Introduction

Plants perceive various external signals, but how signals interact with the extracellular matrix is poorly understood. Pectin is a heteropolysaccharide, primarily composed of galacturonic acid units, contributing to the mechanical strength and integrity of the cell wall. Pectin plays a vital role in growth control with an emerging role in signalling (Wolf, 2022). In the cell wall, PECTIN METHYL ESTERASEs (PMEs) catalyse the removal of methyl groups from the initially esterified pectin, giving rise to negatively charged carboxylic acid groups and altering the interaction of pectin with various extracellular components (Peaucelle et al., 2008). Here we address how the methylation status of pectin in the extracellular matrix contributes to the integration of RAPID ALKALINIZATION FACTORs (RALFs), which are extracellular cysteine-rich plant peptide hormones. RALF peptides are involved in multiple physiological and developmental processes, ranging from organ growth and pollen tube guidance to modulation of immune responses (Stegmann et al., 2017; Abarca et al., 2021). Catharanthus roseus receptor-like kinase 1-like (CrRLK1L) are transmembrane proteins with an extracellular domain, consisting of two adjacent malectin-like domains, for RALF signal perception and a cytoplasmic kinase domain for intracellular signal transduction (Escobar-Restrepo et al, 2007). FERONIA (FER) is currently the best characterized CrRLK1L, contributing to the above-mentioned developmental programs as well as the integration of environmental cues (Feng et al., 2018; Gigli-Bisceglia et al., 2022). On a cellular level, RALF binding to FER leads to rapid apoplastic (extracellular) alkalinisation (Haruta et al., 2014), which modulates cell wall properties, ion fluxes, and enzymatic activities, thereby influencing cellular expansion rates (Haruta et al., 2014; Barbez et al., 2017; Dünser et al., 2019). Besides its role in apoplastic pH, CrRLK1Ls play a role in cell wall sensing (Hematy et al., 2007; Höfte, 2015; Shih et al., 2014) and can bind to extracellular pectin (Feng et al., 2018; Lin et al., 2022). The FER-dependent cell wall sensing mechanism involves the direct interaction with the extracellular LEUCINE-RICH REPEAT EXTENSINs (LRXs) in roots (Dünser et al., 2019), suggesting that LRXs provide a physical link between FER and the cell wall (Herger et al., 2019). On the other hand, root, as well as pollen-expressed LRX proteins, also bind to RALF peptides (Mecchia et al., 2017; Dünser et al., 2019; Moussu et al., 2020), but its contribution to mechanochemical sensing and/or growth control is largely unknown in roots. Here we show that the PME-dependent demethylation status of pectin defines the FER-dependent perception of RALF1, contributing to extracellular regulation of pH, cell wall and plasma membrane remodelling, control of receptor endocytosis as well as organ growth in roots. Our data suggests that negatively charged, demethylated pectin binds to positively charged RALF1 peptides with high avidity. Interference with the demethylation of pectin, out-titrating RALF1 with small charged, demethylated pectin fragments, as well as the disruption of the positive charges in RALF1, abolishes the RALF1 activity in roots. We accordingly propose that the RALF interaction with demethylated pectin is crucial for its FER-dependent perception in roots. Even though root-expressed LRX proteins are strictly required for the FER-dependent mechano-sensing in roots (Dünser et al., 2019), we show here that LRX proteins are not essential for the integration of RALF-pectin signalling in roots. In conclusion, we report on the crucial contribution of pectin demethylation, for FER-dependent RALF peptide hormone signalling in root growth control. Our work proposes that pectin acts as an extracellular signalling scaffold, contributing to signalling dynamics at the plasma membrane.

