Loss of PMR4 callose synthase triggers jasmonic acid–dependent resistance to the clubroot disease in Arabidopsis and Brassica napus

Abstract

Clubroot caused by the protist pathogen Plasmodiophora brassicae Woronin is a major disease in Brassicaceae crops. Growing clubroot-resistant cultivars remains the most effective and practical control strategy, but diverse or emerging pathotypes can often overcome resistance genes deployed. In this study, we identified POWDERY MILDEW RESISTANT 4 (PMR4) as a potential host susceptibility factor that can be targeted as a novel genetic resource for durable clubroot resistance. Recessive PMR4 mutations in Arabidopsis thaliana conferred strong resistance to multiple P. brassicae pathotypes, independent of salicylic acid–mediated plant immunity. Using CRISPR/Cas9, we edited 2 PMR4 orthologs in the Brassica napus genome, and the resulting homozygous mutants exhibited resistance to both powdery mildew and clubroot diseases. This study reveals that P. brassicae infection induces callose deposition in Arabidopsis and B. napus roots in a PMR4-dependent manner, similar to wound- and powdery mildew–induced callose deposition in Arabidopsis leaves. The Arabidopsis pmr4-1 mutation does not affect P. brassicae primary infection but blocks the pathogen entry from epidermal cells to the stele in a jasmonic acid signaling–dependent manner, which coincides with accumulation of lignin-like and suberin compounds in periderm cell walls. Together, these findings establish PMR4-encoded callose synthase as a host susceptibility factor facilitating secondary infection by P. brassicae and demonstrate its potential as a gene-editing target for enhancing resistance to powdery mildew, clubroot, and possibly other biotic and abiotic stresses in Brassicaceae crops.

Introduction

Brassicaceae crops, especially canola/oilseed rape (Brassica napus L.), are cultivated globally with significant economic and nutritional values. However, most of these crops are affected by clubroot, a severe soil-borne disease caused by a protist Plasmodiophora brassicae (Pb) Woronin (Cardoza and Stewart 2006; Pérez-López et al. 2018). Resting spores, as the primary inoculum of the pathogen, can survive for up to 17 yr in the soil (Wallenhammar 1996). Primary zoospores released from resting spores initially infect root hairs, producing primary plasmodia during this infection stage that do not cause macroscopic symptoms (Howard et al. 2010; Hwang et al. 2012). The secondary zoospores released from root hairs can penetrate root epidermis and invade cortical and inner tissues of main roots, resulting in club-shaped root galls (Fig. S1) (Hwang et al. 2012; Liu et al. 2020a). The gall formation impairs water and nutrient uptake, leading to plant wilting and premature death (Hwang et al. 2012; Schwelm et al. 2015).

The most efficient disease management for clubroot to date is growing clubroot-resistant (CR) cultivars in various rotations (Hwang et al. 2012; Peng et al. 2015). Although there are several commercial CR oilseed and cabbage cultivars available, their resistance is controlled by a single limited number of dominant CR genes, which can be broken quickly (Matsumoto et al. 2012), especially given the diverse and evolving pathogen populations found in western Canada (Strelkov et al. 2018). Distinct from the R-gene strategy, an emerging alternative approach based on susceptibility (S) genes has been gaining interest due to its potential to be more durable than the conventional R-gene-driven disease resistance in the field (Zaidi et al. 2018). These S genes may facilitate host–pathogen compatibility by promoting one or more steps of the infection process (van Schie and Takken 2014). Increasing evidence has shown promise to create such novel disease-resistant crops through the alteration of S genes. With the development of genome-editing techniques like CRISPR/Cas9 in plants (Doudna and Charpentier 2014; Gao 2021), the target S genes can be precisely modified, leading to transgene-free crops (Soda et al. 2018; Boubakri 2023) with potentially durable resistance to plant diseases.

We reasoned that Arabidopsis thaliana serves as an ideal model system for screening and identification of clubroot S (CS) genes. Firstly, A. thaliana is genetically and physiologically related to the rapeseed species. Secondly, the Arabidopsis ecotype Col-0 is susceptible to all known Pb pathotypes. Thirdly, many T-DNA insertion and mutagen-induced mutants in Col-0 are available in the public domain, making the screening of diverse genotypes possible. In this study, we screened a range of mutants and identified POWDERY MILDEW RESISTANT 4 (PMR4) as a promising CS gene. PMR4 encodes a putative callose synthase (Vogel and Somerville 2000) and was previously identified as an S gene in the context of powdery mildew (PM) disease on Arabidopsis because the loss of PMR4 callose synthase activities resulted in salicylic acid (SA)–dependent plant defense responses against PM (Vogel and Somerville 2000; Nishimura et al. 2003). Unexpectedly, our experimental data showed that disrupting the SA signaling in the pmr4 mutant did not restore its susceptibility to clubroot disease; instead, pmr4-conferred clubroot resistance was dependent on jasmonic acid (JA) signaling and correlated with the biosynthesis of lignin-like and suberin compounds in root periderm cell walls. These findings agree with a recent report (Wu et al. 2025). Targeted genome editing of B. napus PMR4 orthologs confers dual resistance to both PM and clubroot, highlighting its potential for novel disease resistance in Brassicaceae crops.

Results

Arabidopsis pmr4-1 mutant confers strong resistance against multiple Pb pathotypes

To identify CS genes, Arabidopsis mutants carrying reported S genes that conferred enhanced resistance against biotrophic/hemi-biotrophic pathogens were collected. These mutant plants were inoculated with the Pb field isolate of pathotype 3H, the most prevalent pathotype in western Canada with high virulence (Hollman et al. 2023), followed by the assessment of severity against clubroot infection at 21 d postinoculation (dpi). Among 26 mutant lines tested to date (Table S1), a mlo2,6,12 triple mutant and a ubc13a,b double mutant displayed moderate resistance to clubroot (Fig. 1a). MILDEW RESISTANCE LOCUS O (MLO) genes encode plant-specific calcium channels (Gao et al. 2022), and loss of MLO family genes confers resistance to fungal PM in the model plant Arabidopsis (Consonni et al. 2006) and crop species including barley, wheat, pea, and tomato (Jørgensen 1992; Kusch and Panstruga 2017). UBC13s encode ubiquitin conjugation enzymes that mediate noncanonical K63-linked polyubiquitination (Hofmann and Pickart 1999; McKenna et al. 2001). The Arabidopsis ubc13a,b double mutation compromises auxin signaling, root growth (Wen et al. 2014), and immune responses (Wang et al. 2019). Surprisingly, a sweet11,12 double mutant, which was reported to be clubroot resistant (Walerowski et al. 2018), and a corresponding sweet11,12,15 triple mutant did not display an apparent reduction in the disease severity index (DSI) under our experimental conditions (Fig. 1a). In sharp contrast, the pmr4-1 mutant consistently displayed strong resistance to clubroot with DSI close to 0% and almost no visible gall formation across 6 independent experiments (Fig. 1a).

Figure 1.

Arabidopsis pmr4-1 and gsl5-1 are highly resistant to Pb. a) Pathogenicity assay on selected Arabidopsis mutants against the clubroot disease. DSI reflects quantitative assessments to the disease. Clubroot symptoms were rated at 21 dpi after Pb pathotype 3H inoculation for at least 4 biological replicates with each line containing 15 plants. Data are means ± Sd obtained from 4 independent experiments. Statistical differences were analyzed by 1-way ANOVA taking Col-0 as a control. ****P ≤ 0.0001. b) Representative shoot (upper row) and root (lower row) images of Pb pathotype 3H-inoculated wild-type (Col-0), pmr4-1 and gsl5-1 mutants captured at 21 dpi. Scale bar = 2 cm. c) Quantitative analysis of Arabidopsis clubroot disease severity as shown in b). DSI values are means ± Sd obtained from 4 biological repeats with 15 plants for each replicate. ****P ≤ 0.0001. d) Relative Pb pathogen biomass by quantitative PCR as described (Wen et al. 2020). AtUBQ10 gene was used as host reference. Values are relative to the inoculated plants at 0 dpi. At least 5 plant roots were pooled as 1 biological replicate. Data represent means ± Sd. n = 3. Two-way ANOVA analysis for relative Pb biomass was performed to compare the pathogen load differences between Col-0 and pmr4-1, and asterisk(s) represent statistical significance (α = 0.05). **P ≤ 0.01; ****P ≤ 0.0001. e) Representative root images of Col-0 and pmr4-1 inoculated with various Pb pathotypes captured at 21 dpi. Scale bar = 2 cm. f) DSI (means ± Sd) of Col-0 and pmr4-1 inoculated with various Pb pathotypes. n = 4. ****P ≤ 0.0001.

