Genetic and physiological analysis of the relationship between partial resistance to clubroot and tolerance to trehalose in Arabidopsis thaliana


Author for correspondence:
Antoine Gravot
Tel: + 33 223 485139


  • In Arabidopsis thaliana the induction of plant trehalase during clubroot disease was proposed to act as a defense mechanism in the susceptible accession Col-0, which could thereby cope with the accumulation of pathogen-synthesized trehalose. In the present study, we assessed trehalose activity and tolerance to trehalose in the clubroot partially resistant accession Bur-0.
  • We compared both accessions for several trehalose-related physiological traits during clubroot infection. A quantitative trait loci (QTLs) analysis of tolerance to exogenous trehalose was also conducted on a Bur-0xCol-0 RIL progeny.
  • Trehalase activity was not induced by clubroot in Bur-0 and the inhibition of trehalase by validamycin treatments resulted in the enhancement of clubroot symptoms only in Col-0. In pathogen-free cultures, Bur-0 showed less trehalose-induced toxicity symptoms than Col-0. A QTL analysis identified one locus involved in tolerance to trehalose overlapping the confidence interval of a QTL for resistance to Plasmodiophora brassicae. This colocalization was confirmed using heterogeneous inbred family (HIF) lines.
  • Although not based on trehalose catabolism capacity, partial resistance to clubroot is to some extent related to the tolerance to trehalose accumulation in Bur-0. These findings support an original model where contrasting primary metabolism-related regulations could contribute to the partial resistance to a plant pathogen.


Clubroot is a plant disease caused by the obligate biotroph protist Plasmodiophora brassicae, which affects all Brassicaceae including Arabidopsis thaliana. This disease is associated with worldwide agronomic losses with significant economic incidences (Dixon, 2009). The life-cycle of the soil-borne P. brassicae consists of two phases. In the primary phase events are confined to the root hairs whereas the secondary phase occurs in both the cortex and the stele of the hypocotyl and roots of infected plants. In the secondary phase, multinucleate plasmodia cause cell hypertrophy (abnormal cell enlargement) and cell hyperplasia (uncontrolled cell division) leading to the development of tumors (clubs) that obstruct nutrient and water transport (Ingram & Tommerup, 1972). Successful management of clubroot disease relies on an integrative combination of strategies, including adapted cropping practices as well as chemical, biological and cultivar control methods (Donald & Porter, 2009). Clubroot resistance under oligogenic control recently introduced into commercial cultivars of cabbage, cauliflower and oilseed rape crops led to complete resistance, but could be overcome by rapidly evolving pathogen populations, after large-scale use (Diederichsen et al., 2009). Polygenic resistance is regarded as potentially ‘harder to break through’ and should thus be considered for the long-term management of host plant resistance (Brun et al., 2010).

Consistent information is now available about the genetic architecture of quantitative partial resistance to clubroot in different Brassica species (Manzanares-Dauleux et al., 2000; Rocherieux et al., 2004; reviewed in Diederichsen et al., 2009). In Arabidopsis, a screen of 57 accessions identified three genotypes – Bur-0, Tsu-0 and Kn-0 – harboring partial resistance to clubroot (Alix et al., 2007). Four additive quantitative trait loci (QTLs) contributing to resistance to the P. brassicae eH isolate were identified in the progeny of the Bur-0 (resistant) × Col-0 (susceptible) cross (Jubault et al., 2008b). However, mechanisms underlying partial resistance remain poorly understood and there is no report on the nature and/or function of genes underlying partial resistance QTLs. This gap in knowledge contrasts with our increasing understanding of the physiopathological mechanisms involved in the infection of susceptible hosts by P. brassicae. The involvement of auxin and cytokinins in the development of symptoms has been widely investigated. ‘Omics’ and targeted metabolic profiling approaches associated with phenotyping of Arabidopsis mutants or overexpressor lines have recently contributed to refine a detailed model for hormonal control of clubroot development (reviewed by Ludwig-Müller & Schuller, 2008).

Several studies have also focused on the large and characteristic perturbations in carbohydrate metabolism which occur in Brassicaceae during clubroot infection. The development of galls is associated with a continuous accumulation of starch, glucose and fructose in roots, a process that is related to an enhanced carbon remobilization from leaves (Keen & Williams, 1969; Ludwig-Müller et al., 2009). Among the metabolic traits involved in the interaction, trehalose, a disaccharide generally found only in trace amounts in higher plants, reaches strikingly high levels in clubs during the secondary phase, when the secondary plasmodia are developing in cortical cells. The pathogen was proposed to synthesize this trehalose – a very common transitory storage molecule in microbes, fungi and insects – and indeed a P. brassicae transcript coding for trehalose-6-phosphate synthase (TPS) accumulates in developing galls during infection (Brodmann et al., 2002). This disaccharide, however, does not accumulate exclusively within the plasmodial compartment, as it was also measured at substantial levels in leaves of root-infected plants (Brodmann et al., 2002).

