The FERONIA Receptor Kinase Maintains Cell-Wall Integrity during Salt Stress through Ca2+ Signaling.

Cells maintain integrity despite changes in their mechanical properties elicited during growth and environmental stress. How cells sense their physical state and compensate for cell-wall damage is poorly understood, particularly in plants. Here we report that FERONIA (FER), a plasma-membrane-localized receptor kinase from Arabidopsis, is necessary for the recovery of root growth after exposure to high salinity, a widespread soil stress. The extracellular domain of FER displays tandem regions of homology with malectin, an animal protein known to bind di-glucose in vitro and important for protein quality control in the endoplasmic reticulum. The presence of malectin-like domains in FER and related receptor kinases has led to widespread speculation that they interact with cell-wall polysaccharides and can potentially serve a wall-sensing function. Results reported here show that salinity causes softening of the cell wall and that FER is necessary to sense these defects. When this function is disrupted in the fer mutant, root cells explode dramatically during growth recovery. Similar defects are observed in the mur1 mutant, which disrupts pectin cross-linking. Furthermore, fer cell-wall integrity defects can be rescued by treatment with calcium and borate, which also facilitate pectin cross-linking. Sensing of these salinity-induced wall defects might therefore be a direct consequence of physical interaction between the extracellular domain of FER and pectin. FER-dependent signaling elicits cell-specific calcium transients that maintain cell-wall integrity during salt stress. These results reveal a novel extracellular toxicity of salinity, and identify FER as a sensor of damage to the pectin-associated wall.

INTRODUCTION

Growth places a cell in a precarious situation. During cell expansion, structural components that limit the size of the cell must be weakened, while at the same time, homeostatic mechanisms must act to prevent a loss in cell integrity. In plant roots, these delicate processes occur in organs fully exposed to soil, where the physicochemical properties of this environment can change dramatically over the micron-length scale [1]. Excessive salinity occurs in natural and agricultural field conditions and can impose both osmotic and ionic stress, which limit the ability of cells to take up water from the environment and cause cytoplasmic and organellar toxicity [2]. The root system responds to salt stress by dynamically regulating growth rate and direction at the organ level. Cells in the elongation zone enter a quiescent stage upon salt treatment for several hours before growth ultimately resumes [3]. The efflux carrier PIN2-mediated redistribution of auxin results in reorientation of the root tip away from the saline environment [4]. At the cellular level, salinity causes excessive radial cell expansion in roots [5], similar to genetic or chemical disruption of cell-wall integrity [6, 7]. Conversely, mutants defective in cell-wall organization are hypersensitive to salt stress [5, 8]. These data highlight a potential apoplastic toxicity of salinity stress that has not been well characterized and for which a signal transduction pathway has not been defined.

In plants, accumulating evidence suggests the existence of cell-wall integrity sensing pathways that monitor changes in wall properties [9]. Among potential cell-wall sensors, plasma-membrane-localized receptor-like kinases (RLKs), such as the FERONIA (FER)-related malectin-domain-containing THESEUS1 [10] and ANXUR1/2 [11], as well as the leucine-rich repeat (LRR) RLK, MALE DISCOVERER 1-INTERACTING RECEPTOR LIKE KINASE 2 [12], have been suggested to be involved in cell-wall sensing. Loss-of-function mutations in FER cause pleiotropic mutant phenotypes, including defects in female fertility, cell elongation, root-hair development, mechanosensing, and responses to hormones and pathogens [13-16]. One emerging feature of FER-dependent signaling is the downstream induction of Ca2+ responses [17-19]. The spectrum of fer phenotypes suggests that the protein may be involved in sensing a cue that is common to these biological pathways, such as a change in the properties of the cell wall.

RESULTS

FER Is Essential for Seedling Viability under Salt Stress

To investigate the potential role of FER in cell-wall sensing during salt stress, we challenged loss-of-function fer mutants (fer-2, fer-4 [20, 21]) with high salinity. Compared to WT, fer seedlings displayed significant root growth defects within 24 hr of salt stress (Figures 1A and 1B) and were not able to fully recover their growth rate (Figure 1C). These growth defects were fully complemented by a trans-gene containing GFP-tagged FER cDNA driven by its own promoter (Figures 1A and 1B) and are consistent with the reduced root length observed for fer mutants under salinity in a previous study [22]. The root growth defect of fer seedlings is also associated with a low survival rate (12.7% ± 4.3%) compared to WT (92.2% ± 2.5%) at 2 days after salt treatment, highlighting the essential function of FER in salt tolerance. Salt hypersensitivity in fer mutants was most severe at concentrations above 100 mM NaCl, and was not observed for hyper-osmotic media containing mannitol or sorbitol with comparable osmolarity (e.g., 280 mM mannitol is iso-osmotic with 140 mM NaCl) (Figures S1A and S1B). Growth defects were observed with Na2SO4 (Figure S1C), suggesting that fer mutants are hypersensitive to the stress imposed by sodium ions, rather than the associated osmotic stress.

Figure 1

FER Is Required for Root Growth Recovery during High Salinity

(A) Seedlings of WT, fer-4, and a FER complemented line 24 hr after transfer to media. Black lines mark the position of the root tip at time of transfer.

(B) Root growth of WT, two fer alleles, and two independent FER complemented lines 24 hr after transfer to media. n ≥ 15.

(C) Quantification of hourly growth rate of WT and fer-4 roots transferred to media containing 140 mM NaCl. n = 16.

(D) Root growth of WT, fer-4, ralf1-2, RALF1 RNAi, and RALF1 overexpression lines 24 hr after transfer to media. n ≥ 17.

Error bars indicate SEM. *p < 0.05, **p < 0.01, ***p < 0.001, two-way ANOVA compared to WT and control treatment. See also Figure S1.

FER interacts with a peptide growth regulator, RAPID ALKALINIZATION FACTOR 1 (RALF1), and causes inhibition of the proton transport capacity of plasma-membrane-localized H+-ATPase pumps, which presumably affects growth [23]. RALF1 does not appear to have important functions in regulating growth under salt stress, as the ralf1 mutant [23] and a RALF1 RNAi transgenic line [24] did not show a significant growth defect (Figure 1D). These data suggest the role of FER during salt stress is largely independent of RALF1, although this does not preclude the possibility that related RALF1-like peptides are necessary signaling components in a salt-dependent pathway. Interestingly, overexpression of RALF1, which results in seedlings with shorter roots [24], conferred some relative resistance to salt stress (Figure 1D), which suggests that hyperactivation of the FER pathway may provide protective effects.

