is indispensable for root curvature in response to moisture gradients (24). Analysis of the [Ca2+]cyt distribution along the root tip of control and hydrostimulated miz1 mutants harboring the Cameleon Ca2+ sensor revealed lower basal [Ca2+]cyt levels under control conditions, which were not elevated nor asymmetrically distributed at the EZ following hydrostimulation (Fig. 1, , and Movie S3), suggesting that a functional MIZ1 is required for generating the long-distance Ca2+ signal in response to hydrostimulation. Indeed, the expression of an active MIZ1 (under the transcriptional regulation of the native MIZ1 promoter) in the miz1/NES-YC3.6 plants fully restored the [Ca2+]cyt signal and root bending (Fig. 1 ). These results identified MIZ1 as a cellular component in mediating Ca2+ signaling in response to hydrostimulation. Since MIZ1 is associated with the ER membrane (25), we also examined whether Ca2+ levels in the ER ([Ca2+]ER) might be altered in hydrostimulated root tips by analyzing Arabidopsis plants expressing the ER localized CRT-D4ER Cameleon sensor (26). Interestingly, [Ca2+]ER significantly decreases in hydrostimulated root tips in parallel with the elevation in [Ca2+]Cyt (). These data strongly suggest that, in root tip cells, the ER serves as a Ca2+ reservoir that functions in generating the hydrotropic [Ca2+]cyt signal.

Fig. 1.

Asymmetric MIZ1-mediated [Ca2+]cyt signal is required for root tip response to moisture gradient. (A) NES-YC3.6 confocal microscopy visualization in roots of Col-0, miz1, miz1/empty pPZP RCS2 BAR (vector), and miz1/MIZ1pro:MIZ1 seedlings. FRET/CFP-based images were pseudocolored, where red indicates higher [Ca2+]cyt levels. EZ, elongation zone; MZ, meristematic zone; RC, root cap; V, vasculature. g represents gravity vector, and Ψ represents water potential gradient. (Scale bar, 50 µm.) (B) Measurements of root curvature following control or 1 h of hydrostimulation. (C) Quantification of FRET/CFP intensity ratio of two longitudinal halves of 50-µm root segments, 50–500 µm above the apex of NES-YC3.6–expressing Col-0 seedlings. (D) Quantification of FRET/CFP intensity ratio (lines) of the two longitudinal halves of the EZ (250–350 µm above apex) of control and 1-h–hydrostimulated NES-YC3.6–expressing Col-0 seedlings. Curvature (columns) was measured at each time point. In C and D, C, control; H, hydrostimulated. Error bars represent mean ± SD (three biological independent experiments; 10 seedlings each); *P < 0.01, **P < 0.001, Student’s t test versus FRET/CFP value measured in the concave side.

Cameleon ratiometric analysis from the meristem to the EZ (50–500 µm above apex) in two sides of control and hydrostimulated Col-0 and miz1 roots revealed the formation of a statistically significant asymmetric [Ca2+]cyt distribution at ∼250- to 400-µm segment above the root apex of the EZ of hydrostimulated Col-0, which was not observed in control Col-0 roots nor in control or hydrostimulated miz1 roots (Fig. 1, , and Movies S1–S3). This relatively slow, long-distance signaling pattern, occurring in the range of an hour, is distinct in its kinetics and tissue specificity from previously reported Ca2+ waves or propagations in response to abiotic or biotic stimuli that occur in seconds to minutes (14, 27, 28).

Next, we determined whether the kinetics of the asymmetric distribution of [Ca2+]cyt coincides with the root bending time course by measuring the NES-YC3.6 signal at the EZ and the root bending every 10 min, 0–120 min from the start of hydrostimulation. In Col-0, the results show a maximum difference of [Ca2+]cyt across the root EZ at about 50–80 min from the onset of hydrostimulation, with an estimated [Ca2+]cyt peak of 244.5 ± 22 nM (conversion of the FRET/CFP ratio value to molar concentration was performed as in ref. 20) at the forming convex side following 70 min, whereas significant curvature toward higher water potential was observed after about 60–70 min (Fig. 1). No asymmetric signal was apparent in miz1 roots in which [Ca2+]cyt levels did not rise above ∼102.9 ± 15 nM in either root side at any time point during 120 min of hydrostimulation (). These data clearly show that [Ca2+]cyt elevation and its asymmetric distribution at the EZ of the root precede root bending and thus most likely regulate it in a mechanism requiring the activity of MIZ1.

