Xyloglucan deficiency leads to a reduction in turgor pressure and changes in cell wall properties affecting early seedling establishment

Summary

Xyloglucan is believed to play a significant role in cell wall mechanics of dicot plants. Surprisingly, Arabidopsis plants defective in xyloglucan biosynthesis exhibit nearly normal growth and development. We investigated a mutant line, cslc-Δ5, lacking activity in all five Arabidopsis CSLC genes responsible for xyloglucan backbone biosynthesis. We observed that this xyloglucan-deficient line exhibited reduced cellulose crystallinity and increased pectin levels, suggesting the existence of feedback mechanisms that regulate wall composition to compensate for the absence of xyloglucan. These alterations in cell wall composition in the xyloglucan-absent plants were further linked to a decrease in cell wall elasticity and rupture stress, as observed through atomic force microscopy and extensometer-based techniques. This raised questions about how plants with such modified cell wall properties can maintain normal growth. Our investigation revealed two key factors contributing to this phenomenon. Firstly, measurements of turgor pressure, a primary driver of plant growth, revealed that cslc-Δ5 plants have reduced turgor, preventing the compromised walls from bursting while still allowing growth to occur. Secondly, we discovered the conservation of elastic asymmetry (ratio of axial to transverse wall elasticity) in the mutant, suggesting an additional mechanism contributing to the maintenance of normal growth. This novel feedback mechanism between cell wall composition and mechanical properties, coupled with turgor pressure regulation, plays a central role in the control of plant growth and is critical for seedling establishment in a mechanically challenging environment by affecting shoot emergence and root penetration.

eTOC Blurb

Bou Daher et al show that Arabidopsis mutants lacking xyloglucan, have reduced cellulose crystallinity, altered cell wall mechanical properties and a reduction in turgor pressure. Bou Daher et al also demonstrate the importance of xyloglucan in early seedling establishment.

Introduction

A pivotal aspect of plant cells is their stiff outer layer known as the cell wall, which confers shape and structure. Loosening this shell allows plant cells to expand while adopting specific shapes and functions. Aside from water, their major component, cell walls consist predominantly of polysaccharides, including cellulose, hemicellulose, and pectin, alongside proteins and minerals. Primary cell wall polysaccharides display high levels of interactions and are constantly remodeled and deposited to allow expansive growth ¹. Feedback mechanisms monitoring the composition, structure and mechanics of the cell wall, also known as “cell wall integrity”, are of high importance to prevent cell damage ^1,2.

The mechanics of plant cell growth were initially described by Lockhart ³, where it was presented as plastic deformation resulting from tensile stress generated by turgor pressure above a stress threshold. Lockhart’s equation presents limitations and shortcomings since it is only valid for equilibrium growth and does not account for elastic deformations in the walls of growing cells. This model was later expanded by Ortega ⁴ to incorporate the elastic properties of the cell wall. In recent years, more sophisticated models at the cellular and tissue levels, considering water fluxes, cell geometries, cell wall polymer physical properties and other complex wall characteristics, have emerged ^5–11. Cellular turgor pressure is an isotropic force and cell geometry and/or local changes in the mechanical properties of walls allow for differential growth to happen and complex shapes to be generated. In the case of the etiolated hypocotyl, a highly elongating plant organ, a difference in the mechanical properties between the axial and transverse anticlinal walls referred to as elastic asymmetry together with anisotropic mechanical properties allow the cells to acquire and maintain a cylindrical shape and the organ to grow anisotropically ^12,13.

Polysaccharides in the cell wall contribute differentially to its mechanical properties. Cellulose, the primary load-bearing material, exhibits varying mechanical strength depending on its crystallinity index. Cellulose fibers with higher crystallinity possess superior tensile strength and higher elastic modulus ¹⁴.

Pectin, a complex polysaccharide, acts as the embedding material for other cell wall components, including cellulose. Interactions between pectin and other polysaccharides have been reported ^15–18. Pectin can modulate the wall’s mechanical properties through changes in its chemistry. Homogalacturonan (HG), the dominant type of pectin in the primary wall, is secreted in a predominantly esterified form, rendering it more fluid. Upon de-esterification, HG residues can be cross-linked with calcium ions, creating a stiffer, more gel-like material ^19,20. The degree of esterification-induced changes in wall mechanical properties is involved in the anisotropic growth of various cell and organ types ^12,13,21.

Xyloglucan is the major hemicellulosic polysaccharide in dicotyledonous plants. It is synthesized in the Golgi apparatus and transported by secretory vesicles to the apoplast ^22–24. Cellulose Synthase Like-C (CSLC) genes encode the β-glucan synthase responsible for the synthesis of the β-1,4 glucan backbone of xyloglucan ^25,26. Xyloglucan backbones are decorated with D-xylosyl groups which in turn can be linked with D-galactosyl groups. An additional L-fucose residue can be added to the galactose. These side chain sugars are added through the activity of several glycosyl transferases such as xyloglucan xylosyltransferases (XXT) which are responsible for the addition of the xylose residues ^27–29.

Xyloglucan has long been believed to tether cellulose microfibers and regulate expansion through proteins and enzymes involved in wall loosening, which can be caused by reduction of yield stress and/or by increasing plastic extensibility. Experimental and modeling data from the last decade suggest the existence of mechanical hotspots of xyloglucan connecting cellulose microfibrils. These mechanical hotspots are hypothesized to play important roles in stress relaxation in the wall ^30–32 despite representing an extremely minor fraction of the total pool of xyloglucan in the cell wall.

Xyloglucan is the most abundant hemicellulose, constituting ~ 20–25% of the dry primary cell wall in dicots ³³. Despite its abundance, the absence of xyloglucan does not dramatically alter plant development or organ growth in mutants that reduce xyloglucan biosynthesis, such as mutations in the CSLC or XXT genes ^{26,30,34–36}. These findings raise questions about the role of xyloglucan in the wall. Analysis of the xxt1/xxt2 mutant walls revealed an increase in HG content and a possible rise in glucomannan, a low-abundance hemicellulose in the primary walls of dicots ^34,37. Mechanical testing on this mutant demonstrated a reduction in wall stiffness (elastic modulus) and plastic extensibility compared to the wild type ^34,36,38. Therefore, understanding the mechanisms behind the maintenance of growth in the absence of xyloglucan would be of great importance in fundamental research targeting growth and development and would benefit more applied research dealing with field crops and horticultural applications.

Turgor pressure serves as the tensile force driving cell expansion, and heterogeneity in turgor pressure at the tissue level has been reported ^5,8. The relationship between cell wall integrity and turgor pressure remains poorly understood. Mutations in cellulose biosynthesis genes (CESA) or in the CESA-interacting protein KORRIGAN1 (KOR1) cause an increase in jasmonic acid (JA). This JA spike can be mitigated by treatment with hyperosmotic solutions, which reduce turgor pressure ³⁹. Moreover, THESEUS1 (THE1) (a Catharanthus roseus receptor-like kinase 1-like (CrRLK1L)) has been shown to influence cell wall stiffness and contribute to changes in turgor pressure, involving both JA and abscisic acid ⁴⁰. Additionally, THE1, together with the mechanosensitive channel MCA1, have been linked to maintaining cell wall integrity in an osmosensitive manner ⁴¹. These findings suggest the existence of feedback mechanisms between cell wall integrity and turgor pressure and highlight the importance of unraveling these mechanisms towards building a clearer understanding of plant growth and development.