Consistent with our genetic analysis, treatment of fer mutants with the H+-ATPase inhibitor sodium orthovanadate did not rescue growth under salt stress (Figure S1D). Instead, root growth was consistently inhibited by sodium orthovanadate treatment regardless of salt stress, suggesting that excessive proton-pump activity is unlikely to be the basis for the salinity-dependent mutant phenotype observed in fer.

fer Root Cells Lose Cell-Wall Integrity and Burst during Growth Recovery under Salt Stress

To determine the cellular basis for the growth defects of fer roots under high salinity, we performed live imaging with seedlings expressing a plasma-membrane-localized YFP marker [25]. Consistent with previous studies [5], 4D confocal imaging followed by MorphoGraphX analysis [26] revealed that WT root cells in the elongation zone began to display a loss of expansion anisotropy at the end of the quiescent stage 6–8 hr after the beginning of high-salinity treatment, and resulted in radial swelling of the outer tissue layers [5] (Figures 2A–2C; Movie S1). In the fer mutant, a similar loss of expansion anisotropy was initiated but quickly followed by rupturing of cells (seen as a break in the continuity of the plasma-membrane marker, and subsequently a loss of YFP fluorescence; Figures 2D–2F; Movie S1). Cell bursting occurred to such an extent that it sometimes caused a complete loss of intact cells in the outer tissues of the growth zone (Figure 2F). Tissue-specific quantification of these events in fer seedlings revealed that cells in the epidermal and cortical layers showed a loss of cell integrity starting 3–4 hr after the beginning of salt treatment and peaking in frequency by 7–8 hr (Figure 2H; Figures S2A and S2B). In contrast, whereas a few epidermal cells in WT also lost cell integrity after salt treatment, the frequency was much lower and was rarely observed in cortical cells (Figure 2G). We also traced the volume change of fer epidermal cells and found a rapid increase before cell bursting (Figure S2C), suggesting that excessive loosening of the wall may cause the loss of cell integrity. Taken together, these data reveal that FER promotes growth recovery and seedling survival during salt stress, most likely by maintaining cell-wall integrity.

Figure 2

FER Maintains Cell Integrity during Growth Recovery after Salt Stress

(A–F) Time-lapse images of LTI6-YFP-expressing WT (A–C) and fer-4 (D–F) roots after transfer to media containing 140 mM NaCl. Individual epidermal cells at each time point are identified by the same color after image segmentation in MorphoGraphX. Asterisks mark two WT cells that display loss of anisotropy during recovery. “x” marks a fer cell that burst during growth. Scale bars, 50 μm.

(G and H) Quantification of burst cells from time-lapse imaging of LTI6-YFP-expressing WT (G) and fer-4 (H) roots. n = 15.

(I–K) Heatmaps showing the average number of burst cells of WT (I), fer-4 (J), and sos1-1 (K) roots in each 750-μm longitudinal zone during each 1-hr time window after transfer to media containing 140 mM NaCl. Quantification is performed with a single Z section, 16 μm from the epidermis. Gray areas indicate average cell death rate higher than 4 per hr. n ≥ 10.

Error bars indicate SEM. See also Figures S2 and S3 and Movie S1.

The ionic, rather than osmotic, sensitivity of the fer mutant is similar to the salinity overly sensitive (sos) class of mutants, which are defective in excluding sodium from the cytoplasm [27, 28]. Despite this similarity, fer and sos mutants display different phenotypes under salt stress. Previous studies have shown that sos1 and sos2 mutant roots exhibit hypersensitivity to Na+ and Li+, but not K+ [29]. This contrasts starkly with fer, which is hypersensitive to K+; root growth was retarded even at 50 mM KCl or KNO3 (Figure S2D). fer and sos mutants also displayed different sensitivities to the concentration of NaCl in the media; sos1 and sos2 showed a severe reduction in growth starting at concentrations of 50 mM NaCl, whereas a considerably higher concentration (greater than 100 mM) was required to induce a similar growth defect in fer (Figure S3A). Furthermore, unlike fer, whose salt-induced loss of cell viability is restricted to the elongation zone, sos mutants displayed an extensive wave of cell death that spread from the elongation zone to the root tip, excluding the root cap (Figures 2I–2K; Figure S3B). These differences point to a unique growth-dependent function of FER in the salt response that distinguishes it from other known salt-tolerance pathways.

Previous work has shown that the glycosylphosphatidylino-sitol-anchored protein (GPI-AP) LORELEI-like GPI-AP1 (LLG1) interacts directly with FER and is required for FER localization to the plasma membrane and signaling [30]. Similar to other llg1 mutant phenotypes, which are indistinguishable from fer mutant seedlings, llg1 roots also showed the same ionic, but not osmotic, sensitivity as fer and similar cellular damage arising from a loss of cell-wall integrity (Figure 4C; Figure S3A). These results provide further evidence that the FER/LLG1-mediated salt-tolerance pathway is distinct from the SOS pathway.

Figure 4

Borate Treatment Prevents Cell Bursting in fer and mur1 Mutants during Salt Stress

(A) Root growth of WT and fer-4 mutant 2 days after transfer to NaCl media with or without supplementary CaCl2 or boric acid. Salt (S), 140 mM NaCl; Ca, 5 mM CaCl2; B, 3 mM boric acid. n ≥ 15. (**p < 0.01, ***p < 0.001, two-way ANOVA compared to WT and salt-only treatment.)

(B) Root growth of WT and two mur1 alleles 24 hr after transfer to media with or without 140 mM NaCl. n ≥ 16. (***p < 0.001, two-way ANOVA compared to WT and control treatment.)

(C) Number of burst cells observed within the first 16 hr after salt treatment in WT, fer-4, a FER complemented line, two llg1 alleles, two mur1 alleles, and mur1-1 supplemented with 5 mM CaCl2 and 3 mM boric acid using wide-field microscopy. Significantly different groups are denoted with different letters (p < 0.05). n ≥ 7.

(D–G) Quantification of burst cells from time-lapse imaging of LTI6-YFP-expressing WT (D and E) and fer-4 (F and G) roots transferred to NaCl mediawith or without supplementary CaCl2 and boric acid. n ≥ 10.

Error bars indicate SEM. See also Figure S4.

FER Is Necessary to Reverse the Wall-Softening Effects of Salinity

To examine whether salinity and FER affect the mechanical properties of the cell wall at the root tip region, atomic force microscopy (AFM) was used. Previous studies using AFM on elongating hypocotyl tissues had plasmolyzed cells before probing, which allowed the mechanical properties of the cell cross-walls to be interrogated [31]. Here, the surface wall of epidermal cells was of greatest interest, because we observed rupturing of cells in this region; therefore, un-plasmolyzed roots were used (see STAR Methods). Within the indentation range used (less than 500 nm; Figure S4B), the force-by-indentation curves of both turgid and plasmolyzed cells similarly fit a paraboloid model (Figures S4C and S4D), suggesting that there was minimal contribution from cellular turgor within our measurements. After salt treatment, an overall drop in wall stiffness was observed in samples collected 2–4 hr after the beginning of salt treatment in both WT and fer (Figures 3A, 3B, 3D, and 3E), represented by a shift in the stiffness distribution curve toward lower values (Figure 3G) as well as an overall decrease in the average force measured in the elongation zone (Figure S4A). Interestingly, WT roots were able to recover wall stiffness at the beginning of the growth recovery phase, as the stiffness distribution curve is indistinguishable from non-salt-treated roots (Figures 3C and 3G). However, fer roots failed to restore wall stiffness before entering the growth recovery phase (Figures 3F and 3G), which may underlie the loss of cell-wall integrity.

