•Cysteine is the metabolic precursor of essential biomolecules such as vitamins, cofactors, antioxidants and many defense compounds. The last step of cysteine metabolism is catalysed by O-acetylserine(thiol)lyase (OASTL), which incorporates reduced sulfur into O-acetylserine to produce cysteine. In Arabidopsis thaliana, the main OASTL isoform OAS-A1 and the cytosolic desulfhydrase DES1, which degrades cysteine, contribute to the cytosolic cysteine homeostasis.
•Meta-analysis of the transcriptomes of knockout plants for OAS-A1 and for DES1 show a high correlation with the biotic stress series in both cases.
•The study of the response of knockout mutants to plant pathogens shows that des1 mutants behave as constitutive systemic acquired resistance mutants, with high resistance to biotrophic and necrotrophic pathogens, salicylic acid accumulation and WRKY54 and PR1 induction, while oas-a1 knockout mutants are more sensitive to biotrophic and necrotrophic pathogens. However, oas-a1 knockout mutants lack the hypersensitive response associated with the effector-triggered immunity elicited by Pseudomonas syringae pv. tomato DC3000 avrRpm1.
•Our results highlight the role of cysteine as a crucial metabolite in the plant immune response.
Plants are continuously subjected to attack by a plethora of pathogens and herbivores. Only a small proportion of them are able to invade the plant and exploit it as a source of energy because the plants are resistant to most of them. In addition to nonhost resistance, plants are able to recognize pathogen-associated molecular patterns (PAMPs) and induce a response leading to a basal or PAMP-triggered immunity (PTI). However, pathogens have evolved to avoid recognition by the host and delay the defense responses until they are no longer effective, resulting in a compatible plant–pathogen interaction. For their defense, plants have developed specific resistance mechanisms based on gene-for-gene interaction (Staskawicz et al., 1995; Jones & Dangl, 2006). During the so-called effector-triggered immunity (ETI), an avirulence protein from the pathogen is recognized by a resistance protein from the plant, and this interaction rapidly starts the plant’s defense program strongly enough to halt pathogen spreading (Whalen et al., 1991; Hammond-Kosack & Jones, 1997; Dangl & Jones, 2001). Interestingly, most of the defense response is activated in both PTI and ETI, but the timing makes the difference (Tao et al., 2003).
One of the earliest responses to pathogen infection is the production of reactive oxygen species (ROS, reviewed in Lamb & Dixon, 1997). The production of ROS is weakly induced in both PTI and ETI at the very early stages of the infection, but only in the ETI is there a second more prolonged and extensive induction of the oxidative burst. This second phase of ROS induction is therefore characteristic of the ETI response and is accompanied by an accumulation of nitric oxide (NO) and salicylic acid (SA), which, together with ROS, promote the hypersensitive response (HR) (Delledonne et al., 1998; Alvarez, 2000). The HR is therefore typical for the ETI, and it causes a rapid death of the infected cells (Heath, 2000). In addition to limiting the spread of biotrophic pathogens, HR contributes to the activation of defense in adjacent cells and to the activation of systemic acquired resistance (SAR), a broad-spectrum form of disease resistance mediated by the action of SA, which is accompanied by the systemic activation of some defense responses (Vlot et al., 2008).