Results and discussion

The application of RALF1 peptides induces strong repression of main root growth in Arabidopsis thaliana (Haruta et al., 2014; Figure 1A, B), which we initially used here as a bioassay to visualise RALF1 activity. To assess if demethylated pectin may provide some conceptualising information to peptide signalling, we pharmacologically interfered with the demethylation status of pectin, using epigallocatechin gallate (EGCG). EGCG is a natural inhibitor of PME activity (Lewis et al., 2008) and its application lowered main root growth but its co-treatment markedly interfered with the RALF1-induced reduction in root growth (Figure 1A, B). In contrast, the structurally similar epicatechin (EC) is not an inhibitor of PMEs (Lewis et al., 2008) and did not block RALF-induced root growth repression (Figure 1C, D). In agreement, EGCG but not EC application reduced the labelling of de-methyl esterified pectin in roots (Figure 1E, F) as visualised by the fluorescent probe COS⁴⁸⁸ (Mravec et al., 2014). This set of data suggests that the EGCG-induced interference with PME activity correlates with reduced RALF1 responses in roots. To consolidate this assumption, we employed the overexpression of PECTIN METHYLESTERASE INHIBITOR (PMEI3), which is well characterised to interfere with PME activity (Peaucelle et al., 2008) and accordingly also reduced the extracellular demethylation of pectin in epidermal root cells (Figure 1G, H). In agreement with the EGCG application, PMEI3, as well as PMEI5, gain-of-function reduced the RALF1-induced root growth repression (Figure 1I, J).

Figure 1.

We accordingly conclude that pharmacological and genetic interference with PME activity leads to the repression of RALF1 effects on root growth. To test the specificity of this effect, we subsequently used the peptide flagellin22 (flg22). The flg22 peptide is derived from the N-terminus of bacterial flagellin and is known to elicit root growth repression via the innate immune receptor FLAGELLIN SENSITIVE2 (FLS2) (Chinchilla et al., 2006). EGCG treatments were not able to suppress, but additively enhanced the FLG22-induced root growth repression (Figure 1K, L), suggesting a rather specific effect of PME inhibition on balancing RALF1 activity.

Next, we addressed how PME activity affects RALF1-dependent root growth control. FER-dependent perception of RALF peptides may affect cell wall integrity and it is hence conceivable that the pharmacological and genetic interference with extracellular PME activity could counterbalance some extracellular effects of RALF1 signalling. Using electron microscopy, we indeed observed a RALF1-induced effect on cell wall integrity, showing swollen cell walls, but also revealed plasma membrane invaginations (Figure 2A, B). These pronounced RALF effects were to our knowledge never reported and hence we confirmed these effects independently, using live cell confocal imaging. RALF1-induced invaginations of plasma membrane markers, such as Low-Temperature inducible protein 6b (LTi6b) (Figure 2C) and Novel Plant SNARE 12 (NPSN12) (SFigure 1A), were readily detectable. Similarly, confocal imaging of propidium iodide (PI)-stained cell walls of wild-type seedlings also depicted the RALF1-induced swelling of the apoplastic signal and revealed additional ectopic accumulations (Figure 2D, F). These RALF1-induced alterations were absent in PI-stained fer-4 loss-of-function mutants (Figure 2E, F), suggesting that FER activation is required for these cellular effects of RALF1. The pharmacological (Figure 2G, H) as well as genetic (Figure 2I-K) interference with PME activity abolished the RALF1-induced alterations in cell wall labelling, phenocopying fer-4 mutants.

Figure 2.

PME activity could hence either indirectly counterbalance RALF1 effects on the cell wall and/or it could directly define RALF1 signalling output in root cells. Accordingly, we next tested the RALF1 impact on apoplastic pH regulation, which is a primary readout of RALF1 perception by FER (Haruta et al., 2014). We used 8-hydroxypyrene-1,3,6-trisulfonic acid trisodium salt (HPTS) as a fluorescent pH indicator for assessing apoplastic pH in epidermal root cells (Barbez et al., 2017). As expected, the application of RALF1 peptides induced an extracellular pH alkalization in wild-type epidermal root cells (Barbez et al., 2017; Figure 3A, B). In contrast, EGCG application (Figure 3A, B) as well as the overexpression of PMEI3 (Figure 3C, D), abolished the RALF-induced alkalization of the apoplastic pH. We accordingly conclude that interference with PME activity disrupts the primary output signalling of RALF1 in root cells.

Figure 3.