To determine whether the asymmetric [Ca2+]cyt signal across the root indeed results from an asymmetric water potential distribution at the root tip, rather than a general, or predisposed response to osmotic stress, we compared the kinetics of the distribution of [Ca2+]cyt across the root in a diagonal split-agar/sorbitol system, with that caused by a horizontal split-agar/sorbitol system (Fig. 2). After 30 min of stimulation, an asymmetric Ca2+ signal was observed in the diagonal split-agar/sorbitol assay. In contrast, no asymmetric [Ca2+]cyt signal was observed in Col-0 roots even after 1 h of hydrostimulation in the horizontal split-agar/sorbitol system (Fig. 2). Moreover, the direction of root bending in the diagonal system was unified toward high water potential, whereas following stimulation in the horizontal system, nonunified root growth direction was observed (Fig. 2). These data indicate that the formation of the asymmetric [Ca2+]cyt signal directly results from root tip exposure to asymmetric water potential. In contrast, examination of the [Ca2+]cyt in gravistimulated roots revealed no generation of a Ca2+ signal along the root, before or after an observed root curvature (), in agreement with previous findings (17). Thus, the slow, asymmetric Ca2+ signal from the root tip to the EZ is a specific response to water potential distribution across the root. Collectively, these data, obtained in intact roots, strongly support the importance of the root cap and meristem in perceiving changes in water potential pressure and in generating a Ca2+ signal transmitted to the EZ to promote bending, as also suggested in previous studies (1, 4).

Fig. 2.

Asymmetric [Ca2+]cyt signal in the root tip is generated in response to exposure to asymmetric water potential gradient. (A) Schematic presentation of horizontal and diagonal hydrostimulation assay systems. g represents gravity vector, and Ψ represents water potential gradient. (B and C) Quantification of FRET/CFP intensity ratio (B) and root curvature (C) of two longitudinal halves of the EZ (250–350 µm above apex) of NES-YC3.6–expressing Col-0 seedlings hydrostimulated for the indicated times, using the two systems depicted in A. In B and C, error bars represent mean ± SD (three biological independent experiments, 10 seedlings each).

The Shootward Asymmetric Ca2+ Signal Is Transmitted Through the Root Phloem to the EZ Where It Is Laterally and Asymmetrically Distributed Across the Root.

To pinpoint the specific vascular tissue in which the slow, long-distance Ca2+ signal is transmitted, we visualized the NES-YC3.6 fluorescence ratio (signal intensity) in the background of the bright-field images of root segments from two perspective angles relative to the phloem and xylem poles (Fig. 3). To optimize the resolution of our visual inspection, we adjusted the maximum and minimum ratio values such that the cells with the highest signal intensity could be identified. Clearly, when the ratio signal in the protoxylem was found to be below the intensity cutoff, a prominent continuous longitudinal [Ca2+]cyt signal was visualized in the phloem tissue (Fig. 3), a result that is in agreement with previous reports of Ca2+ transport and function in phloem sieve tubes in response to osmotic or biotic stresses (16, 29).

Fig. 3.

[Ca2+]cyt signal is transmitted via the root phloem to form an asymmetric lateral Ca2+ gradient in response to hydrostimulation. (A) Confocal microscope visualization of NES-YC3.6 EZ of Col-0 roots from two perspective angles relative to the xylem and phloem poles. The maximum and minimum FRET/CFP ratio values were adjusted so that the cells with the highest signal could be identified by superimposing the ratio images on bright-field images. (Scale bar, 50 µm.) (B) Schematic presentation of the in-tube hydrostimulation assay designed for root imaging using light-sheet fluorescence microscopy. (C) Visualization of radial EZ root sections of control and hydrostimulated NES-YC3.6–expressing Col-0 seedlings treated as in B. (Scale bar, 50 µm.) Images were created based on FRET/CFP ratio and pseudocolored when red indicates higher [Ca2+]cyt. (D) Quantification of the FRET/CFP intensity ratio of control and hydrostimulated roots radial halves that were obtained as in B. Error bars represent mean ± SD (three biological experiments; five seedlings each). Bars with different letters represent statistically different values by Tukey’s HSD post hoc test (P < 0.01).

