Background

Plant cell walls are involved in all interactions between plants and their environment. Examples include pathogen infection or exposure to drought, where wall composition and structure change to prevent water loss, pathogen susceptibility or at least limit further pathogen spread [1, 2]. These changes of the walls are exemplified by reinforcement with Callose during infection or modifications of pectic polysaccharides to prevent water loss during exposure to drought stress [3, 4]. Cell walls are extremely plastic, undergoing dynamic changes to enable plant cells to expand and differentiate during growth and development [5]. Controlled deposition of cellulose microfibrils through interactions between cellulose synthases and microtubules during cell expansion exemplify the changes in cell wall organization, permitting tightly controlled cell expansion [6, 7]. Deposition of suberin and lignin during formation of the Casparian strip in pericycle cells of the primary root exemplifies modifications of cell walls during cell differentiation [8,9,10]. These examples illustrate processes active during plant-environment interactions and development, enabling cell walls to fulfill their respective biological functions.

How do cell walls perform these various functions, which sometimes involve opposite performance requirements, while simultaneously maintaining their functional integrity? The available evidence supports the existence of a dedicated mechanism, which is monitoring the functional integrity of the plant cell wall and initiates adaptive changes in cellular and cell wall metabolism to maintain cell wall integrity (CWI) [11,12,13,14,15]. Studies of the mode of action of the CWI maintenance mechanism often investigate the responses to cell wall damage (CWD), which can be generated by cell wall degrading enzymes (cellulase, pectinase etc.) or compounds like isoxaben (ISX) [16]. ISX inhibits specifically cellulose production during primary cell wall formation in elongating plant cells [17,18,19]. Established responses to CWD include growth inhibition involving cell cycle arrest, changes in the levels of phytohormones like jasmonic acid (JA), salicylic acid (SA) and cytokinins (CKs) as well as changes in cell wall composition involving pectic polysaccharides, lignin and Callose deposition [20,21,22,23,24].

The available evidence implicates receptor-like kinases (RLK) like MALE DISCOVERER 1-INTERACTING RECEPTOR LIKE KINASE 2 (MIK2), FEI1, FEI2, THESEUS 1 (THE1) and FERONIA (FER) in CWI maintenance [16, 25,26,27,28]. THE1 and FER belong to the Catharanthus roseus RLK1-like kinase (CrRLK1L) family, which has 17 members. These RLKs consist of an intracellular Serine / Threonine-kinase domain, a transmembrane domain and an extracellular domain exhibiting similarity to the malectin domain originally identified in Xenopus laevis [29]. Currently it is not clear if malectin domains in CrRLK1Ls are either required for binding to cell wall epitopes, mediate protein-protein interaction or actually do both [28, 30,31,http://signal.salk.edu/cgi-bin/tdnaexpress). Plants homozygous for the insertions were isolated using PCR-based genoty** as well as insertion positions in the individual gene and their effects on transcript levels determined. For WSR1 the first insertion is located in the 5′ (wsr1–1, Salk_046217) and the second (wsr1–2, Salk_092919) in the 3′ untranslated region of the gene (Additional file 1: Figure S3a). The insertions in WSR2 were mapped to the first intron (wsr2–2, Salk_123509) and the third exon (wsr2–1, Salk_058271) (Additional file 1: Figure S3b). For WSR3 the insertions were located either in the promoter region (wsr3–2, SALK_079548) or in the 10th intron (wsr3–1, SALK_039575) (Additional file 1: Figure S3c). In the wsr4–2 (SALK_121365) allele the insertion is located in the 1st while the one giving rise to the wsr4–1 (SALK_082484) allele is located in the 2nd exon (Additional file 1: Figure S3d). To determine if the insertions affect transcript levels of the genes, we performed qRT-PCR using total RNA isolated from mock-treated 7 days-old seedlings. These experiments identified four bona fide KO- (wsr2–1, wsr2–2, wsr3–1, wsr4–2) and four KD-alleles (wsr1–1, wsr1–2, wsr3–2 and wsr4–1) for the candidate genes (Additional file 1: Figure S3a-d).