PMR4, also known as GSL5 and CalS12, encodes a predicted trans-membrane callose synthase (Zaveská Drábková and Honys 2017). PMR4 was initially identified because its mutant alleles conferred PM resistance (Vogel and Somerville 2000). The Arabidopsis pmr4-1 mutation contains a single G-to-A base substitution in the second exon of the PMR4 gene, resulting in the conversion of codon TGG for Trp-687 to a stop codon TAG and the formation of a prematurely truncated polypeptide of 686 amino acids in length (Nishimura et al. 2003). This truncated PMR4 variant retains the N-terminal 6 transmembrane helices but lacks the putative catalytic domain and the subsequent 7 transmembrane helices at the C terminus. To independently assess whether the enhanced clubroot resistance is caused by the nonsense mutation in PMR4, a T-DNA insertion line gsl5-1 (GABI-KAT 089H05), in which a T-DNA is inserted to the second exon of PMR4 close to the nonsense mutation in pmr4-1 (see Fig. S8a), was acquired and inoculated with Pb pathotype 3H. Typical swollen root galls were observed in wild-type roots, and the plants were severely affected at 21 dpi, while pmr4-1 and gsl5-1 plants remained healthy without visible symptoms both underground and aboveground (Fig. 1b), with DSI values near 0% (Fig. 1c). Similarly, Pb biomass in Col-0 increased steadily from 10 to 21 dpi, while no measurable increase was observed in Pb-infected pmr4-1 during this period (Fig. 1d). This result indicates that the enhanced clubroot resistance observed in pmr4-1 and gsl5-1 mutants is exclusively caused by mutations in the PMR4 gene.

Since both pmr4-1 and gsl5-1 can potentially produce truncated proteins, it is unclear whether they are complete loss-of-function mutations. To determine the nature of mutations for clubroot resistance, a heterozygous gsl5-1 line and its self-pollinated segregants were assessed by Pb inoculation as above. The gsl5-1/+ heterozygous plants were largely susceptible to the clubroot pathogen, and their self-pollinated progenies segregated into 16 resistant and 43 susceptible individuals. A chi-square test showed that it fits a phenotypic segregation ratio of 3:1 (χ²_3:1 = 0.142), indicating that the gsl5-1-mediated clubroot resistance is recessively inherited. Hence, we conclude that the enhanced clubroot resistance is conferred by a recessive mutation at the PMR4/GSL5 locus.

The involvement of PMR4 in clubroot susceptibility encouraged us to further determine if the resistance observed in the pmr4-1 mutant is race/pathotype dependent. In addition to pathotype 3H, pmr4-1 mutant plants were also highly resistant to 3 additional pathotypes/isolates found in western Canada, including 3A, 5X-LG2, and 5X-LG3 (Fig. 1e and f). Hence, the loss of PMR4/GSL5 resulted in robust clubroot resistance against multiple P. brassica pathotypes under our experimental conditions.

pmr4-1-conferred clubroot resistance is independent of SA signaling

pmr4-1 has been extensively studied for its resistance to the fungal disease PM (Vogel and Somerville 2000; Jacobs et al. 2003; Nishimura et al. 2003). A widely accepted explanation for this resistance is that callose or the callose synthase encoded by PMR4 negatively regulates the SA pathway–mediated immunity. Consequently, PM resistance in the pmr4-1 mutant is attributed to enhanced activation of the SA signaling pathway (Nishimura et al. 2003). Given the substantial biological and ecological differences between this foliar fungal pathogen and the root-infecting protist that causes clubroot, we investigated whether the enhanced clubroot resistance in the pmr4-1 mutant is also dependent on the functional SA pathway.

To determine whether pmr4-1-mediated clubroot resistance depends on the SA pathway, we tested a pmr4-1 pad4-1 double mutant and a transgenic pmr4-1 NahG line, both with compromised SA signaling, for response to the clubroot infection. PAD4 acts upstream of SA accumulation in a positive feedback loop required for defense activation (Zhou et al. 1998; Jirage et al. 1999), while bacterium-derived enzyme NahG specifically degrades SA (Delaney et al. 1994). These lines have been previously used to demonstrate the SA dependency of pmr4-1-mediated PM resistance (Nishimura et al. 2003). To our surprise, both pmr4-1 pad4-1 and pmr4-1 NahG plants remained resistant to clubroot, exhibiting phenotypes indistinguishable from those of pmr4-1 single mutant (Fig. 2a). In contrast, severe galling was observed on Col-0, NahG, and pad4-1 plants following Pb inoculation (Fig. 2a). Clubroot DSI (Fig. 2b) further confirmed that blocking the SA signaling pathway did not compromise pmr4-conferred clubroot resistance.

Figure 2.

Arabidopsis pmr4-1 confers dual resistance to Pb and Ec. a) Representative root images of plants from indicated genotypes with and without Pb pathotype 3H inoculation at 21 dpi. Scale bar = 2 cm. b) Quantitative analysis of Arabidopsis clubroot disease severity as shown in a). DSI values are means ± Sd obtained from 3 groups of biological repeats, each containing 15 individual plants. One-way ANOVA was performed, and different letters represent various disease severity levels as well as statistical significance. c) Relative Pb biomass as measured by qPCR of a Pb internal transcribed spacer 1 region (Wen et al. 2020). AtUBQ10 was used as a host tissue normalization reference. At least 3 plant roots were pooled to form 1 mixed sample with 3 biological replicates for each line. Data are means ± Sd obtained from 3 technical replicates. One-way ANOVA analysis for Pb relative biomass at 21 dpi was performed. *P ≤ 0.05, ****P ≤ 0.0001. d) Representative leaf images of plants from indicated genotypes with and without Ec inoculation at 15 dpi. Scale bar = 2 cm. e) Quantitative analysis of Ec conidiophore density on Ec-inoculated leaves of indicated genotypes at 5 dpi. One-way ANOVA was performed, and different letters represent statistical significance. Data represent means ± Sd. f) Transcriptome profiling of the gene expressions involved in SA and N-hydroxypipecolic acid biosynthesis, signaling, and response in the roots before or after Pb inoculation at 14 dpi. Heatmap was generated based on the Z-score from 3 biological repeats for each group of samples.

To avoid bias in the disease severity assessment, we performed quantitative real-time PCR (qPCR) to monitor pathogen biomass in infected root tissues postinoculation. In agreement with the disease severity result, Pb biomass was barely detectable in all pmr4 mutant backgrounds at all time points examined (Fig. 2c), indicating that the SA signaling is not required for pmr4-mediated clubroot resistance. Interestingly, NahG and pad4-1 roots showed lower Pb biomass than Col-0 (Fig. 2c), indicating that SA signaling could enhance Pb susceptibility.

To test whether SA independence of pmr4-1-mediated clubroot resistance is Pb pathotype specific, we inoculated wild-type, pmr4-1, pmr4-1 pad4-1, and pmr4-1 NahG plants with Pb pathotypes/isolates 3A, 5X-LG2, and 5X-LG3. Resistance in both pmr4-1 pad4-1 and pmr4-1 NahG plants was consistent across all pathotypes/isolates and comparable to pmr4-1, while Col-0 and all SA-deficient single mutants remained susceptible (Fig. S2).

To further address the distinct requirements of SA defense pathway in pmr4-1 and determine whether pmr4-1 can confer dual resistance against PM and clubroot diseases, we simultaneously inoculated the same plants with conidiospores of the host-adapted Erysiphe cruciferarum (Ec) on the foliage and Pb resting spores in soil. As expected, typical PM symptoms were observed in Col-0, NahG, pad4-1, pmr4-1 NahG, and pmr4-1 pad4-1. In contrast, only pmr4-1 showed leaf lesion spots (Fig. 2d), indicative of SA-mediated hypersensitive reaction. In addition, parasitic structures of Ec were monitored by staining inoculated leaf samples with acidic aniline blue staining to Ec-inoculated leaf samples, which revealed enhanced resistance to Ec by the pmr4-1 mutation, as evidenced by decreased hyphal growth rate of Ec with fewer hyphal branches and a lower density than Col-0 in conidiophores and disease symptoms on the leaf surface at 7 dpi in pmr4-1 plants (Fig. S3). Furthermore, inactivation of the SA signaling pathway in pmr4-1 mutants led to increased conidiophore formation in Ec-inoculated leaves (Fig. 2e), confirming that SA is required for pmr4-1-mediated PM resistance. Taken together, we conclude that pmr4-1 exhibited dual resistance to PM and clubroot diseases in SA-dependent and SA-independent manners, respectively.