Endogenous trehalose is assumed to be involved in osmoprotection mechanisms in desiccation-tolerant organisms. Biotechnological approaches for enhancing tolerance to drought stress have attempted to engineer the artificial accumulation of trehalose in transgenic crops or associate symbionts (Romero et al., 1997; Garg et al., 2002; Karim et al., 2007; Suárez et al., 2008). However, and paradoxically, in plants that do not naturally accumulate endogenous trehalose (most higher plants), treatment with exogenous trehalose causes phytotoxic effects (Veluthambi et al., 1981). This trehalose toxicity is linked to accumulation of starch and anthocyanins and root growth inhibition (Wingler et al., 2000; Schluepmann et al., 2003), both of which also typically occur during clubroot disease. The initial discovery that exogenous trehalose treatment leads to growth impairment and starch accumulation is now interpreted as a retrocontrol exerted by artificially accumulated trehalose on the trehalose-6-phosphate (T6P) pathway, leading to T6P accumulation (Schluepmann et al., 2004). Indeed, T6P is a molecule maintained at very low levels in plants, but whose variation has a strong regulatory effect on primary metabolism (Kolbe et al., 2005; Lunn et al., 2006; Zhang et al., 2009; Smeekens et al., 2010).

In the Col-0 accession, trehalase activity and transcription of the unique trehalase encoding gene are both induced during the earliest steps of P. brassicae infection. This mechanism is assumed to reduce trehalose accumulation in clubroot-infected plant tissues and therefore would be involved in defense (Brodmann et al., 2002) even though this accession is highly susceptible to clubroot. However, it is unknown if this mechanism could play a more prominent role in naturally occurring resistant plant accessions. To elucidate this point, we investigated a possible relationship between trehalose metabolism and partial resistance to clubroot in Bur-0. We assessed trehalose and trehalase activity levels during clubroot infection and the effect of the trehalase inhibitor validamycin on clubroot symptoms in Col-0 and Bur-0. In vitro experiments were performed to quantitatively compare the respective sensitivity to treatments with exogenous trehalose in Bur-0 and Col-0. A QTL approach was then used to identify relationships between genetic factors involved in tolerance to trehalose and resistance to clubroot.

Materials and Methods

Plant material and genetic map construction

Col-0 and Bur-0 accessions of A. thaliana which are susceptible and partially resistant to clubroot respectively (Alix et al., 2007) were used in clubroot tests and for the evaluation of the tolerance to exogenous trehalose treatments. The Bur-0 × Col-0 recombinant inbred line (RIL) population was obtained from the Versailles Resource Centre for A. thaliana (VNAT, INRA, France,; the production of this progeny is described in detail in Simon et al. (2008). Of the 347 available RILs, 250 were used for QTL mapping: 164 RILs that constitute the core population (Simon et al., 2008) and 86 other RILs, which were randomly selected. The linkage map was created using 87 single nucleotide polymorphisms (SNP) genotyped with SNPlex technology (Applied Biosystems, Life Technology Corporation, Carlsbad, California, USA); details are given in Jubault et al. (2008a). The genetic map was established using mapmaker/exp 3.0 with the Kosambi mapping function.

Heterogeneous inbred family (HIF) lines 10351 and 13351 were also provided by the Versailles Resource Centre for A. thaliana (VNAT, INRA, France). These near-isogenic lines were derived from the RIL accession 351 of the Bur-0 × Col-0 segregating population, through the exploitation of residual heterozygosity. The HIF lines 10351 and 13351 exhibit Bur-0 and Col-0 alleles, respectively, in the region of the clubroot resistance QTL At-Pb5.1 reported in Jubault et al. (2008b) (between the markers c5_04011 and c5_10428, see the Supporting Information, Fig. S2). For clarity, 10351 and 13351 are thereafter called 351-Bur and 351-Col.

Clubroot tests

Susceptibility to clubroot was estimated by resistance tests at 21 d post inoculation (dpi) as described in Jubault et al. (2008a), using the isolate eH, the accessions Bur-0 and Col-0 and the HIF lines 351-Bur and 351-Col. For each replicate (three to eight biological replicates depending on experiments), data was expressed as the means of the evaluation of 12 plants. Susceptibility to clubroot was quantified by gall area evaluation (Ga in mm2) using image analysis. As Bur-0 is a fast-growing accession with large rosette and hypocotyls, we expressed the size of the gall relative to the surface of leaves, roughly evaluated by the square of the longest leaf length (maximal leaf length2 = leaf area index = La, in cm2). The ratio Ga : La was multiplied by 5000 to give the Ga : La index to obtain values in the range of those obtained with the classical disease index. For this purpose, every plant was photographed with a scale and image analyses were performed using imagej software (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA,, 1997–2011). Tissues from those plants was then collected and pooled for additional biochemical characterization. Roots were washed, and then root fragments of 3 cm from the root crown were sampled and frozen in liquid nitrogen before freeze drying and metabolite quantification. In some experiments, validamycin was applied by inoculation of 1 ml of 20 μM validamycin A (Duschefa, aqueous solution in the soil, near each plant hypocotyl. These applications started from 7 dpi, up to the estimation of symptoms at 21 dpi. Data for clubroot resistance in the Bur-0 × Col-0 RIL population – evaluated by a classical disease index method – were from Jubault et al. (2008b).