If indeed the asymmetric Ca2+ increase in the EZ vasculature (phloem) regulates root bending, it should either reach the cells of the peripheral layers (e.g., cortex) by lateral mobilization, where differential elongation takes place, or it should be conveyed to the peripheral cells by a different signal. To address this issue, we used light-sheet fluorescence microscopy to visualize the radial root EZ of NES-YC3.6–expressing seedlings upon hydrostimulation (30). For this purpose, we designed a special “in-tube” hydrostimulation system (Fig. 3). Under control conditions, we found the highest [Ca2+]cyt level in phloem cells and much less in peripheral tissue layers, mainly at the phloem pole outer layers, with no apparent asymmetry (Fig. 3 ). Strikingly, in hydrostimulated roots, [Ca2+]cyt levels were elevated in the peripheral tissues, with substantially higher levels in the phloem and cortex of the evolving convex root side (Fig. 3 ). Interestingly, lateral Ca2+ mobilization, possibly through plasmodesmata (31, 32), from phloem sieve tubes to peripheral tissues, was previously proposed (16). Collectively, these data strongly suggest that, in response to asymmetric water potential distribution across the root tip, an asymmetric long-distance Ca2+ signal spreads via the phloem, followed by lateral Ca2+ mobilization to peripheral cells at the EZ, required for differential root elongation and bending. Nevertheless, how Ca2+ at the EZ affects differential cell elongation remains an open question.

Elevation of [Ca2+]cyt in the Root Is Essential for Hydrotropic Bending.

To test the effect of Ca2+ on the root’s tropic response to moisture distribution, we treated wild-type (Col-0) plants with the highly selective, cell-permeant Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis–acetoxymethyl ester (BAPTA-AM) before hydrostimulation using the split-agar/sorbitol system. Control plants (without BAPTA-AM) displayed normal root bending, as described in a similar published experimental setup (13, 33, 34), whereas BAPTA-AM–treated roots displayed arrested bending in response to the change in water potential in their microenvironment even after 12 h, while continuing their growth toward the sorbitol-containing media (Fig. 4 ). In contrast, pretreatment of seedlings with the Ca2+ ionophore, Br-A23187, significantly enhanced root curvature by up to ∼50% more than in control roots (Fig. 4 ). The effectiveness of BAPTA-AM and Br-A23187 treatments on [Ca2+]cyt levels before hydrostimulation was monitored in roots of NES-YC3.6–expressing seedlings, and it was found to reduce and elevate [Ca2+]cyt levels, respectively (), as expected. In addition, these treatments were not found to significantly affect root growth under normal conditions (). These data suggest that the tropic response of Arabidopsis roots to a water potential gradient requires the elevation of [Ca2+]cyt levels.

Fig. 4.

Cytosolic Ca2+ levels determine root tip response to moisture gradient. (A, C, and E) Split-agar/sorbitol system to determine root curvature in response to hydrostimulation (). (A–D) Wild-type Arabidopsis (Col-0) seedlings were treated for 2 h with the cell-permeant Ca2+ chelator BAPTA-AM (10 µM) or the ionophore Br-A23187 (20 µM) before 12 or 6 h of hydrostimulation (), respectively, and curvature was scored. (E and F) eca1-3 seedlings (versus Col-0 as control) were hydrostimulated for 6 h, and curvature was scored. In B, D, and F, error bars represent mean ± SD (three biological independent experiments, 10 seedlings each); *P < 0.01 and **P < 0.005, Student’s t test versus control (B and D) or Col-0 (F). (G) Confocal microscope visualization of NES-YC3.6 signals in root tips of Col-0 and eca1-3 under control conditions and following 1 h of hydrostimulation. Images were created based on FRET/CFP ratio and pseudocolored; red indicates higher [Ca2+]cyt level. g represents gravity vector, and Ψ represents water potential gradient. (Scale bar, 50 µm.) All hydrostimulation assays were performed using the split-agar/sorbitol system.

To corroborate the effect of [Ca2+]cyt in roots responding to water potential gradients, mutants with aberrations in type 2B Ca2+ pumps, including ACA2 (AT4G37640), ACA8 (AT5G57110), and ACA10 (AT4G29900), were subjected to hydrostimulation in the split-agar/sorbitol system. None of the tested mutants exhibited appreciable differences from WT in root bending in response to hydrostimulation. On the other hand, mutant seedlings of the type 2A Ca2+-ATPase ECA1, a pump that imports Ca2+ into the ER lumen (35), and which is related to the mammalian sarco/ER Ca2+-ATPase (SERCA Ca2+ pump) (36), displayed enhanced bending toward higher water potential (Fig. 4 ). The growth rate of eca1 roots was similar to that of the WT roots ().