Our data suggests that PME inhibition interferes with RALF signalling output, ultimately affecting cell wall properties and root growth. To assess if PME activity affects the extracellular perception of RALF1, we next investigated the ligand-induced cellular dynamics of its receptor FER. An early event of receptor-mediated signalling often involves the ligand-induced endocytosis of the receptor as has been previously demonstrated for the RALF1-receptor FER (Yu et al., 2020). Therefore, we next tested whether the RALF1 ligand-induced endocytosis of FER is altered upon changes in the methylation status of pectin. In agreement with previous reports, we observed the RALF1-induced internalization of functional FER::FER-GFP in epidermal root cells (Figure 3F, G). On the other hand, the ligand-induced internalization of FER-GFP was abolished when we pharmacologically or genetically suppressed PME activity (Figure 3F-J). These findings propose that the PME activity is required for the earliest events of RALF1 perception by FER in roots.

It is hence conceivable that the association of positively charged RALF1 peptides with demethylated, negatively charged pectin is required for its FER-dependent perception. To indirectly assess if RALF1 peptides bind demethylated pectin, we tested if RALF1 peptides bind fragments of demethylated oligogalacturonides (OGs).

We initially assessed potential OG binding to RALF1, using the RALF1 impact on root cells as a bio-assay. The addition of free, demethylated OGs into the media did not visibly affect the cell wall or FER endocytosis on its own (SFigure 2A-D). On the other hand, free OGs in the medium strongly interfered with RALF1-induced cell wall swelling and FER receptor endocytosis in roots (SFigure 2A-D). This effect was concentration-dependent (Figure 4A, B), suggesting that excess of demethylated OGs in the medium can bind and out-titrate RALF1 peptides and thereby limit the RALF1 activity at the root surface.

Figure 4.

To characterise this binding in vitro, we next used biotinylated RALF1 in biolayer interferometry (BLI), which is an optical technique for measuring macromolecular interactions by analyzing interference patterns of white light reflected from the surface of a biosensor tip. The equilibrium dissociation constant (Kd) in the range of 105 nanomolar (nM) (Figure 4C) suggests a physiologically relevant binding affinity. Notably, we however observed an initial OG association to RALF1 with a constant (Kon) in the micromolar (µM) range, but the dissociation kinetics revealed a remarkably slow rate (Figure 4C, D). This is indicative of relatively low affinity at the individual RALF1-OG binding event level but multiple intermolecular binding events may eventually lead to a high accumulative binding strength. We hence propose that RALF1 peptides and demethylated OGs undergo low affinity but high avidity binding assemblies. This finding complements recent data, indicating that the binding of RALF peptides and demethylated OGs leads to phase separation in vitro (Liu et al., 2024).

In pollen, the LRX8-RALF4 complex binds demethylated OGs in a charge-dependent manner (Moussu et al., 2023). In agreement, RALF1 contains in total 8 positively charged amino acids (3 lysine (K) and 5 arginine (R)) (Figure 5A, B). This is reminiscent of the well-characterised demethylated pectin-binding protein POLYGALACTURONASE-INHIBITING PROTEIN (PGIP), which binds to demethylated pectin through a positively charged binding site formed by clustered residues of lysine (K) and arginine (R) (Spadoni et al., 2006). We next assessed if the positive charges in RALF1 are essential for its bioactivity in roots. We synthesized a RALF1-like peptide by substituting the K and R residues (RALF1^-KR), creating a slightly negatively charged peptide in the pH range of the cell wall (Figure 5A, C). RALF1^-KR peptides are not bioactive, because they did neither affect root growth, nor cell wall integrity, nor did they induce the ligand-induced endocytosis of FER in epidermal root cells (Figure 5D-I). These findings suggest that the positively charged residues in RALF1 are essential for its activity in roots.

Figure 5.