In vitro plant growth and trehalose treatments

Surface-sterilized seeds were germinated on agar (1%) solidified Hoagland’s medium. After 3 d of stratification at 4°C, plates were incubated in a culture chamber under 130 μE illumination, 80% hygrometry with a 16 h/8 h photoperiod at 22°C/20°C. In a first approach, seeds of parental lines, Col-0 and Bur-0, were germinated on trehalose-free medium for 4 d and then transferred aseptically to trehalose (or saccharose in the control plates) containing media at 40 mM or 80 mM. Plantlets were then photographed after an additional culture of 10 d. Next, trehalose toxicity symptoms were evaluated in parental lines germinated directly on trehalose-containing media, through the analysis of root growth inhibition and accumulation of anthocyanins in plantlets in response to trehalose, using various trehalose concentrations in the medium (20–80 mM). Root growth was followed by ink-dot labeling at 7 and 10 d post stratification (dps). Depending on experiments, root growth was evaluated directly on plates or using image analysis and ImageJ software. The proportion of root growth inhibition was calculated from the number of roots which grew < 1 mm in the interval between 7 and 10 dps. This variable was then referred to as TIRGI (trehalose induced root growth inhibition). Plant leaf rosettes were cut at 10 dps, weighed and frozen in liquid nitrogen, and then freeze-dried before metabolite extractions for the quantification of anthocyanins and trehalose. Anthocyanin accumulation (expressed as OD532 ml−1 mg DW−1) at 10 dps on 40 mM exogenous trehalose was referred to as the variable TIAA (trehalose induced anthocyanin accumulation). Evaluation of tolerance in RIL and HIF lines was performed using direct sowing on media containing 40 mM trehalose.

Metabolite extractions and quantifications

Freeze-dried samples were ground to a fine powder using a ball mill and then metabolites were extracted as follows: samples were incubated for 15 min with agitation in 400 μl methanol and 400 μM adonitol (internal standard); (200 μl of chloroform was added, followed by 5 min agitation; 400 μl of ultrapure water was added, tubes were vortexed vigorously for 1 min then centrifuged for 10 min (14 000 g), leading to separation of a lower apolar phase containing chlorophyll and an upper polar phase containing water/methanol and metabolites. A 400-μl sample from this upper phase was combined with 100 μl HCl 0.3 M to reveal purple anthocyanin derivatives. Anthocyanins were quantified by measuring OD532 in these acidified extracts and data were expressed as OD532 ml−1 mg DW−1. A 100 μl sample of the remaining nonacidified methanolic extracts was evaporated under vacuum for subsequent analysis of metabolites. Trehalose and glucose were quantified following a procedure adapted from the metabolic profiling procedure described in Lugan et al. (2009). The dry residue was dissolved in 50 μl of 20 mg ml−1 methoxyamine hydrochloride in pyridine at 30°C for 90 min. Afterwards, 50 μl of N,O-bis(trimethylsisyl)trifluoroacetamide (BSTFA) was added and samples were incubated at 37°C for 30 min. One microliter of the mixture was injected in a Trace 2000 GC-flame ionization detector (FID) (Thermo-Fisher Scientific, Waltham, CA, USA) fitted with an AS2000 Autosampler (Thermo-Fisher Scientific), a J&W DB5 30 m × 0.32 mm × 0.25 μm column and a FID. The gradient temperature was: 5 min at 70°C, 5°C min−1 until 220°C, 2°C min−1 until 260°C, 20°C min−1 until 300°C and finally 5 min at 300°C.

Real-time RT-PCR experiments

The RNA extractions and DNase treatments were performed as described in Jubault et al. (2008a). First-strand complementary DNA (cDNA) synthesis was performed in a 20 μL total reaction volume using 250 ng DNAse-digested total RNA, oligo(dT)12–18 primers (Invitrogen,, 1 mM dNTPs, 1× first strand buffer (Invitrogen), 20 mM dithiothreitol (DTT; Invitrogen), 40 U RNaseOUT recombinant ribonuclease inhibitor (Invitrogen) and 200 U SUPERSCRIPT II reverse transcriptase (Invitrogen) by incubating for 2 h at 42°C. The reaction was terminated by incubation for 15 min at 70°C. The relative transcript abundance was quantified using Light Cycler 480 Cyber GREEN I Master (Roche Applied Science) and two technical replicates for each of two biological replicates were performed using the Light Cycler 480 (Roche Applied Science, A PCR-amplification of TRE1 and PP2AA3 was performed on 1/10 diluted cDNA with the primers described in Table 1 at a final concentration of 0.5 μM, using the following cycle parameters: 95°C 5 min, 95°C 15 s/60°C 30 s. Negative controls were carried out using water instead of cDNA template. Relative quantification of gene expression was calculated following Pfaffl (2001) using protein phosphatase 2A subunit PDF2 encoding gene (At1g13320) as the reference (Czechowski et al., 2005). We checked that this reference gene exhibited low variation (delta CT < 2) among the samples analysed.