To ensure that the SA pathway was indeed inactivated in pad4-1 and NahG transgenic lines, we examined expression levels of the SA signaling marker gene PR1 in root tissues of above genotypes following Pb inoculation. Consistent with previous reports (Chen et al. 2016; Lovelock et al. 2016), PR1 was strongly induced in Col-0 but not in pmr4-1 roots at 21 dpi, while in early time points, PR1 expression in response to the Pb infection was delayed and then diminished in pmr4-1 roots (Fig. S4a). Furthermore, pad4-1 significantly reduced and NahG abolished PR1 expression in response to the Pb infection (Fig. S4a). These observations collectively confirmed that the pad4-1 mutation and NahG transgene blocked the SA signaling.

To investigate the influence of Pb infection on the entire SA pathway, we examined the expression of genes involved in SA biosynthesis, signaling, and responses in root tissues of Col-0 and pmr4-1 with or without Pb inoculation by RNA-seq. The results showed that the SA pathway genes in Col-0 roots were largely suppressed by Pb infection at 14 dpi, consistent with previous reports (Ludwig-Müller et al. 2015; Bulman et al. 2019; Djavaheri et al. 2019), while FMO1, PR1, PR2, and PR5 were slightly induced (Fig. 2f). In contrast, the transcript levels of many SA pathway genes increased moderately in the corresponding pmr4-1 roots, and they were not suppressed by Pb infection (Fig. 2f). These RNA-seq data were further supported by reverse transcription quantitative PCR (RT-qPCR) analysis of selected genes (Fig. S4b to g). Together with our phenotypic analyses, we conclude that although the SA pathway activity is not suppressed in pmr4-1 roots upon Pb infection, the pmr4-conferred clubroot resistance does not require functional SA signaling.

Genome-edited B. napus PMR4 mutants are dually resistant to clubroot and PM diseases

The pmr4-1 mutation did not show significant impact on Arabidopsis plant growth and development (Fig. S5), making PMR4 an ideal target for engineering clubroot resistance. Many crops belonging to the Brassicaceae family suffer from the clubroot disease (Hwang et al. 2012). Since loss of the PMR4 activity in Arabidopsis confers clubroot resistance, we investigated whether the PMR4 gene is conserved among these crops and whether targeting PMR4 orthologs in B. napus also leads to resistance to clubroot. The Arabidopsis PMR4 amino acid sequence was used to blast the Ensembl Plants database (https://plants.ensembl.org/Multi/Tools/Blast) focusing on Brassicaceae crops with high economic values. A phylogenetic analysis of candidate PMR4 orthologs (Fig. S6) shows that these genes are highly conserved and widely present across Brassicaceae species. Two genes in the B. napus genome, annotated as BnaC09g00800D (BnaPMR4.C09) and BnaA09g01630D (BnaPMR4.A09), encode proteins with 91% amino acid sequence identity to Arabidopsis PMR4 (E-values at or below 10⁻⁵⁶) (Fig. S7). Similar to AtPMR4, the predicted BnaPMR4.C09 and BnaPMR4.A09 are polytopic integral membrane proteins with 10 and 12 transmembrane helices, respectively (Fig. S8) and contain 2 conserved potential catalytic domains: 1,3-β-glucan synthase component FKS1-like domain (FKS1_dom1) and 1,3-β-glucan synthase domain (Hsieh et al. 2024).

To create B. napus PMR4 mutants that mimic Arabidopsis pmr4-1 and gsl5-1, 2 sgRNA constructs in Exon 2 capable of targeting both PMR4 genes (Fig. 3a and b) were designed and separately assembled into a CRISPR/Cas9 vector. These CRISPR/Cas9 sgRNA constructs were transformed into a B. napus doubled haploid line DH12075 by using Agrobacterium-mediated hypocotyl transformation (Zhou et al. 2002). While no mutation was yet found in sgRNA4 transformed plants, 2 sgRNA3 transformed lines were found to carry +1 frameshift mutations in both target genes. Line pmr4-3g37 harbored biallelic homozygous (T/T) insertions in BnaPMR4.C09 (Fig. S9a) and biallelic heterozygous (T/G) insertions in BnaPMR4.A09 (Fig. S9b), making it a T0 homozygous double mutant. On the other hand, line pmr4-3g41 carried monoallelic (T/−) insertions in both BnPMR4 genes (Fig. S9a and b), and was therefore a T0 heterozygous double mutant. The gene-edited homozygous pmr4-3g37 (T3 generation) and pmr4-3g41 (T2 generation) lines were evaluated for the clubroot resistance against Pb pathotype 3H alongside reference lines including DH12075 (parental) and sg4 (an sgRNA4 transgenic DH12075 line with unedited PMR4 genes). Pb-infected pmr4-3g37 and pmr4-3g41 plants remained healthy, displaying no aboveground symptoms such as chlorosis or stunting (Fig. 3c). Furthermore, no visible root swelling or galling was observed in these lines up to 28 dpi. At 35 dpi and beyond, these 2 B. napus pmr4 mutants formed much smaller galls and only portion of their roots compared to the susceptible lines, which formed large root galls on the main root of all plants (Fig. 3c). Importantly, the aboveground portion of Pb-infected pmr4-3g37 and pmr4-3g41 plants grew normally even at 42 dpi, whereas DH12075 and sg4 plants displayed stunted growth (Fig. 3c) and eventually died. Disease severity ratings (Fig. 3d) supported the observed resistance phenotypes. Additionally, droplet digital PCR (ddPCR) analysis showed significantly lower Pb biomass in the infected roots of pmr4-3g37 and pmr4-3g41 compared to DH12075 and sg4 (Fig. 3e). Similar resistance was observed against Pb pathotypes 3A and 5X-LG2, with pmr4-3g37 displaying resistance comparable to a CR control NAS6. In contrast, susceptible controls Westar (susceptible commercial line), DH12075 (parental) and 45H29 (a formerly resistant cultivar subsequently overcome by newer Pb pathotypes) (Jiang et al. 2020) were all susceptible under the same experimental conditions (Fig. S10a).

Figure 3.

Creation and clubroot disease test of B. napus pmr4 mutants. a) Comparison of predicted Arabidopsis and B. napus PMR4 genes. Filled and open boxes indicate exons and introns, respectively. Blue asterisk indicates the point mutation site in pmr4-1; green triangle indicates the T-DNA insertion site in gsl5-1; red arrows indicate target sites of sgRNAs for BnaPMR4 genes. b) B. napus sequences to show sgRNA target sites. PAM sequences are in bold. c) Representative root and shoot phenotypes of B. napus plants infected with Pb pathotype 3H. Scale bar = 5 cm. d) The clubroot disease symptom as shown in c) was recorded by a 0 to 3 class-scoring system. Values were obtained from at least 10 to 30 individual roots per genotype at each time point. Distinct colors indicate the proportions of different disease classes. e) Pb pathogen load quantified by ddPCR targeting on a segment of the internal transcribed spacer 1 region of Pb relative to B. napus housekeeping gene BnActin. Plants were inoculated with 1 × 10⁸ spore/mL Pb pathotype 3H resting spore inoculum and were assessed at the indicated dpi. Two-way ANOVA was performed, and different letters represent statistical significance (α = 0.05). Data represent means ± Sd. At least 5 individual roots from each genotype were collected as 1 biological repeat. Three repeats were assessed for each treatment.

The T0 heterozygous pmr4-3g41 line was self-pollinated, and the T1 segregants were tested for resistance against the newly-reported Pb pathotype 3A. Out of a population of 141 plants assessed, 10 were resistant while 131 were susceptible (Fig. S10b), which would fit well with the expected 2-gene segregation ratio of 15:1, provided both mutations are recessive. All 10 resistant plants contained homozygous mutations in both PMR4 genes. Similarly, when pmr4-3g41 T1 segregants were inoculated with Pb pathotype 3H, 8 out of 107 plants were resistant (Fig. S10c), and all resistant plants carried homozygous mutations in both PMR4 genes. These results demonstrate that B. napus homozygous PMR4 mutants are resistant to multiple Pb pathotypes and that both BnaPMR4s function as CS genes.

We also inoculated pmr4-3g37 plants with host-adapted Ec to determine whether B. napus pmr4 mutants are also resistant to PM. Indeed, pmr4-3g37 plants exhibited strong resistance to PM as canonical whitish powdery symptoms can be observed on parental DH12075 leaves, while no such disease feature was found on pmr4-3g37 leaves (Fig. 4a). In contrast, a double CR gene resistant line CPS14 (Tu et al. 2024) also showed susceptibility to the Ec infection (Fig. 4a). Microscopic observations of Ec parasitic structures following acidic aniline blue staining revealed fewer Ec hyphal branches at 7 dpi on pmr4-3g37 leaves than on DH12075 leaves (Fig. 4b, upper row). Furthermore, while Ec-inoculated DH12075 and CPS14 leaves showed prominent papillae-associated callose deposition at 2 dpi, the response was dramatically reduced in pmr4-3g37 leaves (Fig. 4b, lower row). This reduction correlated with decreased conidiophore development in Bnapmr4 mutants (Fig. 4c), consistent with phenotypes reported for the Atpmr4-1 mutant (Jacobs et al. 2003). These findings support the conclusion that BnaPMR4 is a promising candidate gene target for creating novel CR B. napus lines with the added benefit of PM resistance.

