Clathrin-Mediated Endocytosis Regulates Root Endodermal Suberization via ROS

Summary

• Endomembrane trafficking plays a crucial role in plant adaptation to environmental stresses, yet its involvement in endodermal root suberization remains poorly understood.

• Here, we show that disruption of clathrin-mediated endocytosis (CME) or canonical exocytosis led to an ectopic suberin deposition in the Arabidopsis root endodermis towards the root tip. Genetic disruption of endocytosis phenocopied the effects of the CME inhibitor ES9–17, while genetic disruption of clathrin-independent endocytosis led to reduced suberization, suggesting distinct, pathway-specific roles in regulating suberin deposition.

• Ectopic suberization upon CME inhibition required the CIFs-SGN3-SGN1-RBOHF/D signaling axis, independent of ABA. Notably, CME disruption led to accumulation of RBOHF in the plasma membrane, driving NADPH oxidase-dependent H₂O₂ accumulation in endodermis. Scavenging H₂O₂ or inhibiting NADPH oxidases abolished ET disruption-induced suberization, while exogenous H₂O₂ promoted it. Conversely, peroxidase activity inhibition reduced basal suberization but failed to suppress ET disruption-induced enhanced suberization, implicating ROS as a dominant driver.

• Our findings reveal a dual ET regulatory mechanism: exocytosis inhibition leads to suberization independently of known pathways, while CME impairment acts via RBOHF-mediated ROS to increase suberization on the endodermis. This study reveals that ET can control endodermal root suberization in Arabidopsis, linking membrane trafficking to apoplastic barrier formation through reactive oxygen species.

Introduction

Endomembrane trafficking (ET) is a fundamental cellular process conserved in eukaryotes(Zhang et al., 2019). In plants, it is indispensable for the response and adaptation to abiotic stresses, like drought and nutrient starvation(Žárský, 2016; Zhang et al., 2019, 2024). The major ET routes in plant cells include the secretory(exocytic) and endocytic pathways. The first one relies mainly on the exocyst complex, an octameric protein complex essential for the targeting and fusion of secretory vesicles to cell membranes(Elias et al., 2003; Žárský, 2022; De la Concepcion, 2023). Proteins of the EXO70 family and particularly EXO70A isoforms are involved in the canonical exocyst function in polarized exocytosis that is important for polar growth and cell-wall biogenesis(Synek et al., 2006; Žárský et al., 2020). The EXO70A1 isoform is involved in the auxin carrier PIN2 recycling(Drdová et al., 2013), root hair polar growth(Campanoni & Blatt, 2006; Synek et al., 2006) and localization of the CASPs proteins, playing a key role in the formation of the Casparian strip(CS) through lignification in the endodermis(Kalmbach et al., 2017). The endocytic pathway in plant cells can be divided into two pathways based on the need for assembly of a clathrin protein lattice around the forming vesicle for cargo internalization. Clathrin-mediated endocytosis (CME) is the major route for internalization of plasma membrane proteins and molecules from the extracellular environment(Narasimhan et al., 2020; Dahhan & Bednarek, 2022). Clathrin-independent endocytosis (CIE) relies on the Flotillin1 (Flot1) protein associated with sterol- and sphingolipid-enriched membrane microdomains(Žárský, 2016; Khalilova et al., 2023).Both endocytosis processes cooperatively regulate the dynamics of plasma membrane localized proteins and are in balance with the secretory pathway(Zhang et al., 2019; Dahhan & Bednarek, 2022), all under a tight cellular regulation.

Suberin, a major natural lipophilic transcellular barrier, is deposited between the cell wall and the plasma membrane in Arabidopsis root endodermal cells in what is called a suberin lamellae(Shukla & Barberon, 2021; Serra & Geldner, 2022). The suberin layer serves multiple functions, including controlling the uptake and passive diffusion of molecules from the apoplast into the symplast of endodermal cells, as well as acting as a barrier against nematode penetration and pathogen invasion into the vascular tissue(Grünhofer et al., 2021; de Silva et al., 2021; Chen et al., 2022). Due to its properties, suberin has been proposed as an engineering target for plant resilience and due to the durable nature of suberin, as a way to promote soil carbon sequestration, aiding in carbon dioxide removal from the atmosphere(Harman-Ware et al., 2021; Eckardt et al., 2023). Suberin is found in almost all seed plant species, but important progress related with its synthesis, transcriptional regulation and polymerization has been made so far in potato periderm, Arabidopsis roots and more recently in the tomato suberized exodermis(Serra & Geldner, 2022; Cantó-Pastor et al., 2024; Jo et al., 2025). The chemical complexity of the structural suberin lamellae—a glycerol-based heteropolymer composed of a polyaliphatic polyester domain (SPAD) linked to a polyphenolic domain (SPPD)—along with its tissue-specific localization, environmental regulation of biosynthesis, and plant species-dependent variation(Thomas et al., 2007; Grünhofer et al., 2021; Serra & Geldner, 2022; Woolfson et al., 2022), makes it a highly diverse plant biopolymer.

Current hypotheses about suberin polymer assembly outside the plasma membrane suggest the need of transport and secretion of suberin monomers from the cytosol to the apoplast(Serra & Geldner, 2022; Woolfson et al., 2022). However, there is limited evidence on how this transport occurs and whether it involves exocytosis, endocytosis, or both. Several ATP-binding cassette (ABC) transporters, especially from the ABCG family, are proposed to transport suberin monomers (aliphatic precursors, ferulic acid esters, glycerol derivatives) across the plasma membrane(Landgraf et al., 2014; Yadav et al., 2014). Extracellular lipid transfer proteins(LTPs) may also help shuttle hydrophobic suberin precursors through the hydrophilic cell wall toward the site of polymerization(Deeken et al., 2016; Lee & Suh, 2018; Chen et al., 2025). In addition, it is known that lipophilic metabolites and their precursors are transported within the hydrophilic cytoplasm toward the PM(Ichino & Yazaki, 2022). Vesicle-mediated trafficking, which sequesters these lipophilic compounds from the cytosol, is likely involved in transporting these molecules and may represent an additional route for the movement of suberin monomers.