The LRX-RALF complex and its interaction with pectin exerts a condensing effect on the cell wall in pollen (Moussu et al., 2023). Moreover, LRX proteins also bind and link FER with a cell wall-sensing mechanism in roots (Dünser et al., 2019). Hence, we next tested if LRX proteins are required for the joint RALF1/pectin-dependent signalling in roots. We observed that RALF1 applications still induced cell wall alterations in lrx1 lrx2 lrx3 lrx4 lrx5 quintuple mutant roots (Figure 6A-D). These findings confirm that LRX proteins are not essential for the RALF activity in roots. Similar to wild-type roots (Figure 6E, F), pharmacological interference with PME activity still repressed RALF1 activity in the lrx1 lrx2 lrx3 lrx4 lrx5 quintuple mutant roots (Figure 6G, H). These findings indicate that the LRX proteins are not required to define the demethylated pectin-dependent signalling output of RALF1 peptides in roots.

Figure 6.

Our data proposes that RALF1 peptides associate with high avidity to demethylated pectin, which is required for the RALF perception by FER. We hence conclude that the methylation status of pectin provides a conceptualising input to RALF peptide hormone signalling.

Concluding remarks

Here we address a function of the cell wall component pectin for integrating RALF1 peptide root growth signals (SFigure 3). Intriguingly, the exogenous application of RALF1 does not only affect cell wall characteristics but also induces prominent plasma membrane protrusions into root epidermal cells. Plasma membrane invagination is a highly controlled process and often relates to plant endosymbiotic events (Su et al., 2023). The developmental importance of this RALF1 effect remains to be addressed but could pinpoint pectin signalling playing a central role in the structural coordination of cell wall and plasma membrane dynamics. We illustrate that the pharmacological and genetic interference with PME activity specifically abolishes RALF1, but not flg22 peptide, activity in roots. PME activity is required for all hallmarks of RALF1 signalling, including the rapid alkalinisation of the cell wall and the ligand-induced endocytosis of FER. We hence propose that the generation of demethylated, negatively charged pectin in the cell wall is required for the FER-dependent perception of positively charged RALF1 at the root cell surface. We show that RALF1 binds to demethylated OGs in vitro and that the application of free OGs can out-titrate RALF1 peptides. Our in vitro data proposes low affinity but a high avidity binding of OGs and RALF1. This interaction mechanism could be key in the spatial enrichment of RALF peptides at the root surface because receptor interactions at the plasma membrane could contribute to the accumulated strength of multiple binding interactions.

After we pre-printed a previous version of this manuscript, Moussu and colleagues proposed that the LRX8-RALF4 complex and its charge-dependent binding of demethylated OGs controls extracellular condensations in pollen tubes (Moussu et al., 2023). The contribution of FER signalling remains unknown in this process. Other very recent findings suggest on the other hand that RALF peptide binding to demethylated pectin is sufficient to lead to its extracellular phase separation, subsequently inducing the clustering of FER receptors in the plasma membrane (Liu et al., 2024). It remained however unknown if LRX proteins are implied in this response. LRX proteins directly interact with FER and thereby contribute to cell wall sensing in roots (Dünser et al., 2019), but we show here that the root-expressed LRX proteins are not required for the here addressed pectin/RALF signalling in roots. This finding reflects on possibly distinct FER-dependent and RALF-dependent roles of LRX proteins. Additionally, distinct modes of cell wall integrity control may be in place in root cells and tip-growing pollen tubes.

Based on our findings, we envision that negatively charged, demethylated pectin could function as a signalling scaffold for positively charged RALF peptides, which seems crucial for its signalling via the FER receptor. This could also lead to self-emerging feedback properties because high extracellular PME activity is expected to strongly acidify the apoplast (Wolf et al., 2009), which in turn would be counteracted by the demethylated pectin-induced activation of RALF peptides and its consequences on apoplast alkalinisation (Haruta et al., 2014; Barbez et al., 2017; this study). The here uncovered mechanism links the pectin-related cell wall status with RALF peptide hormone signalling, providing a conceptualising extracellular signalling scaffold.