Table 1.   List of primers used for Real-time PCR experiments
Locus (gene name)PrimersAmplicon length (bp)

Trehalase activity

Trehalase activity was assayed based on the method described in Kendall et al. (1990). Enzymes were extracted from 100 mg of liquid nitrogen-ground biological material with 400 μl of 0.2 M potassium phosphate buffer, pH 5.8. After centrifugation, the protein content was quantified in the supernatant using the Bradford procedure. Trehalase assays were started with the addition of trehalose (or water for background) to a final concentration of 10 mM. After incubation for 1 h at 30°C, 100 μl were sampled then boiled for 10 min and centrifuged. Glucose concentrations were measured in the supernatants using GC-FID as described earlier. The amount of glucose released from trehalase activity was inferred taking into account the background glucose content in control reactions.

Starch quantification

Starch was quantified in c. 5–6 mg samples of ground dry plant matter. Samples were first incubated in 80% ethanol for 1 h at room temperature (RT), centrifuged and the supernatant containing soluble sugars was discarded. This washing step was repeated once. Pellets were air-dried, then loosened from the bottom of microtubes using a spatula and resuspended in 150 μl MOPS (3-(N-morpholino)propanesulfonic acid)) 50 mM pH 7.5 with 5 μl of thermostable amylase (Sigma) at 4 mg ml−1. Samples were incubated at 100°C for 6 min then cooled on ice. Then 200 μl of sodium acetate buffer 0.2 M pH 4.8 and 35 U of amyloglucosidase (Sigma) were added and the samples were incubated for 3 h at 50°C to quantitatively digest starch. Finally, samples were centrifuged and glucose was quantified in supernatants using the Glucose (HK) Assay Kit (Sigma).

Statistical methods and QTL analysis

The data obtained from each test were statistically analysed using a generalized linear model (PROC GLM of Statistical Analysis System (SAS) software, SAS Institute Inc., 2000). The number of independent biological replicates, the number of plants used for each replicate and statistical tests are specified in figure legends. The QTL detection was performed by composite interval mapping (CIM) using qtl cartographer version 2.5 (Wang et al., 2010) following the procedure described in Jubault et al. (2008b).


Comparison of trehalose contents and trehalase activities in infected Col-0 and Bur-0 roots

Inoculation of 7-d-old plantlets with eH spore suspensions resulted, 21 d later, in typical pronounced clubroot symptoms in the Col-0 accession, contrasting (as expected) with the low-level symptoms observed in Bur-0 (Fig. 1a,b). Small amounts of trehalose were detected in roots of uninoculated plants of both accessions at 21 dpi. At this stage, trehalose levels were strongly enhanced in both infected roots of Col-0 and Bur-0, although to a lesser extent for the latter (Fig. 1c). Expression of the unique trehalase encoding gene TRE1 was monitored using real-time PCR. TRE1 expression was induced at similar levels in both genotypes at both 14 dpi and 21 dpi (Fig. 1d). Enzymatic trehalase activity was assayed in root sample extracts, through the evaluation of resulting glucose following an in vitro incubation with a large amount of trehalose substrate. Inoculation-dependent enhancement of trehalase activity was only observed in crude extracts of Col-0 infected roots (Fig. 1e), and they remained at basal levels in inoculated Bur-0 roots. Trehalase activity in infected Col-0 roots was found to be within the range of that described by Brodmann et al. (2002).

Figure 1.

Clubroot symptoms and trehalose metabolism in Bur-0 and Col-0. (a) and (b) Illustration of clubroot symptoms at 21 d post inoculation (dpi) in Bur-0 and Col-0 ecotypes after inoculation by the eH isolate of Plasmodiophora brassicae. (c) Trehalose content, (d) Real-time PCR quantification of TRE1 gene expression levels, using PP2AA3 as reference gene (open bars, 7 dpi; light tinted bars, 14 dpi; dark tinted bars, 21 dpi). (e) Trehalase activity in roots of Bur-0 and Col-0 following inoculation (I) or in uninoculated controls (NI). Data are means of six (c) or two (d,e) independent biological replicates, each one consisting of a pool of root samples from 12 plants. Asterisks indicate a significant difference between two treatments according to Student tests (< 0.05). Bars, 1 cm.

Effect of validamycin on clubroot symptoms in Col-0 and Bur-0

Validamycin is a trehalase inhibitor with a soil half-life of only few days, as reported by Xu et al. (2009). A solution of validamycin (20 μM in water) was applied at 7, 10, 14, 17 and 20 dpi to the soil near the hypocotyls of inoculated (eH isolate) and uninoculated plants. This treatment did not have an obvious visible effect on plant growth and development of uninoculated controls. For inoculated plants of the susceptible accession Col-0, application of validamycin resulted in enhanced clubroot symptoms evaluated at 21 dpi by the Ga : La index (Table 2). By contrast, in the partially resistant Bur-0 accession, validamycin treatment did not induce significant changes in clubroot symptoms.

Table 2.   Effect of validamycin treatment during infection on clubroot symptoms
 Ga : La pathological index
  1. Means are estimated from three independent experiments, each one consisting of the evaluation of 12 infected plants (eH isolate) at 21 dpi with or without validamycin (see the Materials and Methods section). The Ga : La pathological index reflects the ratio between gall extent and rosette leaf area and was evaluated using image analysis as described in the materials and methods section. Results are reported ± SE.

  2. aSignificant effect of the validamycin treatment according to Student’s pair t-test (< 0.05).