In this study, we characterized mutant lines with reduced xyloglucan content. We demonstrate that these lines exhibit increased pectin content and reduced cellulose crystallinity, along with diminished turgor pressure compared to the wild type, which allow plant growth despite reduction in the mechanical strength of the wall. We also demonstrate that the lack of xyloglucan negatively affects early seedling establishment. These findings add to the known processes controlled by feedback between the cell and its wall with consequences for plant development.

Results

Reduced xyloglucan induces an increase in pectin and a reduction in cellulose crystallinity.

Arabidopsis mutant lines completely lacking xyloglucan have been shown to display close to normal growth and development ^26,28,34,35. CSLC enzymes are responsible for the biosynthesis of the xyloglucan backbone ^25,26, with five genes encoding CSLC in the Arabidopsis genome. The cslc quintuple mutant (cslc-Δ5) lacks detectable xyloglucan ²⁶. We analyzed the early stages of seedling development. Apart from the previously reported short root hair length²⁶, we noticed no significant difference in light grown seedlings between the mutant and the wild type (WT) plants (Figure S1A). We then examined the kinematics of rapidly elongating hypocotyls and roots in the dark (Video S1). While cslc-Δ5 hypocotyls displayed slightly reduced elongation over time, root growth kinematics were indistinguishable during the first three days of growth (Figures S1B–S1D and Video S1).

Previous studies have demonstrated that xyloglucan-deficient cells of the xxt1/xxt2 and xxt1/xxt2/xxt5 mutants exhibit alterations in the biochemistry and structure of other cell wall components including pectin levels and cellulose bundling ^34,37,42,43. To characterize cell wall composition in dark-grown Arabidopsis seedlings, we utilized different order cslc mutants. Both the cslc-Δ5 and the xxt1/xxt2 mutants lacked detectable fucosylated xyloglucan epitopes in the cell wall fractions as assayed by immunolabeling with CCRC-M1 antibody that binds to fucosylated xyloglucan ⁴⁴ (Figures 1A and 1B).

Figure 1. cslc mutations confer a decrease in xyloglucan and an increase in pectin.

(A) Immunolabelling of dot blots of xyloglucan fraction 1 (XYGL-I) extracted with 0.67 M KOH and fraction 2 (XYGL-II) extracted with 4 M KOH, prepared from 3 days post germination (DPG) dark grown seedlings with CCRC-M1 (fucosylated xyloglucan) antibody. (B) Quantification of the immunolabelling signal. (C) Quantification of uronic acid in 3DPG dark grown seedlings. (D) Immunolabelling with LM19 (de-esterified HG) and LM20 (esterified HG) antibodies of plastic embedded sections from 24HPG dark grown hypocotyls at the level of the third cell from the bottom. The inset represents the intensity on these false colored micrographs. (E) Quantification of the immunolabelling signal. In B and C, bars represent means ± standard error of the mean (SEM) from three biological replicates of at least 200 seedlings for each. In E and F, the horizontal lines represent the median and the boxes represent the interquartile range. Data points represent individual cells from sections coming from three different hypocotyls for each. Asterisks represent a significant difference with p < 0.05. Scale bars = 50 µm. See also Figure S1 and Video S1.

Cell wall extracted xyloglucan is composed of two fractions. Xyloglucan fraction I (XYGL-I) is the fraction easily extractable from the cell wall with low concentration of KOH (0.67 M) whereas fraction II (XYGL-II) is the strongly bound and protected fraction of xyloglucan requiring high concentration of KOH (4 M) to separate it from the cellulosic components of the wall. Loss of activity of CSLC4, the most highly expressed CSLC in vegetative tissues ²⁶, was associated with a 70% decrease in XYGL-I and a 30% decrease in XYGL-II. XYGL-I in the cslc4/cslc5 double mutant was down to only 5% of the wild-type (WT) level, whereas XYGL-II was at 41% of the WT level (Figures 1A and 1B).

Using a biochemical approach, we observed an increase in the relative amount of uronic acid (UA), a major pectin component of the Arabidopsis primary cell wall, as the cslc mutation order increased (Figure 1C). The cslc-Δ5 and xxt1/xxt2 mutants had an approximately 47% higher ratio of UA to alcohol insoluble residue (AIR) than the WT seedlings, while the cslc4 and cslc4/cslc5 seedlings exhibited intermediate levels per unit of AIR weight extracted (Figures 1C and S1E). Immunolabeling on plastic-embedded sections revealed an increase in both esterified and de-esterified homogalacturonan (HG) in the cslc-Δ5 line (Figures 1D and 1E), consistent with prior findings with xxt1/xxt2 mutants ³⁴. We also noticed that the ratio of esterified to de-esterified pectin was conserved between WT and cslc-Δ5 samples (Figure S1F) indicating that the total amount of pectin per unit of AIR weight extracted and not the degree of esterification changed in the mutant. We further confirmed the conservation of the degree of esterification using a biochemical method that showed that the amount of methanol released by saponification per unit of UA was similar in the WT and the quintuple mutant (Figure S1G).

When using a potassium iodide quantification approach, a method traditionally used to quantify xyloglucan ^45,46, we detected a substantial signal even in the cslc-Δ5 mutant seedlings (Figure S2A). This method is not specific for xyloglucan, however, as it can also detect other polysaccharides with glucose backbones ⁴⁷. Further investigation, coupled with the elimination of the possibility of the signal coming from starch since our cell wall extracts were heated, cooled and then treated with α-amylase, led us to suspect that the signal originated from globular or non-crystalline cellulose extracted in the soluble xyloglucan fraction. We quantified the amount of crystalline cellulose remaining in the AIR fraction of the cell wall after amylase treatment, pectin, and xyloglucan extraction. We observed a 20% decrease in crystalline cellulose in the cslc-Δ5 line compared to the WT (Figure 2A). Additionally, using bacterial cellulose binding modules CBM-2a, with high affinity to crystalline cellulose and CBM-28 which binds to amorphous cellulose ⁴⁸, we found that, on hypocotyl cross sections, CBM-2a showed more binding to the WT cell wall, while CBM-28 bound to more epitopes in the cslc-Δ5 line (Figures 2B and 2C). Since the lack of xyloglucan induces a relative increase in pectin, we wanted to exclude the possibility of epitope masking. We extracted the pectin fraction from the AIR and proceeded to probe the cellulosic fraction with the CBMs. Consistent with our CBM labeling of hypocotyl cross sections, we found that the quintuple mutant exhibits a decrease in crystalline cellulose and an increase in amorphous cellulose (Figure S2B–D).

Figure 2. Xyloglucan deficient seedlings have reduced cellulose crystallinity.