Early discoveries in potato wound healing suberization identified an apoplastic anionic peroxidase with a substrate preference for phenylpropanoid derived monomers(Bernards et al., 1999). Historically the suberin SPPD have been described as “lignin-like” because their chemical and structural deposition in the endodermis resembles the Casparian strip and this has been used as a proxy to predict how the phenylpropanoid derived monomers assembled into the SPPD which together with the SPAD contribute to the formation of the suberin lamellae(Woolfson et al., 2022). Those previous studies also suggested that once the phenolics monomers are in the space between the plasma membrane and the cell wall, they undergo apoplastic peroxidases (PRXs) mediated polymerization, a process that consumes the hydrogen peroxide generated by the combined activities of NADPH oxidases (also referred to as Respiratory Burst Oxidase Homologs, RBOHs) and superoxide dismutases (SODs)(Bernards et al., 1999; Bernards & Razem, 2001; Razem & Bernards, 2002). While this process remains unclear in Arabidopsis roots, however some enzymes with apoplastic location have been directly related with suberin synthesis and degradation(probably via the SPAD modification), for instance some members of the GDSL-lipases(GELPs) family(Ursache et al., 2021). Other enzymes, such as PRX64, have been suggested to participate in a compensatory suberization mechanism arising from an impaired lignification(Yi Chou et al., 2018; Rojas-Murcia et al., 2020).

In this study, we used a combination of pharmacological and genetic approaches to disrupt endomembrane trafficking and found that it strongly affects suberin deposition in the Arabidopsis root endodermis. This was specifically pronounced for CME. We demonstrate that CME inhibition leads to enhanced suberization through a pathway that is independent of ABA but dependent on CIF peptides, SGN signaling components, and the NADPH oxidase RBOHF, an event that resembles the CIFs/SCHENGEN pathway role during the monitoring and completion of the CS establishment. Furthermore, we show that ROS homeostasis in endodermal cells, particularly elevated hydrogen peroxide (H₂O₂) levels, plays a central role in promoting suberin polymerization. Our findings uncover a previously unrecognized link between membrane trafficking and suberin biosynthesis via ROS and position endocytosis as a key regulatory node in the control of root barrier formation.

Materials and Methods

Plant Material and Growth Conditions.

The Arabidopsis thaliana (L.) Heynh., Col-0 reference accession was used for most of the experiments. All mutant lines used in this study have been published previously. Mutant lines flot1(SALK_205125C)(Khalilova et al., 2023), rbohF, rbohD and abi1–1 were obtained from the Arabidopsis Biological Resource Center (https://abrc.osu.edu/). Seeds were surface sterilized with a 30% solution of commercial bleach for 10 min then washed with sterile distilled water under sterile conditions for at least 5 times. Seeds were sowed in square plates containing ½ MS media with MES buffer pH 5.75, 1% sucrose and 0.8% Phytagel. Plates were sealed and placed in the dark at 4°C for 3 days then moved to light. The seeds germinate under long-day conditions (16 h light/8 h dark) in a walk-in growth chamber (Conviron, US) maintained at 21°C, with a light intensity of 120–150 μmol/m²/s and 60% humidity.

Pharmacological Treatments.

List of all chemicals, solvents and working concentrations used in this work are summarized in Table S1. For most of the treatments, 5 days-old plants were transferred to square plates containing ½ MS media with MES buffer pH 5.75, 1% sucrose, 0.8% Phytagel plus the specific chemical(s) according to the experiment, for a period of 24 h. For the tamoxifen(tam)-inducible CHC-HUB1 line, 4 days old plants were transferred for 48 h to plates containing 2 μM tamoxifen (MedChemExpress, US). For fluridone (Milipore-Sigma, US) treatment only, 4 days old plants were transferred for 48 h to plates containing 10 μM fluridone. For fluridone and endomembrane trafficking inhibitors combined treatment, 4 days old plants were pretreated for 24 h in plates containing 10 μM fluridone then transferred for another 24 h to plates with 10 μM fluridone and 30 μM ES-2 or ES9–17. All treated and non-treated plants analyzed for root suberin occupancy were 6 days old.

Confocal laser scanning microscopy and image analysis.

All imaging acquisition experiments were done in the Waitt Advanced Biophotonics Core Facility of the Salk Institute. The Olympus Fluoview FV3000 inverted confocal microscope (Olympus, Japan) with a Multi Area Time Lapse (MATL) module incorporated and a UPlanApo 4x/0.16 dry objective was used for tile (512×512 tile scan size) scanning the whole plant. A galvanometer scanner was used for high-speed imaging. Around 12–15 plants of similar size were mounted each time in distilled water and the Fluorol Yellow (FY) 088(Santa Cruz Biotechnology, US) was excited with a 488 nm laser at 1% intensity and its emission recorded between 520–550 nm. For the imaging of line GPAT5::GFP-NLS, untreated and ES9–17 or ES-2 treated plants were incubated for 5 min at RT in the dark in a 1μg/mL propidium iodide (PI) solution (Thermo Fisher, US). Plants were rinsed with distilled water and imaged. A combination of tile scanning and Z-stack per tile was used to image the initial 4 mm from the root tip using a UPlanApo 20x/0.8 dry objective. GFP was excited at 488 nm and PI at 561 nm, the emission was recorded between 520–550 nm and 630–700 nm, respectively. For the visualization of plants overexpressing RBOHD-GFP and RBOHF-GFP in the endodermis, ES9–17 treated plants for 24 h and non-treated plants were imaged using a UPlanApo 40x/1.5 Sil objective. GFP was excited at 488 nm and its emission was recorded between 500–580 nm. For the quantitative image analysis of GFP fluorescence intensity, regions of interest (ROI) were selected and measured for plasma membrane and the entire cytoplasmic region. The plasma membrane / cytoplasm ratio was calculated by first subtracting the mean fluorescence intensity of the region outside of the root and then dividing the PM mean fluorescence intensity by the mean fluorescence intensity of the entire cytoplasmic region. All image analysis, measurements and intensity quantifications were done using the Fiji software.

Hydrogen peroxide HyPer7 sensor imaging and quantification.

Plants expressing the cytosolic Hyper 7 sensor were treated with ES-2, ES9–17, DPI+ES9–17, SHAM, SHAM+ES9–17 or SHAM+ES-2 for a 24 h period, untreated plants were used as control. Plants were mounted in distilled water and imaged using an Olympus Fluoview FV3000 inverted confocal microscope (Olympus, Japan) and a UPlanApo 40x/1.5 Sil objective. Sequential excitation of the sensor at 405 nm and 488 nm was used and the emission was recorded for both channels between 508–535 nm. The quotient of the mean 488 and 405 channels was used to quantify HyPer7 oxidation in the endodermis. To illustrate root HyPer7 oxidation in roots, the 488 and 405 channels were filtered with a Gaussian Blur of sigma radius set to 1 px and a ratiometric image was produced using the Ratio Plus plugin in Fiji and false colored to visually show the sensor changes.