In addition to its structural role in proteins, cysteine functions as a precursor for essential biomolecules, such as vitamins and cofactors (Droux, 2004; Wirtz & Droux, 2005), antioxidants, such as glutathione (Meyer & Hell, 2005; Mullineaux & Rausch, 2005), and many defense compounds, such as glucosinolates, thionins or phytoalexins (Rausch & Wachter, 2005). Cysteine is synthesized in plants in the cytosol, plastids and mitochondria by the sequential action of the enzymes serine acetyltransferase (SAT, EC 184.108.40.206), which synthesizes the intermediary product O-acetylserine (OAS), and O-acetylserine(thiol)lyase (OASTL, EC 220.127.116.11), which combines a sulfide molecule with an OAS molecule to produce cysteine. In A. thaliana plants, several OASTLs and SAT encoding genes have been identified by sequence homology, and the functionality of some of them has been demonstrated by several studies (Haas et al., 2008; Heeg et al., 2008; Lopez-Martin et al., 2008a; Watanabe et al., 2008a,b; Alvarez et al., 2010; Bermudez et al., 2010). Most of the cysteine is formed and accumulated in the cytosol (Krueger et al., 2009, 2010) by the action of the major cytosolic OASTL, encoded by OAS-A1 (Barroso et al., 1995; Lopez-Martin et al., 2008a). It has been demonstrated that OAS-A1 is involved in the defense responses against abiotic stresses (Barroso et al., 1999; Dominguez-Solis et al., 2001, 2004) and is essential for maintaining the antioxidant capacity of the cytosol (Lopez-Martin et al., 2008a,b). In knockout oas-a1 plants, intracellular cysteine and glutathione levels are significantly reduced, and the glutathione redox state is shifted toward its oxidized form. Moreover, oas-a1 knockout mutants accumulate ROS in the absence of external stress, and show spontaneous cell death lesions in the leaves. In addition, cysteine can be degraded in the cytosol by the action of l-cysteine desulfhydrase (DES, EC 18.104.22.168), which catalyses the formation of sulfide, ammonia and pyruvate from cysteine in a stoichiometric ratio of 1 :1 :1. In A. thaliana, the only cytosolic desulfhydrase described to date is encoded by DES1 (previously known as CS-LIKE, At5g28030). Knockout des1 plants show enhanced antioxidant defenses and tolerance to conditions promoting oxidative stress (Alvarez et al., 2010). Thus, des1 and oas-a1 knockout mutants show opposite phenotypes, suggesting that they exert opposing functions by regulating cytosolic cysteine homeostasis.
To investigate the role of cysteine in response to stress further, we have explored its role in the plant–pathogen response by analysing the responses of the different oas-a1 and des1 alleles under pathogen attack.
Materials and Methods
Plant material and growth conditions
Arabidopsis (A. thaliana) wild-type ecotypes Col-0 and No-0, the SALK_103855 and RATM13-2715-1_G (des1-1 and des1-2, Alvarez et al., 2010), the SALK_072213 and SAIL_94_E12 (oas-a1.1 and oas-a1.2, Lopez-Martin et al., 2008a) and cad2-1 (Cobbett et al., 1998) mutants were used in this work. The plants were grown in soil under a photoperiod of 8 h of white light (120 μE m−2 s−1) at 20°C/16 h dark at 18°C. Plants were cultivated for 6- to 7-wk. Under these conditions, size, developmental stage, leaf size and thickness of mutant plants were undistinguishable of their respective wild type.
RNA extraction and microarray hybridization
For microarray of the des1-1 mutant, plants were grown in soil under a photoperiod of 16 h of white light (120 μE m−2 s−1) at 20°C/8 h dark at 18°C. Rosette leaves of 20-d-old plants were used for total RNA isolation with Trizol reagent (Invitrogen) and cleaning with the RNeasy Plant Mini Kit (Qiagen). This was used to synthesize biotinylated cRNA for hybridization to Arabidopsis ATH1 arrays (Affymetrix, Santa Clara, CA, USA), using the 3′ Amplification One-Cycle Target Labeling Kit. Briefly, 4 mg of RNA was reverse transcribed to produce first-strand cDNA using a (dT)24 primer with a T7 RNA polymerase promoter site added to the 3′ end. After second-strand synthesis, in vitro transcription was performed using T7 RNA polymerase and biotinylated nucleotides to produce biotin-labeled cRNA. The cRNA preparations (15 μg) were fragmented into fragments of 35–200 bp at 95°C for 35 min. The fragmented cRNAs were hybridized to the Arabidopsis ATH1 microarrays at 45°C for 16 h. Each microarray was washed and stained in the Affymetrix Fluidics Station 400 following standard protocols. Microarrays were scanned using an Affymetrix GeneChip Scanner 3000.