(A) Cellulose quantification from alcohol insoluble residue (AIR). (B) Hypocotyl sections probed with CBM2a (crystalline cellulose) and CBM28 (amorphous cellulose) and false colored for signal intensity. The inset represents the intensity of the signal. (C) Signal intensity of hypocotyl sections tagged with CBM2a and CBM28. (D) Representative images of 3DPG seedlings treated with isoxaben. (E) Length inhibition and (F) swelling of the hypocotyls of 3DPG dark grown seedlings treated with isoxaben. In A, horizontal bars represent the median and data points are average signal from cells coming from a minimum of 8 seedlings from each and in 3 biological replicates. In C, E and F, the horizontal lines represent the median, boxes represent the interquartile range, horizontal bars represent the median, and data come from 3 biological replicates. Asterisks represent a significant difference with p < 0.05. Scale bars = 50 µm in C and 1 mm in E. See also Figure S2.

To assess the effect of reduced crystalline cellulose, we grew seedlings on media containing a low concentration of isoxaben (0.5 nM), which would slightly perturb cellulose biosynthesis. While this low dose induced only a 7% reduction in hypocotyl length and a 13% increase in diameter in WT seedlings, the effects on the growth anisotropy of the quintuple mutant were more dramatic, with a 31% decrease in length and a 30% increase in diameter compared to the mock treatment (Figures 2D–F, S2E and S2F). This indicates that the cslc-Δ5 mutant has compromised wall integrity and is hypersensitive to further reductions in cell wall cellulose content.

Taken together, these data emphasize the importance of xyloglucan in maintaining cellulose crystallinity and suggest the existence of a mechanism through which cells increase pectin levels when xyloglucan levels in the cell wall (or cellulose crystallinity) are reduced.

cslc mutations affect the elasticity and strength of the cell wall.

Xyloglucan is believed to tether cellulose microfibrils and create bundling hotspots that hypothetically could influence the mechanical properties of the cell wall ³⁰. Given that cslc mutations lead to an increase in HG, a pectin component involved in the mechanical properties of hypocotyl cell walls ^12,21,49, and a reduction in cellulose crystallinity, we sought to investigate the mechanical properties of the wall and the effects of these biochemical and structural modifications on wall elasticity, viscosity and strength.

Using Atomic Force Microscopy (AFM)-based microindentations of plasmolyzed hypocotyl cells, we observed a decrease in the elastic modulus of the cell walls in cslc-Δ5 compared to the WT. The indentation moduli (IM) of both the axial and transverse walls were approximately 54% lower in the quintuple mutant compared to the WT (Figures 3A and 3B), while the ratio of IM of the axial to the transverse walls, which represents the elastic asymmetry in the hypocotyl cylindrical cells, was maintained (IM transverse/axial walls = 1.3 in both lines).

Figure 3. cslc-Δ5 seedlings have reduced elastic modulus and are weaker than the WT.

(A) Representative micrographs from atomic force microscopy (AFM) scans of 24HPG dark grown seedlings. The arrow points at the axial wall and the arrowhead points at the transverse wall. (B) Plots of the indentation modulus (IM) of the AFM processed images. (C) Hypocotyl elastic modulus of 48HPG dark grown seedlings stretched using an automated confocal micro-extensometer (ACME) at 5 mN (left) and 10 mN (right). (D) Break strength of 48HPG dark grown seedlings. In B and D, horizontal lines represent the median, boxes represent the interquartile range. Data come from 3 biological replicates and at least 8 seedlings from each line in B and 10 in D. In C, horizontal bars represent the median and data come from 3 biological replicates. Asterisks represent a significant difference with p < 0.05. Scale bar = 20 µm. See also Figure S3.

Employing a microextensometer-based approach to measure the elastic modulus at the tissue level ⁵⁰, we found that cslc-Δ5 exhibited 32% and 31% lower modulus than the WT when subjected to forces of 5 mN and 10 mN, respectively (Figures 3C, S3C and S3D). We also observed an increase in the elastic modulus by 16% and 17% in the WT and the mutant, respectively, when subjected to a 10 mN force, indicating the possibility of strain stiffening in the cell wall although this difference was only statistically significant in the mutant but not in the WT.

The plant cell wall is considered a viscoelastic material ⁵¹. Although the elastic properties have been extensively studied, little is known about the viscous component and the contribution of cell wall constituents or their interactions to it. We applied a constant force with the AFM tip and extracted the decay coefficient using three independent modeling approaches, an exponential model, a visco-elastic model and a power law model (refer to the methods section for details).

To validate the fit, we analyzed the residuals or goodness of the fit, which are inversely proportional to the accuracy of the fit. All models showed very low residuals, with the exponential and linear viscoelastic models having slightly better fits than the power law model (Figure S3B).

None of the modeling approaches revealed any difference in viscosity between the mutant and WT seedlings (Figure S3A), indicating that the reduction in crystalline cellulose content or hemicellulose does not significantly contribute to cell wall viscosity in these lines. And since the ratio of esterified/de-esterified HG remained constant in the quintuple mutant (Figure S1F), it is plausible that the balance of stiff to soft pectin is the key factor controlling viscosity in the plant cell wall while cellulose and xyloglucan have a higher contribution to the elastic properties.

To further characterize the mechanical properties of the cslc-Δ5 hypocotyls, we measured the break strength which is the tensile force a hypocotyl can withstand before rupturing. We found that cslc-Δ5 hypocotyls were significantly weaker than the WT (Figure 3D).

cslc mutants have reduced turgor pressure.

Plant cell growth is governed by the interplay between turgor pressure pushing on the cell wall and the cell wall’s resistance/yielding to this pressure ⁴. Despite significant differences in cell wall composition and a substantial reduction in elastic modulus and rupture stress observed in the cslc-Δ5 mutant compared to the WT, the growth of the mutant did not reflect these disparities (Figures S1A–S1C). Under such conditions, if turgor pressure remained constant, we would expect the mutant to show either increased volumetric growth through more elongation and/or swelling or with possible cell bursting or delamination if the wall structural integrity does not resist the pressure, neither of which occurred. This led us to hypothesize that the turgor pressure in the cslc mutant lines might be lower than in the WT.

When dark-grown seedlings were treated with a mannitol solution, we observed that the xyloglucan-deficient lines started to plasmolyze earlier and exhibited more extensive plasmolysis compared to the WT under the same osmoticum concentration (Figure 4A and Videos S2–S4). Additionally, the plasmolysis began in the bottom cells of the hypocotyl of all the lines but extended higher in the double mutant and even higher and more dramatically in the quintuple mutant line (Figure 4B and Videos S2–S4). To further investigate this, we employed a microextensometer to measure the forces induced by plasmolysis in light-grown seedlings. After achieving an osmotic balance in the liquid growth media, we applied a high concentration of NaCl to induce an almost instantaneous and complete plasmolysis (Figure S4C). We then extracted the forces and calculated the turgor pressure loss resulting from the treatment (see methods). Our findings revealed that the turgor pressure was approximately 40% lower in the quintuple mutant compared to WT (Figures 4C, S4A and S4B). These data indicate that the absence of xyloglucan triggers changes in the cell wall that lead to a decrease in turgor pressure.