Col eH44.9 ± 7.3
Col eH + VAL68.7 ± 12.2a
Bur eH31.3 ± 4.2
Bur eH + VAL30.3 ± 7.0

Evaluation of tolerance to exogenous trehalose in Bur-0

Experiments were performed to evaluate trehalose toxicity symptoms in Bur-0 and Col-0. In a first approach we germinated seedlings on trehalose-free media and then transferred plantlets to media containing 40 mM or 80 mM trehalose. A treatment with 80 mM sucrose was used as control. At 40 mM trehalose, after 10 d, first symptoms of toxicity were apparent only for Col-0 plantlets (Fig. 2). At 80 mM trehalose, plantlets of both accessions were stunted. Col-0 plantlets, however, exhibited symptoms of higher toxicity (curled leaves with high anthocyanin content and very short root systems), while for Bur-0 plants, leaves remained green and their root system was more developed (Fig. 2). Sucrose treatment did not lead to growth inhibition, giving an indication that trehalose effect was not related to any osmotic shock. A second experiment was then carried out, with seedlings germinated in the presence of 20–60 trehalose concentrations. A GC-FID analysis in samples from plantlets at 10 d post stratification (dps) revealed an accumulation of trehalose. This accumulation increased with the concentration of trehalose in the medium (Fig. 3a). The mean trehalose content was slightly higher in Col-0 than in Bur-0 after growth on 40 mM. A more pronounced contrast was observed for the 60 mM treatments, as the Col-0 plantlets were strongly stunted (data not shown) with high content of anthocyanins, while Bur-0 seedlings were still green and apparently alive. We also determined root-associated and leaf-associated criteria to quantify trehalose toxicity symptoms. The most relevant criteria were the percentage of root growth inhibition between 7 dps and 10 dps (Fig. 3b), and anthocyanin accumulation in plantlets (Fig. 3c). Application of these criteria suggested that Bur-0 exhibited increased tolerance to exogenously supplied trehalose for 60 mM and 40 mM trehalose treatments, and even for 20 mM when considering only anthocyanin accumulation. Plotting trehalose toxicity symptom indicators against plant trehalose contents showed that, for a plant trehalose content higher than 10 μmol g DW−1, Bur-0 exhibited less symptoms than Col-0 plantlets (Fig. 4).

Figure 2.

Trehalose toxicity symptoms in Bur-0 and Col-0 in the transfer technique. Seeds of accessions Bur-0 (a,c) and Col-0 (b,d) were germinated on trehalose-free Hoagland’s media for 4 d then transferred to media supplemented with trehalose 40 mM, 80 mM or sucrose 80 mM (control) for an additional 10-d period. Scale bars represent 1 cm.

Figure 3.

Trehalose toxicity symptoms in Bur-0 (open bars) and Col-0 (closed bars) in the direct germination technique. Seeds were germinated directly on trehalose supplemented media. (a) Trehalose content in plantlets at 10 dps. (b) Root growth inhibition as a percentage of roots exhibiting growth < 1 mm between 7 d post stratification (dps) and 10 dps. (c) Anthocyanin content in plantlets at 10 dps. Data are means of two (a,c) or four (b) biological replicates, each one consisting of eight plantlets. Error bars represent standard errors. Asterisks indicate a significant difference between two genotypes according to Student’s tests (< 0.05).

Figure 4.

Relationships between trehalose content in plantlets and toxicity symptoms in Bur-0 (squares) and Col-0 (circles) accessions. Toxicity was estimated from root growth inhibition (a) and anthocyanin content (b). Both figures were constructed using data presented in Fig. 3. Error bars represent standard errors.

QTL analysis for tolerance to exogenous trehalose

Trehalose sensitivity was evaluated in 230 Recombinant Inbred Lines (RILs) issued from the cross Bur-0 × Col-0 (Simon et al., 2008) using both root (TIRGI) and leaf-associated (TIAA) criteria. Trehalose toxicity was evaluated through direct germination tests at 40 mM trehalose, because it was assumed that at this concentration RILs with higher tolerance than Bur-0 or greater sensitivity than Col-0 could be potentially identified. This analysis resulted in valuable data for 182 RILs for TIRGI and 178 RILs for TIAA. Both traits showed in the RIL progeny a continuous distribution pattern, suggesting a quantitative and polygenic control of the trehalose tolerance (Fig. S1).