Responses to ISX-induced CWD are modified in seedlings with insertions in WSR1 or 4

ISX-induced CWD leads to lignin deposition in seedling root tips as well as increased JA and SA production [20, 21, 48]. Here we investigated if the KO / KD alleles isolated for the candidate genes affect these responses by performing experiments with mock or ISX-treated seedlings in liquid culture. After treating the seedlings for 7 hours we measured both JA and SA levels (Fig. 3a, b). In ISX-treated wsr4–2 seedlings, we detected reduced production of JA while SA levels were lower in wsr1–2 seedlings. Lignin was detected after 24 h of ISX treatment using phloroglucinol and quantified using image analysis [16]. Mock-treated seedling roots did not show any lignin deposition (Additional file 1: Figure S4). After ISX-treatment, we detected a significant reduction in lignification of wsr4–2 seedlings (Fig. 4, Additional file 1: Figure S4). These results suggested that WSR1 and 4 might be involved in the response to ISX-induced CWD.

Fig. 3
figure 3

Relative jasmonic acid and salicylic acid accumulation in wsr seedlings after ISX treatment. a Jasmonic acid and (b) salicylic acid were quantified in Col-0, wsr1–1, wsr1–2, wsr2–1, wsr2–2, wsr3–1, wsr3–2, wsr4–1 and wsr4–2 seedlings after 7 h of mock (DMSO; empty bars) or ISX (filled bars) treatment. Bars represent mean values from 3 to 4 independent experiments and error bars indicate SD. Asterisks indicate statistically significant differences to the ISX-treated wild type according to Student‘s t test (*p < 0.05)

Full size image
Fig. 4
figure 4

Relative lignification in wsr root tips after ISX treatment. Lignification in root tips of Col-0, wsr1–1, wsr1–2, wsr2–1, wsr2–2, wsr3–1, wsr3–2, wsr4–1 and wsr4–2 seedlings was quantified after 24 h of ISX treatment. Bars represent mean values from 3 independent experiments and error bars indicate SD. Asterisks indicate statistically significant differences to the wild type according to Student‘s t test (*p < 0.05)

Full size image

WSR1, 2, 3 and 4 contribute to cell wall formation during stem growth

To determine whether the genes of interest affect cell wall metabolism in general, the levels of cellulose, uronic acids and neutral cell wall sugars were determined in cell wall preparations from rosette leaves and mature stems of Col-0 and wsr mutant plants. The mutant plants did not exhibit any pronounced growth phenotypes compared to Col-0 plants (Additional file 1: Figure S5). Here, we only investigated the strongest KO or KD allele for each candidate gene. We detected a reduction in cellulose content only in leaves of wsr4–2 plants compared to Col-0 (Fig. 5a). Our quantification of cellulose in mature stems detected increased amounts in wsr1–2 and wsr2–1, while cellulose was reduced in wsr4–2 stems (Fig. 5b). Analysis of lignin content in stems of adult plants did not detect any differences between Col-0 and mutant plants (Additional file 1: Figure S6). Analysis of neutral cell wall sugars and uronic acids in leaves detected only for the low-abundant glucuronic acid in wsr1–2 significant differences to Col-0 controls (Fig. 6a). In stem-derived material, we observed enhanced levels of rhamnose and xylose in wsr1–2 (Fig. 6b). In wsr3–1 stems, fucose, rhamnose, arabinose and galactose contents were elevated. In wsr4–2 glucose amounts were reduced while mannose was slightly enhanced compared to Col-0 controls. To summarize, our cell wall analyses detected differences in cellulose and different neutral cell wall sugar contents with effects detected most pronounced in stems. The similarities and variability observed regarding cell wall phenotypes suggest that the different genes may be involved in distinct but also overlap** aspects of cell wall metabolism in adult plants.