Figure 4. cslc mutant hypocotyls have reduced turgor pressure.

(A) 2-minute interval time point images from Col-0, cslc4/cslc5 and cslc-Δ5 hypocotyl cells response to treatment with 600 mM mannitol. Red circles represent the onset of plasmolysis. (B) Hypocotyl plasmolysis 18 minutes after the beginning of imaging. Arrowheads indicate the highest point in the hypocotyl where plasmolysis can be observed. The bottom panels represent zoomed images from the top part of the hypocotyls 18 minutes after the beginning of imaging (red boxes). (C) ACME-based quantification of turgor pressure at the hypocotyl level in 2DPG light grown seedlings. Horizontal bars in C represent the median. Scale bars = 50 µm in A and 100 µm in B. See also Figure S4 and Videos S2–4.

We then wanted to determine the effects of these changes in wall composition and turgor on plant behavior in a mechanically challenging environment. To do so, we tested the seedling emergence when the seeds were covered with a layer of sand either 3 mm or 5 mm thick (Figure 5A). We noticed that under a thin layer of sand, 80% of the WT seedlings emerged whereas only 38% of the quintuple mutants managed to do so. When the physical challenge was increased, 52% of the WT seedlings managed to emerge and only 8% of the mutants were capable of breaking through the layer of sand (Figures 5A and 5B). We also noticed that the quintuple mutant has a modest apical hook defect (Video S1). When we measured the hook angle at 48HPG, we found that the WT had a 165° hook angle whereas the mutant hook was 122° on average (Figure S5). Since hook angle affects seedling emergence ^52,53, this complicates interpretation of the above emergence experiments. Therefore, as an alternative means of mechanically challenging young seedlings, we tested the ability of roots to penetrate stiff media. While 52% of the WT roots were able transition from soft (0.6% agar) to hard (1.2% agar) media, only 24% of the cslc-Δ5 roots managed to penetrate the harder media (Figures 5C and 5D). These data highlight the importance of xyloglucan and presumably its effect on turgor pressure and cell wall mechanics during seedling emergence and root growth in the soil.

Figure 5. Lack of xyloglucan affects seedling emergence and root penetration.

(A) Representative images of Col-0 and cslc-Δ5 seedlings emergence from a sand layer of 3mm or 5mm thick after 3 days in the dark and 2 days in the light for greening. (B) Percentage of seedling emergence. (C) Representative images of seedlings germinating on a 5mm thick layer of 0.6% Agargel^™ media before reaching the growth media containing 1.2% Agargel^™. Arrowhead points at an emerged root and arrow points at a root that failed to penetrate the stiffer media. Data were generated from 3 biological replicates with 15 or 20 seedlings in each for seedling emergence and 21 to 27 for root penetration. Bars represent the median in B and D. Asterisks represent a significant difference with p<0.05. Scale bar = 1 cm in A and 5 mm in C. See also Figure S5.

Discussion

Plant cell walls play essential roles in plant growth, structure, defense, and the provision of metabolizable energy and signaling molecules ^54,55. Over time, our understanding of the relationship between cell wall structure and function has progressed from conceptual to more concrete, thanks to the development of sophisticated imaging and modeling tools ⁵⁶. Despite these advances, numerous crucial questions concerning the contributions and interactions between cell wall components remain unanswered. Particularly intriguing is the role of the major wall component xyloglucan, as its absence from the cell wall has been shown to have limited effects on plant growth and development ^{26,28,30,34,35}.

Our findings indicated that pectin levels in Arabidopsis seedlings were inversely proportional to xyloglucan content (Figure 1). This suggests that plants can sense the absence of xyloglucan at some molecular or mechanical level and respond by increasing the relative amount of pectin in the wall through either an increase in HG production and likely rhamnogalacturonan or a reduction in pectin recycling to maintain higher pectin levels in the cell wall. The latter hypothesis may be more likely since Kim and coworkers ²⁶ did not find a significant increase in the transcripts of genes involved in pectin biosynthesis in the cslc-Δ5 mutant. Sowinski and coworkers ³⁴ also demonstrated that xxt1/xxt2 mutants had increased HG and glucomannan (GM), a hemicellulose that resembles xyloglucan in structure and function but is less abundant in dicot walls ³⁷. However, Yu et al. (2022) did not observe an increase in the abundance of glucomannan in the xxt1/xxt2 mutant. The question of whether glucomannan is involved in complementing the xyloglucan defect in the cslc-Δ5 mutant and at which developmental stage this might occur raises interesting points for further exploration. Notably, no changes in HG were observed in the meristem of xxt1/xxt2 compared to the WT ⁴³, indicating that the changes might be tissue-specific or related to the required degree of cell anisotropy, the intensity of pressure governing growth in specific cell types, or the amount of secondary cell wall present in the tissue.

The mechanical properties of cellulose are influenced by its crystallinity, with higher crystallinity positively correlated with increased stiffness ¹⁴. We found a reduction in cellulose crystallinity and an increase in amorphous cellulose in the cell wall of cslc-Δ5 (Figures 2 and S2). This structural change in the load-bearing material is likely to substantially impact the mechanical properties of the cell wall, as we have shown (Figure 3). Changes in cellulose organization and content were also observed in the xxt1/xxt2 mutant ^42,43. It is possible that xyloglucan is either necessary for the formation of crystalline cellulose during the extension of the nascent microfibrils or that its absence exposes more of the amorphous cellulose since solid-state NMR analysis on the cell wall of xxt1/xxt2/xxt5 showed more dynamic cellulose compared to WT ⁵⁷. It was previously found that cellulose content could be affected by the level of crystallinity since cellulose with lower degree of crystallinity was found to be more susceptible to enzymatic hydrolysis by cellulases ⁵⁸. One other possibility is that in the absence of XYGL, members of the endotransglucosylase/hydrolase (XTH) family such as XTH3 capable of generating cellulosic insoluble material ⁵⁹ might have an effect on cellulose and its crystallinity. Proteins from another member of the CSL family, CSLD were shown to display a UDP-glucose β-1,4-glucan synthase activity and to be able to complement the catalytic activity of the membrane localized CESA. Although most of the csld mutants display phenotypes in tip growing cells, some of the higher order mutants display hypocotyl phenotypes ^60,61. Despite no significant changes in the CSLD transcript levels between cslc-Δ5 and Col seedlings ²⁶ and the fact that CSLD movements are insensitive to isoxaben ⁶², the possibility that enzyme activity of members of CSLD family is elevated in cslc-Δ5 and thereby increasing the cell wall amorphous cellulose pool, cannot be totally discarded.