In view of the known association of MIZ1 with the ER membrane (25), we pursued investigating ECA1 and MIZ1 regarding Ca2+ signaling, water tracking, and their possible interaction. Visualization of [Ca2+]cyt in NES-YC3.6–expressing Col-0 and eca1-3 roots under control and hydrostimulation conditions revealed higher concentrations of [Ca2+]cyt in eca1-3 than in Col-0 under both conditions (Fig. 4) and an enhanced [Ca2+]cyt signal following 1 h of hydrostimulation (), which most likely explains the rapid response of eca1 to hydrostimulation. Interestingly, examining the expression pattern of ECA1 in ECA1:ECA1-GFP (in the genetic background of eca1-1) revealed ECA1 expression in all root tip tissues, with a higher abundance in the region between the apex and the EZ, and particularly high expression in the phloem (). Measuring the hydrotropic root bending of ECA1:ECA1-GFP-harboring eca1-1 seedlings indicated full restoration of the normal tropic response (). These data suggest that inhibition of ECA1 may be required for generating the phloem-transmitted long-distance [Ca2+]cyt signal in response to hydrostimulation.

MIZ1 Directly Interacts with ECA1 and Regulates Its Activity.

To further study the possible involvement of ECA1 and MIZ1 in generating the [Ca2+]cyt signal, we visualized the NES-YC3.6 ratios in Col-0 and miz1 roots following treatment with cyclopiazonic acid (CPA), which was previously found to inhibit the Arabidopsis ECA1 (35). CPA treatment for 1 h elevated the [Ca2+]cyt levels in Col-0 (in accordance with ref. 26) but not in miz1 roots. This indicates that MIZ1 is required for CPA-mediated inhibition of ECA1 (Fig. 5), possibly by direct interaction of MIZ1 with ECA1 or with an ECA1-associated protein complex. To further explore this possibility, we performed immunoprecipitation assays to isolate MIZ1-interacting proteins by using a GFP isolation kit to trap MIZ1-citrine or miz1-citrine from extracts of the corresponding transgenic plants. Interestingly, Western blot analysis of MIZ1-associated proteins indicated that ECA1 was indeed precipitated with MIZ1-citrine, but to a significantly lesser extent with miz1-citrine (Fig. 5 ), suggesting that ECA1 interacts with active MIZ1 in vivo (either directly or within a protein complex). To determine whether ECA1 and MIZ1 interact directly, we expressed ECA1 fused to the C terminus of ubiquitin (Cub) and MIZ1 or miz1 fused to the N terminus (Nub) to perform split-ubiquitin yeast two-hybrid assays. Remarkably, direct interaction of ECA1 with MIZ1, but not with miz1, was observed in this system (Fig. 5). Expression of ECA1, MIZ1, and miz1 in yeast was confirmed using Western blot analysis (). In addition, to study the possible effect of MIZ1 on ECA1 function, we performed complementation analysis in which ECA1 was expressed alone or was coexpressed with MIZ1 or miz1 in the yeast K616 triple mutant, which lacks functional endogenous Ca2+-dependent ATPases and is thus unable to grow on a Ca2+-depleted medium (37). All transformants grew similarly on nonselective medium (10 mM Ca2+) (Fig. 5). Functional complementation assays on selective media (100 µM Ca2+ or even 20 mM EGTA, which reduces the free Ca2+ concentration in the medium to the nanomolar range), showed that expression of the Arabidopsis ECA1 alone completely restored yeast growth and complemented the K616 phenotype (Fig. 5), as previously described (35, 38). Remarkably, coexpression of ECA1 with MIZ1 substantially reduced yeast growth under these selective conditions, suggesting an inhibitory effect of MIZ1 on ECA1. To assess the specificity of ECA1 inhibition by MIZ1, ECA1 was coexpressed with the miz1 mutant (Fig. 5), which binds weakly to ECA1 both in yeast and in plant-derived microsomes (Fig. 5 ). Indeed, the growth of yeast coexpressing ECA1 with the miz1 mutant was much less inhibited compared with the growth of yeast expressing the WT proteins. Western blot analysis confirmed the expression of ECA1, MIZ1, and the miz1 mutant in the relevant yeast transformants (). These results suggest that MIZ1 has an inhibitory effect on ECA1 function, consistent with the in planta interaction of the two proteins (Fig. 5 ). Finally, to determine whether the MIZ1/ECA1 mechanism functions in the root tip, we quantified the [Ca2+]cyt in root tips (the specific examined root tip part is indicated in ) of Col-0, miz1, and eca1-3 under control conditions and following 1 h of hydrostimulation. The levels of [Ca2+]cyt in the root tips of miz1 did not change appreciably in response to hydrostimulation and did not exceed the levels in Col-0 under control conditions (). On the other hand, the [Ca2+]cyt levels in the root tips of eca1-3 were found to be higher than those of Col-0 under control conditions and to be slightly elevated in response to hydrostimulation (). Collectively, these data suggest that both MIZ1 and ECA1 are required for generating the Ca2+ signal in the root tip in response to hydrostimulation.