Figure 4.

Susceptibility of B. napus plants to the PM disease. a) Leaf phenotypes of DH12075, CPS14, and pmr4-3g37 at 40 d after Ec inoculation (lower row). Corresponding uninfected controls are also shown in the upper row. Scale bar = 4 cm. b) Upper row: representative images of hyphae growth and conidiophore production of Ec on plant leaf surfaces at 7 dpi. Fungal structures were stained with acidic aniline blue and examined by light microscopy. h, hyphae; cp, conidiophores. Scale bar = 100 μm. Lower row: callose deposition at the Ec-penetration sites stained with basic aniline blue and revealed with epifluorescence microscopy. Scale bar = 50 μm. c) Quantitative measurement of Ec conidiophore formation on Ec-inoculated leaves at 5 dpi. Data are means ± Sd from 30 colonies collected from 3 leaves. One-way ANOVA was performed, and different letters represent statistical significance (α = 0.05).

Pb secondary infection is compromised in pmr4-1

To understand why PMR4 inactivation confers resistance to the clubroot disease, we monitored the Pb development in Col-0 and pmr4-1 roots using multiple methods, including confocal laser-scanning microscopy (CLSM) (Liu et al. 2020a), cytological examination based on cross-section (Walerowski et al. 2018), and transmission electron microscopy (TEM), to track primary and secondary infections. Since 7 dpi typically marks the end of primary Pb infection, with stele colonization and symptom development beginning around 14 dpi (Liu et al. 2020a), we collected samples at 7, 10, and 14 dpi to time primary, primary-to-secondary and secondary Pb infection stages.

The primary infection appeared to be successful and completed in both Col-0 and pmr4-1 roots, as zoospores were released from zoosporangia in their epidermal cells at 7 dpi observed by CLSM (Fig. S11a; Videos S1 and S2). Zoospore release was assessed at 14 dpi by examining infected lateral roots near the hypocotyl in both Col-0 and pmr4-1, which showed comparable levels of zoospore release at the later infection stage (Fig. S11b). Notably, Z-stack videos captured the successful Pb secondary plasmodium colonization in the Col-0 cortex adjacent to an empty zoosporangium in the neighboring epidermal cell (Video S3). In contrast, although numerous empty zoosporangia were detected in pmr4-1 epidermal cells, secondary plasmodium was barely found in the inner cell layers (Video S4), suggesting dramatically reduced secondary infection in pmr4-1.

For cross-sectional microscopic analysis of secondary infection, lateral roots from the maturation zone, where pathogen invasion is initiated, were sampled. In Arabidopsis, periderm replaces epidermis during secondary growth in the hypocotyl, upper primary root, and old lateral roots, a process initiated by pericycle cell division, followed by endodermal programmed cell death and eventual detachment of epidermal and cortical cells (Wunderling et al. 2018). Consistent with our CLSM observations, comparable number of zoosporangia were found in epidermal cells from both Col-0 and pmr4-1 at 7 dpi (Fig. 5a), indicating successful primary colonization. In Col-0, cell deformation and swelling were evident by 10 dpi and became more pronounced at 14 dpi, accompanied by increased secondary plasmodia and the absence of a mature periderm (Fig. 5a). During the same period, Pb-infected pmr4-1 roots exhibited no visible cell deformation but instead showed an intensified periderm staining pattern, in which secondary plasmodia were rarely detected (Fig. 5a), suggesting that secondary infection is largely absent in pmr4-1 roots.

Figure 5.

Pb disease progression in the Col-0 and pmr4-1. a) Representative transverse sections (0.5 µm) of Col-0 and pmr4-1 roots stained with toluidine blue to show changes in response to the Pb infection, with enlarged one-quarter views in corresponding lower rows. Scale bars represent 100 µm. Black and red arrows indicate Pb zoosporangium and secondary plasmodium, respectively. Red brackets indicate periderm. b) Proportion of Col-0 and pmr4-1 cells harboring Pb secondary plasmodium (sp). At each time point, 3 independent sections from 3 individual plants were examined to determine secondary infections using Zeiss AxioPlan light microscopy. Data represent means ± Sd. n = 3. Two-way ANOVA analysis was performed to compare differences between Col-0 and pmr4-1 at 7, 10, and 14 dpi (α = 0.05). c) Expression of Pb pathogenesis marker genes to quantify secondary infection stages in Col-0 and pmr4-1. All data were normalized to the expression of PbActin1 and relative to the expression level of inoculated plants at 0 dpi. Two-way ANOVA was performed, and different letters represent statistical significance. Data represent means ± Sd. n = 3. d) Overview comparisons of Col-0 and pmr4-1 under TEM before and after Pb infection at 7, 10, and 14 dpi under the same magnification from outermost layer to xylem cells. Images of Pb-infected Col-0 at 10 and 14 dpi were stitched by serial overlap views due to the root swelling upon infection. Scale bar = 20 µm. Red asterisks indicate the Pb secondary plasmodium.

To quantify the establishment of Pb secondary infection, we calculated the proportion of host cells colonized by Pb secondary plasmodia. As expected, the proportion of colonized cells in Col-0 increased over time, whereas pmr4-1 roots consistently showed few if any infected cells (Fig. 5b). To further assess the secondary infection phase, we analyzed the expression of a previously characterized secondary phase-specific Pb marker gene intraflagellar transporter 172 (IFT172) (Tu, 2018), encoding a putative cilium- and microtubule-associated protein homolog essential for protist flagellar assembly and length maintenance, which was markedly induced in Col-0 upon infection but remained low in pmr4-1 during the entire infection cycle (Fig. 5c), indicating impaired secondary infection in this mutant.

TEM (Fig. 5d) further revealed that Pb growth was arrested in pmr4-1 root tissues while induced enlarged and deformed cells in Col-0 characteristic of advanced cortical and stele infection. We also measured the average size of healthy and Pb-infected root cells from both lines and observed that cell size gradually increased in Col-0 roots at 10 and 14 dpi, corresponding to 20-fold increase in cell volume. In contrast, pmr4-1 root cells did not change in volume following Pb infection (Figs. 5d and  S12a). Meanwhile, pmr4-1 root periderm cell walls were thicker than Col-0 in the absence of infection and became even thicker after Pb infection at 14 dpi (Fig. S12b), whereas cell walls from equivalent Col-0 root layers became thinner following Pb infection (Fig. S12b). Together, these observations indicate that loss of PMR4 prevents Pb secondary infection and thereby blocks completion of the disease cycle.

Pb infection leads to PMR4-dependent callose deposition in roots

PMR4 is required for wound-induced callose deposition at wound sites (Jacobs et al. 2003; Nishimura et al. 2003), which is thought to mimic plant response to pathogen invasion. A wound-induced callose accumulation assay showed robust callose deposition in Col-0 leaves, whereas this response was absent in pmr4-1 leaves (Fig. S13a), consistent with previous reports (Jacobs et al. 2003; Nishimura et al. 2003). Col-0 and pmr4-1 therefore served as positive and negative controls, respectively, for the wound-induced callose deposition response in B. napus leaves. Under the same experimental conditions, DH12075 and heterozygous pmr4-3g41 T0 leaves displayed clear callose deposition at the wound site 24 h after wounding. In contrast, callose accumulation was absent at the wound site in both pmr4-3g37 and homozygous pmr4-3g41 T1 mutant leaves (Fig. S13a). Because wound-induced callose deposition remained detectable in heterozygous pmr4-3g41 leaves, both Bnapmr4 alleles are considered loss-of-function mutations. These results indicate that PMR4 is required for wound-induced callose deposition in both Arabidopsis and B. napus.

Considering that PMR4 functions as a callose synthase responsive to fungal pathogens and wounding in leaves (Nishimura et al. 2003), we asked whether the Pb infection also induces callose deposition in B. napus root tissues and whether PMR4 is required for this process. To test this, DH12075 and pmr4-3g37 roots were examined by aniline blue staining, with or without Pb pathotype 3H inoculation. At 14 dpi, transverse root sections were prepared and stained with aniline blue to visualize callose (Fig. S13b). In uninoculated roots, only background signals, namely phloem callose and xylem lignin autofluorescence, were observed in both DH12075 and pmr4-3g37, consistent with previous reports (Truernit et al. 2008; Xie et al. 2011). Following Pb inoculation, DH12075 roots showed clear callose accumulation (white arrows), whereas corresponding pmr4-3g37 root tissues lacked callose deposition (Fig. S13b).