Using AFM-based approaches to evaluate the mechanical properties of the epidermis and a microextensometer-based approach that extracts tissue-level elasticity, we found that the cslc-Δ5 mutant was softer than the WT (Figure 3). This was expected since xxt1/xxt2 mutants were previously found to have more extensible cell walls using extensometer-based approaches ^34,36,38. We also observed that an elastic asymmetry, by which the ratio of elasticity of the axial to the transverse walls, was conserved, which explains the similar elongation pattern of the mutant hypocotyls compared to the WT. Additionally, the esterification level of pectin was maintained between the quintuple mutant and the WT (Figure S1F and S1G). This esterification level was shown to be involved in maintaining growth anisotropy during germination and hypocotyl elongation ¹² and could be the reason behind the conservation of elastic asymmetry in these xyloglucan-lacking mutants. The difference in the absolute elastic modulus values between the AFM and ACME-based measurements could be attributed to the fact that with AFM, we are extracting the mechanical properties of the epidermis whereas with the extensometer we are pulling on the whole tissue including the vasculature that contains secondary wall formations. Another possibility is that extensometer-based forces are applied parallel to the hypocotyl axis whereas AFM-based indentations are perpendicular to it.

We found that cslc mutant plants have reduced turgor pressure proportional to the levels of cellular xyloglucan, an indication that, if reduced xyloglucan leads to weaker walls, a feedback mechanism linking xylogucan levels and internal cellular pressure could prevent the mutant wall from exceeding its rupture stress. In fact, xyloglucan chemistry as well as cellulose content were shown to affect the rupture stress of Arabidopsis hypocotyls ⁶³. We also noticed that seedlings lacking xyloglucan have hypocotyls 44% weaker than the WT plants (Figure 3D).

The difference in turgor pressure between the cslc lines and the WT brings another important point into consideration; the effect of turgor pressure on the properties of the wall. The possibility that cell wall extensibility or the yield threshold could be variables and depend on the pressure should not be excluded ⁶⁴. This means that the mechanical properties are the result of not only the biochemistry but also the tension/compression generated by turgor in a complex multicellular system. The xyloglucan-lacking mutants represent an excellent tool to further investigate these hypotheses by applying treatments or using growth conditions that are known to affect turgor and then analyze the mechanical and structural properties of the walls.

Pectin was previously shown to affect seedling emergence in the context of hook formation and ethylene signaling ^53,65. In our study, we observed a pronounced adverse effect on seedling emergence and root capacity to penetrate mechanically challenging environments stemming from a reduction in xyloglucan content within the cell wall matrix (Figure 5). These findings hold particular relevance in the context of contemporary challenges associated with global shifts in temperature, precipitation patterns, soil nutrient levels, and evolving cultivation techniques. Factors such as soil compaction, the formation of soil crusts, and variations in nutrient availability have all emerged as critical issues, with demonstrably detrimental impacts on seedling emergence ^66–70. Understanding the multifaceted determinants of seedling emergence and early seedling establishment and leveraging this knowledge to enhance crop performance and refine agricultural practices offer promising avenues to mitigate the losses incurred due to suboptimal seedling establishment.

No differentially expressed genes related to turgor control or drought related stress appeared in the cslc-Δ5 mutant RNAseq data ²⁶. In fact, only around 60 genes were shown to be differentially expressed highlighting the fact that plants could cope with the complete loss of xyloglucan without undergoing dramatic transcriptional reprogramming. This also indicates that the modifications leading to the changes in wall composition and turgor in the xyloglucan-lacking lines are most likely posttranslational and involve signaling pathways through proteins that can either sense cell wall integrity and/or the mechanical tension in the wall or the plasma membrane.

Plasma membrane-localized protein kinases with potential carbohydrate-binding domains have been reported in plants ^71–74. Some of these kinases have been shown to interact physically with cell wall carbohydrates ^71,75–78, making them potential sensors of wall integrity that could feed into downstream regulators to control or modulate physiological responses, including turgor pressure, to protect cell integrity. For instance, the theseus1 mutant (the1) was able to attenuate the defects caused by a reduction in cellulose biosynthesis ⁷³ and has been implicated in maintaining cell wall integrity together with the mechanosensitive channel MCA1 in an osmosensitive manner ⁴¹. Recently, the perception of RALF1 (rapid alkalinization factor 1), a hormone peptide that binds to FERONIA ⁷⁹, was shown to depend on the level of de-esterified HG and subsequently affecting FERONIA internalization ⁸⁰. RALF4 was also found to play an important role in the structure of the pollen tube cell wall by interacting with the negatively charged de-esterified HG of the wall ⁸¹. Therefore, the mutants lacking xyloglucan and having increased levels of de-esterified HG could serve as tools to test if and how these potential sensors contribute to the modulation of cell wall integrity and turgor pressure to protect the cells and promote normal plant growth and development.

Overall, our study sheds light on the complex relationship between xyloglucan, cell wall structure, mechanical properties, turgor pressure and plant development, uncovering insights into the interplay of cell wall components and turgor pressure and their roles in growth regulation. Further investigations are warranted to unravel the molecular mechanisms underlying these observations, potentially leading to a deeper understanding of plant cell wall biology and its implications for developmental processes. Our findings also hold the potential for biotechnological applications in agriculture and plant engineering.

Star Methods

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Firas Bou Daher (boudaher@umn.edu).

Materials availability

The materials generated in this study are available upon request.

Data and code availability

Original data have been deposited in the University Digital Conservancy and are available at https://conservancy.umn.edu/handle/11299/166578/. Accession number is listed in the Key resource table.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
LM19	Plant Probes, UK	ID_PlantProbes: LM19; RRID: AB_2734788
LM20	Plant Probes, UK	ID_PlantProbes: LM19; RRID: AB_2734788
CCRC-M1	CarboSource Services, University of Georgia	NA
SpectraDye, goat-anti-mouse IgG IR800	Advansta	R-05061-250
Alexa Fluor 647 goat anti-rat IgG	Life Technologies	A21247
Bacterial and virus strains
BL21 (DE3) E. coli	NA	NA
Biological samples
Col-0	NASC	N60000
cslc4	Arabidopsis Biological Resource Center	SALK_083490
cslc5	Arabidopsis Biological Resource Center	CS490585
cslc4,5,6,8,12	Arabidopsis Biological Resource Center	CS72438
cslc4/cslc5	This manuscript
cslc4/cslc5-MYR-YFP	This manuscript
cslc4,5,6,8,12-MYR-YFP	This manuscript
MYR-YFP	Willis et al 82
xxt1/xxt2	Cavalier et al 36
Chemicals, peptides, and recombinant proteins
CBM2a-mGFP	Roberts et al 48	NA
CBM28-mNeptune	Roberts et al 48	NA
Isoxaben	Sigma-Aldrich	36138
Citifluor CFM-3	Electron Microscopy Sciences	NA
formaldehyde	Acros	11969
sucrose	Macron	8360-06
sucrose	Sigma-Aldrich	S9378
Murashige & Skoog	Duchefa Biochemie	M0221.0050
Murashige & Skoog	PhytoTech Labs	M524
Agargel™	Sigma-Aldrich	A3301
plant agar	Duchefa Biochemie	P1001
KOH	Fisher Chemical	P250-1
LR White	SPI supplies	02646-AB
α-amylase	Sigma-Aldrich	10065
Alcohol oxidase	Sigma-Aldrich	A2404
Sulfamic acid	Sigma-Aldrich	242772
Sodium tetraborate	Alfa Aesar	40114
Sulfuric acid	Sigma-Aldrich	258105
3-Phenylphenol	Sigma-Aldrich	262250
Galacturonic acid	Sigma-Aldrich	48280
bovine serum albumin	Sigma-Aldrich	2153
Critical commercial assays
Deposited data
Original data	University of Minnesota Digital Conservancy	https://conservancy.umn.edu/handle/11299/166578/
Experimental models: Cell lines
Experimental models: Organisms/strains
Oligonucleotides
Recombinant DNA
Software and algorithms
jpkfile	https://gitlab.gwdg.de/ikuhlem/jpkfile
scipy.optimize	https://docs.scipy.org/doc/scipy/reference/optimize.html
Other
Tough Tags tape	Diversified Biotech	TTSW-1000

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

This paper does not report original code.