Figure 3

Regulation of Cell-Wall Mechanics Is Impaired in the fer Mutant after Salt Treatment

(A–F) Apparent Young’s modulus (EA, or “stiffness”) map of WT (A–C) and fer-4 (D–F) root epidermal cells in the elongation zone. Samples were assayed in control condition (A and D) or 2–4 hr (B and E) and 5–8 hr (C and F) after 140 mM NaCl treatment. Bright-field images show the region where the analysis was performed. Scale bars, 100 μm.

(G) Range distribution of the EA of WT and fer-4 root epidermal cells before and during salt treatment. A local polynomial regression fitting with the 95% confidence interval (gray) was used to represent the distribution curve. n ≥ 5. KPas, 1000 Pascal units.

See also Figure S4.

Fortification of Pectin Cross-Links Restores Growth and Cell-Wall Integrity of fer Seedlings under Salt Stress

We hypothesized that the initial decrease in wall stiffness after salt treatment was due to the effects of sodium ions directly disrupting load-bearing ionic interactions within the wall. A particularly important component in growing cells is the pectin matrix, in which the cellulose network is embedded [32]. As a group of highly complex and dynamically modified heteropolysaccharides, pectin has been studied for its role in controlling cell-wall extensibility. The basic building block of pectin is polygalacturonic acid (PGA), which forms a hydrogel through cross-linking mediated by divalent cations. In the presence of Ca2+, gelation occurs due to the formation of intermolecular junction zones within de-esterified homogalacturonan to form the so-called egg box structure, which has been proposed to be load bearing [32]. Monovalent metal ions such as Na+ and K+ can displace pectin-bound Ca2+ and disrupt pectin cross-linking in vitro [33], which may underlie the hypersensitivity of fer to both of these ions (Figure 1B; Figures S1C and S2D).

RG-II is a sub-type of pectin unique for its use of boron to facilitate covalent dimerization that adds strength to the pectin matrix and reinforces Ca2+-mediated pectin cross-linking [34]. Interestingly, whereas supplementing with borate alone only partially, but significantly, complemented the root growth defects of fer under salt stress, adding both Ca2+ and borate together nearly completely restored growth (Figure 4A; Figure S4E) and substantially reduced the cell-wall integrity defects observed in the epidermal and cortical cell files (Figures 4D–4G). These data suggest that disruption of RG-II dimerization by sodium ions may be the basis for the cell-wall integrity defects observed in fer mutants, as only RG-II in the wall is dependent on borate. Consistent with this hypothesis, two independent mutant alleles of murus1 (mur1), which have a constitutive defect in RG-II dimerization due to loss of a sugar moiety near the dimerization site [35], exhibited a fer-like growth defect and cell-bursting phenotype upon salt stress, although weaker than fer (Figures 4B and 4C). Supplementation of borate and Ca2+ in the media also largely rescued the cell-bursting phenotype of mur1 mutants (Figure 4C), indicating that the hypersensitivity to salinity of mur1 was due to a defect in pectin cross-linking.

We also tested the contribution of other cell-wall components during salt stress. Disrupting the cellulose synthase complex by isoxaben treatment had a significant interaction with salt treatment (Figures S4F–S4H), suggesting that root growth under salinity is especially sensitive to defects in wall biosynthesis. The observed non-additive effect of isoxaben treatment is consistent with salinity impacting a non-cellulose-based component of the wall. The xyloglucan xylosyltransferase 1,2 (xxt1,2) double mutant, which has a reduced level of xyloglucan [36], appeared largely normal under salt stress (Figure S4I). Together, these data suggest that salinity causes a loss in cell-wall integrity during salt stress through effects on RG-II-type pectin, and these can be exacerbated through defects in other wall components.

The FER Extracellular Domain Binds Pectin In Vitro

The presence of the tandem malectin-like domains A and B (MALA, MALB) in the extracellular domain of FER (Figure S5A) and related CrRLK1L proteins suggests a potential to interact directly with cell-wall polysaccharides, such as pectin [13, 37]. To test this hypothesis, we examined the ability of a tagged FER extracellular domain (FERecd) (Figure S5B), and MALA and MALB alone, to interact with PGA, the backbone polymer of pectin. In several independent assay systems, hemagglutinin (HA)-FERecd consistently showed dose-dependent interactions with PGA (Figure 5). In affinity electrophoresis, HA-FERecd mobility was retarded in native polyacrylamide gels embedded with increasing concentrations of PGA (Figure 5A). Similarly, HA-FERecd showed a PGA-dependent mobility shift when pre-mixed with different concentrations of PGA prior to electrophoresis on a native gel (Figure 5B). The electrophoretic mobility of similarly treated BSA was not affected in these assays (Figures 5A and 5B). In dot blots and ELISA using a monoclonal antibody that recognizes homogalacturonan (JIM5), immobilized HA-FERecd, but not several control proteins, retained JIM5-detected epitopes after incubation with increasing concentrations of PGA (Figure 5C; Figure S5C). Furthermore, when PGA was bound to membranes, maltose binding protein (MBP)-tagged FERecd, MALA, and MALB also bound to the immobilized pectin in a dose-dependent manner (Figures 5C–5E; Figures S5C and S5D). Similarly, when MBP-MALA or MBP was immobilized on amylose beads, significantly higher levels of pectin were bound to MBP-MALA-coated beads relative to control MBP-coated beads (Figure S5E). Inclusion of Ca2+ in the binding buffer had no discernible impact on FERecd binding (Figure S5G). The ability to interact with pectin extends to the extracellular domains of a number of FER-related receptor kinases that were tested, including BUDDHAS PAPER SEAL (BUPS1) [38], ANXUR1/2 (ANX1/2) [11], as well as the unrelated WALL ASSOCIATED KINASE1 (WAK1), as was shown previously [39] (Figures S5F and S5H). On the other hand, although all receptor kinases, with their extracellular domains located in the cell-wall milieu, have the potential to interact with its components, some react poorly with pectin, such as the LRR-domain-containing BAK1 [40] (Figure S5I).

Figure 5

FER Extracellular Domain Interacts with Pectin In Vitro

(A) In-gel affinity electrophoresis. Equal amounts of purified HA-FERecd (Figure S5B) or BSA were electrophoresed in native polyacrylamide gels polymerized in the presence of different PGA concentrations.

(B) Electrophoretic mobility shift assay. Purified HA-FERecd or BSA was mixed with different concentrations of PGA prior to electrophoresis on a native polyacrylamide gel.

(C and D) Representative images of dot-blot assays.

(C) HA-FERecd or control proteins (GFP-His6 and MBP) were immobilized on polyvinylidene fluoride (PVDF) filters and incubated with increasing concentrations of PGA. Pectin bound to the protein dots was detected by JIM5 antibody.

(D) PGA were immobilized on nitrocellulose filters, which were then incubated with FER-derived protein domains or control (MBP). Proteins bound to pectin dots were detected by anti-MBP antibody.