Fig. 5
figure 5

Cellulose content in adult wsr plants. Cellulose content was quantified in cell wall preparations from 5 weeks-old Col-0, wsr1–2, wsr2–1, wsr3–1 and wsr4–2 plants. a Leaf cellulose, (b) stem cellulose. Bars represent mean values and error bars indicate SD (n = 4). Asterisks indicate statistically significant differences to the wild type according to Student‘s t test (*p < 0.05; **p < 0.01)

Full size image
Fig. 6
figure 6

Cell wall matrix monosaccharide composition in adult wsr plants. Relative amounts of the monosaccharides Fucose, Rhamnose, Arabinose, Galactose, Glucose, Xylose, Mannose, Galacturonic acid and Glucuronic acid were quantified in cell wall matrix hydrolysates of 5 weeks-old Col-0, wsr1–2, wsr2–1, wsr3–1 and wsr4–2 plants. a Leaf monosaccharides, (b) stem monosaccharides. Bars represent mean values and error bars indicate SD (n = 4). Asterisks indicate statistically significant differences to the wild type according to Student’s t test (*p < 0.05; **p < 0.01)

Full size image

Plants with insertions in WSR1 exhibit pathogen response phenotypes

It has been shown previously that the CWI monitoring RLKs THE1 and MIK2 affect pathogen susceptibility [27]. Here we investigated if the mutations in WSR genes also affect the outcome of plant-pathogen interactions by inoculating adult plants carrying insertions in the four genes with the necrotrophic fungus Plectosphaerella cucumerina BMM (PcBMM) and quantifying fungal biomass 5 days post inoculation (Fig. 7). Arabidopsis Gβ 1 (agb1–1) and irregular xylem 1 (irx1–6) plants were included as controls since they exhibit reduced (agb1–1) or enhanced resistance (irx1–6) to PcBMM infection [77, 78]. Fungal growth on infected wsr2–1, 3–1 and 4–2 plants was similar to Col-0 controls. However, growth was significantly enhanced on wsr1–2 plants, suggesting that reduction of WSR1 gene expression affects resistance to PcBMM infection and implicating this gene in disease resistance.

Fig. 7
figure 7

Relative susceptibility of wsr plants to Plectosphaerella cucumerina. Three weeks-old Col-0, wsr1–2, wsr2–1, wsr3–1, wsr4–2, irx1–6 (resistance control) and agb1–1 (susceptibility control) plants were infected with the necrotrophic leaf pathogen isolate PcBMM. The relative fungal biomass was determined 5 days post infection (dpi) by qPCR analysis of the PcBMM β-tubulin gene. Bars represent mean values and error bars indicate SD (n = 4–6). Asterisks indicate statistically significant differences to the wild type accoding to Student‘s t test (*p < 0.05; **p < 0.01).