Microarray data analysis
Microarray analysis was performed using the affylmGUI R package (Wettenhall et al., 2006). The Robust Multi-array Analysis (RMA) algorithm was used for background correction, normalization and for summarizing expression levels (Irizarry et al., 2003). Differential expression analysis was performed using Bayes t-statistics using the linear models for microarray data (Limma), which is included in the affylmGUI package. The P-values were corrected for multiple testing using Benjamini–Hochberg’s method (False Discovery Rate) (Benjamini & Hochberg, 1995; Reiner et al., 2003). A cutoff value of a 1.5-fold change and P value of < 0.05 were adopted to discriminate expression of genes that were differentially expressed in the mutant plant with respect to the wild type. Gene classification into functional groups was obtained from the Bio-Array Resource for Arabidopsis Functional Genomics (http://www.bar.utoronto.ca) and the mapman software (http://gabi.rzpd.de/projects/MapMan/). The microarray data for the oas-a1.1 previously published (Lopez-Martin et al., 2008a) and des1-1 mutants were meta-analysed using the Bio-Array Resource for Arabidopsis Functional Genomics (http://www.bar.utoronto.ca, Toufighi et al., 2005).
Bacterial pathogen infections
The bacterial strains used in this study were Pseudomonas syringae pv. tomato (Pst) DC3000 and Pst DC3000 bearing a plasmid containing the avrRpm1 avirulence gene (Grant et al., 1995). For treatment of the plants, bacterial cultures were collected from plates in 10 mM MgCl2 and their concentrations were adjusted to 5 × 106 bacteria ml−1 (DO600 = 0.01, Pst DC3000 avrRpm1) or 2.5 × 106 bacteria ml−1 (DO600 = 0.005, Pst DC3000). Sterile 10 mM MgCl2 was used as a mock solution. The bacterial suspension or the mock solution was then pressure infiltrated into the abaxial side of the leaves of 6- to 7-wk-old plants, both wild types and mutants grown at the same time and conditions, using a syringe without a needle (Swanson et al., 1988).
Bacteria growth tests
Pst DC3000 avrRpm1 bacteria were collected from Luria–Bertani (LB) plates supplemented with rifampicin (50 μg ml−1) in 10 mM MgCl2, and their concentration was adjusted to 5 × 106 bacteria ml−1 (DO600 = 0.01). Six series of 1 : 10 dilutions were done and 10 μl of the resulting suspensions were plated in parallel in LB supplemented with rifampicin and LB supplemented with rifampicin and 0.5 mM l-cysteine. Plates were grown for 48 h at 28°C and photographed.
In planta growth of virulent or avirulent P. syringaeDC3000
The protocol for measuring the growth of bacteria was adapted from (Tornero & Dangl, 2001). Wild-type and mutant plants were grown for 6–7 wk at the same time and conditions and inoculated with bacterial pathogens as described earlier. One hour after the inoculation, the samples for day zero were taken. To determine bacterial growth, 100 mg of leaves were ground in 500 μl of 10 mM MgCl2 and gently vortexed. From each sample, 20 μl was added to the wells of a microtiter plate containing 180 μl of 10 mM MgCl2, and serial 10-fold dilutions were plated on Petri dishes containing 50 μg ml−1 rifampicin. The plates were incubated at 30°C and after 30 h the number of colonies was counted. The number of colony forming units (CFU) mg−1 fresh weight was determined by the formula:
where N is the number of colonies counted in the dilution number d, and the constant k (500 in our case) represents the number of CFU present in the sample per colony appearing in the first dilution (Tornero & Dangl, 2001).
The Botrytis cinerea strain ME4 was grown in a solid strawberry broth for 12 d, and spore suspensions were prepared at a concentration of 5 × 105 spores ml−1 in 12 g l−1 potato dextrose broth (PDB). Six- to seven-week-old wild-type and mutant plants grown at the same time and conditions were pulverized with a Preval sprayer (Coal City, IL, USA) with spore suspension. Approximately 2 ml of spore suspension per plant was used. The plants were covered with a transparent film to keep 100% humidity. The plants were collected for PCR analysis after 5 d.
