Screen for proteins promoting persistent plasma membrane – cell wall connections in Nicotiana benthamiana.

A. Confocal image of Arabidopsis root expressing p35S::eGFP-LTI6B, that shows a plasmodesmata-unrelated Hechtian strand traversing the plasmolytic space in a root hair cell plasmolyzed with sorbitol 0.6M. Scale bar = 10μm. B. Schematic representation of the working hypothesis, illustrating potential scenarios where plasma membrane proteins act as molecular glue between the plasma membrane and the cell wall. C. Scheme (Illustrated with BioRender) and brief description of candidate proteins located at the plasma membrane – wall interface. RLK: Receptor Like Kinase. RLP: Receptor Like Protein. GPI-AGP: Glycosyl-Phosphatidyl-Inositol – ArabinoGalactan Protein. GAP: GPI- Anchored Protein. D. Transient transformation of N. benthamiana followed by plasmolysis of pavement cells with glycerol 10%, and quantification of density of Hechtian strands labeled with LTI6B-mCitrine. Proceedings for the screening detailed in Methods section. E. Boxplot showing the density of Hechtian strands (HS), expressed as number of HS per μm of detached plasma membrane (PM), for each tested candidate compared to the control LTI6B. Boxes indicate the interquartile range (IQR), the line inside each box represents the median, whiskers show the data range within 1.5×IQR, and black dots indicate outliers. Statistical significance was determined using the Kruskal–Wallis test followed by Dunn’s post hoc test with Benjamini–Hochberg correction. Sample size (N = number of experiments, n = number of images with 1 -3 cells averaged) and p-values are indicated below the plot. Asterisks denote statistically significant differences compared to the LTI6B control. F. Confocal image showing a N. benthamiana pavement cell (at the middle-plane) expressing LecRK-I.9*-TagRFP (magenta), after plasmolysis with glycerol 20%. The cell walls were stained with calcofluor (cyan). Scale bar = 12 μm. Inset showing the Hechtian strands in detail. Scale bar = 10μm.

© 2026, BioRender Inc. Parts of this image were created with BioRender are not made available under the same license as the Reviewed Preprint.

Analysis of LecRK-I.9 contribution to plasma membrane – cell wall connections in Nicotiana benthamiana.

A. Quantification of percentage of the cell surface occupied by the Hechtian reticulum after plasmolysis (% Occupancy). 1. SurfCut projection of the surface. 2. Classified image in plasma membrane (green), area with H. reticulum (red) and background (violet). 3. Mask (black) of plasma membrane and background. 4. Selection (yellow) of area with H. reticulum (white). 5. Binary and skeletonized image. Boxplot representing the % Occupancy for LTI6B and LecRK-I.9* within plasmolyzed areas up to 1000 μm2 in size. Statistical significance was determined using the Welch two-sample t-test. Sample size (N = number of experiments, n = number of images with 1 -3 cells averaged) and p-value are indicated on the plot. Asterisks denote statistically significant differences compared to the LTI6B control. B. Alphafold-predicted structure of LecRKI.9 from Uniprot highlighting the protein domains and the approximate location of the mutations. The boxplot displays the density of Hechtian strands (HS), expressed as number of HS per μm of detached plasma membrane (PM), for LecRK-I.9*, LecRK-I.9*ΔLec and the control LTI6B. Statistical significance was determined using the Kruskal–Wallis test followed by Dunn’s post hoc test with Benjamini–Hochberg correction. No significant difference was observed between LTI6B and LecRK-I.9*ΔLec. Sample size (N = number of experiments, n = number of images with 1 -3 cells averaged) and p-values are indicated on the plot. C. Transient transformation of N. benthamiana followed by plasmolysis of pavement cells with sorbitol 0.6M (30 min.). In green, LTI6B-mCitrine labelling the plasma membrane (and some cytoplasmic streaming). In magenta, the extracellular lectin domain (ECD) of LecRK-I.9 fused to TagRFP (as shown by the scheme), exported to the apoplast by its endogenous signal peptide (SP). In red, the signal of TagRFP exported to the apoplast by the chitinase signal peptide (SP-TagRFP). 1. Entire pavement cell and neighboring cells showing plasma membrane detached from the wall. Scale bar = 15 μm. 2. Inset corresponding to the yellow ROI in 1 showing the signal of ECD-LecRK-I.9 predominantly in the wall. Scale bar = 10 μm. 3. Inset corresponding to the yellow ROI in 2 showing plasmolyzed spaces (*) without ECD-LecRK-I.9 and putative microdomains in the wall that accumulate ECD-LecRK-I.9 (arrowheads). Scale bar = 5 μm. 4. Negative control showing that a protein free in the wall diffuse (arrowheads) into the plasmolized space (*). Scale bar = 10 μm. (Brightness and contrast adjusted in all the images for better observation).