Suberin Staining and Suberin Occupancy Quantification.

Root suberin staining was performed as previously described(Marhavý & Siddique, 2021). Briefly, 6 days old plants were incubated in the dark with a freshly prepared 0.03% Fluorol Yellow 088 in lactic acid solution for 20 min in a water bath at 70°C, then incubated in distilled water for 1 min at RT. Plants were transferred to a 0.5% aniline blue solution(prepared fresh from a 2.5% aniline blue in 2% acetic acid solution,(Millipore-Sigma, US)) and kept for 20 min at RT in the dark. Plants were transferred to distilled water for 10 min in the dark before imaging. For suberin quantification tiles were stitched into a single image in Fiji using the plugin “Stitching”(Preibisch et al., 2009) and a pairwise stitching process of the individual tile files. Suberin Occupancy was reported as the percentage of the ratio between the root suberized zone length (patchy and fully suberized)/total root length, both lengths measured using the segmented line tool in the Fiji software.

Statistical Analysis.

All data values were expressed as the mean ± SD using the GraphPad Prism 10 (GraphPad Software, US) statistical analysis software. ANOVA followed by a post-hoc Tukey-Kramer test (p<0.05) or two-tailed (p<0.05) unpaired t-test was used to determine statistical differences among treatments and/or genotypes. No power analysis was used to estimate sample size for any of the experiments.

Results

Endomembrane trafficking inhibition induces ectopic suberization in the Arabidopsis root endodermis.

To assess the role of the ET in the suberization process we used two chemical membrane trafficking inhibitors. Endosidin 2(ES-2) that specifically binds and blocks the EXO70A1 isoform(Zhang et al., 2016; Li et al., 2023b), thus inhibiting the canonical exocytosis, and ES9–17 that targets both CHC1 and CHC2 proteins disrupting the CME(Dejonghe et al., 2019), the main cell cargo internalization pathway in Arabidopsis. Col-0 plants treated with both inhibitors showed an increase in suberization towards the root tip, where it never occurred in the wild type, with the highest endodermal suberin occupancy in ES-2 treated plants (Fig. 1a). Endodermal suberin deposition is a well described process in Arabidopsis roots, with a non-suberized zone (root tip up to 2–3 mm), followed by a patchy zone of ongoing suberization with intercalated non-suberized passage cells (between 4 and 7 mm), and a fully suberized zone (after 7 mm)(Barberon et al., 2016; Shukla & Barberon, 2021; De Bellis et al., 2022). Previously it has been shown that plants harboring mutant alleles of the EXO70A1 gene exhibit an increased number of suberized endodermal cells, along with an extended fully suberized root zone and higher suberin occupancy(De Bellis et al., 2022). The lotr2/exo70a1 mutant not only shows enhanced suberization but also extracellular vesiculo-tubular structures associated with suberin deposition in the plasma membrane of endodermal cells(De Bellis et al., 2022). This is consistent with the suberization phenotype of ES-2 treated plants shown here.

Figure 1. Disruption of CME and exocytosis pathways lead to an ectopic root endodermal suberization.

(a) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT plants with 30 μM ES9–17 or ES-2, (right) representative images of 6 days old treated and untreated Col-0 WT plants showing Fluorol Yellow (FY) suberin staining endodermal pattern. Data are the mean ± SD (N = 14–16 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(b) (Left) Boxplot of the lengths from root tip up to first endodermal cell showing endodermal GFP nuclear signal from 30 μM ES9–17 or ES-2 treated vs untreated GPAT5::NLS-GFP plants, (top right) representative median view images of the unsuberized zone from 6 days old treated and untreated GPAT5::NLS-GFP plants showing the endodermal nuclear GFP expression pattern (GFP signal in yellow, propidium iodide staining in cyan hot), scale bars = 100 μm., (bottom right) zoomed images from dotted rectangles on top images showing GFP patterning, ep(epidermis), co(cortex), en(endodermis), scale bars= 50 μm. Data are the mean ± SD (N = 5–7 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (**) p < 0.01, (***) p < 0.001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values.

Next, we wanted to know whether ET disruption induces de novo suberin biosynthesis in the unsuberized zone ( root tip up to 3 mm), where in normal conditions it does not occur(Shukla et al., 2021; De Bellis et al., 2022). To achieve this, we utilized a reporter for the GPAT5 enzyme, which is required for suberin synthesis in roots and is used often as a marker for active endodermal suberization(Beisson et al., 2007; Barberon et al., 2016; Gully et al., 2024).We measured the length from the root tip up to the first endodermal cell showing a nuclear GFP signal in plants expressing GPAT5::NLS-GFP(Shukla et al., 2021). Untreated GPAT5::NLS-GFP plants showed no endodermal nuclear signal in the unsuberized zone (Fig. 1b). This correlates with the well described GPAT5 expression pattern: absent in the unsuberized zone(Beisson et al., 2007) and starting to express in the patchy suberized zone in the Col-0 wild type which is around or after the initial 3mm of root length(Barberon et al., 2016; Shukla et al., 2021; De Bellis et al., 2022). In contrast, ES-2 and ES9–17 treated plants showed endodermal GFP signals in the unsuberized zone at distances within 1 to 3 mm from the root tip (Fig. 1b), which is consistent with the ectopic suberization pattern close to the root tip shown for those chemicals (Fig. 1a). These results support the notion that endomembrane trafficking disruption promotes an enhanced suberization in the root endodermis through active suberin biosynthesis in the otherwise unsuberized zone of the endodermis.