Material and methods

Plant material and growth conditions

All experiments were carried out in Arabidopsis thaliana, ecotype Col-0. The following plant lines were described in previous publications: PMEI3-OE (Peaucelle et al., 2008), PMEI5-OE (Wolf et al., 2014), lrx1/lrx2/lrx3/lrx4/lrx5 (Dünser et al., 2019), LTI6b-GFP (Cutler et al., 200), UBQ10::NPSN12-YFP (Wave131Y, Geldner et al., 2009), FER::FER-GFP (in fer-4, Shih et al., 2014), fer-4 (Shih et al., 2014), FER::FER-GFP x PMEI3-OE (in fer-4, obtained by crossing). After surface sterilization in 70% and 100% ethanol, seeds were vernalized at 4°C for 2 days in darkness, afterwards grown vertically on ½ strength Murashige and Skoog (MS) medium plates containing 1% sucrose in a long-day regime (16 h light and 8 h darkness) at 21 °C.

Chemicals and peptides

All chemicals were dissolved in dimethyl sulfoxide (DMSO) and served as solvent control, unless indicated otherwise. We used Propidium iodide (PI) in a working concentration of 0.02 mg/mL for cell wall counterstaining (obtained from Sigma (MO, USA)). Epigallocatechin gallate (EGCG) was used to interfere with PME activity in liquid treatments (obtained from Cayman Chemical (MI, USA)), epicatechin (EC) was used as negative control (from Cayman Chemical (MI, USA)) and flg22 peptides as a peptide control (obtained from Sigma (MO, USA)). The RALF1 peptide (mature RALF1, with the amino acid sequence: ATTKYISYQSLKRNSVPCSRRGASYYNCQNGAQANPYSRGCSKIARCRS) and the non-charged RALF1^-KR peptide version (amino acid sequence: ATTSYISYQSLTSNSVPCSDTGASYYNCQNGAQANPYSDGCSYIASCRS) were both synthesized by PSL GmbH (Heidelberg, Germany) and dissolved in water. The galacturonan oligosaccharides OG^10-15 and OG^25-50 were obtained from Elicityl (Crolles, France) and Biosynth (Staad, Switzerland), respectively.

Confocal microscopy

For image acquisition, an upright Leica TCS SP8 FALCON FLIM confocal laser scanning microscope, equipped with a Leica HC PL APO Corr 63 x 1.20 water immersion CS2 objective, was used. GFP was excited at 488 nm (fluorescence emission: 500 - 550 nm), PI at 561 nm (fluorescence emission: 640 - 750 nm). Experiments assessing apoplastic pH were carried out as previously described (Barbez et al., 2017) using 8-hydroyypyrene-1,3,6-trisulfonuc acid trisodium salt (HPTS; Sigma-Aldrich). In brief, roots of 6-day-old seedlings were treated for 3 h in liquid ½ MS medium containing 1 µM RALF1 and/or 50 µM EGCG or the appropriate amount of solvent control (water or DMSO, respectively) for 3 h. If not mentioned otherwise, pharmacological treatments have been performed in liquid medium in multiwell plates. Prior to imaging, the roots were mounted on a solid block of ½ MS medium containing 1 mM HPTS. The block was transferred to a microscopy slide and imaged with a coverslip on top. Image processing was performed as previously described (Barbez et al., 2017) using a macro for Fiji. Four transversal cell walls were quantified per root. The protonated form of HPTS was excited at 405 nm and 0.5% Diode UV laser intensity, and the deprotonated form at 455 nm with 30% Argon laser intensity. Images here were obtained in sequential scan mode. Z-stacks were recorded with a step size of 420 nm, with a stack containing 20 slices on average. The Gain of the HyD detectors (or PMT for HPTS experiments) stayed constant during imaging. The pinhole was set to 111.4 µm (for HPTS experiments to 196 µm). To counterstain the cell walls, roots were mounted in PI solution (0.02 mg/mL) prior to imaging. The signal intensity of intracellular space or plasma membrane was assessed using Fiji, with four cells analyzed per root. The ROI stayed constant over the experiments and biological replicates. For quantification of the PI staining width (and therefore the thickness of the cell wall) the thickness of transversal and longitudinal sections stained by PI were measured using Fiji, with four to six stained plasma membranes quantified per root. For the RALF1-treated roots, only the invaginated regions were chosen for analysis.