Quantification of B. cinerea DNA accumulation in infected plants
DNA from infected plants was quantified by real-time PCR according to a previous study (Calo et al., 2006). DNA from the B. cinerea creA gene (Tudzynski et al., 2000) was amplified using the oligonucleotides creABOT-F and creABOT-R (see the Supporting Information, Table S1). As an internal standard to normalize the real-time PCR, A. thaliana UBQ10 DNA was amplified using the oligonucleotides UBQ10F and UBQ10R (Table S1). Relative quantifications were performed by subtracting the cycle threshold (CT) value of UBQ10 from the CT value of creA (ΔCT). The percentage of B. cinerea DNA was calculated as .
Measurement of the total SA content
The method for total SA extraction was adapted from a previous study (Verberne et al., 2002). A 100 mg sample of 6- to 7-wk-old plant leaf material was collected and pulverized in liquid nitrogen. After this, 0.5 ml of 90% methanol was added to the powder, which was vortexed for 1 min, sonicated for 5 min and centrifuged for 5 min at maximum speed. The supernatant was collected and the pellet resuspended in 0.25 ml of 100% methanol, and the sonication and centrifugation steps were repeated. The supernatants were combined, centrifuged again and the methanol–water mixtures evaporated in a SpeedVac concentrator (Eppendorf, Hamburg, Germany). To the residue, 250 μl trichloroacetic acid (TCA; 5% solution in water) was added and the mixture was vortex mixed. Partitioning with 800 ml ethyl acetate–cyclohexane (1 : 1, v : v) resulted in the separation of two phases. This partitioning was carried out twice. The aqueous phase was subjected to acid hydrolysis by adding c. 300 μl of 8 M hydrochloric acid to the remaining TCA fraction and heating the sample at 80°C for 1 h (Meuwly & Metraux, 1993). The residue was dissolved in 200 μl of sodium acetate pH 5.5/methanol (9 : 1) and analysed by high-pressure liquid chromatography–mass spectrometry (HPLC-MS)s. The analyses of SA were performed with a linear ion trap mass spectrometer equipped with an electrospray ionization source (Bruker Daltonics, Bremen, Germany) coupled to a liquid chromatograph (Ultimate 3000; Dionex, Waltham, MA, USA). A zwitterionic ZIC-HILIC stationary phase column (200 Å, 3.5 μm particle size, 100 × 2.1 mm; Merck SeQuant, Umea, Sweden) was used at 25°C. The elution gradient was carried out with a binary solvent system consisting of 5 mM ammonium acetate (solvent A) and 0.1% acetonitrile/formic acid (solvent B) at a constant flow rate of 0.2 ml min−1. We used the following gradient profile: a linear gradient from 90% to 10% B (0–16 min), 10% B (16–17 min), a linear gradient from 10% to 90% B (17–18 min) and 90% B (18–28 min). Salicylic acid was detected in negative mode giving the spectra of the deprotonated molecule [M − H]− (m/z 137) and the isotopic form m/z 138 with a retention time of 1.6 min (Fig. S1). The system was controlled with the software package hystar (version 3.2; Bruker Daltonic). A standard solution of SA was used to carry out a calibration curve.
Quantification of thiol compounds
To quantify the total cysteine and glutathione contents, thiols were extracted, reduced with NaBH4, and quantified by reverse-phase HPLC after derivatization with monobromobimane (Molecular Probes, Invitrogen, Paisley, UK) as described previously (Dominguez-Solis et al., 2001). To quantify the reduced cysteine and glutathione, thiols were extracted and directly quantified by reverse-phase HPLC (Lopez-Martin et al., 2008a).
RNA isolation and RT reaction
Total RNA was extracted from Arabidopsis leaves using the RNeasy Plant Mini Kit (Qiagen) and reverse transcribed using an oligo(dT) primer and the SuperScript First-Strand Synthesis System for reverse-transcription (RT)-PCR (Invitrogen) following manufacturer’s instructions.