To further examine Pb infection-triggered callose deposition and PMR4 localization in Arabidopsis roots, we performed an immunohistochemistry (IHC) assay. Both mock and Pb-infected Col-0 and pmr4-1 roots were incubated with β-1,3-glucan- and PMR4-specific primary antibodies, followed by corresponding fluorescent secondary antibodies to detect callose deposition and PMR4 distribution, respectively. In uninfected Col-0 and pmr4-1 roots, only phloem callose signals were frequently observed (Fig. 6a), in agreement with the background callose detected in uninoculated B. napus roots stained with aniline blue (Fig. S13b). Since no PMR4 signal was detected overlaying with such phloem callose signals, the phloem callose synthesis is likely to be independent of PMR4. In contrast, a strong callose signal was observed from Pb-infected host cell wall apoplasts in Col-0 roots, which was proximal to the PMR4 signal found in plasma membrane, whereas no such callose and PMR4 distribution patterns could be observed from Pb-infected pmr4-1 roots (Fig. 6a). Moreover, fluorescence signal correlation analysis of Pb infection-triggered callose and PMR4 suggests a spatial association between the 2 signals (Fig. 6b). Quantification of the mean fluorescence intensity in the periderm revealed elevated callose and PMR4 levels in Pb-infected Col-0 roots but not in pmr4-1 roots (Fig. 6c), supporting the notion that Pb infection induces callose deposition in a PMR4-dependent manner. Interestingly, following Pb infection, a punctate callose deposition pattern was observed in aniline blue-stained pmr4-3g37 roots (Fig. S13b, red arrowheads) and in IHC-visualized pmr4-1 roots (Fig. 6a, white arrowheads). This pattern is characteristic of a plasmodesmata (PD)–associated callose deposition (Vatén et al. 2011) and reminiscent of the callose accumulation seen on Ec-infected pmr4-1 leaves (Jacobs et al. 2003), suggesting possible activity of other callose synthase isoforms at PD, although the biological significance of this pattern remains unclear.

Figure 6.

Immunohistochemical detection of callose and PMR4 in Arabidopsis roots. a) IHC results of callose and PMR4 protein from healthy and Pb-infected Col-0 and pmr4-1 at 14 dpi. Callose and PMR4 antigens were detected by specific primary antibodies derived from mouse and rabbit hosts. Further visualization of 2 signals was performed by incubation of anti-mouse_Alexa_488-conjugated and anti-rabbit_Alexa_594-conjugated secondary antibodies. White arrowheads indicate the punctate PD-localized callose deposition. Regions containing periderm walls were marked by white squares on upper panels and magnified as shown in corresponding lower panels. Scale bar = 20 µm. b) Signal correlation analysis of callose and PMR4 in Pb-infected Col-0 roots. A representative result from 10 measurements is shown. c) Statistical analysis of the mean fluorescence intensity observed from a). Ten to 15 randomly picked periderm fields from at least 5 independent roots were analyzed by Fiji software. Two-way ANOVA was used to show the statistical difference. (α = 0.05). Letters represent significant differences.

Inactivation of JA pathway partially restores pmr4-1 susceptibility to Pb

JA signaling is involved in wound- and pathogen-induced plant immunity (Koo and Howe 2009; Howe et al. 2018; Zhang et al. 2025) but not required for pmr4-mediated resistance to PM (Nishimura et al. 2003). In addition, JA plays a key role in mediating cell wall responses by regulating genes involved in cell wall remodeling and reinforcement during stress and developmental signaling (Howe et al. 2018). As PMR4-mediated callose deposition is a hallmark of cell wall reinforcement during pathogen invasion, we asked whether JA signaling is required for pmr4-conferred clubroot resistance by introducing the dde2-2 mutation known to abolish JA biosynthesis (von Malek et al. 2002) into pmr4-1 plants. Surprisingly, Pb-infected pmr4-1 dde2-2 developed smaller root galls than Col-0 and dde2-2 at 21 dpi, while swollen root tissues were rarely observed in pmr4-1 roots (Fig. 7a). Quantitative analyses revealed a partial restoration of susceptibility with increased DSI (Fig. 7b) and pathogen load (Fig. 7c) in pmr4-1 dde2-2 double mutants compared with pmr4-1 single mutants. Furthermore, cytological examination of Pb-infected pmr4-1 dde2-2 roots revealed massive secondary plasmodia (Fig. 7d) comparable to Pb-infected Col-0 roots (Fig. 5a). Together, these results indicate that pmr4-1-mediated resistance to clubroot requires intact JA signaling, similar to Rcr1-conferred clubroot resistance (Chu et al. 2014).

Figure 7.

Effects of blocking JA signaling on pmr4-1 root response to the Pb infection. a) Representative root images of Col-0, pmr4-1, and pmr4-1-associated mutants with and without Pb pathotype 3H inoculation at 21 dpi. Scale bar = 2 cm. b) Quantitative analysis of Arabidopsis clubroot disease severity as shown in a). DSI values are means ± Sd obtained from 3 groups of biological repeats, each containing 15 individual plants. One-way ANOVA was performed, and different letters represent various disease severity levels as well as statistical significance. c) Relative Pb biomass as measured by qPCR of a Pb internal transcribed spacer 1 region (Wen et al. 2020). AtUBQ10 was used as a host tissue normalization reference. At least 3 plant roots were pooled to form 1 mixed sample with 3 biological replicates for each line. Data are means ± Sd obtained from 3 technical replicates. One-way ANOVA analysis for Pb relative biomass at 21 dpi was performed. Letters represent significant differences. d) Histological analysis of the roots of mock or Pb-inoculated dde2-2 and pmr4-1 dde2-2 plants at 14 dpi. Representative Pb secondary infections were indicated with arrowheads. Representative images were selected from the 5 independent roots with a similar phenotype. Red brackets indicate periderm. Scale bar = 1 mm. e) Transcriptome profiling of the gene expressions involved in JA synthesis, signaling, and response in the roots before or after Pb inoculation at 14 dpi. Heatmap was generated based on the Z-score from 3 biological repeats for each group of samples.

To further elucidate the molecular mechanisms underlying clubroot susceptibility and pmr4-1-mediated resistance, we performed RNA-seq analysis on Col-0 and pmr4-1 plants following Pb inoculation (Fig. 7e). Transcriptomic profiling revealed that Pb infection suppressed the expression of several key genes involved in JA biosynthesis, signaling and response in Col-0 root tissues. In contrast, most of them were derepressed or even further induced (eg, VSP1 and VSP2) in infected pmr4-1 roots (Fig. 7e). Notably, many of these JA-related genes were also upregulated in noninoculated pmr4-1 plants compared with Col-0, suggesting an antagonistic relationship between JA-mediated signaling and PMR4-dependent callose production. Transcriptomic changes in selected JA-related genes were further validated by RT-qPCR, yielding results consistent with the RNA-seq analysis (Fig. S14).

Collectively, these results demonstrate that Pb infection suppresses JA signaling in Col-0 roots, whereas loss of PMR4 function leads to derepressing JA signaling in pmr4-1 roots. The elevated JA pathway activity in pmr4-1 plants likely contributes to their enhanced resistance to clubroot.

JA signaling is required for and positively correlates to lignin-like compound synthesis and possibly suberin deposition in the periderm

Thickened periderm cell walls in pmr4-1 roots, as observed by TEM (Figs. 5d and  S12b), suggest that the altered cell wall architecture plays a role in the clubroot resistance. We investigated potential cellular and molecular mechanisms underlying this phenotype using several approaches. Firstly, UDP-Glc serves as a precursor for both callose and cellulose biosynthesis; hence, the loss of PMR4 callose synthase might tip the balance toward cellulose accumulation in cell walls (Liu et al. 2023). Surprisingly, no significant difference in cellulose abundance, as revealed by calcofluor white (CFW) fluorescence (Ursache et al. 2018), was observed among tested lines regardless of the Pb infection (Fig. 8a and b), indicating that the cellulose content in pmr4-1 cell walls is unaltered and does not contribute to pmr4-mediated clubroot resistance.

Figure 8.