Two QTLs on chromosome 5 were detected using the root-associated criterion for the estimation of tolerance to trehalose. The QTL TIRGI-At5.1 (LOD = 3.4, the allele for higher tolerance to trehalose was derived from Bur-0) was associated with the marker c5_05319 and a confidence interval between 10.2 cM and 21.6 cM (Fig. 5 and Table S1). This marker was previously identified as associated with the clubroot partial resistance QTL Pb-At5.1 (Jubault et al., 2008b). According to the FLAGdb++ v4.5 database (Samson et al., 2004), we found that the confidence interval for TIRGI-At5.1 includes 771 gene identities from At5g13130 to At5g20320. Among them, we did not identify any gene that would have an obvious functional relationship with trehalose metabolism. The second QTL, TIRGI-At5.2 (LOD = 2.8), was mapped to the bottom of chromosome 5, but it did not colocalize with At-PbAt5.2. TIRGI-At5.1 and TIRGI-At5.2 accounted for 8.8% and 5.6%, respectively, of the total variation. Three QTLs were detected using the leaf-associated criterion to evaluate trehalose-induced stress: the major effect QTL TIAA-At1 (chromosome 1, LOD = 4.98, explaining 11.7% of the total phenotypic variation) colocalized with the locus PAP1/MYB75 (At5g56650) previously reported to be involved in the control of anthocyanin biosynthesis in response to sugar status (see the Discussion section). In addition, two low effect-QTLs TIAA-At5.1 (LOD = 2.7) and TIAA-At5.2 (LOD = 2.9) were found on chromosome 5, with confidence intervals recovering those found with the TIRGI method. The RIL plantlets were also analysed for trehalose content after treatment with 40 mM trehalose. No QTLs for trehalose content could be detected using this approach (data not shown).

Figure 5.

The quantitative trait locus (QTL) map for clubroot partial resistance and tolerance to trehalose. Arabidopsis additive QTLs controlling partial resistance to P. brassicae (Pb-At, white bars), trehalose-induced root growth inhibition (TIRGI-At, closed bars) and trehalose-induced anthocyanin accumulation (TIAA-At, hatched bars) in the Bur-0 × Col-0 RIL population. Each marker name consists of its chromosome number and physical position in kb. The bar length is in proportion to the one-LOD likelihood confidence interval.

In parallel, QTLs controlling clubroot partial resistance were recalculated in the present study using phenotypic data from Jubault et al., 2008b, but using exactly same RIL subset as those used for TIRGI and TIAA QTL identification. The resulting confidence interval for Pb-At5.1 between 8.6 cM and 15.6 cM (Table S1) was found to include genes from At5g10580 to At5g16520. The cross-interval defined by At-TIRGI5.1 and At-Pb5.1 includes a subset of 358 genes between At5g13130 and At5g16520.

Evaluation of clubroot symptoms and tolerance to trehalose in HIF lines

The HIF couple 351-Col/351-Bur was obtained from the Versailles Resource Center for A. thaliana. These lines were derived from the RIL 351, which showed residual heterozygosity between the markers c5-04011 and c5-10428, a region that includes the genetic markers within the Pb-At5.1 and TIRGI-At5.1 confidence intervals (Fig. S2). This genetic material is well suited to get a validation of the localization and effects of the two QTLs that had been statistically inferred from the QTL analysis on the RIL segregating population. Clubroot symptoms were quantified at 21 dpi for 351-Col and 351-Bur lines. The 351-Col HIF line displayed stronger clubroot symptoms – more galls and higher levels of starch accumulation in infected roots – when compared with the 351-Bur at 21 dpi (Fig. 6a and Table 3). In vitro cultures also revealed that trehalose-induced root growth inhibition was lower for the 351-Bur line than for 351-Col (Fig. 6b). Experiments with HIF lines confirmed that an allelic variation in the genomic region between c5-04011 and c5-10428 markers, contributes to the control of both tolerance to trehalose and partial resistance to clubroot.

Figure 6.

Heterogeneous inbred family (HIF) validation of quantitative trait loci (QTLs) Pb-At5.1 and TIRGI-At5.1. Parental lines and 351-Bur and 351-Col HIF lines, displaying allelic variation on a region of chromosome 5 that overlaps Pb-At5.1 and TIRGI-At5.1. The QTL confidence intervals were evaluated for susceptibility to clubroot using the Ga : La pathological index (ratio between gall extent and rosette leaf area; a) and for tolerance to trehalose through the evaluation of trehalose-induced root growth inhibition (b). Means were estimated from eight (a) or three (b) independent replicates, each one resulting from the analysis of 12 individual plants. Error bars represent standard errors. Asterisks indicate a significant difference between two treatments according to Student tests (P < 0.05).

Table 3.   Effect of the Bur/Col allelic variation between 351-derived heterogeneous inbred families (HIFs) onto root starch accumulation at 21 dpi for inoculated (I) or uninoculated (NI) plants
GenotypeInoculationRoot starch at the end of light period (μg mg DW−1)Root starch at the end of dark period (μg mg DW−1)
  1. Data are from three biological replicates, each one consisting in 12 plants. Results are reported ± SE.

351-BurI108 ± 15109 ± 14
351-BurNI61 ± 476 ± 11
351-ColI144 ± 21146 ± 15
351-ColNI69 ± 1155 ± 12


Trehalose accumulates in infected roots at levels that can affect plant physiology

In this work, we have shown that differences in tolerance to trehalose between two genotypes, are linked to differences in susceptibility to clubroot. In clubroot-infected tissues, trehalose is accumulated at levels that could modify the plant metabolism. Indeed, in the susceptible Col-0 accession, pathogen infection resulted in the root accumulation of trehalose at concentrations (> 20 μmol g DW−1) that, when resulting from exogenously supplied trehalose treatments, are similar those above which first toxicity symptoms can be detected, especially at the amount of anthocyanin accumulation (Figs 1c, 4). Trehalose was also accumulated in substantial amounts in infected Bur-0 roots, although to a lower extent. This contrast between Col-0 and Bur-0 is not the consequence of higher trehalase activity in Bur-0, because trehalase activity is not induced during the infection in this accession. Instead, lower amounts of trehalose in infected roots are more likely the consequence of a slower development of the pathogen in Bur-0. Plasmodiophora brassicae was indeed proposed to essentially synthesize the trehalose which accumulates in clubroot-infected plants and potentially affect plant metabolism (Brodmann et al., 2002). As root samples from infected Bur-0 plants contained a substantial proportion of gall-free roots, absolute quantification of trehalose relative to whole root sample DW probably underestimated the trehalose content in authentically infected areas.