Quantitative real-time RT-PCR was used to analyse the expression of the OAS-A1 and DES1 genes implicated in the synthesis or degradation of cysteine, respectively, and the RPM1, PR1, WRKY54 and WRKY70 genes implicated in plant immunity. First-strand cDNA was synthesized as described earlier. Gene-specific primers for each gene were designed using the Vector NTI Advance 10 software (Invitrogen; Table S1). The PCR efficiency of all primer pairs was determined to be 100%. Real-time PCR was performed using iQ SYBR Green Supermix (Bio-Rad), and the signals were detected on an iCYCLER (Bio-Rad) according to the manufacturer’s instructions. The cycling profile consisted of 95°C for 10 min followed by 45 cycles of 95°C for 15 s and 60°C for 1 min. A melt curve from 60 to 90°C was run following the PCR cycling. The expression levels of the genes of interest were normalized to that of the constitutive UBQ10 gene by subtracting the CT value of UBQ10 from the CT value of the gene (ΔCT). The fold change was calculated as .
Cell death measurements
The protocol for measuring the electrolyte leakage was adapted from (Dellagi et al., 1998). Six-week-old wild-type and mutant plants grown at the same time and conditions were infiltrated with bacteria in 10 mM MgCl2 as described earlier. At different times, 10 5-mm diameter leaf discs collected from the injected area were placed in a plate with 3 ml of water and incubated for 2 h. Conductivity measurements (four replicates for each treatment) were taken from the plate over time using an Eco Scan CON6 (Eutech, Nijkerk, the Netherlands) conductivity-meter. The units of this measurement are μS cm−1, where cm refers to the distance between electrodes. No increase in the conductivity was observed when plants were injected with the mock solution.
For all the experiments shown, at least three independent samples were analysed (for details, see the legend of the respective figure). ANOVA statistical analysis of data was performed using the program originpro 7.5 (OriginLab Corporation, Northampton, MA, USA).
oas-a1.1 and des1-1 mutants transcriptomes show a high correlation with biotic stresses
Microarrays of mutant Arabidopsis plants lacking the cytosolic OASTL OAS-A1 have been published (Microarray Gene Expression Omnibus database accession number: GSE19245, Lopez-Martin et al., 2008a). Microarrays of plants lacking the cytosolic DES1 gene were also performed on des1-1 mutant and wild-type leaves grown on soil under standard long-day conditions for 20 d. Total RNA was prepared and analysed using the Affymetrix-Arabidopsis ATH1GeneChip array. Three biological replicates were performed for each sample. Restricting the analysis to the genes whose expression was changed at least 1.5-fold as a threshold and a significance level of P < 0.05, we identified 179 genes that exhibited alterations in their transcription level. Among them, 68 genes were upregulated in the des1-1 mutant plant compared with the wild-type plant and 111 were downregulated (Microarray Gene Expression Omnibus database accession number GSE19244). To detect physiologically relevant patterns, the genes with altered expression were assigned to functional categories based on the Classification SuperView tool available in the Bio-Array Resource for Arabidopsis Functional Genomics (BAR, Toufighi et al., 2005, http://bar.utoronto.ca). The resulting group lists revealed that a high proportion of genes with altered expression levels in the des1-1 mutant were associated with the plant’s responses to biotic and abiotic stress (Fig. S2), similar to that described for the oas-a1.1 mutant plant microarray analysis (Lopez-Martin et al., 2008a).
A meta-analysis of the oas-a1.1 and des1-1 transcript profiles data was made by comparing with all databases available in the BAR Expression Browser tools (Toufighi et al., 2005, http://bar.utoronto.ca). These analyses showed that both sets of data had an elevated percentage of co-regulated genes with datasets obtained from Botrytis- and Pst DC3000-infected Arabidopsis plants, as well as with data obtained from other pathogen-infected plants (Figs S3–S6). Moreover, data were analysed using the mapman software. Fig. 1 shows that many genes differentially regulated in both des1-1 and oas-a1.1 mutant compared with wild-type plants fit into the Biotic Stress pathway. We observe that these genes are putatively involved in a variety of responses to pathogen attack. These results indicate that the imbalance of cytosolic cysteine alters the expression of groups of genes involved in the plant response to pathogens, suggesting that this thiol could play a role in the plant immunity network. We aimed then to investigate this hypothesis further.