Our data show that chemically blocking CHC proteins, thereby reducing external cargo and PM-associated proteins internalization, can also cause an increase of endodermal suberization (Fig. 1a). We therefore wanted to genetically disrupt the whole CME to mimic the ES9–17 chemical treatment phenotype. To achieve this, we measured root suberin occupancy in the dominant negative tamoxifen (tam) inducible line CHC-HUB1(Larson et al., 2017) (Fig. 2a). The CHC proteins C-terminal HUB domain contains the trimerization domain and the clathrin light-chain (CLC) proteins binding domain involved in the assembly of the clathrin lattice(Kaksonen & Roux, 2018; Das et al., 2021). Upon its tamoxifen induction, the CHC HUB domain becomes ectopically expressed competing with the endogenous unimpaired CHCs for binding with the CLCs, effectively out-titrating them, leading to a complete disruption of CME(Larson et al., 2017). CHC-HUB1 plants that were treated with tam for 48 h showed an endodermal ectopic suberization towards the root tip, resembling the phenotype of ES9–17 treated plants (Fig. 1a and Fig. 2a).

Figure 2. CME disruption but not CIE disruption induced endodermal suberization.

(a) (Left) Boxplot of root endodermal suberin occupancy values from Col-0 WT and mutants defective in the clathrin mediated endocytosis (CME) pathway, (right) representative images of 6 days old Col-0 WT, chc 1–3, chc 2–1 and the tamoxifen(tam) inducible CHC-HUB1 line showing FY suberin staining endodermal pattern. Tamoxifen-inducible CHC-HUB1 plants were treated with 2μM tamoxifen for 48 h. Data are the mean ± SD (N = 10–12 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between Col-0 and genotypes/treatments. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(b) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated flot1 mutant plants with 30μM of ES9–17 or ES-2, Col-0 WT untreated was included as internal control. (Right) representative images of 6 days old flot1 mutant plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–14 roots), one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

While CHC1 and CHC2 gene sequences are ∼90% identical and are functionally partially redundant in some cellular processes(Dahhan et al., 2022), each gene can have distinct roles. For instance, CHC2 is required to a greater extent than CHC1 for CV protein-induced chloroplast degradation, hypersensitivity to water stress and plant defense(Žárský, 2016; Pan et al., 2023). To assess the role of each individual CHC gene, we used chc single knockout mutants to partially block endocytosis. Only chc1–3 and not chc2–1 mutant showed a higher suberization than wild type (Fig. 2a). Our results confirmed that CME inhibition induces and enhances suberization in Arabidopsis root endodermis and that downregulating CHC1 is sufficient to cause this phenotype.

Plant endocytosis can be divided into 2 pathways based on clathrin dependency for cargo internalization: CME and CIE. ES9–17 blocks CME so we wondered how the disruption of CIE impacts the suberization process. For this, we quantified the suberin occupancy of a flot1 knockout mutant. To our surprise, the mutant had a reduced suberin occupancy compared to WT (Fig. 2b). When flot1 was treated with ES9–17 to block the whole endocytosis pathway or with ES-2, to disrupt exocytosis, plants showed an ectopic suberization like the ES9–17 and ES-2 treated wildtype plants (Fig.1a and Fig. 2b). These data suggest a differential role for CME and CIE in suberization, however further research is needed to clarify the FLOT-1 protein function on this process.

CME disruption dependent suberization relies on the CIFs/SGN pathway.

A combination of exogenous and developmental signals fine tune endodermal suberization in a positive way(Woolfson et al., 2022).Two major independent pathways for this have been mainly described so far: an ABA-dependent pathway and one that relies on the endodermal lignification regulatory hub comprised of the CIF1/2 peptides, its receptors SGN3/SGN1 and RBOHD/RBOHF proteins(Fujita et al., 2020; Shukla et al., 2021). Both pathways converge on the regulation of a set of four endodermal MYB transcription factors (MYB41,MYB53,MYB92,MYB93) whose expression suffices to promote suberization(Shukla et al., 2021).

We therefore sought to explore the connection and dependency between the ectopic suberization caused by ET disruption and these regulatory pathways. For this, we treated plants with 10 μM Fluridone (Flu), an ABA biosynthesis inhibitor(Yoshioka et al., 1998; Kondhare et al., 2014) and with the combination of Flu and ES9–17 or ES-2 for a 24 h period. As previously reported(Shukla et al., 2021), Flu treated plants (Fig. 3a) showed a reduction in suberin occupancy due to a decrease in endogenous ABA levels. Plants treated with Flu and ES9–17 or with Flu and ES-2 (Fig. 3a) showed ectopic endodermal suberization (Fig. 1a). This indicates that the ET disruption-enhanced suberization is not dependent on ABA levels. Because Flu is a potent herbicide that affects root growth even at small concentrations, we decided to corroborate this using the dominant negative ABA insensitive mutant abi1–1 which is defective in the active ABA signaling pathway(Wu et al., 2003) and displays a well described suberin reduction phenotype in the endodermis(Barberon et al., 2016; Shukla et al., 2021). Consistent with the Flu treatment results, both ES9–17 and ES-2 treatments induced a higher endodermal suberin occupancy in the abi1–1 mutant (Fig. 3b), confirming that the ET disruption-dependent suberization process is independent of ABA signaling.

Figure 3. CME disruption-induced endodermal suberization is CIFs/SCHENGEN pathway dependent.

(a) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT plants with 10μM fluridone and 30μM of ES9–17 and/or ES-2, (right) representative images of 6 days old treated and untreated Col-0 WT plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–14 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test;(*) p < 0.05, (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(b) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated abi1–1 mutant plants with 30μM of ES9–17 or ES-2, (right) representative images of 6 days old of abi1–1 treated and untreated plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(c) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT, cif1cif2 and sgn1–2 plants with 30μM of ES9–17 or ES-2, (bottom) representative images of 6 days old of treated and untreated Col-0 WT, cif1cif2 and sgn1–2 plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–16 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (**) p < 0.01, (****) p < 0.0001, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and/or treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

Previous studies have shown that exogenous application of CIF1/2 peptides induces highly localized ROS production through RBOHF and RBOHD NADPH oxidases driving CS formation in roots and also promotes an ectopic endodermal suberization independently of ABA signaling(Fujita et al., 2020; Shukla et al., 2021). SGN3, the receptor for CIF peptides, activates the SGN1 kinase, which in turn phosphorylates both NADPH oxidases at the plasma membrane, promoting lignification of endodermal cells(Fujita et al., 2020; Reyt et al., 2021). To test whether the ET disruption dependent suberization is linked to this pathway we treated the double CRISPR knockout cif1cif2 mutant with the ET inhibitors (Fig. 3c). ES-2 treated plants showed enhanced suberization indicating independent mechanisms for exocytosis disruption and CIF peptide induced suberization (Fig. 3c). Treatment of cif1cif2 plants with ES9–17(Fig. 3c) did not lead to any significant changes in suberin occupancy. To check if there could be a link downstream of CIFs peptides, we also treated the sgn1–2 kinase knockout mutant with the inhibitors (Fig. 3c). Both chemicals elicited the same suberin occupancy pattern as observed in the cif1 cif2 mutant, further supporting the idea that suberization induced by endocytosis disruption—rather than exocytosis disruption—may act through the CIF/SGN3/SGN1 signaling pathway.