Synthesis of COS probes

The COS probe was generated as described in Mravec et al., 2014, using chitosan oligosaccharides conjugated with AlexaFluor 488 hydroxylamine (ThermoFisher, Waltham, Massachusetts, USA). In brief, 1 mg/ml oligosaccharide solution was mixed into 0.1 M sodium acetate buffer, pH 4.9. The AlexaFluor 488 (10 mg/ml in DMSO) was added to the oligosaccharide (½ of the moles of the oligosaccharide). After incubation at 37°C for 48 h with shaking at 1,400 rpm the probe was ready. 6-day-old seedlings were transferred into liquid ½ MS medium and supplemented with DMSO, 50 µM EGCG and EC for 3 hours respectively. After the treatment, seedlings were staining in liquid ½ MS medium supplemented with COS488 at a 1:500 dilution, for 30 min, and washed twice with medium. Imaging of root tips was performed directly afterwards.

Root length analysis and root growth assay

For root growth assays, seedlings were grown for 3 days on solid ½ MS plates. Afterwards, they were transferred into 3 mL of liquid ½ MS medium containing RALF1 and/or EGCG or the appropriate amount of solvent control (water or DMSO, respectively). After 3 days of growth in liquid media with gentle agitation on a platform shaker, the seedlings were placed on solid ½ MS plates and scanned using a flat bed scanner. Root length was measured using Fiji. For statistical analyses, the GraphPad Prism Software (version 9.5.1) was used.

Sample preparation for electron microscopy

6-day-old wild-type seedlings were treated for 3 h in liquid ½ MS media containing solvent control or 1 µM RALF1. After the treatment, roots were cut apart with a razor blade, immediately submerged in MTSB buffer containing 4% p-formaldehyde (PFA) and vacuum infiltrated for 15 minutes in a microwave oven (Pelco BioWave Pro+) at room temperature. The submerged roots were left in the fixative solution for 4 h at room temperature and overnight at 4°C. All further steps were performed according to Dünser et al. (2022).

In silico analysis

For assessment of the charge of the RALF1 and RALF1-KR peptide, the Peptide Calculator Online tool from https://www.biosynth.com/peptide-calculator was used.

Biolayer interferometry assays

Octet@RED 96 was used to perform binding assay between Biotinylated-RALF1 and OG 25-50. Biotinylated-RALF1 (purchased from Peptide Specialty Laboratory GmbH) was solubilized in DMSO as a stock solution in a concentration of 11 mg/ml. The buffer used for the interaction was PBS pH 7.4 containing 25 µg/ml BSA. Test experiments were carried out to rule out the unspecific binding of OG25-50 to streptavidin biosensor and Biotin. Streptavidin biosensor was first dipped in the interaction buffer for 600 s continued by loading with 5 µg/ml Biotinylated-RALF1 as ligand for 600 s. The sensors were washed in the interaction buffer for 600 s followed OG 25-50 in the concentration of 0, 0.5, 1, 2, 3, 4 and 5 µM for association (600 s). Finally, dissociation was monitored for 600 s in the interaction buffer. Three biological replicates were measured and analyzed using Octet Red Data Analysis software version 8.0.2.3. The fitting curve was selected to be 1:1 binding model. The curves were plotted using GraphPad Prism6 software.

Acknowledgements

We are grateful to Niko Geldner, Gabriele Monshausen, Grégory Mouille, and Jozef Mravec for sharing published material; our team members for helpful discussions; and the LIC Imaging Center Freiburg for expertise and support. We would like to thank Rosula Hinnenberg of the EM facility at the Faculty of Biology, University of Freiburg, for her assistance with the generation of EM data. The TEM (Hitachi HT7800) was funded by the DFG grant (project number 426849454). This work was supported by the Austrian Science Fund (FWF) (P33044 to J.K-V.,), and the German Science fund (DFG; 470007283 to J.K.-V. and CIBSS – EXC-2189 Project ID 390939984 to J.K.-V. and to E.B.).