Effects of pmr4-1 mutation and JA pathway on periderm root cell wall composition. a) CLSM-based cellulose and lignin-like phenolic compound detection in Col-0, pmr4-1, and JA-deficient-associated mutants. Cellulose and lignin-like compounds were stained with CFW and BFu, respectively. White squares marking periderm were enlarged below each corresponding image. Scale bars = 20 µm. A pseudocolor scheme was given to visualize the spatial distribution and relative abundance of the detected cell wall components, with colors ranging from dark to bright indicating increasing fluorescence intensity. BF, bright field. b) Statistics of the relative deposition amount of cellulose and lignin-like phenolic compound observed from a). The periderm was selected for mean fluorescence intensity analysis by Fiji software. Ten randomly selected fields from 5 independent roots were calculated, and ordinary 1-way ANOVA was used to show the statistical significance (α = 0.05). Letters represent significant differences. c) Heatmap of transcriptome profiling of the gene expressions involved in phenylpropanoid synthesis, phenolic polymer assembly, suberin synthesis, and fatty acid synthesis processes in the roots with or without Pb inoculation at 14 dpi from Col-0 and pmr4-1. Heatmap was generated based on the Z-score from 3 biological repeats for each group of samples. d) Suberin detection from Pb-infected Col-0 and pmr4-1 at 14 dpi. Cryostat-sections of Pb-infected Col-0 and pmr4-1 roots at 14 dpi were stained with FY088 and observed under CLSM. Representative images focusing on the periderm phellem cell wall (upper row) were shown along with corresponding 3D surface plots (lower row). Scale bars = 100 µm. e) TEM images of periderm phellem cell wall. CW, polysaccharide cell wall; SL+, suberin layer with a lamellated substructure; SL−, a largely amorphous suberin layer with little substructure; ph, phellem. Scale bars = 0.5 µm. f) Suberin layer thickness. The thickness of the suberin layer visualized in TEM was measured in at least 4 phellem cells by measuring the thickness of the suberin lamellae. Values represent the means ± Sd, n ≥ 50. Different lowercase letters indicate significant differences, as determined by 1-way ANOVA, α = 0.05.

Secondly, wounding triggers JA signaling and promotes lignin deposition around wound sites in plant tissues (Xu et al. 2024). Since JA signaling is required for pmr4-mediated clubroot resistance (Fig. 7), we also assessed the deposition of lignin-like phenolic compounds using a histochemical approach. Basic fuchsin (BFu) staining, which specifically labels lignified cell walls (Ursache et al. 2018), revealed increased lignin-like phenolic components in noninfected pmr4-1 roots compared with Col-0, with a further increase following Pb infection. In contrast, BFu fluorescence was barely detectable in JA-deficient dde2-2 mutants, regardless of PMR4 genotype or infection status (Fig. 8a and b). As a complementary approach, cell wall autofluorescence detected using 405-nm laser-based confocal microscopy (Perez-de-Lis et al. 2024) showed consistent patterns of lignin-like phenolic polymer deposition (Fig. S15a and b).

Thirdly, RNA-seq transcriptomic analysis revealed that genes associated with phenylpropanoid biosynthesis, as well as the assembly and deposition of phenolic polymers, were significantly suppressed in Col-0 roots following Pb infection, whereas these genes remained active or were even induced in pmr4-1 roots. Lignification in developing Arabidopsis root tissues is closely linked to the suberin layer formation at the inner cell walls of periderm cells, which involves cross-linking between aromatic phenolic compounds and fatty alcohols (Wunderling et al. 2018; Gully et al. 2024). Consistent with this relationship, genes involved in fatty acid biosynthesis, which provide specific precursors for suberin production, exhibited expression patterns similar to those of phenolic compound biosynthesis and polymerization genes. Similarly, genes associated with suberin biosynthesis and deposition showed comparable expression trends (Fig. 8c), and these patterns were further validated by RT-qPCR analysis (Fig. S16). Collectively, genes involved in phenylpropanoid biosynthesis and polymerization, fatty acid biosynthesis, and suberin formation and deposition were coordinately regulated during clubroot infection and in the pmr4-1 mutant. These results suggest that suberization is downregulated in root tissues during susceptible clubroot disease development but is upregulated upon loss of PMR4 function.

To investigate suberization in root tissues during Pb infection, we utilized fluorol yellow 088 (FY088) staining to visualize suberin deposition (Chen et al. 2024). Following infection, strong FY088 fluorescence was observed at the phellem cell walls of pmr4-1 roots, whereas weaker signals were detected in Col-0 (Fig. 8d, upper row). A 3D surface plot further confirmed enhanced suberin accumulation in pmr4-1 compared with Col-0 (Fig. 8d, lower row). TEM imaging (Fig. 8e) of the same root regions revealed distinct suberin lamella structures between Col-0 and pmr4-1. In Col-0 roots, cell walls displayed high electron density with fewer lamella layers, indicative of predominantly aliphatic-rich suberin. In contrast, Pb-infected pmr4-1 roots exhibited multiple layers of a lamellated, low electron density suberin, consistent with an aromatic-enriched structure (Gully et al. 2024). Notably, while Pb infection compromised the suberin layers in Col-0 roots, pmr4-1 roots showed substantial suberin thickening (Fig. 8f). The periderm tissues of pmr4-1, enriched in phenolic polymers and suberin, resemble the cork layer found in woody plants (Teixeira 2022) and could serve as a physical barrier restricting pathogen spread within root tissues during secondary infection.

Discussion

The Brassicaceae family includes many valuable crops, among which canola varieties of B. napus are cultivated worldwide for edible oil and other applications due to their high nutritional value and energy outputs. Clubroot disease poses a major threat to oilseed production worldwide. Growing CR cultivars remains the most effective disease management strategy; however, resistance conferred by CR genes can be overcome readily due to the pathogen diversity and evolution. CR genes have been found mainly in other Brassica species that often need to be introduced into B. napus varieties through traditional breeding, which is time consuming and difficult to keep pace with the evolving pathogen population (Hwang et al. 2012). Recently, a novel gene WeiTsing (WTS) from Arabidopsis was identified and characterized. B. napus plants carrying WTS display a broad-spectrum clubroot resistance (Wang et al. 2023), making it a promising CR gene for clubroot management. In this study, we developed a strategy to screen available Arabidopsis mutants aiming at identifying candidate CS genes that can be edited in Brassicaceae crops with potential durable clubroot resistance against multiple Pb pathotypes.

Our genetic screen to date has identified several Arabidopsis mutant lines displaying various levels of clubroot resistance, and one of them (pmr4-1) showed strong resistance. Several pieces of evidence in this study support the conclusion that PMR4, a previously reported S gene for PM (Nishimura et al. 2003), is also an S gene for clubroot. Firstly, 2 mutant lines carrying independent alleles, pmr4-1, containing a premature stop codon in Exon 2, and gsl5-1, a T-DNA insertion in the same exon, showed strong resistance to clubroot. Secondly, although it is unclear whether pmr4-1 and gsl5-1 are null mutations, they behave as loss-of-function alleles and are indeed recessive mutations. Thirdly, in B. napus, resistance was observed only when both PMR4 orthologs were disrupted by gene editing, indicating functional redundancy. Finally, wound- and pathogen-induced callose deposition experiments confirmed that both Arabidopsis and B. napus pmr4 mutants are loss-of-function mutations. Although other CS genes have been previously reported (Chen et al. 2016; Walerowski et al. 2018), PMR4, to the best of our knowledge, is the first CS gene whose inactivation results in such strong clubroot resistance. Since the PMR4 gene is highly conserved within the Brassicaceae family, we created the corresponding pmr4 double mutant lines in B. napus by genome editing and demonstrated that they were dually resistant to both clubroot and PM diseases. Since the PMR4 callose synthase activity is stress-induced (Jacobs et al. 2003; Nishimura et al. 2003) and pmr4-1 mutant plants do not exhibit apparent penalties under normal growth conditions, we anticipate that PMR4 orthologs in Brassicaceae crops can be targeted via gene editing to enhance clubroot resistance while maintaining desired agronomic traits, although field-level fitness remains to be confirmed.

Perhaps the most striking findings from this study were that, while the PM resistance conferred by pmr4-1 depended on the functional SA but not JA pathway (Nishimura et al. 2003), its resistance to the clubroot disease was dependent on JA but independent of SA response, suggesting that pmr4-1 confers resistance to these 2 pathogens through distinct mechanisms. In the case of PM resistance, the elevated SA signaling in pmr4 mutants was regarded as an underlying mechanism (Nishimura et al. 2003). Our studies convincingly demonstrated that, unlike observations in leaves, the pmr4-mediated clubroot resistance was largely due to derepressed JA response in roots. Our experimental data favor a notion that PMR4 functions as a host susceptibility factor required for the completion of Pb infection cycle. Firstly, Pb induced PMR4-dependent callose deposition in Arabidopsis and B. napus roots, reminiscent of wound- and Ec-induced callose deposition in Arabidopsis leaves (Jacobs et al. 2003). Secondly, we showed that pmr4-1 blocks Pb secondary infection, resulting in an incomplete Pb disease cycle. Notably, primary infection often proceeds in CR canola varieties (Xu et al. 2018; Wang et al. 2023; Tu et al. 2024) and even in some nonhosts (Liu et al. 2020b), reinforcing the notion that restricting Pb secondary infection is the primary target to achieve clubroot resistance. Finally, quantitative pathogen biomass analysis confirms that disrupting PMR4 callose synthase suppresses Pb proliferation and gall formation.