Fig. 5.

Direct interaction between ECA1 and MIZ1 facilitates hydrotropic [Ca2+]cyt signaling. (A) Confocal microscope visualization of NES-YC3.6 signals in roots of Col-0 and miz1 roots treated with 0 (control) or 10 µM CPA for 1 h. g represents gravity vector. (Scale bar, 50 µm.) (B and C) Immunoprecipitation assay. (B) Microsomal membranes were isolated from 5-d-old whole seedlings expressing the MIZ1-citrine or miz1-citrine (Input), and protein was purified using GFP isolation kit (IP), followed by Western blot analysis with anti-ECA1 (38) or anti-GFP antisera. (C) Quantification of ECA1 and citrine signals. Error bars represent mean ± SD (three biological experiments). (D) Split-ubiquitin yeast two-hybrid assay. Vectors harboring ECA1 fused to the ubiquitin C terminus coding sequence (ECA1-Cub) and MIZ1 or miz1 fused to ubiquitin N terminus coding sequence were cotransformed to NMY51 Saccharomyces cerevisiae strain on selective medium supplemented with X-gal for detection of β-galactosidase activity. For positive and negative controls, cotransformation was performed with pAI-Alg5 (Nub-Alg5) or pDL2-Alg5 (NubG-Alg5, mutated Nub), respectively (). (E) Yeast growth complementation assay (). S. cerevisiae strain K616 was transformed with empty pESC-URA vector or with the vector harboring ECA1, ECA1 plus MIZ1, or ECA1 plus miz1, and spotted onto SG-URA plates supplemented with 10 mM CaCl2 (Ca2+, not selective), 100 µM CaCl2 (selective), or 20 mM EGTA (highly selective). Representative results are from six independent drop tests with at least four independent biological samples. Expression of the proteins described in D and E are shown in .