To assess the extent to which endocytosis disruption may require additional molecular components downstream of the CIFs/SGN pathway to enhance endodermal suberization, we performed chemical treatments on rbohD and rbohF single knockout mutants, as well as the rbohDF double knockout. ES-2 treatment increased the suberin occupancy in all the three mutants (Fig. 4a), providing further evidence for an ABA and CIFs/SGN independent suberization mechanism with no requirement for RBOHD/F activity. Conversely, the ectopic suberization phenotype induced by ES9–17 appears to require the activities of both RBOHF and RBOHD, as well as SGN1 and the CIF1/2 peptides. However, RBOHF may play a more central role than RBOHD, as ES9–17–treated rbohF plants showed significantly lower suberin occupancy compared to rbohD plants under the same conditions. In addition, rbohF and rbohDF suberin occupancy did not show any significant differences in our conditions upon ES9–17 treatment (Fig. 4a).

Figure 4. ES9–17 treatment leads to accumulation of RBOHF at the PM, promoting a ROS burst in the endodermis.

(a) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT, rbohD, rbohF and the double rbohDF mutant plants with 30μM of ES9–17 or ES-2, (right) representative images of 6 days old treated and untreated Col-0 WT, rbohD, rbohF and rbohDF plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–12 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (**) p < 0.01 (*) p < 0.05, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and/or treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(b) (Left) Boxplot of the endodermal cells GFP PM/cytosol ratio values from untreated and 30μM ES9–17 treated RBOHD-GFP and RBOHF-GFP expressing plants, (right) representative images showing the GFP signal distribution in the suberized root region, for each expressing line, between PM and cytosol in the endodermis upon ES9–17 treatment. Data are the mean ± SD (N = 8–10 roots), two-tailed unpaired t-test, (****) p < 0.0001, (**) p < 0.01. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 20 μm.

(c) (Left) Boxplot of the endodermal 488/405 nm ratio values from untreated and treated plants expressing the hydrogen peroxide HyPer7 sensor in the cytosol. Plants were treated with 30μM of ES9–17 and with 5μM DPI + 30μM ES9–17, (right) representative images showing the HyPer7 sensor changes upon treatments in root epidermis(ep), cortex(co) and endodermis(en) from suberized root region. Data are the mean ± SD (N = 10 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and/or treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 25 μm. Calibration bar indicates reduced(red) or oxidized(ox) sensor state.

ES9–17 impairs RBOHF endocytosis promoting an endodermal ROS burst.

Plasma membrane (PM)-localized RBOHD is expressed in several root tissues while RBOHF is expressed mainly in root endodermis and both produce ROS via post-translational activation upon abiotic and biotic stresses(Lee et al., 2013; Morales et al., 2016). A recent study showed that phosphorylation of RBOHD at the C-terminal region leads to ubiquitination, a signal for endocytosis(Lee et al., 2020, 2022). RBOHD is constantly internalized via CME and then trafficked to the vacuole for degradation, which may be a mechanism to keep PM-localized RBOHD at certain levels(Lee et al., 2020, 2022), thereby regulating apoplastic ROS accumulation. However, in the case of RBOHF it is unclear if it is trafficked via endocytosis in a similar way to RBOHD. Above, we have shown that CME disruption-dependent suberization requires the presence of both RBOHF/D activities (Fig. 4a). To gain insights on how the CME disruption would affect the recycling and secretion of both PM- localized oxidases in endodermal cells, we treated plants overexpressing GFP tagged RBOHD or RBOHF proteins(Smokvarska et al., 2020) with ES9–17 and quantified the PM/cytosol GFP signal ratio (Fig. 4b). Upon ES9–17 treatment the RBOHD-GFP signal is nearly absent in the cytosol while it shows a reduced but detectable GFP signal in the PM of endodermal cells compared to untreated plants (Fig. 4b). These data indicate that while the endomembrane trafficking is altered, disrupting endocytosis causes RBOHD to locate mainly at the PM of endodermal cells, however at a lower level than untreated plants according to the PM/cytosol ratio values (Fig. 4b). In contrast, for RBOHF-GFP we found an increase of the PM/cytosol ratio upon ES9–17 treatment, indicating that RBOHF may undergo internalization via CME and blocking this pathway causes an increase of the RBOHF-GFP signal at the plasma membrane of endodermal cells (Fig. 4b). These findings suggest that RBOHF plays a primary role in suberization induced by CME disruption, while RBOHD would play a secondary role.

NADPH oxidase activity at the PM is one of the main sources of ROS and these proteins are relevant for the regulation of the homeostasis between the apoplastic and cytosolic ROS pools in plant cells(Foreman et al., 2003; Monshausen et al., 2007; Podgórska et al., 2017; Rohman et al., 2024). To assess whether RBOHF—and to a lesser extent RBOHD—contribute to an endodermal ROS burst associated with endocytosis impairment, particularly through elevated H₂O₂ levels, we employed the genetically encoded HyPer7 sensor, which specifically detects hydrogen peroxide in the cytosol. Although measuring apoplastic ROS levels in endodermal cells is technically challenging, HyPer7 served as a suitable proxy to quantify cytosolic H₂O₂ levels in these cells (Fig. 4c). HyPer7 is a highly pH-insensitive hydrogen peroxide specific ratiometric probe that has been successfully used to detect, quantify and visualize H₂O₂ gradients, even at small amounts, in mammals and plants(Ugalde et al., 2021; Hoehne et al., 2022; Dopp et al., 2023). ES9–17 treatment led to an increased 488/405 nm ratiometric quotient in the endodermis and in the whole root compared to untreated plants expressing HyPer7 (Fig. 4c). This indicates high cytosolic H₂O₂ levels leading to HyPer7 oxidation. To test whether the increase of H₂O₂ levels originates from an active NADPH oxidase activity in the PM, we treated HyPer7 plants with DPI, which is a NADPH oxidase inhibitor, and with ES9–17. DPI addition blocked ES9–17 mediated HyPer7 oxidation (Fig.4c). Together our data indicate that CME disruption induced-suberization requires RBOHF to be preferentially located at the PM of endodermal cells, and RBOHD, to positively regulate the hydrogen peroxide levels.