Induction of trehalase is a basal defense mechanism in the susceptible accession Col-0

When considering Col-0, our results corroborate the previous work of Brodmann et al., 2002: in parallel with trehalose accumulation, trehalase activity was highly induced in infected roots both at the transcriptional (AtTRE1) and enzymatic level. In addition, our study revealed that the inhibition of trehalase activity during clubroot infection (by treatment with validamycin) resulted in enhanced symptoms in Col-0. This suggests that trehalase induction is a basal defense mechanism, that is, a disease resistance activated by a virulent pathogen on a susceptible host, in agreement with the definition of Jones & Dangl (2006). In clubroot disease, trehalase induction would thus restrict the deleterious effect of trehalose in plant, as proposed by Brodmann et al. (2002), and trehalose can be seen as a pathogen-effector, using the broad definition of Hogenhout et al. (2009).

There are two indications that the observed effect of validamycin treatment is directly related to the inhibition of plant trehalase AtTRE1, and not an artifact caused by the inhibition of Plasmodiophora trehalase enzymatic activity. First, during the cortical infection steps, P. brassicae is found as an intracellular inclusion, whereas plant trehalase TRE1 is targeted to plant plasma membranes with an apoplastic active site (Frison et al., 2007). We can thus expect that validamycin would first come into contact with plant extracellular enzymes before perturbing P. brassicae trehalose-related enzymes. Second, contrary to what we observed, the inhibition of pathogen trehalase activity is expected to reduce its fitness. For example, validamycin is used in agriculture as a fungicide for the control of Rhizoctonia associated plant diseases (Bai et al., 2006). In that case, however, the fungi do not form cellular inclusions inside plant cells. Nevertheless, additional work is now required to confirm our view through complementary approaches, including, for example, the downregulation of TRE1 expression in Col-0 and HIF lines.

Partial resistance to clubroot in Bur-0 is not associated with enhanced trehalase activity but could be linked to tolerance to trehalose

In contrast to the situation observed for Col-0, the accumulation of trehalose in infected Bur-0 roots was not accompanied by the induction of trehalase activity, suggesting that in this partially resistant accession, apoplastic catabolism of trehalose by the plant enzyme TRE1 is not a major component of plant resistance. Corroborating this idea, clubroot symptoms were not enhanced in Bur-0 when trehalase activity was inhibited by validamycin. The discrepancy between the induction of TRE1 RNA and enzyme levels in Bur-0 suggests that, at least at the transcriptional level, early plant responses to the infection partly overlap in the two accessions.

Bur-0 exhibits a significant level of tolerance to exogenous trehalose, using both leaf and root-associated stress criteria (TIAA and TIRGI respectively) and various phenotyping conditions (direct germination or transfer experiments), suggesting a whole-plant level mechanism. The ability of Bur-0 to tolerate exogenous applications of trehalose could be the result of reduced trehalose absorption/translocation kinetics or higher trehalose catabolism in this accession. This hypothesis would fit with the contrasted accumulations of trehalose when comparing both accessions after trehalose treatments, although mostly for the highest dose – 60 mM (Fig. 3). Indeed, at 20 mM and 40 mM trehalose in the medium, accumulations of trehalose in plants are only slightly different (although statistically significant) and cannot account fully for the contrast between stress symptoms in Bur-0 and Col-0. Fig. 4 further illustrates the idea that for a given trehalose accumulation level in plant, Bur-0 exhibited less toxicity symptoms than Col-0. A constitutive mechanism of tolerance to the accumulation of trehalose would therefore contribute to the tolerance to trehalose in Bur-0. Accordingly, we did not detect any QTL for trehalose accumulation in tissue using the same experimental samples that allowed the detection of QTL controlling TIAA and TIRGI. We can then reasonably assume that TIRGI-At5.1 (the detected QTL that exhibits the highest additive effect, Table S1) does not depend on differential trehalose accumulation. Variation among Arabidopsis ecotypes for trehalose sensitivity has not been previously described, but the abi4 (aba insensitive 4, At2g40220) mutant (Col-0 genetic background) is tolerant to trehalose (Ramon et al., 2007). This phenotype was attributed to a defect in ABI4-directed starch synthesis, a cellular signaling mechanism that would be specifically involved in trehalose toxicity. One hypothesis could have been that Bur-0 displays an abi4-like metabolic phenotype. However, the abi4 mutation leads to glucose tolerance whereas Bur-0 is highly sensitive to glucose in germination tests (Fig. S3), suggesting that different physiological processes underlie tolerance in those two genotypes.