Discussion

In this work, we revisited Darwin’s assumption that a signal is transmitted from the root cap to the EZ in response to moisture differences across the root tip (1). Previously, forward genetics approaches revealed only two genes (MIZ1 and MIZ2) mediating hydrotropism (24, 39); however, how the encoded proteins are involved in hydrotropism is not yet understood. We demonstrate that root curvature in response to hydrostimulation requires long-distance [Ca2+]cyt mobilization from the root cap to the EZ (Figs. 1 and 3 and ) that is mediated by the interaction of MIZ1 with the type 2A Ca2+-ATPase isoform ECA1, an ER-localized Ca2+ efflux carrier. Furthermore, we provide evidence for the importance of the root tip in generating the hydrostimulated long-distance Ca2+ signal through the interaction of these proteins (). The expression pattern of ECA1 and MIZ1 in the root tip, stele, and peripheral tissues of the EZ () (25, 40, 41) raises the possibility of an active MIZ1/ECA1 mechanism in different root tissues. However, the mechanism that underlies the propagation of the long-distance hydrotropic-driven [Ca2+]cyt signal remains unknown. [Ca2+]cyt propagation may involve other intermediate signals such as ROS or electric signals, as described for long-distance [Ca2+]cyt signals associated with different physiological responses (27, 42, 43). We show that such Ca2+ signals do not occur in the miz1 mutant (Fig. 1 and ). Importantly, we found that in response to cross-root water potential differences, the long-distance Ca2+ signal is transmitted asymmetrically through the phloem to the EZ where it is laterally distributed asymmetrically to peripheral cells (Figs. 1 and 3 ), and where it most likely promotes differential cell elongation underlying curvature. Furthermore, manipulations of [Ca2+]cyt confirmed that transient elevation of [Ca2+]cyt is required for root curvature toward water. Specifically, in the absence of the functional ER-localized Ca2+-ATPase pump ECA1, [Ca2+]cyt levels are elevated (Fig. 4 and ) and root curvature toward water is enhanced (Fig. 4 ). The role of the ER in regulating the homeostasis of [Ca2+]cyt is reminiscent of the previously reported elevation of [Ca2+]cyt levels in tobacco plants in which the type 2B ER-localized Ca2+-ATPase NbCA1 was silenced, resulting in an enhanced hypersensitive immune response (44). Recently, the lack or overexpression of the ER-localized CCX2 (a Ca2+/cation exchanger) was found to affect both cytosolic and ER Ca2+ dynamics and tolerance to salt and osmotic stress, demonstrating a role for the ER Ca2+ reservoir in the regulation of [Ca2+]cyt homeostasis (45). Furthermore, our study provides several lines of evidence for the regulation of ECA1 by MIZ1. First, CPA, a known inhibitor of ECA1, was unable to inhibit ECA1 in a miz1 mutant (Fig. 5), suggesting that MIZ1 is associated with ECA1. Indeed, coimmunoprecipitation experiments confirmed this association (Fig. 5 ). Furthermore, protein–protein interaction assays in yeast revealed the direct interaction of ECA1 with MIZ1 but not with the miz1 mutant protein (Fig. 5), which in planta abrogates hydrotropism (24). However, although ECA1 appears to function as a major player in the mechanism of root hydrotropism, the involvement of other Ca2+ transporters and/or channels, which function in maintaining [Ca2+]cyt homeostasis or in generating a change from homeostasis in response to a signal, could not be ruled out. Collectively, these results raised the possibility that MIZ1 is a negative regulator of ECA1 and that the interaction of MIZ1 with ECA1 is essential for elevating [Ca2+]cyt levels, which underlies root hydrotropic curvature. To test this, we employed a previously described functional complementation assay of yeast strain K616 by expression of ECA1 (35, 38). The results confirmed the ability of MIZ1, and to a much lesser extent miz1, to attenuate the complementation of yeast strain K616 by ECA1 (Fig. 5), consistent with the proposed inhibitory effect of MIZ1 on ECA1 function. The experimental evidence demonstrating a mechanism underlying the elevation of [Ca2+]cyt levels by inhibition of an ER type 2A Ca2+-ATPase is unique, yet is consistent with the previously reported elevation of [Ca2+]cyt levels resulting from artificially silencing a tobacco ER-localized type 2B Ca2+ ATPase (44). The simultaneous decline of [Ca2+]ER and the elevation of [Ca2+]cyt levels in the root tip upon hydrostimulation () further support this scenario. Hence, our study offers an example of a long-distance Ca2+ signal that is generated by inhibiting a Ca2+ efflux carrier, in agreement with previous theoretical considerations of the essential role of Ca2+ efflux carriers in stress signaling (46).

Another important question that we addressed is the spatial distribution of [Ca2+]cyt reaching the EZ. We used light-sheet fluorescence microscopy to obtain an image of [Ca2+]cyt distribution across the root in response to hydrostimulation. Remarkably, in response to hydrostimulation, [Ca2+]cyt was distributed asymmetrically across the root at the EZ, suggesting that this asymmetric distribution may facilitate asymmetric root cell elongation that results in root curvature. However, the mechanism downstream of this Ca2+ signal remains to be identified. The possible involvement of ABA in this Ca2+ effect cannot be ruled out (6).

However, several open questions regarding hydrotropism still remain unanswered. How is a water potential gradient across the root detected? How does the detector transduce the sensed water potential gradient to MIZ1 to generate an asymmetric Ca2+ signal? How does [Ca2+]cyt propagate from the root tip to the EZ? How is the Ca2+ signal mobilized across the root at the EZ? Finally, how does asymmetric [Ca2+]cyt, which is distributed across the root, mediate asymmetric cell elongation and root curvature? Further interdisciplinary investigations are required to answer these open questions.

Materials and Methods

Plant materials, methodologies of microscopy work, yeast two-hybrid system, yeast functional complementation assays, and biochemical procedures are described in .