Cysteine content is correlated with plant resistance to pathogens
To investigate the possible role of cytosolic cysteine in plant defense against pathogens, the two alleles defective in the cytosolic l-cysteine desulfhydrase DES1, des1-1 and des1-2 (Alvarez et al., 2010), and the two alleles defective in the cytosolic OASTL OAS-A1, oas-a1.1 and oas-a1.2 (Lopez-Martin et al., 2008a) were challenged by a necrotrophic (B. cinerea) and a hemibiotrophic (P. syringae pv. tomato DC3000 –Pst DC3000) compatible pathogen. When challenged with the fungus, des1-1 accumulated 36% less B. cinerea DNA than wild-type plants of the Col-0 ecotype, whereas des1-2 accumulated up to 56% less B. cinerea DNA than its corresponding No-0 ecotype wild type (Fig. 2a). Conversely, oas-a1.1 and oas-a1.2 accumulated, respectively, 86% and 130% more B. cinerea DNA than their corresponding wild type of the Col-0 ecotype (Fig. 2a). Similarly, both oas-a1.1 and oas-a1.2 showed an earlier infection by Pst DC3000 than the Col-0 wild type, as 2.5 times more Pst DC3000 CFU mg−1 FW was found at 1 d after infection (dpi) than Col-0 and the difference was 1.5 times at 3 dpi (Fig. 2b). The des1-1 mutant was more resistant to Pst DC3000 infection than Col-0 at both 1 dpi (eight times less CFU mg−1 than Col-0) and 3 dpi (2.5 less infected than Col-0). In addition, the des1-2 mutant showed a behavior very similar to the des1-1 mutant; at 1 dpi, it showed five times fewer CFU mg−1 than its corresponding No-0 wild type and at 3 dpi it showed two times fewer CFU mg−1 than No-0. In summary, both kinds of mutants presented opposite phenotypes: whereas des1-1 and des1-2 were more resistant to either B. cinerea or Pst DC3000 infection, oas-a1.1 and oas-a1.2 were more sensitive to both pathogens.
An increase in SA elicits a SAR phenotype in many instances (Vlot et al., 2008). Because des1-1 and des1-2 were more resistant to both a biotrophic and a necrotrophic pathogen, we measured the total SA content in wild-type and des1 mutant plants. As expected, the total SA content in des1-1 mutant was significantly higher (1.5 times) than in wild-type Col-0 plants (Fig. 3a). However, the SA content in des1-2 plants was not statistically different from its corresponding wild type (No-0), although the difference was 1.3 times the No-0 levels (Fig. 3a). We also measured the total amount of SA in the both oas-a1 mutants and no differences were observed in comparison with wild-type Col-0 plants (Fig. 3a). When we further measured thiol content of wild type and mutants, we found that des1 mutants showed increases of 20–30% in Cys content and 14–15% in glutathione content when compared with the respective wild-type plants, whereas oas-a1 mutants showed 24–31% less Cys and 24–28% less glutathione that wild-type plants (Fig. 3b), similar to previously published data (Lopez-Martin et al., 2008a; Alvarez et al., 2010). Furthermore, we calculated the degree of oxidation of each thiol in the different plant lines (Table S2). The values for the degree of glutathione oxidation calculated for both ecotypes of wild type Arabidopsis plants are in accord with those previously reported (Meyer et al., 2007). While the degree of glutathione oxidation in the des1 mutant alleles was slightly lower, it was significantly greater in the oas-a1 mutant alleles, as reported previously (Lopez-Martin et al., 2008a; Alvarez et al., 2010). With respect to the degree of cysteine oxidation, to our knowledge there is no published report of this, and when compared with the wild-type plants we observed similar behavior in the mutants as for glutathione. While the des1 mutants showed a lower degree of oxidation, the oas-a1 mutants showed a greater degree, but this increase was not as strong as that observed for glutathione. Collectively, these data suggest that a 20–30% increase in the cytosolic cysteine or 14–15% glutathione contents, mainly in their reduced forms (as reflected by their lower degree of oxidation), could elicit basal resistance to pathogens via SA signaling.