This study suggests that pmr4-mediated and JA signaling–dependent periderm cell wall reinforcement plays a critical role in the protection against Pb infection. Firstly, TEM image analysis revealed thickened root periderm cell walls in pmr4-1 and particularly in response to the Pb infection; in contrast, the mature periderm was not observed in Pb-infected Col-0 roots, suggesting that Pb infection is associated with incomplete periderm maturation in roots of a susceptible host. Secondly, lignin-like and suberin compounds were accumulated more in pmr4-1 than in Col-0 cell walls and likely formed specialized structures. Thirdly, the above changes were positively correlated with transcriptomic data showing increased transcript levels in related biosynthetic pathways. Finally, the increased lignin-like compounds in pmr4-1 roots depended on functional JA signaling, which was also correlated with the JA pathway transcriptomic profile. Based on the above analyses, we propose a working model as shown in Fig. 9, in which the Pb pathogen suppresses JA signaling and prevents periderm maturation to facilitate its secondary infection. Loss of PMR4 reinstates JA signaling, leading to increased phenolic compound and suberin synthesis to promote root periderm maturation, which in turn forms a barrier to prevent Pb secondary infection. Apparently, several questions remain to be addressed by future investigation. For example, is enhanced JA signaling alone sufficient to promote phenolic compound biosynthesis and root periderm maturation? Is lignin-like or suberin, or both in periderm, required for the protection against Pb infection? How does the Pb pathogen utilize PMR4 callose synthase to inhibit JA signaling?

Figure 9.

A working model depicting mechanisms of pmr4-mediated resistance to Pb in roots. Pb zoospores are released from zoosporangium in root hairs or epidermal cells in both Col-0 and pmr4-1, indicating successful primary infection. However, while Pb pathogen can complete its disease cycle in Col-0 roots and form resting spores in cortical cells and the stele, secondary infection is rarely found in pmr4-1 roots. Consequently, Pb infection causes giant galls in Col-0 roots, while pmr4-1 roots barely develop symptoms. Diagrams on top depict the host periderm cell wall or equivalent layers enlarged from dashed circles. In Col-0 roots, Pb invasion triggers the PMR4-dependent callose deposition. JA signaling is repressed upon pathogen infection, leading to compromised lignin-like phenolic compound and suberin syntheses. In pmr4-1 periderm, lack of PMR4-dependent callose deposition correlates with derepressed JA signaling and enhanced synthesis of lignin-like phenolic compounds and suberin, which cross-link other cell wall components to form a physical barrier to prevent Pb zoospore penetration into the stele.

While this study was in progress (initial manuscript was submitted and posted on bioRxiv on 19 September 2024), Dr Liu and his colleagues reported similar findings (Wu et al. 2025). Hence, 2 research teams independently found that loss of PMR4/GSL5 confers strong clubroot resistance in a JA-dependent but SA-independent manner. Our studies report that JA-dependent root periderm cell wall homeostasis plays a critical role in pmr4-mediated clubroot resistance. Since the JA signaling plays important roles in secondary cell wall development by promoting lignin and suberin biosynthesis (Hyde et al. 2018; Im et al. 2024; Xu et al. 2024; Jiang et al. 2025), which are characteristic compounds found in periderm tissues (Wunderling et al. 2018; Serra et al. 2022) and have been implicated in defense against both biotic (Du et al. 2011; Thangavel et al. 2016) and abiotic (Lux et al. 2011; Ranathunge et al. 2016; de Silva et al. 2021) stresses, the pmr4 mutation may constitute even broader stress responses in higher plants.

Materials and methods

Plant materials and growth conditions

A. thaliana wild-type (Columbia-0) and mutant lines used in this study were either purchased from ABRC (https://abrc.osu.edu/) and NASC (https://arabidopsis.info/BasicForm) or kindly provided by other research groups. Detailed information is listed in Table S1. All oligonucleotides used as primers for genotyping, real-time PCR and ddPCR analyses are listed in Table S2. A doubled haploid canola (B. napus) cultivar DH12075 was used to generate transgenic pmr4-3g37 and pmr4-3g41 plants as described below. For Arabidopsis seedling culture, seeds were sterilized in 10% bleach for 15 min, rinsed in sterile ddH₂O 3 times, and then germinated on a 1/2 strength MS medium containing 0.6% agar. For clubroot or PM disease tests, Arabidopsis and B. napus seeds were sown and grown in the soil Sunshine Mix #5 Natural & Organic (Sun Gro Horticulture Canada Ltd). Arabidopsis and B. napus plants were grown at 22 °C under 16 h ∼125 μE m⁻² s⁻¹ light and 8 h dark photoperiod conditions.

Pb and Ec inoculation

Resting spore inocula from Pb pathotypes 3H, 3A, 5X-LG2, and 5X-LG3, the designation based on Canadian Clubroot Differential system (Tso et al. 2021), were prepared using galled roots of infected B. napus plants. The galled roots were homogenized in a blender with distilled water, and then plant residues were removed from the suspension by filtering through 8 layers of cheesecloth. The resting spore density was measured by hemocytometry and then diluted in ddH₂O to a final density of 1 × 10⁷ and 1 × 10⁸ spores per mL for Arabidopsis and B. napus inoculation, respectively. For the Arabidopsis inoculation, each 14-d-old seedling was inoculated with 1 mL of resting spore inoculum. B. napus were inoculated twice with 1 mL spore inoculum each, once at the seeding and a second time at 7 d after planting when seedlings emerged. For both Arabidopsis and B. napus plants, the clubroot disease severity was rated using a modified 0 to 3 scoring system (Kuginuki et al. 1999). A score of 0 indicates no disease; 1, very small galls on the lateral roots but without deformation on the primary roots; 2, medium size galls on primary roots; and 3, severe giant gall formations in the primary roots with or without very few swollen lateral roots left. The clubroot symptoms were rated for 3 to 4 biological replicates, and every 15 plants were pooled and rated as 1 biological repeat. The DSI was calculated according to the formula:

$$\begin{gathered} \mathrm{DSI}=\frac{\sum (\mathrm{rating}\mathrm{class})\times (\mathrm{number}\mathrm{of}\mathrm{plants}\mathrm{in}\mathrm{the}\mathrm{rating}\mathrm{class})}{\mathrm{total}\mathrm{number}\mathrm{of}\mathrm{plants} \\ \mathrm{times}3}\times 100 \end{gathered}$$

Ec was maintained and propagated on B. napus plants. Approximately 4-wk-old Arabidopsis plants were inoculated with Ec conidiospores at a density of 5 to 10 conidia/mm². For B. napus, tested plants were placed in settling towels and inoculated with conidia by tapping the infected B. napus leaves above the tested plants. The inoculated plants were kept away from air flows for at least 3 h after the conidia inoculation to obtain the desired germination rates of the pathogen.

ddPCR and RT-qPCR

For the sampling of Pb-inoculated root samples, 5 root galls or 6 to 7 nonswollen roots were pooled as 1 biological repeat, with at least 3 repeats for each treatment. Root samples were collected and cut by scissors and ground by Lysing Matrix Tubes (FastPrep) or by liquid nitrogen. The homogenized mixture was used either for genomic DNA extraction by using a DNeasy Plant Mini Kit (QIAGEN) or for total RNA extraction using a RNeasy Plant Mini kit (QIAGEN) following the instructions.

The biomass of Pb was quantified by qPCR or ddPCR using the QX200 ddPCR System (Bio-Rad) following the manufacturer's instructions. For each ddPCR sample, a 20-µL reaction contained 10 µL of probe supermix (Bio-Rad), 1 µL each of 22.5 µm forward and reverse primers, 1 µL of 6.25 µm probe, and 2 µL genomic DNA. The mixture was partitioned into droplets using a droplet generator, transferred to a 96-well PCR plate and followed by a standard PCR reaction. Droplets were read using the droplet reader, and data were analyzed with QuantaSoft Software (Bio-Rad). A fluorescence threshold of 2,000 was applied to distinguish positive droplets from background signals.