Full size image

Discussion

In plants a mechanism is existing, which monitors and maintains the functional integrity of the cell walls [12, 14]. This mechanism seems to exhibit similarities to the one described in S. cerevisiae and is capable of detecting CWD and initiating adaptive changes in cellular and cell wall metabolism to maintain the functional integrity of the wall [11]. Understanding of the molecular mechanisms underlying CWD perception and the signaling cascades involved in regulating the CWD response in plants is increasing [16, 28]. However, our knowledge of the genes and molecular processes bringing about changes in cell wall metabolism in response to CWD and their function during growth and development is very limited. Here, we have determined if we can identify genes both mediating responses to CWD in seedlings and cell wall metabolism in adult plants by combining transcriptomics data from ISX-treated seedlings with FTIR-based cell wall analysis of seedlings carrying T-DNA insertions in candidate genes identified from the transcriptomics data. This approach identified 17 genes (out of 23 original candidates), where T-DNA insertions in the candidate genes seem to lead to cell wall defects based on FTIR results. The functions of these genes, belonging to different gene families, are often not well-understood (Additional file 1: Table S1). Very little is known about the biological function of At1g74440 beyond that it encodes an ER membrane protein. The gene has been implicated in biotic and abiotic stress responses mediated by Plant Natriuretic Peptides (PNPs) based on co-expression with AtPNP-A [62]. ATFIM2 encodes a protein belonging to the Fimbrin family and seems to modulates the organization of actin filaments [70]. SQE2 encodes a squalene epoxidase converting squalene into oxidosqualene, which forms the precursor of all known angiosperm cyclic triterpenoids [59]. Triterpenoids are required for production of membrane sterols and brassinosteroids. PSAT1 encodes an amino transferase required for serine biosynthesis taking place in the chloroplast [69]. Serine biosynthesis in turn is required during photorespiration, a prerequisite for carbohydrate metabolism and plant growth. While AGPs have been implicated in cell wall remodeling, very little information is available regarding the specific function of AGP7 in this context [72]. AtBXL1 encodes an enzyme acting during vascular differentiation as a β-D-xylosidase while acting as an α-L-arabinofuranosidase during seed coat development [60]. PRX47 encodes a putative peroxidase, is apparently expressed in differentiating vascular tissue in seedling roots and stems and involved in lignification [68]. UGT71C3 and UGT76B1 encode UDP-glycosyltransferases (UGTs), which have been implicated in glycosylation of phytohormones and / or metabolites during the response to biotic and abiotic stress [61]. UGT76B1 in particular seems to glycosylate isoleucic acid, which is required for coordination of SA- and JA-based defence responses active during infection by pathogens like Pseudomonas syringae and Alternaria brassicicola [67]. ADR1-L1 encodes a coiled-coil nucleotide-binding leucine-rich repeat protein and forms an important element of the effector-triggered immunity in plants [64, 79]. Reviewing the available knowledge provides further evidence that several of the genes are probably required for processes relevant for cell wall (PRX47, ATBXL1, AGP7), membrane (SQE2), photosynthetic (PSAT1) metabolism or light perception (PKS1). Intriguingly UGT76B1, UGT71C3, ADR1-L1, SERK4 and At3g16560 have been implicated before in the responses to abiotic or biotic stress, which also involves plant cell walls [61, 67, 79]. In our experimental conditions the seedlings are exposed to CWD but not biotic / abiotic stress, thus raising the possibility that these genes are actually responding to cell wall-related events, which may also occur during biotic and abiotic stress. More importantly the results suggest that the approach pursued here enables us to identify amongst the many genes in the Arabidopsis genome those that contribute to the responses to CWD and regulation of relevant aspects of cell wall and membrane metabolism.

We characterized four candidate genes in more detail. These had been selected based on the FTIR phenotypes apparently caused by insertions in the candidate genes and the limited detailed knowledge regarding their biological functions. qRT-PCR-based expression analysis of the four genes in ISX-treated seedlings yielded results similar to the data from the transcriptomics experiment. Experiments with loss- and gain-of-function alleles of THE1 showed that ISX-induced changes in the transcript levels of WSR1, 2, 3 and 4 are sensitive to an increase in the activity of THE1 (the1–4) while effects of reductions (the1–1) are less pronounced [75]. These results are to be expected since complete loss of THE1 results in reduced responses to CWD but not complete losses, suggesting that the THE1-mediated CWI maintenance mechanism is either redundantly organized or other signaling mechanisms exist [16]. However, the results support the notion that WSR gene expression is regulated by the THE1-mediated CWI maintenance mechanism and that WSR activity might be controlled on the transcriptional level.