(E) Quantification of a typical dot-blot assay (Figure S5D) showing dose-dependent binding of the FERecd to heat-fragmented PGA. Error bars indicate SEM. (***p < 0.001, Student’s t test compared to MBP control.)

See also Figure S5.

Together, these data provide strong evidence for the MAL domains as underlying the ability for the FERecd to interact with pectin in vitro. Although these data do not preclude the possibility that the FERecd interacts with other components of the wall, they suggest the potential of a direct sensing role for FER to cell-wall perturbation. The observation that pectin-gelling agents mitigate the severity of fer’s response to salt stress (Figures 4A and 4D–4G) also supports the hypothesis that pectin-FER interactions contribute biologically to mediating salinity-induced responses.

FER Induces Cell-Specific [Ca2+] Transients that Maintain Cell-Wall Integrity during Growth Recovery

FER is necessary to induce a transient increase in cytosolic [Ca2+] upon mechanical stimulation, which is necessary to regulate downstream processes [17, 18]. Osmotic shock also induces a rapid calcium transient upon initial exposure to the stress; however, it is not clear whether calcium transients are also associated with late-stage responses such as growth recovery. To determine whether FER is necessary for salinity-induced [Ca2+] transients, we utilized WT and fer seedlings expressing the Yellow Cameleon 3.6 (YC3.6) fluorescence resonance energy transfer (FRET) [Ca2+] sensor [41] and imaged seedlings once every 15 s at the root tip region for up to 15 hr during salt treatment. These extended imaging sessions revealed never-before-described cell-specific [Ca2+] transients occurring mostly in the early elongation zone (Figures 6A–6C; Movie S2), which were absent in roots transferred to standard media (Figure S6A). These transients were localized to individual cells and persisted for less than a minute (Figures 6A–6C; Movie S2). Kinematic analysis of growth at the root tip clearly showed that these [Ca2+] transients were correlated spatially and temporally with growth recovery (Figures 6D and 6F; Movie S3). Importantly, these late-stage calcium transients were strongly reduced in frequency in fer mutants (Figure 6E; Figure S6D; Movie S3), which correlates well with their inability to recover growth (Figure 6G; Figure S6E). WT roots treated with 280 mM mannitol (same osmolarity as 140 mM NaCl) or 100 mM NaCl showed far fewer [Ca2+] transients, which is consistent with the lack of a fer-dependent growth defect under these conditions (Figures S6B–S6E).

Figure 6

FER-Dependent Cell-Specific Calcium Transients Occur during Growth Recovery and Are Necessary to Maintain Cell-Wall Integrity

(A) Maximum-intensity projections of 120 FRET: CFP ratio images (30 min) of a YC3.6-expressing WT root after transfer to media containing 140 mM NaCl. Scale bar, 50 μm.

(B) Δ R:R0 ratio change of the 4 cells labeled in (A) over the course of 30 min during salt treatment.

(C) Time-lapse images of cell 3 in (A) showing the FRET:CFP ratio, FRET, and CFP images.

(D and E) Heatmaps showing the average number of [Ca2+] spikes in cells of WT (D) and fer-4 (E) roots in each 50-μm longitudinal zone during each 30-min time window after transfer to media containing 140 mM NaCl. n = 10. Purple lines mark the early elongation zone used for analysis in Figure S6D.

(F and G) Heatmaps showing the average elemental growth rate of WT (F) and fer-4 (G) roots in each 50-μm longitudinal zone during each 30-min time window after transfer to media containing 140 mM NaCl. n = 10. Purple lines mark the early elongation zone used for analysis in Figure S6E.

(H) Time-lapse images of LTI6-YFP-expressing WT roots pre-treated with 140 mM NaCl for 5 hr, and then transferred to media containing 140 mM NaCl and 0.5 mM LaCl3. Individual epidermal cells at each time point are identified by the same color after image segmentation in MorphoGraphX. Scale bar, 50 μm.

(I) Quantification of burst cells from time-lapse imaging with LTI6-YFP-expressing WT roots during different treatments. Imaging starts 5 hr after the first treatment. Salt, 140 mM NaCl; LaCl3, 0.5 mM LaCl3; Std, standard. n ≥ 6. Error bars indicate SEM.

See also Figures S6 and S7 and Movies S2, S3, and S4.

To test whether these FER-dependent [Ca2+] transients are necessary to maintain cell-wall integrity, we transferred WT seedlings that were pre-treated with salt for 5 hr to media containing a plasma-membrane Ca2+ channel inhibitor, LaCl3. This treatment greatly reduced the number of cell-specific [Ca2+] transients (Figure S6F) and resulted in severe cell bursting (Figures 6H and 6I; Movie S4). Similarly, EGTA, a Ca2+ chelator, also induces WT cell bursting after salt treatment, although less severely and with a slightly different temporal profile (Figures S6G–S6J). These data are consistent with the hypothesis that the loss of cell-wall integrity observed in fer mutants is a consequence of the reduction in late-stage cell-specific [Ca2+] transients. [Ca2+] transients induced upon immediate exposure to NaCl or mannitol treatment occurred outside of the growth zone in WT, and these were generally weaker in fer mutants (Figure S7); however, the lack of exclusively salinity-specific defects in these particular transients suggests that they are not directly involved in the maintenance of cell-wall integrity that occurs several hours later. Ectopic [Ca2+] transients were observed in the fer elongation zone, and suggest localized hypersensitivity of fer mutants to salt stress.

DISCUSSION

Here we report that salinity has a direct impact on the mechanical properties of the cell wall, revealing a previously unknown apoplastic toxicity of sodium ions and, more broadly, monovalent metal cations. We show that the RLK FER is necessary for the acclimation of seedlings under high salinity and that lack of FER-dependent signaling leads to cell rupture, most likely due to the inability of cells to sense and compensate for the weakening of the wall. FER-dependent defects under high salinity can be counteracted by treatment with pectin-gelling agents, and mutants that disrupt pectin cross-linking show fer-like defects in maintaining cell-wall integrity. Together, these data support a model in which high salinity weakens the wall by disrupting pectin cross-linking. Damage to the wall is most likely exacerbated when turgor pressure is restored and tension on the pectin matrix increases, as cells in the elongation zone reactivate cell expansion during the growth recovery stage of the salt-stress response [3]. Such growth-dependent damage may be sensed by FER through its pectin-binding extracellular domain and lead to a rapid cell-autonomous increase in cytosolic [Ca2+], which regulates downstream cellular events necessary to prevent a loss of integrity (Figure 7).

Figure 7

A Model for How FER Signaling Regulates Cell-Wall Integrity under High Salinity

Sodium ions disrupt wall structure, which is sensed by the extracellular domain of FER together with its co-receptor, LLG1. This triggers the activation of an unknown calcium channel resulting in cell-specific [Ca2+] transients during the growth recovery phase. A cell-wall repair mechanism is activated to reinforce wall strength, allowing cells to maintain integrity during cell expansion. See also Table S1 and Data S1.