A plausible model is therefore emerging from our results, where trehalase induction does not act as a defense mechanism to delay clubroot symptoms in Bur-0, despite substantial trehalose accumulation in infected roots, and presumably in line with the fact that this accession exhibits a higher tolerance to trehalose. If correct, this hypothesis necessarily involves at least some overlaps between the genetic architectures of tolerance to trehalose and clubroot resistance. In the idea of testing this hypothesis, we took advantage of the availability of a Bur-0 × Col-0 segregating population to perform a QTL analysis of the tolerance to trehalose.

TIAA-At1 co-localizes with the anthocyanin-controlling locus PAP1

The major QTL controlling anthocyanin accumulation in response to trehalose (TIAA-At1) is located on chromosome 1. This QTL did not colocalize with any QTL controlling the extent of root growth under trehalose treatment, or with previously detected QTL for clubroot resistance. The confidence interval for this QTL contains the gene encoding PAP1 (PRODUCTION OF ANTHOCYANIN PIGMENT1), the MYB75 transcription factor previously identified as a positive regulator of anthocyanin accumulation in response to sucrose and originally identified through a QTL approach (Teng et al., 2005). Although it is plausible that allelic variation on MYB75 is underlying TIAA-At1, this would also suggest that Bur-0 and Col-0 display contrasted anthocyanin accumulation in response to sugar status, whatever the stress. This would mean that there is a significant risk that this variable offers a biased estimation of stress intensity and is not as relevant as expected for the evaluation of trehalose tolerance.

Tolerance to trehalose is genetically linked to a QTL for clubroot resistance

In addition to the specific case of TIAA-At1, the two other QTLs detected, TIAA-At5.1 and TIAA-At5.2, colocalize with the two QTLs TIRGI-At5.1 and TIRGI-At5.2, which control trehalose-induced root growth inhibition. Our genetic analysis therefore infers that these two genomic regions exert a control on tolerance to trehalose, whether we use a root or a leaf-associated phenotyping method. This suggests that the results are somewhat robust, even if anthocyanin accumulation was – a posteriori – not the best toxicity marker to study trehalose response (see the previous paragraph).

TIRGI-At5.1 and TIAA-At5.1 are located near the marker c5-05319, and the confidence intervals for these two QTLs overlap with QTL Pb-At5.1, which controls clubroot resistance, as described by Jubault et al. (2008b). These statistically significant results inferred from the phenotypes observed in a RIL population were confirmed by the phenotypes of HIF lines. This colocalization brings a substantial support to our initial hypothesis about a possible link between tolerance to trehalose and resistance to clubroot.

One explanation of the toxic effects of exogenous trehalose in plants is the induction of a massive allocation of carbon into starch synthesis, in leaves (Wingler et al., 2000). We have shown that starch is accumulated more in clubroot-infected 351-Col roots than in 351-Bur. It would therefore be interesting now to utilize those HIF lines to investigate if higher starch accumulation in infected roots of the 351-Col line is the consequence or the cause of the higher susceptibility to clubroot of this line.

The other QTLs controlling root growth inhibition in response to trehalose (TIRGI-At5.2) and anthocyanin accumulation (TIAA-At5.2) colocalize in a region devoid of QTL for clubroot partial resistance. Because of their small effect on tolerance to trehalose, this lack of colocalization does not dismiss our previous conclusion. Small-effect QTLs for clubroot resistance in this region could have been missed in Jubault et al. (2008b) study.

Tolerance to trehalose shows parallels with previously described atypical carbohydrate metabolism in Bur-0

Bur-0 is an atypical accession when considering several aspects, including for example flowering time, growth rate and carbohydrate metabolism (Werner et al., 2005; Cross et al., 2006). An attractive hypothesis would be that specific features of carbohydrate metabolism in Bur-0 are partly related to an idiosyncratic fine-tuning of primary metabolism control by the T6P/trehalose pathway. A comprehensive analysis (in appropriate HIF pairs) of carbohydrate metabolism and trehalose or T6P-controlled transcriptional networks will be of great interest to gain a better understanding of cellular mechanisms underlying tolerance to trehalose and clubroot resistance. Further positional cloning is also required to clearly assess if those phenotypes are under the control of the same genetic polymorphism and to obtain insight into the putative underlying coding sequence.


Our initial goal was to test if high trehalase activity could contribute to clubroot partial resistance in Bur-0. Our experimental data infirmed this hypothesis, but revealed a surprising level of tolerance to exogenous trehalose in this accession, that appears to be at least in part associated to a better ability to tolerate trehalose accumulation. A link between this metabolic feature and clubroot resistance is supported by a colocalization between genetic factors controlling clubroot resistance and tolerance to trehalose. These results open new perspectives on the long-known outstanding trehalose accumulation in clubroot infected plants, and calls for taking into account the genetic determinants of primary metabolism features in the study of partial quantitative resistances to plant pathogens.


The authors thank Christine Camilleri (Versailles Genetics and Plant Breeding Laboratory, Arabidopsis thaliana Resource Centre, INRA, France) for providing seeds of RIL and HIF lines, Vincent Bouguennec, Morgane Havard, Laurent Charlon and Pascal Glory for their technical assistance, and François Lahrer for the critical reading of the manuscript.