Table 2 provides a global overview of the phenotypes observed for the insertion lines in the four genes. Reduction of WSR1 and WSR2 activity seemed to cause similarly pronounced FTIR phenotypes in an area where diagnostic signals for cellulose and pectins are normally found (Fig. 1). For wsr1–2, we detected increased amounts of cellulose, rhamnose and xylose in stem-derived material while glucuronic acid was reduced in leaf material (Figs. 5, 6). ISX-induced SA production in seedlings and resistance to PcBMM in adult plants were reduced (Figs. 3, 7). In wsr2–1 plants, we also detected an increase in cellulose in stem-derived material while responses to CWD, PcBMM susceptibility and non-cellulosic cell wall matrix composition were similar to the controls (Figs. 3, 5, 6, 7). Reductions in WSR3 and 4 expression seemed to result in FTIR phenotypes related to cellulose elements, hemicelluloses and pectins (Fig. 1). In wsr3–1 stem-derived cell wall material, we detected significant differences in the amounts of fucose, rhamnose, arabinose and galactose compared to controls (Fig. 6). In wsr4–2, cellulose content was reduced both in stem- and leaf-derived cell wall material, while the amounts of glucose and mannose in stem-derived material were reduced and increased, respectively (Figs. 5, 6). Analysis of responses to CWD found reduced JA and lignin production in ISX-treated wsr4–2 seedlings and no differences to wild type in wsr3–1 seedlings (Figs. 3, 4). The phenotype observed in wsr2–1 in combination with the limited available protein information provides unfortunately no new insights regarding the function of WSR2 [55]. The specific effects on neutral cell wall sugars in wsr3–1 plants suggest that WSR3 could contribute to cell wall polysaccharide metabolism possibly by mediating transport between the Golgi (where non-cellulosic cell wall polysaccharides containing fucose, rhamnose, arabinose, galactose are synthesized) and the plasma membrane [80]. This would make sense bearing in mind that WSR3 / SFH19 belongs to the SEC14-protein family, whose members have been implicated in phosphoinositide production required for membrane homeostasis and signaling processes regulating cellular processes like vesicle transport [71, 81]. These results implicated WSR2 and WSR3 in cell wall metabolism but not in the representative responses to CWD examined here. Both WSR1/DIR7 and WSR4/PXC3 seem to be required for CWD responses on the seedling levels and for specific aspects of cell wall metabolism in adult plants, suggesting that they are both involved in CWD-induced signaling processes regulating changes in cell wall composition. WSR1 seems only required for increased SA production in response to ISX-induced CWD, whereas WSR4 is required for both JA and lignin production. The wsr4–2 phenotypes were similar to those described for CWD responses in mik2 seedlings where also only JA and lignin production differ from controls while SA amounts are similar [16, 27]. These results suggest that both RLKs might be required for the same aspects of CWI maintenance. The THE1-dependent reduction in WSR4 transcript levels in response to CWD suggest it could repress CWD-induced responses. This would be similar to the CWD response phenotypes of FER, where a FER KD leads to enhanced production of JA, SA and lignin [16]. However, the WSR4 loss of function phenotypes suggest the RLK is required for ISX-induced JA and lignin production. This implies the existence of additional regulatory elements interacting with WSR4 to give rise to the observed mutant phenotypes. Since the related RLK PXC1 has been implicated in vascular development, a function for WSR4 in coordination of CWD perception with cell wall metabolism is conceivable [57].

Table 2 Overview of the phenotypes observed for mutant lines of the different candidate genes examined. Statistically significant differences compared to the wild type are indicated with blue (increased) or red (decreased) arrows
Full size table

Conclusions

To summarize, seedlings with T-DNA insertions in 17 of the 23 candidate genes that were selected in this study seemed to exhibit FTIR phenotypes. Gene expression analysis showed that WSR gene expression is modulated in response to ISX-induced CWD, with the modulation apparently sensitive to changes in THE1 activity. This connected the genes identified to the THE1-dependent CWI maintenance mechanism, suggesting that our approach has identified new components mediating CWI maintenance in Arabidopsis. Follow up studies with KO or KD lines for four candidate genes found cell wall phenotypes in adult plants for all four and effects on CWD responses for WSR1 and 4. These results also suggest strongly that a more detailed analysis of the remaining candidate genes identified, will probably yield interesting novel insights into the mode of action of the CWI maintenance mechanism and cell wall metabolism in general.