FER, as well as other members of the CrRLK1L family, also maintain cell integrity in tip-growing cells, such as root hairs and pollen tubes [11, 20]. Tip-growing cells expand through localized softening of the wall, a process that makes these cells susceptible to rupture. Whereas other cell types normally lack such defects in fer, salt stress induces similar bursting in non-tip-growing cells. That salt stress induces bursting even in non-tip-growing fer cells reflects enhanced susceptibility of the mutant cell wall to salt-induced weakening. Interestingly, pectin is particularly prominent at the apex of tip-growing cells, and its cross-linking dynamics are crucial for tip growth [42]. Thus, we speculate that FER-based signaling may be particularly important in regulating cell growth that heavily relies upon pectin cross-linking.

A physical interaction between the extracellular domain of FER and pectin was detected in vitro, suggesting that cell-wall properties may be directly monitored by FER. However, it is important to emphasize that such binding does not necessarily demonstrate direct sensing, as it still remains to be tested whether pectin fragments can directly activate FER signaling. WAKs are another group of RLKs that can bind to pectin, especially oligogalacturonides (OGs), which are released upon pathogen infection or wounding [43]. However, the role of WAKs in abiotic stress response is largely unknown. Recently, very short OGs were found to serve as a dark-associated signal to regulate hypocotyl elongation during skotomorphogenesis [44], suggesting a more general role of OGs as signaling molecules. It is possible that OGs of different lengths can be recognized by different receptors and trigger unique responses. It will be interesting to test whether such pectin fragments are also released during salt stress and can activate FER signaling.

Pectin-FER interactions themselves may be sensitive to monovalent cations in muro, or perhaps changes in wall integrity may affect the interaction of FER with other signaling partners such as RALF-like peptides [23]. Although the molecular details of how such interactions trigger signaling remain to be determined, our results clearly demonstrate that salt stress impacts a FER-dependent signaling pathway that leads to a rapid cell-autonomous increase in [Ca2+] during the growth recovery stage and prevents cell bursting under high salinity (Figure 7). These rapid cell-scale [Ca2+] transients reveal spatial heterogeneity in the acclimation status of cells within the organ, and highlight the important role that cell-autonomous signaling plays in enabling cells to keep pace with growth of the tissue as a whole.

Our work suggests that regulation of wall chemistry may be an important salt-tolerance mechanism. Recent characterization of the seagrass (Zostera marina) genome revealed an expansion in the number of genes encoding pectin methylesterases, consistent with the relatively low degree of methylation in seagrass pectin [45]. It will be interesting for future studies to determine whether pectin modification (e.g., by providing a mechanism to modulate/regulate cross-linking in the wall matrix) also plays a role in enabling salt tolerance in glycophytes such as Arabidopsis.

FER functions in a diverse number of pathways and mediates sensing of abiotic and biotic cues [13, 46]. We have found through transcriptome analysis of fer mutant roots that of the 415 WT-specific salt-induced genes there is substantial enrichment for biotic stress and defense responsive genes, such as those induced by chitin (Table S1; Data S1). Indeed, Espinoza et al. recently demonstrated that salt stress induces a chitin-like transcriptional signature that is dependent on CERK1, the chitin receptor [47]. Induction of biotic-stress-associated transcriptional responses under salt stress may be due to the coupling of FER, a DAMP (damage-associated molecular pattern) receptor, with MAMP (microbial-associated) or PAMP (pathogen-associated) receptors at the plasma membrane. The stress hormone abscisic acid (ABA), which is induced by salt stress [3], dampens biotic-stress signaling [48], and it may be that specificity in the pathway is established through other abiotic stress cues. Alternatively, lack of coherent DAMP and MAMP/PAMP cues could cause these responses to be short lived. This model of FER signaling may apply to other functions that this receptor plays, where coherence of different signaling cues refines the downstream functions that wall sensing ultimately elicits.

STAR★METHODS

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, José R. Dinneny (dinneny@stanford.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Seeds of wild-type Arabidopsis and T-DNA insertion mutants fer-4 (GABI_106A06 [20]), llg1-1, and llg1-2 (CS66105 and CS66106 [30]) were obtained from the Arabidopsis Biological Research Center (ABRC). fer-2 seeds were obtained from Dr. Paul Larsen [21]. Mutants xxt1(CS67834), xxt2(CS67835), xxt1,2 (CS16349) [36], mur1-1 (CS6243) and mur1-2 (CS6244) [50] were also obtained from the ABRC. sos1-1 and sos2-2 mutants are as described [29, 51]. Seeds from homozygous T3 plants expressing RALF1 RNAi construct (irAtRALF1) and 35S::RALF1 were provided by Dr. Tabata Bergonci and Dr. Daniel Moura [24]. The 35S::YC3.6-expressing Col-0, fer-4 T3 seeds were provided by Dr. Gabriele Monshausen (Penn State) [17]. To eliminate position effects of the reporter T-DNA insertion site, 35S::YC3.6-expressing Col-0 was crossed to fer-4. F3-generation seedlings homozygous for the YC3.6 transgene were used for fluorescence imaging. The plasma-membrane marker line LTI6-YFP [25] in Col-0 background was obtained from Dr. David Ehrhardt (Carnegie Institution for Science) and crossed into fer-4, sos1-1, and sos2-2 background to generate homozygous fer-4, sos1-1, and sos2-2 plants expressing LTI6-YFP. To generate FER complemented lines, a wild-type version of FER including 1.3kb of its native promoter was cloned, ligated with GFP at its 3′ and a modified dpGreen-based binary vector, which contained a 35S:PM-mCherry selection marker using Golden Gate Cloning strategy [52]. The construct was introduced to fer-4 plants through agrobacterium-based transformation using floral dip method. Plants from two independent GFP-expressing homozygous T3 lines were used for this study.

METHOD DETAILS

Plant growth conditions

Surface-sterilized Arabidopsis thaliana (Col-0) seeds were plated on 0.7% gellan gum (Gelrite, Sigma G1910) containing 4.33 g/L Murashige and Skoog salts (Caisson MSP01-50LT) and 1% sucrose, pH 5.7, stratified for 2 d at 4°C and grown in a Percival CU41L4 incubator at a constant temperature of 22°C with long-day conditions (16 h light and 8 h dark). Seedlings were grown for 5 d before being transferred to media supplemented with NaCl or other chemicals. To prepare media for different chemical treatment, chemicals were added to the MS media to their final concentrations before adjusting pH.

For short-term imaging of intracellular cytosolic [Ca2+] levels, surface-sterilized Arabidopsis thaliana (Col-0) seeds were plated on 0.8% Agar (Duchefa, Netherlands) containing 0.5X strength Murashige and Skoog salts (Duchefa, Netherlands), pH 5.8, stratified for 2 d at 4°C and grown in a Percival CU41L4 incubator at a constant temperature of 22°C with long-day conditions (16 h light and 8 h dark) at 120–150 μmol m−2 s−1 light intensity for five days.