Experimental model and study participant details

All Arabidopsis lines used in this study are in the Col-0 background. cslc4 (SALK_083490), cslc5 (GABI_944E09) and cslc4,5,6,8,12 quintuple (CS72438) seeds were acquired from the Arabidopsis Biological Resource Center. cslc4/cslc5 was generated by crossing the cslc4 and cslc5 mutants. The mutant lines with membrane markers were generated by crosses with the MYR-YFP line ⁸². xxt1/xxt2 line was previously described in Cavalier et al (2008).

Dark grown seedlings were grown in foil double-wrapped petri dishes at 22°C, and light grown seedlings were grown in plates at 22°C with 16H light/8H dark cycles and under ~ 70 μEm^-2s^-1 fluorescent lighting. All plates were grown vertically.

Seeds were surface sterilized with 70% EtOH for 5 minutes followed but a quick wash with 100% EtOH and left to dry on sterile filter paper in a flow hood prior to sowing.

Sterile seeds were sown on half-strength Murashige & Skoog (MS, PhytoTech Labs, M524) plates containing 0.6% Agargel^™ (Sigma-Aldrich, A3301). Media pH was adjusted to 5.7 using KOH (Fisher Chemical, P250-1). After sowing, the seeds were left at 4°C for 2–3 days in darkness, then transferred to white light to stimulate germination, and subsequently to the experimental growth media and conditions. Germination was defined as the time when the radicle broke through the endosperm and testa (0HPG). At this time, seedlings were selected and aligned vertically on ½ MS plates containing 1% sucrose (Macron, 8360–06) and 0.6% Agargel^™ (pH 5.7).

For micro-extensometer and AFM experiments, Murashige & Skoog (Duchefa Biochemie M0221.0050), sucrose (Sigma S9378) and plant agar (Duchefa Biochemie, P1001) were used.

Method details

Probing hypocotyls with carbohydrate binding modules.

Carbohydrate binding modules CBM2a-mGFP with high affinity for crystalline cellulose and CBM28-mNeptune with high affinity for amorphous cellulose were previously described ⁴⁸ and were kindly provided by Professor Nigel Scrutton. They were introduced into BL21 (DE3) E. coli, cultured at 37°C to an optical density of 0.4, and then induced with 1mM IPTG for four hours prior to protein extraction.

Bacteria were lysed using a French press and protein extract was used to incubate hand sectioned 24HPG dark-grown hypocotyls at the level of the third cell for three hours. Samples were then washed four times with phosphate buffer saline (PBS) solution and mounted on an agar pad to expose the cut area for confocal imaging.

Thin section preparation

Seedlings were fixed in PBS solution containing 2% formaldehyde (Acros 11969) and 2.5% glutaraldehyde (Sigma-Aldrich, G6257) under vacuum for one hour and left at 4°C overnight. Samples were washed twice in PBS then dehydrated in a series of solutions with increasing ethanol concentrations. Samples were embedded in medium grade LR White (London Resin Company) which was left to polymerize at 60°C overnight. 0.5 µm thin sections were generated using glass knives on a Leica UltraCut microtome.

Immunolabelling

Sections were blocked in PBS solution containing 2% bovine serum albumin (BSA) (Sigma-Aldrich, 2153) for one hour. The primary antibodies (LM20 and LM19, Plant Probes) were diluted 200 times in PBS-BSA and left on the samples for 3 hours. Samples were then washed four times for five minutes each in PBS-BSA and incubated for two hours in Alexa Fluor 647 goat anti-rat IgG (Life technologies A21247) which was diluted 400 times in PBS-BSA. Samples were then washed six times for 5 minutes each in PBS and mounted in citifluor CFM-3 (Electron Microscopy Sciences). Slides were sealed with nail polish, kept in the dark and imaged the same day.

Cell wall cross section area

To quantify the surface area of the cell wall on a cross section, thin sections were generated from the central area used to conduct the microextensometer experiments. Sections were stained with propidium iodide and imaged immediately on a confocal microscope. Using ImageJ, background signal was subtracted before applying a Gausian filter and thresholding all the images with identical settings. Area was calculated from the pixels occupying the cell wall.

Isoxaben treatment

10 mM stock solution of isoxaben (Sigma-Aldrich, 36138) was prepared in DMSO. This solution was further diluted 1000 times in water and the resulting 10 µM stock solution was used to prepare the media. 0HPG seedlings were transferred to standard growth media containing 0.5 nM isoxaben, grown in the dark for 72 hours prior to imaging.

Confocal imaging

Sections with pectin immunolabelling were imaged with a Leica SP8-iPhox inverted DMI8 microscope (Leica Microsystems, Wetzlar, Germany). A 20X, 0.75NA objective was used. The fluorophore was excited with a 552 nm laser, and the signal was collected between 570 and 700 nm on a hybrid detection (HyD), single molecule detection (SMD) detector.

Imaging of the plasmolysis experiment was conducted on a Leica SP8-iPhox inverted DMI8 microscope (Leica Microsystems, Wetzlar, Germany). A 10X, 0.4NA objective was used. A 514 nm laser was used to excite the membrane marker and the signal was collected between 520 and 570 nm with a photomultiplier tube (PMT) detector. And on a Nikon AX R using a 10X, 0.45 NA objective and a 514 nm laser excitation and 534–558 nm emission detection.

CBM imaging was conducted on a Nikon AX R confocal microscope set on a Nikon NiE upright microscope. A 20X, 0.7NA objective was used. For CBM2a a 488 nm excitation laser was used, and signal was collected between 500 and 550 nm. For CBM28 a 561 nm excitation laser was used, and signal was collected between 581 and 634 nm with a PMT detector.

CBM treated dots were imaged with a Nikon A1plus confocal microscope set on a Ti2 inverted microscope using a 4X (0.2NA) objective. For CBM2a a 488 nm excitation laser was used, and signal was collected between 500 and 550 nm. For CBM28 a 562 nm excitation laser was used, and signal was collected between 575 and 625 nm with a PMT detector.

Propidium iodide-stained sections were imaged with a Nikon A1plus confocal microscope set on a Ti2 inverted microscope. A 20X 0.75NA objective was used. The fluorophore was excited with a 562 nm laser, and the signal was collected between 575 and 625 nm with a PMT detector.