In Arabidopsis, LecRK-I.9* form clusters under hyperosmotic treatments, the mobility of which is lectin dependent.

A. Confocal images showing LTI6B (green), LecRK-I.9* (grey) and LecRK-I.9*ΔLec (magenta) plasma membrane localization in Hechtian strands of plasmolyzed hypocotyl epidermal cells (at the cell middle-plane) from stably transformed 7 days-old Arabidopsis seedlings (homozygote T3 generation) after sorbitol 0.6M 30-minutes incubation. Scale bar = 10 μm. (Brightness and contrast adjusted in all the images for better observation). B. 2-minutes time-lapse series of cotyledon pavement cells surfaces showing the mobility of clusters after incubation in sorbitol 0.6M. Frames are color-coded: Time = 0 seconds (red), Time = 60 seconds (green), and Time = 120 seconds (blue). Clusters that did not move appear white in the merge images. Scale bar = 5 μm. (Brightness and contrast adjusted in all the images for better observation).

LecRK-I.9* cluster distribution correlates with predicted differences in wall properties, only when the lectin domain is present.

A. Confocal images showing outer and inner surfaces of cotyledon and hypocotyl epidermal cells from 7-days-old seedlings incubated in water (control condition) or sorbitol 0.6M during 30 minutes (hyperosmotic condition). The patterns of clusters are shown for LTI6B (in green), LecRK-I.9* (in grey) and LecRK-I.9*ΔLec (in magenta). Scale bars = 25 μm; except for LecRK-I.9* and LTI6B cotyledons in sorbitol and hypocotyls of LecRK-I.9*ΔLec in sorbitol where scales bars = 10 μm. All images of LecRK-I.9* and LTI6B cotyledons were acquired with Airyscan detector. (Brightness and contrast adjusted in all the images for better observation). B. Boxplots presenting the number density of clusters for LTI6B, LecRK-I.9* and LecRK-I.9*ΔLec in the hyperosmotic condition normalized by the control condition ((Sorbitol 0.6M – Water)/ Water). Inner vs. Outer surfaces within each genotype were compared by Wilcoxon test. N = 3 experiments, n = ∼3-9 plants/genotype/treatment (average 1-3 cells per image). Asterisks denote significant differences. p-value is indicated on the cotyledon plot. Images on the hypocotyl plot highlight the strong trend of asymmetrical pattern displayed by LecRK-I.9* clusters in the inner and outer surface of the hypocotyl cell. Scale bar = 10 μm.

LecRK-I.9* overexpressing lines can expand their cotyledons and leaves despite the low water potential stress.

A. Images showing WT (Col-0) Arabidopsis seedlings, LecRK-I.9* and LecRK-I.9*ΔLec overexpressing lines after 10 days from being transferred to mock plates (0.06 Osmol/L) or PEG-infiltrated plates (0.2 Osmol/L). See Material and methods for details about the assay and plate preparation. Blue arrows show examples of resistant plants. Scale bars = 5mm. B. Scheme depicting the criteria for considering a plant as resistant: expansion of cotyledons and leaves achieving a rosette architecture comparable to the one in mock conditions. C. Boxplot showing the frequency of resistant plants for each genotype across 7 experiments (N), each one with 2 to 3 plates per treatment (Mock/ PEG) containing 7 -10 seedlings of each genotype per plate (n = 20 – 30 seedlings per genotype per experiment). Statistical significance was determined using the Kruskal–Wallis test followed by Dunn’s post hoc test with Benjamini–Hochberg correction. Sample size and p-values are indicated on the plot.