Root growth analysis

For end-point analysis of root length, seedlings were grown on standard media for 5 d then transferred to media supplemented with various concentration of NaCl, mannitol or sorbitol, position of root tips marked on the plate and grown for 1 d. Images were captured using a CanonScan 9000F flatbed scanner (Canon). The change in root length was quantified using ImageJ.

For live-imaging analysis of primary root growth, 5-d-old seedlings were transferred to media supplemented with 140 mM NaCl and imaged using a custom macroscopic imaging system as previously described [53]. Growth rates were quantified using a semi-automated image analysis algorithm, RootTip Trace, written as a plugin for ImageJ as previously described [3]. In brief, a line along the midline of the root from the tip to the base was drawn in the last frame of a stack of time-lapse images. The algorithm would calculate the elongation of the root along the line in each frame. Growth measurements made during 1 h were summed together to calculate hourly growth rate.

Fluorescence microscopy

Imaging chambers were made from a cut-open 60 × 15-mm Petri dish (Fisher, AS4052) with a 22 × 40 (mm) No. 1.5 micro cover glasses (VWR, 48393-172) taped to the bottom. Seedlings were transferred to the center of the chamber with the roots sandwiched between the cover glass and a 0.5-cm-thick slab of gelrite media with appropriate supplements. Water-soaked kimwipes were places around the edge of the chamber to maintain humidity and the chamber was sealed with Parafilm.

Bright-field images were taken from a Leica DMI6000 inverted compound microscope with HC PL APO 10× / 0.4 CS objective (Leica) and LAS software (Leica).

For long-term 4D confocal imaging, roots of LTI6-YFP-expressing seedlings were imaged on a Leica TCS SP8 microscope with an HC PL APO 20× / 0.75 n.a. objective. A super-continuum white light laser was tuned to 515 nm for excitation. Emission was detected from 525 to 575 nm by a HyD SMD detector. Imaging data were acquired using LAS-X software (Leica).

For long-term imaging of intracellular cytosolic [Ca2+], roots of YC3.6-expressing seedlings were imaged on an inverted epifluorescence microscope (Leica DM IRE2). The objective used was an HC PL APO 10× / 0.4 n.a. (Leica). A Sutter Lambda illumination system was used for excitation. CFP/FRET was excited at 102 μW for 200 ms using a xenon lamp and a Chroma ET 436/20 emission filter. Control YFP was excited at 149 μW for 100 ms using a Chroma ET 500/20 emission filter. CFP and YFP emissions were detected through ET470/22 and ET550/49 filters respectively. Imaging data were acquired using SlideBook software (Intelligent Imaging Innovations).

Short-term imaging of intracellular cytosolic [Ca2+] levels was performed in a custom-made flow-through chamber as previously described [54]. For NaCl or mannitol treatment, the seedlings were washed with incubation buffer (5 mM KCl, 50 μM CaCl2, 10 mM MES, pH 5.8 adjusted with TRIS) for 3 min by continuous circulation (1.5 mL min−1) by a peristaltic pump. Then the incubation buffer was exchanged with buffer (5 mM KCl, 50 μM CaCl2, 10 mM MES, pH 5.8 adjusted with TRIS) supplemented with 150 mM NaCl or 300 mM mannitol. In this setup, buffer exchange in the flow-through chamber occurs at 54 s after valve activation for initiation of buffer exchange. After 7 min treatment, seedlings were washed with incubation buffer. An inverted ZEISS Axio observer (Carl Zeiss Microimaging, Goettingen, Germany) equipped with an emission filter wheel (LUDL Electronic Products, Hawthorne, NY, USA) and a Photometrics cool SNAPHQ2 CCD camera (Tucson, Arizona) with a Zeiss EC Plan-NEOFLUAR 5×/0.16 dry objective was used for imaging. A xenon short arc reflector lamp (Hamamatsu) through a 440 nm filter was used to excite the CFP of YC3.6. For emission detection, 485 nm (CFP) and 535 nm (YFP) emission filters were used. CFP and YFP (FRET) Images were obtained with Metafluor software (Meta Imaging series 7.7 from Molecular devices). An exposure time of 300 ms was used for both CFP and YFP (FRET) image acquisition with binning 2.

MorphographX analysis

MorphographX analysis is performed according to the user manual [26] with some modifications. LTI6-YFP image stacks were filtered using Gaussian Blur with a radius of 0.3 μm. “Edge Detect” was used to extract the root shape with a threshold between 300 and 600 followed by “Fill Holes” to follow the raw signal as closely as possible. A mesh was generated using the “Marching Cubes Surface” with a cube size of 5 μm. The mesh was trimmed at the bottom and subsequently smoothed 5 times by “Smooth Mesh” and refined 2 times by performing the “Subdivide the Mesh” function. The LTI6-YFP signal was projected onto the mesh. Root epidermal cells were identified by manual seeding using “Add New Seed” tool and the mesh was segmented using a “Watershed Segmentation” algorithm. Same cells in a time series imaging were associated and labeled with the same color using “Parent labeling.”

Quantification of cell bursting and volume change

4D confocal imaging data of roots expressing LTI6-YFP were 3D reconstructed and segmented to hourly movies using ImageJ. Events that cells loss shape and YFP signal were counted manually in the root epidermal and cortical layers. For comparing cell death between WT, fer, and sos mutants, single plane LTI6-YFP images were aligned to the root tip using the “StackReg” plugin and reoriented using the “Rotate” function. A 750 μm region of the root tip was divided into 10 sections for analysis. Single cell death events were counted as a loss in LTI6-YFP signal. Due to the large number of cells in the lateral root cap and meristematic region, a total number of cell in the first three sections from the tip were estimated to be 25, 60, 60. For cell volume measurement, 4D confocal image series of roots expressing LTI6-YFP were pre-processed (rolling ball background subtraction 50, Gaussian blur 1–2) in ImageJ. Cell segmentation and volumetric analysis was performed in the Cell module of Imaris (Bitplane). For counting cell bursting events using bright-field images, events that cause root cells to rapidly lose volume or appear opaque were quantified as a cell death event.

Atomic force microscopy

The AFM data were collected as previously described [31, 55]. Roots were immobilized on glass slides and surrounded by stiff agarose after dissection. Approximate elongation zone was defined based on the visual landmark observed through a bright field microscope. To measure wall properties at normal turgor pressure, the roots were immersed in water with or without salts (140 mM NaCl). In order to extract the mechanical properties of only the outer cell wall, the maximum indentation force was set for a maximum indentation of 60 nM. The following cantilever was used: “Nano World” (Nanosensors Headquarters, Neuchâtel, Switzerland) SD-R150-T3L450B tips with a spring constant of 0.15–1.83N/m (the one used was estimated to be 0.85 N/m) with silicon point probe tips of a 150-nm radius.