Image processing and analysis

Confocal images were processed using ImageJ (NIH, Bethesda, MD, USA). Maximum projection images were generated, and background noise was subtracted using a 50-pixel rolling ball radius. For immunolabelling images, a 16-color look up table was applied. A 3-pixel wide line was used in ImageJ to plot the fluorescence intensity profile on the walls of the epidermal cells. For processing images to measure cell wall area on a cross section, after background removal, a Gaussian blur was applied with a 2.0 sigma radius. A threshold image was generated, and the area covered by the wall was calculated. For cellulose dot blots probed with CBMs, maximum projection images of Z-stacks were generated, background noise was subtracted using a 50-pixel rolling ball radius, and a Gaussian blur was applied with a 1.0 sigma radius. All data were then exported to and analyzed in Microsoft Excel.

Atomic force microscopy (AFM)

24HPG seedlings were quickly fixed to a double-sided tape and plasmolyzed with 0.6 M mannitol solution 15 minutes prior to imaging. Hypocotyls were indented at the level of the third cell from the bottom with a Nano Wizard 3 AFM (JPK Instruments, DE). A spherical tip (Biosphere B500-NCH, HighDensityCarbon) with 500 nm radius, on a 125 µm length cantilever with an approximate 40 N/m spring constant and a resonance of 330 KHz was used for indentation. It was calibrated using thermal tuning. Indentations were performed with a 500 nN force at 100 µm/s speed which generated an indentation depth of about 300 nm. Data were collected from a 100x50 µm area at a resolution of 64x32 points. For viscosity data collection, three points on the axial wall and two on the transverse wall were poked with 500 nN and at 10 µm/s indentation speed and each held for 5 seconds.

For extracting quantitative visco-elastic parameters from dynamic AFM data, raw data in the form of voltage vs. time curves were converted to force vs. time curves using the optical lever sensitivity and the spring constant of the indenter. These force indentation curves were fitted to a number of viscoelastic mechanical models using linear regression. Values for time constants and goodness of fit were extracted for comparison between the different genotypes.

We chose three different dynamic mechanical models to try and extract relevant viscoelastic parameters from the data, an exponential decay in force against time ⁸³, a linear viscoelastic model ⁸⁴ and a power law formulation ⁸⁵ a goodness of fit value was also extracted for each regression to give an indication of model suitability.

$$a{e}^{\frac{-\left(x-d\right)}{b}}+c$$

In the exponential decay model, the parameter “$\text{b}$“ gives a characteristic relaxation time.

$$a\left(1-b{e}^{-cx}\right)$$

In the linear viscoelastic model, the parameter “$\text{c}$“ gives a characteristic relaxation rate.

$$\left\{\begin{array}{l}\hspace{1em}{b}_1{x}^{a_1},\hspace{1em}\hspace{0.33em}\hspace{0.33em}\hspace{0.33em}\hspace{0.33em}\text{if}\hspace{0.33em}x<x_{th}\hfill \\ b_2{x}_{th}^{a_1-a_2}{x}^{a_2},\hspace{0.33em}\hspace{0.33em}\hspace{0.33em}\text{if}\hspace{0.33em}x>x_{th}\hfill \end{array}\right.$$

In the power law formulation, the parameter “${\text{a}}_1$“ gives the power of the initial relaxation phase.

Data was parsed using the jpkfile package (https://gitlab.gwdg.de/ikuhlem/jpkfile) and fitted using the scipy.optimize package in Python 3.10.11 (https://docs.scipy.org/doc/scipy/reference/optimize.html).

Mechanical measurements with ACME

Setup, hardware and software for the automated confocal micro-extensometer (ACME) has been previously described in detail ⁵⁰.

For sample mounting, entire seedlings were attached to the extensometer’s arms without damaging the tissue using Tough Tags tape (Diversified Biotech, TTSW-1000). For consistency, extra care was taken to ensure that a similar area of hypocotyls was located in the 1 mm gap between the two arms of the extensometer. For measurement of hypocotyls cell wall extensibility, samples were flash-frozen in liquid nitrogen, thawed on ice, and attached to the tape with the addition of a thin layer of cyanoacrylate glue. Extensometer experiments were conducted in distilled water to prevent sample drying. Samples were held at 0 force for 60 seconds, then 20 oscillations of stretching and relaxation were made (5 mN for 1 second, 0 mN for 1 second, 20 times) the samples were then held at 0 force for 60 seconds, and new batch of oscillations was performed at 10 mN. To assess potential slippage during the experiments, a movie was recorded with a Leica SP8 equipped with a HCX APO L 20×/0.5-W objective in a single z-plane using the bright-field detector for rapid imaging. The size of the gap and the force applied were recorded in a csv file for every timepoint. The average strain for each batch of oscillation was then calculated using R script previously described ⁵⁰.

We measured the area occupied by the cell wall on cross sections of the hypocotyl at the level of the stretched region to calculate the elastic modulus (EM). The EM was calculated as follows: EM = Stress/Strain. Stress = Force/Area. The Force in Newtons, the cell wall area in mm² and the strain in mm.

Turgor pressure assessment was performed using live, two-day old hypocotyls of light grown seedlings. For a detailed movie of sample mounting see the supplemental data of ⁵⁰. Samples were attached to the extensometer and held at 0 position in 50 ml liquid media while monitoring the force exerted by the sample. Once the force has reached a plateau, 25 mL of 20% NaCl was added to the media which was sufficient to trigger instantaneous plasmolysis of the sample. The force was still monitored, and the experiment stopped when it reached a new lower plateau. The recorded force was exported as a csv file and used to extract the differential of force between the turgid and plasmolyzed state of the sample. To calculate the turgor pressure, we measured the area occupied by the cell wall in the middle of the hypocotyl. The pressure was calculated as follows: Pressure=(Force/Area) * 0.00981. The pressure is in MPa, the force in grams, the cell wall area in mm², and 0.00981 is used to convert gram-force to Newtons.

Break strength testing

48HPG dark grown seedlings were frozen in liquid nitrogen and stored at −80°C. Seedlings were thawed on ice and each end was sandwiched between two tiny pieces of paper glued with cyanoacrylate glue exposing 1 mm of the hypocotyl between the two sandwiches which were attached to the ends of a 15 lb braided line (Kastking Destron) made of ultra-high molecular weight polyethylene (UHMWPE) material. The bottom line was attached to a 3.5 g lead weight and placed on a balance and the top end was hand pulled to determine the hypocotyl break point. The balance display was recorded with a phone camera to determine the exact rupture weight for each seedling.

Alcohol insoluble residue (AIR) and cell wall analysis

The tissue was ground in a glass mortar in 70% EtOH. Samples were then incubated in solutions of 70% EtOH, 100% EtOH, methanol:chloroform (50%:50%) and finally with acetone. After each incubation, the samples were spun and resuspended and incubated for one hour. After removing the acetone, samples were left to dry overnight. Samples were then resuspended in water and heated to 70°C for 30 minutes, then left to cool down to room temperature.10 units of α-amylase (Sigma-Aldrich, 10065) per mg AIR were added and samples were incubated for six hours. CDTA was then added to a final concentration of 50 mM and left on a rotator for 4 hours. Samples were then spun, and the supernatant was discarded. Fraction I of xyloglucan was extracted with a solution of 0.67 mM KOH after resuspending the pellet and incubating the extract in that solution overnight. Fraction II of xyloglucan was extracted the same way with a 4 M KOH solution. Xyloglucan solutions were stored at −20°C.