ET disruption requires an active NADPH oxidase activity and H₂O₂ for suberization.

Arabidopsis endodermal root suberin comprises two structural domains: SPAD and SPPD. The SPPD is commonly referred to as a “lignin-like” domain because of the presence of phenylpropanoid-derived monomers, including monolignols, cross-linked to the cell wall(Woolfson et al., 2022). Data from potato tuber periderm wounding and the Arabidopsis lignification pathway have prompted the hypothesis that SPPD assembly might involve apoplastic peroxidase-mediated crosslinking of phenolic compounds(Bernards et al., 1999; Bernards & Razem, 2001; Razem & Bernards, 2002; Lee et al., 2013). Such a process would require H₂O₂ generated by the activities of a membrane bound NADPH-dependent oxidase and apoplastic superoxide dismutase systems (Fig. 5a), as occurs during Arabidopsis CS formation(Lee et al., 2013; Woolfson et al., 2022). However, there has not been strong evidence so far indicating that this happens for the Arabidopsis SPPD during the suberin lamellae deposition. To gain clarity on this current hypothesis in the context of ET disruption mediated enhanced suberization, we employed a combined chemical and genetic approach.

Figure 5. H₂O₂ generated via NADPH-dependent oxidase activity is required for ET disruption dependent suberization.

(a) Schematic representation of the hypothetically required steps between NADPH oxidase activity and suberin polyphenolic domain (SPPD) assembly during suberin polymerization. In red, inhibitors of the different steps, Diphenyleneiodonium chloride (DPI, NADPH oxidase inhibitor), N,N’-Dimethylthiourea (DMTU, H₂O₂ scavenger), Salicylhydroxamic acid (SHAM, peroxidase inhibitor).

(b) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT plants with 5μM DPI, 5μM DPI + 30μM ES9–17 and 5μM DPI + 30μM ES-2, (right) representative images of 6 days old treated and untreated Col-0 WT plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (**) p < 0.01, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and/or treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(c) (Top) Boxplot of root endodermal suberin occupancy values from untreated and treated chc1–3 and exo70a1–2^+/− mutant plants with 5μM DPI, (bottom) representative images of 6 days old treated and untreated chc1–3 and exo70a1–2^+/− mutant plants FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(d) (Left) Boxplot of root endodermal suberin occupancy values from untreated and treated Col-0 WT plants with 20mM DMTU, 20mM DMTU + 30μM ES9–17 and 20mM DMTU + 30μM ES-2, (right) representative images of 6 days old treated and untreated Col-0 WT plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–11 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (ns) non-significant differences. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and/or treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(e) (Left) Boxplot of root endodermal suberin occupancy values from 20mM DMTU treated vs untreated chc1–3 and exo70a1–2^+/− mutant plants, (right) representative images of 6 days old treated and untreated chc1–3 and exo70a1–2^+/− mutant plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

Combined treatment of DPI and ES9–17 of HyPer7 expressing plants, kept the sensor in its reduced state, most likely due to a decrease in the endodermal hydrogen peroxide levels (Fig. 4c). To assess the consequences of this on the suberization process we treated Col-0 WT plants with DPI and measured endodermal root suberization. This led to a significant reduction of root suberization (Fig. 5b) confirming the positive role of the RBOHs activity in this process. Furthermore, concomitant treatment of plants with DPI and ES9–17 or ES-2 blocked the characteristic ectopic suberization phenotype induced by these chemicals, with plants showing a reduced or a similar suberin occupancy compared to non-treated plants (Fig. 5b). To further test this using a genetic approach, we treated with DPI two mutants exhibiting enhanced suberization: The chc1–3 mutant (Fig. 2a) was used as a genetic equivalent to the ES9–17 treatment, while the heterozygous exo70a1–2^+/− mutant (Fig. S1) served as a genetic proxy for ES-2–mediated inhibition. Both mutants showed a significant decrease in the suberin occupancy upon DPI treatment (Fig. 5c). To determine the impact of the reduction in ROS levels on this suberization phenotype, in particular a H₂O₂ reduction, we decided to treat plants with DMTU, a hydrogen peroxide and hydroxyl radicals (^•OH) scavenger(De Zacchini & De Agazio, 2001; Li et al., 2023a). DMTU caused a significant reduction of the suberin occupancy in Col-0 WT plants and blocked the enhanced suberization phenotype when plants were co-treated with ES9–17 or ES-2 (Fig. 5d). DMTU treatment also significantly reduced the suberin occupancy in the chc1–3 and exo70a1–2^+/− mutants (Fig. 5e). Finally, we exogenously treated Col-0 WT plants with 0.5mM H₂O₂ for 24 h and we detected a significant increase in suberin occupancy (Fig. S2). Together, this evidence reflects the major role of ROS, particularly hydrogen peroxide on the general suberization process and in the ET disruption dependent suberization.

Every peroxidase (PRX) currently described being involved in the Arabidopsis lignification process belongs to the same superfamily of apoplastic Class-III peroxidases (Cosio & Dunand, 2009; Rojas-Murcia et al., 2020). A common feature among this family of secreted proteins is their highly redundant genetic functionality in the oxidative polymerization of various apoplastic components (Cosio & Dunand, 2009; Shigeto & Tsutsumi, 2016) typically associated with the cell wall, including pectins (Francoz et al., 2019), extensins (Pacheco et al., 2022), and notably monolignols in the apoplast of lignifying cells in the xylem (e.g., PRX71 (Hoffmann et al., 2020)) and in the root endodermis (e.g., PRX64(Lee et al., 2013)). The association between suberin and these PRXs emerged from the characterization of a Class III secreted anionic peroxidase that accumulates in suberizing tissues during potato tuber wound healing and oxidizes monomers derived from the phenylpropanoid pathway (Bernards et al., 1999). To gain insights into a potential role of these PRXs during the general suberization and the ET disruption dependent suberization process in Arabidopsis, we utilized SHAM to chemically block these proteins. While SHAM is a general Hem iron-containing peroxidases inhibitor, the Class III apoplastic PRXs multigenic family are the largest group of these proteins in Arabidopsis with around 73 members(Cosio & Dunand, 2009).