All force spectroscopy experiments were performed as previously described [31, 55]. Briefly, stiffness of samples was determined as follows: an AFM cantilever loaded with a spherical tip was used to indent the sample over a 40 × 100 μm square area, within the area 20 × 50 measurements were made resulting in 1000 force-indentation experiments; each force-indentation experiment was treated with a Hertzian indentation model to extract the apparent Young’s modulus (EA); each pixel in a stiffness map represents the apparent Young’s modulus from one force-indentation point. The EA was calculated using the JPK Data Processing software (ver. Spm - 4.0.23, JPK Instruments AG, Germany), which allows for a more standardized analysis than the estimation of the EA using a standard Hertzian contact model [31, 55]. Only the retraction curve was used in our analyses as is typically the case in nano-indentation experiments. Similar stiffness differences were also observed when using the deformation curve (data not shown), yet this measurement appeared more sensitive to the topography leading to aberrant measurements in particular on growing root hairs. The best fit was obtained using a Hertzian model with 50 nm as tip radius. A Poisson ratio of 0.5 was assumed for the material. Range distribution of EA from 4 kPas to 100 kPas in 1-kPa binned groups was calculated using MATLAB.

Recombinant protein production and purification

His6-HA- or MBP-tagged FER extracellular domain (FERecd), MALA or MALB domains were expressed either in insect cell (Tn5, Invitrogen) or E. coli (BL21) based on standard procedure provided by Invitrogen. For protein isolation, insect cell medium was buffer-exchanged and concentrated into Ni2+ binding buffer (50 mM sodium phosphate buffer pH 7.2, 250 mM NaCl with 40 mM imidazole). The samples were loaded onto Ni2+-Sepharose FF column (GE Healthcare) and eluted with an imidazole gradient of 40–400 mM imidazole. His6-HA-FERecd-His6 protein, referred to as HA-FERecd, eluted at approximately 300 mM imidazole as a single Coomassie Blue-stained band with a faint smear after SDS-PAGE, most probably due to glycosylation. It was then buffer-exchanged into 25 mM Tris-HCl pH 7.5, 100 mM NaCl and concentrated to 0.5 mg/mL and stored at 4°C for use in pectin-binding experiments within two to three days. Antibodies against the HA tag were used to detect bound HA-FERecd.

The pMAL-cx5 vector (New England Biolab) was used to express in BL21 E. coli the following proteins: MBP, MBP-FERecd, MBP-MALA, MBP-MALB, MBP-ANX1ecd, MBP-ANX2ecd, MBP-BUPS1ecd and MBP-WAK1ecd. They were purified by affinity chromatography on amylose resin by standard methods as previously described [20]. Briefly, cells were resuspended in binding buffer [40 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1 mM Na2-EDTA] and sonicated with six 10 s pulses at 1-min intervals. After removal of cell debris by centrifugation at 4°C, proteins were applied to amylose resin (NEB) in the same binding buffer. GFP-His6 was purified by elution from Talon metal affinity chromatography (Clontech) by imidazole. His6-HA-BAK1ecd was affinity-purified from Talon resin.

Analysis of protein interaction with pectin

Pectin interaction with FERecd, MALA and MALB were examined by five different assay methods to detect FER-pectin interactions: (1) in-gel affinity electrophoresis, (2) protein-pectin interaction followed by electrophoretic mobility shift assay, (3) two dot-blot assays one with immobilized proteins and one with immobilized pectin, (4) a bead based pull-down assay with protein immobilized on amylose beads, and (5) ELISA. Interactions were based on immunodetection using either monoclonal antibodies, JIM5 or M38 (CCRC or gift from Paul Knox), against pectin [20, 49], or anti-HA, anti-His6 or anti-MBL antibodies against the protein tags. All procedures were carried out at room temperature unless otherwise indicated.

Affinity electrophoresis assays were modeled after previous protein-carbohydrate interaction studies [56]. Briefly, HA-FERecd (2 mg) were electrophoresed in one well of individual 6% native polyacrylamide gels (26:1, acrylamide: bisacrylamide) in a 30 mM MES/30 mM histidine (His) buffer, pH 6.1, each polymerized with different concentrations (0 to 0.5 mg/mL) of PGA (Sigma). Similarly electrophoresed BSA (5 mg) served as controls. Gels run with BSA were stained directly with Coomassie blue. Gels run with HA-FERecd were first soaked in 25 mM Tris-HCl pH 7.5 and 0.1% SDS for 30 min and then transferred to PVDF for immune-detection using mouse anti-HA antibody (Santa Cruz; 1: 5000), followed by incubation with alkaline phosphatase-conjugated goat anti-mouse antibody (Santa Cruz; 1:5000), and developed in alkaline phosphatase developing buffer (20 mM Tris-HCl pH 9.5, 100 mM NaCl, 10 mM MgCl2) in the presence of NBT and BCIP.

Electrophoretic mobility shift assays were modeled after standard gel-shift assays for protein-nucleic acid interaction on native PAGE. HA-FERecd or control BSA were incubated with increasing concentrations of PGA for 15 min prior to gel loading. Running buffer, electrophoretic conditions, protein transfer and detection were the same as those described above for affinity electrophoresis.

For ELISA, 4 μg of HA-FERecd or control proteins were coated on 96-well microtiter plates in coating buffer (25 mM Tris-HCl pH 8, 60 mM NaCl) for 2 h at room temperature. The coated plates were washed in TBST (25 mM Tris-HCl pH 7.5, 100 mM NaCl, 0.05% Tween20), blocked with 2% BSA in TBST for 2 h, and washed in binding buffer (25 mM Bis-Tris-HCl pH 6.5, 100 mM NaCl) before incubation with 0, 1, 2, or 5 mg/mL PGA in binding buffer for 2 h. The plates were washed and then probed for PGA with JIM5 antibody (at 1:1000 dilution) overnight at 4°C. After washing, the plates were probed with (HRP)-conjugated goat anti-rat secondary antibody (Santa Cruz) at 1:5000 dilution for 1 hr, washed and developed with the colorimetric peroxidase substrate o-phenylenediamine dihydrochloride (OPD) in 50 mM citric acid, 50 mM sodium phosphate, pH 5, for 20 min. Absorbance at A450 (A450) was measured on a SpectraMax M5 Multi-Mode Microplate Reader (Molecular Devices) to quantify protein-PGA interaction.

Dot blot assays were either based on proteins (0.25 μg of target HA-FERecd, or controls proteins, GFP-His6, or MBP) immobilized onto PVDF membranes (Perkin-Elmer) or heat-fragmented PGA immobilized onto nitrocellulose membrane (BA85 S&S Protran). Membranes with immobilized proteins were blocked in 2% BSA in TBS (25 mM Tris-HCl pH 7.5, 100 mM NaCl) for 1 h and were then washed in binding buffer (25 mM Bis-Tris-HCl pH 6.5, 100 mM NaCl). They were then incubated with 0, 0.1, 0.5, and 1 mg/mL PGA in binding buffer for 1.5 h, followed by three washes in binding buffer, and three washes in TBST for at least 15 min each. The blots were then incubated with JIM5 antibody (1:1000 dilution), then with horseradish peroxidase (HRP)-conjugated goat anti-rat secondary antibody (Santa Cruz; 1:5000 dilution), with five washes in TBST between each step. Bound PGA was detected by chemiluminescence in standard developing buffer (100 mM Tris-HCl pH 8.5, 85 mM p-coumaric acid, 250 mM luminol, 0.0034% H2O2).