Methods

Reagents

All chemicals and enzymes were purchased from Sigma-Aldrich unless stated otherwise.

Plant material

Wild-type and mutant Arabidopsis thaliana lines used in this study were ordered from the Nottingham Arabidopsis Stock Centre (http://arabidopsis.info/). Detailed information is listed in Additional file 1: Table S1. Seedlings were grown for 6 days in liquid culture (2.1 g/L Murashige and Skoog Basal Medium, 0.5 g/L MES salt and 1% sucrose at pH 5.7) before treatment with 600 nM isoxaben (in DMSO) as described [16]. For cell wall analysis, plants were grown on soil (Pro-Mix HP) in long-day conditions (16 h light, 11000 Lux, 22 °C, 8 h dark, 20 °C, 70% relative humidity). For pathogen infection assays, plants were grown in phytochambers on sterile soil-vermiculite (3,1) under short-day conditions (10 h of light/14 h of dark) at 20–21 °C.

Pathogen infection assays

For Plectosphaerella cucumerina BMM pathogenicity assays, 18 days-old plants (n > 15) were sprayed with a spore suspension (4 × 106 spores/ml) of the fungus as previously described [82, 83]. Fungal biomass in planta was quantified by determining the level of the PcBMM β-tubulin gene by qPCR (forward primer: CAAGTATGTTCCCCGAGCCGT and reverse primer: GGTCCCTTCGGTCAGCTCTTC) and normalizing these values to those of UBIQUITIN-CONJUGATING ENZYME21 (UBC21, AT5G25760).

Quantitative RT-PCR

Total RNA was isolated using a Spectrum Plant Total RNA Kit (Sigma-Aldrich). Two micrograms of total RNA were treated with RQ1 RNase-Free DNase (Promega) and processed with the ImProm-II Reverse Transcription System (Promega) for cDNA synthesis. qRT-PCR was performed with a Roche LightCycler 480 system using LightCycler 480 SYBR Green I Master. Gene expression levels were determined as described [24]. The following gene-specific primers have been used for time course expression analysis in Col-0 and expression analysis in the1 seedlings:

  • ACT2-FOR (5′-CTTGCACCAAGCAGCATGAA-3′),

  • ACT2-REV (5′-CCGATCCAGACACTGTACTTCCTT-3′),

  • WSR1-FOR (5′-TATGGTGATGAACTTTGCGTTC-3′),

  • WSR1-REV (5′-ACTCAACAGTAGCATCTCCTGA-3′),

  • WSR1A-FOR (5′-TACGCTGCTACTGGTCAACG-3′),

  • WSR1A-REV (5′-TTCCTCCAATCACCGGCATC-3′),

  • WSR2-FOR (5′-CTCACTTCCATCGTTTCAAGTG-3′),

  • WSR2-REV (5′-GAAACCAAACGTGGCCTAAA-3′),

  • WSR3-FOR (5′-GAAAGCACGAGACTGGAACG-3′),

  • WSR3-REV (5′-TATCCACCCTCCAACGCAAA-3′),

  • WSR4-FOR (5′-AGCCCTGAGAGATCAAGCATT-3′),

  • WSR4-REV (5′-AGCTCAACTAAGCGATGAAGC-3′).

For the characterization of T-DNA insertion lines, the following primers were employed:

  • ACT2-FOR, ACT2-REV, WSR1-FOR, WSR1-REV, WSR2-FOR, WSR2-REV,

  • WSR3-FOR2 (5′- TCTTATCCGGTTGCGGAAGG-3′),

  • WSR3-REV2 (5′-GTGGTGAGATGACCCAGAGC-3′),

  • WSR4-FOR2 (5′-CTTGATGCAGTTGTGAAAGCA-3′),

  • WSR4-REV2 (5′-TCTTCACCGAAACAATCATCC-3′).