Col-0 WT plants treated with SHAM for 24 h showed a decrease in the suberin occupancy values (Fig. 6a), suggesting a positive role of the peroxidase activity on promoting suberization, probably through the SPPD assembly. The observation that three distinct manipulations—inhibition of NADPH oxidase activity, scavenging of H₂O₂ by DMTU, and inhibition of peroxidases with SHAM—all lead to reduced endodermal suberization provides strong support for a model in which RBOHF, together with RBOHD, generates superoxide that is subsequently dismutated into H₂O₂. This H₂O₂ would then be utilized by PRXs to polymerize suberin lamellae. However, when SHAM treatment was combined with ES9–17 or ES-2 treatments we detected an increase of the endodermal suberization in WT plants (Fig. 6a). These results were corroborated with chc1–3 and exo70a1–2^+/− as both mutants also showed a higher suberization when treated with SHAM (Fig. 6b). We therefore speculated that the cause of this unexpected phenotype would be due to a H₂O₂ root burst that oxidizes the HyPer7 sensor after ES9–17 and ES-2 (Fig. 4c and Fig. S3) treatment, which correlates with higher suberization and somehow is able to overcome the SHAM inhibitory effect in our conditions. Indeed, HyPer7 sensor oxidation increased in plants concomitantly treated with ES9–17 or ES-2 and SHAM, while showing an important reduction upon SHAM treatment only (Fig. 6c), if compared with untreated plants. These results suggest that while peroxidase activity is needed for endodermal suberization, in the particular context of ET disruption induced suberization generating high hydrogen peroxide levels are the driving force.

Figure 6. Peroxidase activity is required for root suberization in Arabidopsis while high H₂O₂ levels play a major role for ET disruption dependent suberization.

(a) (Left) Boxplot of root endodermal suberin occupancy values from 100μM SHAM, 100μM SHAM + 30μM ES9–17 and 100μM SHAM+ 30μM ES-2 treated vs untreated Col-0 WT plants, (right) representative images of 6 days old treated and untreated Col-0 WT plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 10–11 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (***) p < 0.001, (*) p < 0.05. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(b) (Left) Boxplot of root endodermal suberin occupancy values from 100μM SHAM treated vs untreated chc1–3 and exo70a1–2^+/− mutant plants, (right and bottom) representative images of 6 days old treated and untreated plants showing FY suberin staining endodermal pattern. Data are the mean ± SD (N = 7–12 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (*) p < 0.05. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 1 mm.

(c) (Left) Boxplot of the endodermal 488/405 nm ratio values from 100μM SHAM, 100μM SHAM + 30μM ES9–17 and 100μM SHAM + 30μM ES-2 treated vs untreated plants expressing the hydrogen peroxide HyPer7 sensor in the cytosol, (right) representative images showing the HyPer7 sensor changes upon treatments in root epidermis(ep), cortex(co) and endodermis(en) from the suberized root region. Data are the mean ± SD (N = 10–13 roots), ordinary one-way ANOVA followed by a Tukey–Kramer test; (****) p < 0.0001, (*) p < 0.05. Results are representative of two independent experiments showing similar results. Asterisks indicate significant differences between untreated and treated plants. Box = interquartile range between the lower and upper quartiles, center line = median, + = mean, whiskers = min and max values. Scale bars = 25 μm. Calibration bar indicates reduced(red) or oxidized(ox) sensor state.

Discussion

In this study, we reported changes in Arabidopsis root endodermal suberin occupancy in response to endomembrane trafficking disruption and how it relates to known suberization/lignification regulatory pathways. We found that chemical or genetic inhibition of clathrin-mediated endocytosis impairs the internalization of RBOHF and RBOHD, leading to elevated hydrogen peroxide levels in the endodermis and enhanced suberization extending toward the root tip. We also showed that specific chemical inhibition of the EXO70A1 exocyst isoform with ES-2 induces an endodermal suberization towards the root tip (Fig. 1a) as previously described for the lotr2/exo70a1 mutant. The focalized secretory activity and consequent formation of extracellular membrane vesiculo-tubular structures found in the periplasmic space of lotr2/exo70a1 that correlates with the endodermal suberization highlights a link between EXO70A1 and a possible transport of suberin monomers to the apoplast(De Bellis et al., 2022). Although we did not specifically investigate the accumulation of extracellular vesicles (EVs) in response to ET inhibitors treatment in this study, our observation that ES-2–induced suberization was independent of the two mains regulatory suberin pathways—ABA and CIFs/SGNs—under our experimental conditions (Fig. 3b, c and Fig. 4a) suggests the existence of a parallel mechanism controlling suberization. This pathway may involve EV formation and the transport of hydrophobic compounds such as suberin monomers, or potentially monolignols in the context of Casparian strip formation. Many proteins involved in CS formation, such SGN3, SGN1 and RBOHF, are specifically localized to the CS domain or the CS (Fujita et al., 2020), but how these proteins achieve their specific localization is not well understood. Previously it has been described that RBOHD can undergo endocytosis(Lee et al., 2022) and here we provide evidence that CME also plays an important role, at least, in the endodermal PM localization of RBOHF (Fig. 4b). Indeed, ES9–17 treatment fails to induce the enhanced suberization phenotype in the rbohDF, cif1cif2 and sgn1 mutants (Fig. 3b, c and Fig. 4a) suggesting that the endocytic pathway may also contribute to maintaining SGN protein levels. This, in coordination with the CIF peptides, likely facilitates their localization at the plasma membrane, enabling suberization—like the mechanism observed in root endodermal lignification.