For dot blot assays with immobilized pectin, dots of 1.5 μL of PGA at varying concentrations of PGA serially diluted from 20 mg/mL stocks were spotted onto the membrane and allowed to dry. Quantity of PGA in each dot is as indicated in the figures. The pectin dot blots were blocked, incubated with purified recombinant MBP or fusion proteins at 4°C for 2 h (at concentrations indicated in figures or figure legends), and then with anti-MBP antibody (NEB) for 1 h with three to five washes in between. After rinsing, the filters were either incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (Santa Cruz) or for 1 h for colorimetric detection in alkaline phosphatase developing buffer or with anti-MBP-HRP (1:10000 dilution, 1 mg/mL), followed by chemiluminescence detection of bound proteins. Signal quantification was carried out using ImageJ. Pectin immobilized on nitrocellulose membrane was used for binding assays with all other proteins tested. For these dot blot assays, every binding was carried out in triplicate samples, ad each binding assay was repeated at least three times, using independently prepared protein preparations.

For assays with protein immobilized on amylose resin, MALA-MBP or MBP were coated on resin as for affinity chromatography. After washing, the protein-coated resins were incubated with PGA (2 mg/mL, sonicated) for 2 h or overnight. After wash, the resins were incubated with JIM5 (1:500) or M38 (1:100) antibody, followed by FITC-labeled secondary antibody to detect bound pectin microscopically by immunofluorescence. The input protein on the coated beads was detected by anti-MBP antibody (1:500).

[Ca2+] image processing

For the calculation of the change in [Ca2+] concentration over time, the FRET/CFP ratio images were recalculated from the CFP and FRET images with ImageJ software. Normalized change in ratio values (Rt-R0): R0 (abbreviated as ΔR: R0) was calculated, where Rt is the ratio at time point t and R0 is the basal ratio value. A [Ca2+] spike is defined as an event with local [Ca2+] increase of ΔR: R0 > 0.2. A size minimum of 3 pixels was used to exclude sporadic signals.

Kinematic analysis of growth rate

Kinematic analysis was performed using the control YFP images taken at the same time as the FRET and CFP images used for calculating changes in [Ca2+] concentration. Time series of images were opened in ImageJ and aligned to the root tip using the “StackReg” Plugin and reoriented using the “Rotate” function so that the roots were vertical. The tip 600 μm (360 pixels) region was selected for analysis. To calculate elemental growth rate, a kymograph was created by drawing a 1 × 360 pixel box near the midline of the root and performing the “Montage” function with the whole image stack. The kymographs were subsequently gridded into 30 × 12 rectangles where each column corresponded to a 30-min time window and each row representing a 50-μm size region of the root tip. Points were marked where cell boundary lines intercept with the vertical grid lines in each rectangle and elemental growth rate was calculated by dividing the difference in Y axes with time and zone size: (y2-y1) μm /0.5hr * 50 μm.

QUANTIFICATION AND STATISTICAL ANALYSIS

For root growth analysis, all experiment were performed at least three times. Sample size, n, for each experiment are given in the figure legends. For AFM, microscopy and immunolabeling, all experiments were performed at least three times using independent biological samples with similar results being observed. Statistical comparisons (Two-way ANOVA or Student’s t test) were conducted with R software.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
JIM5	CCRC [20] and Dr. Paul Knox [49]	JIM5
M38	CCRC [20] and Dr. Paul Knox [49]	M38
mouse anti-HA antibody	Santa Cruz Biotechnology	SC-7392
alkaline phosphatase-conjugated goat anti-mouse antibody	Santa Cruz Biotechnology	SC-2047
HRP-conjugated goat anti-rat secondary antibody	Santa Cruz Biotechnology	SC-2032
anti-MBP antibody	NEB	E8032S
Chemicals, Peptides, and Recombinant Proteins
Gelrite	Sigma	G1910
MS	Caisson	MSP01-50L
His6-HA-FERecd-His6	This study	N/A
His6-HA-MALA-His6	This study	N/A
His6-HA-MALB-His6	This study	N/A
His6-HA-BAK1	This study	N/A
MBP-FERecd	This study	N/A
MBP-MALA	This study	N/A
MBP-MALB	This study	N/A
MBP-ANX1ecd	This study	N/A
MBP-ANX2ecd	This study	N/A
MBP-BUPS1ecd	This study	N/A
MBP-WAK1ecd	This study	N/A
Experimental Models: Organisms/Strains
fer-4	ABRC	GABI_106A06 [20]
llg1-1	ABRC	CS66105 [30]
llg1-2	ABRC	CS66106 [30]
fer-2	Dr. Paul Larsen [21]	N/A
xxt1	ABRC	CS67834 [36]
xxt2	ABRC	CS67835 [36]
xxt1,2	ABRC	CS16349 [36]
mur1-1	ABRC	CS6243 [50]
mur1-2	ABRC	CS6244 [50]
sos1-1	[29]	N/A
sos2-2	[51]	N/A
RALF1 RNAi in Col-0	Dr. Daniel Moura [24]	N/A
35S::RALF1 in Col-0	Dr. Daniel Moura [24]	N/A
35S::YC3.6 in Col-0	Dr. Gabriele Monshausen [17]	N/A
35S::YC3.6 in fer-4	Dr. Gabriele Monshausen [17], this study	N/A
35S::LTI6-YFP in Col-0	Dr. David Ehrhardt [25]	N/A
35S::LTI6-YFP in fer-4	This study	N/A
35S::LTI6-YFP in sos1-1	This study	N/A
35S::LTI6-YFP in sos2-2	This study	N/A
pFER::FER-GFP in fer-4	This study	N/A
Oligonucleotides
Primer pFER-GG-F: GCTCTTCGAGTAGAAAAGTTAAGAGTGGGAACTGGGAAG	This study	N/A
Primer pFER-GG-R: GCTCTTCGCAGCGATCAAGAGCACTTCTCCGGGA	This study	N/A
Primer gFER-GG-F: GCTCTTCACTGATGAAGATCACAGAGGGACGATT	This study	N/A
Primer gFER-GG-R: GCTCTTCACGCACGTCCCTTTGGATTCATGATC	This study	N/A
Recombinant DNA
pMCY2	[52]	N/A
pFER::FER-GFP	This study	N/A
pMAL-cx5	New England Biolab	N8108S
Software and Algorithms
ImageJ	https://imagej.nih.gov/ij/download.html	N/A
LAS software	Leica	N/A
LAS-X software	Leica	N/A
SlideBook software	Intelligent Imaging Innovations	N/A
Metafluor software	Molecular devices	N/A
MorphographX	http://www.morphographx.org	N/A
Imaris	Bitplane	N/A
JPK Data Processing software	JPK Instruments AG	N/A
MATLAB	https://www.mathworks.com/products/matlab.html	N/A