For RT-qPCR, the first-strand cDNA was synthesized from RNA using an iScript cDNA Synthesis kit (Bio-Rad). qPCR was performed against the first-strand cDNA in a qPCR detection system using the iQ SYBR Green Supermix and a passive reference dye (Bio-Rad). The qPCR was run in a 20 μL mixture containing 0.5 μL each of 10 μm forward and reverse primers and 100 ng of the cDNA template. The calculation for each sample was based on the average of 3 technical replicates. The ΔΔCT method (Livak and Schmittgen 2001) was used to analyze relative transcript abundance. Gene expression was normalized to the endogenous housekeeping genes AtUBQ10 or PbActin1 as indicated in the corresponding figure legends.

Histochemical staining of plant samples

To observe PM structures, Ec-inoculated leaves were detached and collected at different time points and fixed in 1:3 (v:v) acetic acid:ethanol. An acidic aniline blue solution (2 mg/mL, pH 5.0) was used to stain Ec fungal structures.

Pb primary infection and PM-triggered callose deposition were monitored as previously described by Liu et al. (2020a) and Qin et al. (2021), respectively.

To examine root morphology by Toluidine Blue-O, the entire procedure from sample collection to microscope observation was performed as previously described (Tu et al. 2024).

To detect callose deposition in Pb-infected B. napus roots, lateral root samples proximal to hypocotyl were harvested at 14 dpi and fixed overnight with a solution containing 37% formaldehyde, glacial acetic acid, 95% ethanol and deionized water at a volume ratio of 50:5:10:35. Samples were then embedded in the Tissue Plus O.C.T compound (Leica Microsystems), sectioned with a Leica CM1860 cryostat microtome to 50 µm, stained with 0.05% to 0.1% aniline blue solution for 15 min and rinsed 3× 5 min with distilled water before analysis.

To observe cellulose and lignin-like compounds, Electron Microscopy Sciences LR White-embedded roots were cross-sectioned into 0.5-μm thick by an ultramicrotome (Reichert-Jung Ultracut E Ultramicrotome) before being collected on poly-L-lysine-coated slides. 0.1% (w/v) CFW and 0.1% (w/v) BFu dissolved in ClearSee solution were applied onto the root cross-section for overnight staining at 4 °C. Slides were rinsed 3× 5 min with distilled water before mounting.

To observe suberin in Arabidopsis cryostat cross-sections, a 0.4% (w/v) FY088 stock solution in DMSO was prepared. 0.01% (v/v) FY088 diluted in lactic acid was applied to incubate the root samples at 70 °C for 30 min, followed by 2× 5 min rinse with lactic acid at 70 °C before mounting slides.

Microscopy observations

Light microscopic images were taken using a Zeiss Axioplan 2 upright fluorescence microscope. Confocal images were taken on a Zeiss LSM 880 confocal laser scanning microscope. Excitation/emission wavelengths were 405/445 to 480 nm for DAPI and CFW, 405/480 to 515 nm for aniline blue, 405/425 to 523 nm for autofluorescence, 488/499 to 540 nm for HCS LipidTox Green and GFP, 488/490 to 550 nm for FY088, 543/600 to 650 nm for BFu, and 594/599 to 734 nm for RFP.

Immunofluorescence

Both healthy and Pb-infected upper lateral roots ∼1 cm from the hypocotyl were collected, and specimens were prepared as described (Amsbury and Benitez-Alfonso 2022). Vector TrueVIEW autofluorescence quenching kit (MJS BioLynx, VECTSP8400) was used following the instructions to reduce autofluorescence. The specimens were mounted with VECTASHIELD Vibrance Antifade mounting media (Vector Laboratories, H-1700) and examined with Zeiss 880 confocal microscope. Antibodies used in this study include β-1,3-glucan-specific (Biosupplies 400-2, 1:80 dilution) and PMR4-specific (Agrisera AS21 4567, 1:50 dilution) primary antibodies, anti-mouse_Alexa_488 (R&D System IC002G, 1:100 dilution) and anti-rabbit_Alexa_594 (Invitrogen A-21207, 1:100 dilution) secondary antibodies.

Transmission electron microscopy

Polysciences Spurr Low Viscosity Embedding Kit (01916-1) was used to embed root samples. Ultra-thin cross-sections of around 90 nm were collected onto formvar-coated 1 × 2 mm slotted Cu grids and stained with 2% uranyl acetate for 12 min by floating on a drop of stain. After an extensive wash with distilled water, the specimen was stained with Reynold's Lead citrate solution (Ted Pella, Inc.) for 6 min in a closed chamber, dipped in 0.02 N NaOH 20 times, washed in distilled water, and then observed with a Hitachi HT7700 transmission electron microscope.

Cell wall component measurements

Fiji software (Schindelin et al. 2012) was used for analysis. For the measurement of mean fluorescence intensity in a given field, all analyzed CLSM-based images were collected from the same round of assay under the same treatment conditions. The same parameter and threshold were applied. For the measurement of cell wall area or thickness, all analyses were based on the TEM images. TEM images were first fed into Fiji software. The cell area was measured by selecting along the plasma membrane, while cell wall thickness was determined by the distance between adjacent cell plasma membranes.

CRISPR/Cas9 vector assembly

Sequence-specific single-guide RNAs (sgRNAs) were designed by using a web-based tool CRISPR-P (http://crispr.hzau.edu.cn/cgi-bin/CRISPR2/CRISPR). Target sites were selected based on their location in the target genes, GC% content and putative off-targets. sgRNA-F/R primers (Table S2) were annealed and inserted between 2 BsaI sites of a binary vector pHEE401 obtained from Dr Qijun Chen (China Agriculture University, China), resulting in pHEE-BnaPMR4-sgRNAs. Plasmid pHEE401 contains a hygromycin resistance marker driven by a cauliflower mosaic virus 35S promoter, a Cas9 coding sequence driven by an egg cell–specific promoter EC1.2, and the cloned sgRNA is driven by a U6-26p promoter (Wang et al. 2015).

B. napus transformation

The B. napus doubled haploid line DH12075 was used as the transformation recipient in this study. The Agrobacterium tumefaciens–mediated hypocotyl transformation protocol (Zhou et al. 2002) was followed.

Transgenic plant genotyping

Genomic DNA from leaf tissues were used as PCR templates to amplify DNA flanking the CRISPR target sites using specific primers (Table S2). The PCR products were directly sequenced, and the sequencing chromatograms were analyzed to identify and distinguish mutated from wild-type sequences adjacent to the protospacer adjacent motif (PAM).

Transcriptome profiling

The roots of Arabidopsis plants with or without Pb inoculation were harvested at 14 dpi with 3 replicates, in which 5 to 10 individual plants were pooled, and immediately frozen in liquid nitrogen. High-quality RNA was extracted using the RNeasy Kit (QIAGEN #74104). cDNA libraries were prepared using the Illumina TruSeq V1.5 Kit. Paired-end sequencing was performed at the Next-Generation Sequencing Facility, University of Saskatchewan, using NextSeq 2000 sequencing platform. Raw reads prefiltering was performed using Trimmomatic (Bolger et al. 2014) to trim the adaptor sequences, filter low-quality reads and eliminate short reads. STAR (Dobin et al. 2013) was used for fast and accurate alignment of cleaned reads to the A. thaliana TAIR10 reference genome. Reads quantification was performed using the Salmon tool (Patro et al. 2017). To summarize expression levels from the transcript level to the gene level, we used the R tximport package (Soneson et al. 2015) to process the outputs from Salmon. Expression matrixes between genes and all samples were then generated, and differentially expressed gene analysis was performed through R DESeq2 package (Love et al. 2014). Heatmaps were generated based on Z-score, which is calculated as log₂ (TPM + 1) (transcript per million reads [TPM]). The RNA-seq data were submitted to the Gene Expression Omnibus repository (accession number GSE317839).

Author contributions

L.W. developed the mutant screen strategy; R.L. and R.W. conducted the mutant screen. R.L. characterized Arabidopsis pmr4 mutant phenotypes under supervisions of J.T. and Y.W. R.W., K.Y., and X.L. created and characterized B. napus pmr4 mutants under supervisions of T.D. and G.P. P.G. analyzed the RNA-seq data. L.W. and W.X. designed and supervised overall experiments; R.L., L.W., and W.X. wrote the manuscript with inputs from all coauthors.

Supplementary material

Supplementary material is available at The Plant Cell online.

Funding

This work was funded by the Natural Sciences and Engineering Research Council of Canada Alliance grant ALLRP 576795-22 and SaskOilseeds/Western Grains Research Foundation (WGRF) grant ADF 20210893 to W.X.

Data availability

The RNA-seq data were submitted to the Gene Expression Omnibus repository (accession number GSE317839).