Prior evidence together with the data presented in this work still supports the current hypothesis of the major role of the H₂O₂ burst generated through NADPH oxidase activity and used by Class III apoplastic peroxidases for the likely crosslinking of the suberin polyphenolic domain (Fig. 4b, c, Fig. 5b, Fig. 6a, c and Fig. S2). The apoplastic ROS pool is the main source that apoplastic PRXs use for the crosslinking of cell wall compounds and there is a tight balance between the both cytoplasmic and apoplastic H₂O₂ pools (Passardi et al., 2004; Shigeto & Tsutsumi, 2016; Zhu, 2016; Eljebbawi et al., 2022). When H₂O₂ generation in the apoplast exceeds the relatively low scavenging capacity of this space, ROS can accumulate and can diffuse through the plasma membrane or specific aquaporins (Bienert et al., 2007) into the cytoplasm, boosting the intracellular ROS pool and activating symplastic ROS (Sharma et al., 2012; Sachdev et al., 2021). We expect that a similar event is happening upon endomembrane trafficking disruption. For instance, ES9–17 treatment increases RBOHF levels at the plasma membrane (Fig. 4b), likely leading to greater superoxide production and its dismutation into hydrogen peroxide. The resulting apoplastic H₂O₂ accumulation and diffusion into endodermal cells would then contribute to elevated cytoplasmic ROS levels and an overall enhancement of suberization. That could explain why we observed a higher oxidation of the cytosolic HyPer7 sensor after the treatment with the inhibitor and the opposite when we added DPI to block the NADPH oxidase activity (Fig. 4c).

Using SHAM treatment, we demonstrated that peroxidase activity is crucial for suberization in Arabidopsis. Remarkably, elevated hydrogen peroxide levels in the endodermis can bypass SHAM inhibition during endomembrane trafficking disruption, still leading to enhanced suberization (Fig. 6a-c). Previous work had shown that higher order mutants of Class III apoplastic PRXs are impaired in lignification and accumulate hydrogen peroxide in the apoplast of endodermal cells. This was due to suppression of PRX activity, highlighting the role of these proteins for the oxidative polymerization of monolignols via their ROS consumption(Rojas-Murcia et al., 2020). However, the authors did not evaluate the effect on suberization or in the alteration of the overall balance between apoplastic and cytoplasmic ROS pools. In our experiments, the HyPer7 sensor proved to be a valuable tool for detecting apoplastic H₂O₂ disturbances caused by chemical treatments such as ES-2, as it enabled the estimation of changes in cytosolic H₂O₂ levels (Fig. S3). This also allowed us to find that ES9–17 and DPI, which affect plasma membrane–bound RBOHF and RBOHD in distinct ways (Fig. 4c), target the primary sources of apoplastic ROS in the endodermis. Whether SHAM causes apoplastic H₂O₂ accumulation under our conditions remains uncertain, as we currently lack sufficient evidence to confirm either outcome. However, it is known that Class III peroxidases (PRXs) can either consume or produce hydrogen peroxide depending on substrate availability, acting through their oxidative or peroxidative cycles, respectively, and thereby contributing differently to ROS homeostasis (Passardi et al., 2004; Cosio & Dunand, 2009; Shigeto & Tsutsumi, 2016). Based on our HyPer7 and suberin occupancy data, we conclude that the overall effect of SHAM—beyond PRX inhibition—is a reduction in root H₂O₂ levels, which negatively impacts suberization. Notably, this inhibitory effect can be overcome by disrupting endocytosis.

Suberin is a highly complex and plastic biopolymer with its formation and matrix composition varying in form and function across different plant species (Shukla et al., 2021). Current hypotheses and models mostly highlight the positive role of reactive oxygen species, like H₂O₂ and the role of peroxidase activity on the assembly of the Arabidopsis SPPD due the “lignin-like” nature of the domain and the similarity with the Casparian strip assembly process, consequently impacting the establishment of the endodermal suberin lamellae (Serra & Geldner, 2022; Woolfson et al., 2022). Based on the evidence presented here and based on the known dual role that hydrogen peroxide has on the modification of components of the cell wall(Passardi et al., 2004; Cosio & Dunand, 2009; Marzol et al., 2022; Pacheco et al., 2022), we believe that H₂O₂ can function as both a necessary driver of suberization and as a limiting factor. In that way it defines the extent of the process, such that its availability sets the boundary conditions under which suberization can occur, as we showed when adding H2O2(Fig. S2) or when we used a ROS scavenger (Fig. 5d,e). Enzymes of the GDSL-type esterases/lipases (GELPs) family were proposed as strong candidates involved in the SPAD polymerization(Woolfson et al., 2022). Indeed, two clusters of auxin-regulated GDSL-lipases were identified recently in Arabidopsis, where one seems to be required for suberin synthesis, while the other one can drive suberin degradation (Ursache et al., 2021). Overexpression of the Arabidopsis apoplastic lipase GDSL1 in rapeseed is associated with a ROS burst, which is caused by the up-regulation of the NADPH and polyamine oxidases, both markers associated with hydrogen peroxide production (Ding et al., 2020). However, currently there is no strong evidence to associate a specific GELP with a particular process within the suberization deposition or degradation pathway of the SPAD.

In this work we described how the main cell cargo internalization pathway controls endodermal root suberization. Fig. S4 represents a simplified model for this in which CME disruption leads to a plasma membrane RBOHF accumulation, increased apoplastic H₂O₂, and CIFs/SCHENGEN pathway activation, driving ectopic endodermal suberization, whereas normal CME maintains balanced suberization. However, many questions remain unanswered and opened for future research: Do endodermal cells regulate the secretion and recycling of a specific subset of apoplastic PRXs to exert a control over the root suberization process? It is well known that salt stress affects RBOHD endocytosis (Lee et al., 2022) and one of the roles of suberin in roots is to provide an hydrophobic layer to reduce water loss and sodium uptake (de Silva et al., 2021). That leads us to the next question: What other abiotic/biotic/stimuli might regulate the endodermal cell PM levels of RBOHF and/or RBOHD to increase suberization? Although how the enodemembrane trafficking inhibition affects the CS development or endodermal lignification is beyond the scope of this work, it is important to acknowledge that both root lignification and suberization events are tightly linked and work together to maintain an endodermal hydrophobic barrier. A common phenotype of some mutants impaired in Casparian strip (CS) formation is an early or ectopic suberization,e.g. esb1–1, myb36(Kamiya et al., 2015), which compensates for reduced lignification (Barberon, 2017; Shukla & Barberon, 2021; Serra & Geldner, 2022; Woolfson et al., 2022). This raises the question: are the mutations or chemical inhibition of the CME proteins associated with a defective endodermal lignification phenotype? Our findings, along with these open questions, may provide valuable insights not only into the rapidly evolving understanding of suberin—its deposition, and its dynamic and plastic nature—but also into the emerging field of suberin-related engineering aimed at enhancing crop resilience and promoting soil carbon sequestration.

Data Availability:

The data that supports the findings of this study are available in the manuscript.