For xyloglucan quantification using the iodine method, 200 µl of xyloglucan extract, 100 µl KI3 (0.5% I2 and 1% KI) and 1 ml Na2SO4 (20%) were mixed. For the blank, 200 µl of water was used instead of the extract. The mixture was incubated for one hour at 4°C wrapped with foil and 30 minutes at room temperature. Absorbance at 640 nm was measured and samples were compared to the WT.

For xyloglucan immunolabelling, xyloglucan extracts were blotted on a nitrocellulose membrane and left to dry overnight. The membrane was blocked with 2% milk in PBS for 45 minutes before applying CCRC-M1 antibody (CarboSource Services, University of Georgia) with specificity for fucosylated xyloglucan ⁴⁴. The primary antibody was diluted 50 times in PBS, applied to the membrane and incubated for 3 hours. The membrane was then washed 4 times with PBS. The SpectraDye, goat-anti-mouse IgG IR800 (Advansta) secondary antibody with 1:10000 dilution in PBS was then applied for one hour. The membrane was then washed 5 times and imaged using a LI-COR (Odyssey) imaging system and the signal quantified using FIJI. All incubations were done with gentle rocking.

For pectin extraction, 100 µl CDTA (50mM) per 1 mg AIR was added and incubated at 70ºC for 30 minutes with occasional vortexing. Samples were then placed at room temperature overnight with continuous shaking. Samples were then centrifuged at 13000 RPM for 15 minutes and the supernatant containing the pectin fraction was recovered.

For uronic acid quantification, a protocol based on Bethke et al ⁸⁶ was used. 36 µl of five times diluted pectin in water were mixed with 4 µl sulfamic acid (4 M) and 200 µl of sulfuric acid containing 120 mM sodium tetraborate in a flat bottom plate, sealed and incubated at 80ºC for one hour. After cooling samples on ice for 5 minutes, absorbance at 525 nm was measured in a plate reader before adding 40 µl of 80% sulfuric acid containing 2 mg/ml 3-phenylphenol. Absorbance was measured again and the difference between the two measurements was used to quantify uronic acid content based on the data from the calibration curve. The calibration curve was generated by measuring the absorbances for a series of galacturonic acid dilutions from 0 to 2 mM.

For biochemical quantification of pectin esterification level, a protocol based on Müller et al ⁸⁷ was used. To release the methanol, 100 µl of two times diluted CDTA extract was treated with 100 µl NaOH (0.5 M), vortexed and left at room temperature for one hour. NaOH was neutralized with 50 µl HCl (1 M). 50 µl of this solution was used to quantify the methanol released. Methanol was oxidized at room temperature with gentle shaking for 15min with 50 µl of solution-A (0.05 units of alcohol oxidase in 20 mM potassium phosphate buffer pH 7.5). The extract was developed at 60 degrees for 15min with 100ul of solution-B (2 M ammonium acetate, 50 mM acetic acid 20 mM acetyl acetone). The mix was left at room temperature for 15 minutes prior to measuring the absorbance at 412nm. A standard curve was generated with a methanol dilution series between 0 and 25 nmol.

To assess cellulose crystallinity, the residual cell wall after pectin extraction was resuspended in water at 50 µl/mg AIR. 1 µl dots of that cellulose-xyloglucan solution were blotted on nitrocellulose membranes and left to dry overnight. Membranes were blocked with 5 % milk in PBS for 45 minutes and incubated with either CBM2a or CBM28 for three hours. One extra membrane with dots was treated with a control solution not containing any CBM. Membranes were placed on glass slides, covered with a coverslip, and imaged on a confocal microscope.

Plasmolysis

24HPG dark grown seedlings were transferred to a 500 µm deep chamber containing 200 µl mannitol (600 mM), immediately covered with a cover slip and placed on the microscope. Imaging started exactly 1 minute after mounting the samples.

Soil emergence

To eliminate any effect of seed germination or germination time, 0HPG seeds were placed on standard growth media in petri dishes containing custom made cylindrical chambers of either 3 mm or 5 mm height 3D printed from polylactic acid (PLA) material. Seeds were then covered with sand to the top of the chambers. The sand was pre-baked and pre-washed several times with distilled water. Sand was then autoclaved and left to cool prior to saturating it with liquid growth media. Petri dishes were double wrapped with foil and kept in the growth chamber for 3 days after which they were opened, and seedlings were exposed to light for an extra 2 to 3 days to green. Images were acquired with a camera and processed with GIMP (version 2.10.34).

Hypocotyl hook angle

Hypocotyl hook angle was determined by calculating 180° minus the angle formed between the tangential line of the apical part and the axis of the subapical part of the hypocotyl. Imaging was conducted using an Olympus SZX12 dissecting microscope and captured with SPOT Advanced imaging software.

Root penetration

Our standard growth media, supplemented with 1.2% Agargel^™, was poured into a deep (2 cm) petri dish. After solidification, a diametric cut was made using a sterile sharp blade, and half of the media was removed. The resulting gap was filled with 0.6% Agargel^™ media and allowed to set. A cut was made in the soft media, leaving 5mm on top of the stiff media. Sterile seeds were delicately placed just below the surface of the soft media. Subsequently, plates were vertically stratified in the dark for two days, exposed to light for 24 hours, and sealed with two layers of foil. Using an Olympus SZX12 dissecting microscope, imaging was conducted after the plates had been kept in the growth chamber for 80 hours.

Statistical analysis

Statistical significance of differences was determined by student’s t-test for two-group comparisons in Excel and by PlotsOfData ⁸⁸.

Supplementary Material

2. Video S1. Kinematics of seedling growth in the dark. Related to Figure 1 and Figure S1.

Col-0 (green) and cslc-Δ5 (magenta) were transferred at the time of germination and imaged in the dark with an infrared light imaging setup. Scale bar = 5 mm.

3. Video S2. Col-0 hypocotyl response to treatment with 600 mM mannitol. Related to Figure 4.

Arrowheads mark the front of the plasmolysis wave. Scale bar = 100 µm.

4. Video S3. cslc4/cslc5 hypocotyl response to treatment with 600 mM mannitol. Related to Figure 4.

Arrowheads mark the front of the plasmolysis wave. Scale bar = 100 µm.

5. Video S4. cslcΔ5 hypocotyl response to treatment with 600 mM mannitol. Related to Figure 4.

Arrowheads mark the front of the plasmolysis wave. Scale bar = 100 µm.

Highlights.

Normal xyloglucan levels are important for seedling establishment.

Changes in xyloglucan content affect turgor pressure.

Reduced xyloglucan induces a reduction in cellulose crystallinity.

Reduced xyloglucan leads to softer and weaker cell walls.