FTIR spectroscopy and analysis

For FTIR analysis, 4 biological replicates per genotype and 5 technical replicates per biological replicate were collected (i.e. for each genotype 20 spectra were collected). Spectra for each technical replicate were measured from 800 to 5000 cm− 1 with 15 accumulations per measurement on a Bruker Vertex 70. All spectra were measured at 10 kHz, with a 10 kHz lowpass filter and the Fourier transform was carried out using Blackman-Harris 3-term. Atmospheric compensation was carried out on the data using OPUS version 5 (www.bruker.com). The spectra were cropped to the area between 802 cm− 1 to 1820 cm− 1 to cover informative wavenumbers as described in [58]. Linear regression was carried out based on the first 10 points in either end of the spectra and used for baseline correction. The data was normalized to sum 1 with any negative values still present set to 0 for normalization purposes. Biological variation in the Col-0 controls was determined based on three independent experiments carried out with 4 biological replicates and 5 technical replicates per biological replicate (i.e. 20 spectra per experiment). The difference between the insertion lines and Col-0 was calculated by averaging all the technical repeats for a line and subtracting the corresponding average from Col-0. The difference between the insertion lines and Col-0 was plotted by wavelength. Two times the standard deviation of Col-0 was chosen as a cutoff as it would indicate significance if the natural variation is assumed to be symmetrical across Col-0 and the insertion line.

Cell wall analysis

Cell wall preparation and analysis were performed as described [84] with minor modifications. For analysis of stem cell wall composition, major stems of three 5 weeks-old plants per genotype were pooled to form one biological replicate. For analysis of leaf cell wall composition, whole leaf rosettes of three 5 weeks-old plants per genotype were pooled to form one biological replicate. Four biological replicates were analysed in all cases. Plant samples were immediately flash-frozen in liquid nitrogen after sampling and lyophilized. Dried material was ball-milled with zirconia beads in a Labman robot (www.labmanautomation.com), extracted three times with 70% ethanol at 70 °C and dried under vacuum. Starch was removed using a Megazyme Total Starch Kit according to the manufacturer’s instructions. After drying under vacuum, de-starched alcohol insoluble residue (AIR) was weighed out in 2 ml screw caps tubes for cell wall monosaccharide analysis (2 mg AIR) and GC vials for lignin analysis (1.2 mg AIR), respectively, with the Labman robot (0.2 mg tolerance). Cellulose, neutral sugars and uronic acids were determined following the published one-step two-step hydrolysis protocol [84]. High-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) was performed on a Thermo Fisher Dionex ICS-5000 system with CarboPac PA-20 and PA-200 columns as described [84]. Acetyl bromide soluble lignin was quantified as described [85].

Phytohormone analysis

JA and SA were extracted and analysed as described [16]. Briefly, extraction was performed in 10% methanol / 1% acetic acid with Jasmonic-d5 Acid and Salicylic-d4 Acid (CDN Isotopes) as internal standards. Quantification was performed on a Shimadzu UFLC XR / AB SCIEX Triple Quad 5500 system using the following mass transitions: JA 209 > 59, D5-JA 214 > 62, SA 137 > 93, D4-SA 141 > 97.

Lignin detection in roots

Lignification in seedling roots (n > 15) was analysed 24 h after start of treatment. Lignified regions were detected with phloroglucinol-HCl, photographed with a Zeiss Axio Zoom.V16 stereomicroscope and quantified as described [16].

Statistical analysis

Statistical significance was assessed using Student’s t-test in Microsoft Excel (2-tailed distribution, two-sample unequal variance). Statistically significant differences are indicated by * p < 0.05, ** p < 0.01. Boxplots were generated using R package “boxplot” with default settings (range = 1.5*IQR).