One of the effects of SA accumulation is the nuclear localization of NPR1 (Tada et al., 2008), which directly induces the expression of several WRKY transcription factors (Wang et al., 2006). Most of these are activators of PR genes, such as PR1, a gene widely used as a SAR reporter. According to the broad-spectrum disease resistance phenotype shown by des1 mutant plants, the expression of WRKY54 and PR1 were induced in both des1-1 and des1-2 mutants in the absence of infection (Fig. 3c). However, the SA content in oas-a1 mutant plants was indistinguishable from the wild-type concentrations, although they were more sensitive to pathogens. By contrast, the WRKY54 but not the PR1, transcription level was reduced in the oas-a1 mutants compared with wild-type plants (Fig. 3c). Transcriptomic data of oas-a1.1 leaves show that five different WRKY transcription factors (including WRKY54) and three PR proteins are repressed in the mutant compared with wild-type leaves (Lopez-Martin et al., 2008a), suggesting that some components of the immune response are compromised in the oas-a1.1 plants.
Because the oas-a1.1 and oas-a1.2 mutants present not only a 24–31% reduction in cysteine but also a 24–28% reduction in glutathione levels (Fig. 3b), we analysed, in parallel, the behavior of the cad2-1 mutant, which has a 70% reduction in glutathione compared with the wild-type, but the amounts of cysteine are not reduced (Ball et al., 2004). WRKY54 was not repressed in the cad2-1 mutants (Fig. 3c, right panel), demonstrating that the repression of this transcription factor is specifically caused by the decrease in cytosolic cysteine content shown by the oas-a1.1 and oas-a1.2 mutants.
In conclusion, the des1 mutants, which have increased cysteine levels, behave as constitutive SAR mutants, whereas oas-a1 mutant plants, which have reduced cysteine levels, are more sensitive to pathogens.
Cytosolic cysteine is required for the HR in an incompatible interaction
In addition to the basal resistance triggered by PAMPs, Pst DC3000 can elicit an ETI in Arabidopsis when expressing either of two sequence-unrelated type III effectors, AvrRpm1 or AvrB (Bent et al., 1994; Mindrinos et al., 1994; Grant et al., 1995). As the HR is one of the responses associated with ETI, we measured the ion leakage of leaf discs infected with the avirulent bacterium Pst DC3000 expressing avrRpm1 (Pst DC3000 avrRpm1). Pst DC3000 avrRpm1 did not induce ion leakage in oas-a1.1 or oas-a1.2 plants (Fig. 4a), although their basal conductivity is higher than that displayed by the wild-type plants (Table S3) as a result of spontaneous cell death lesions in the oas-a1 mutants (Lopez-Martin et al., 2008a). This indicates that oas-a1 mutants have an impaired ETI response elicited by the avrRpm1–RPM1 interaction. Conversely, des1-1 mutant plants, which have an increased cysteine level of 30% (Fig. 3b), elicited ion leakage similar to that of wild-type plants after 9 h of Pst DC3000 avrRpm1 infection (Fig. 4a). The cad2-1 mutant plants showed an increase in conductivity during the infection with the avirulent Pst DC3000 avrRpm1 strain, which was indistinguishable from the increase in wild-type plants under the same conditions. This response demonstrates that the lack of a HR response during an incompatible interaction in the oas-a1.1 and oas-a1.2 mutants was caused by its cysteine deficit and its decreased glutathione content.
To demonstrate that cysteine is a molecule mediating the HR response in A. thaliana, we carried out experiments of reversion of the phenotype in vivo with cysteine. The addition of 500 μM of cysteine at the site of infection reverted the phenotype shown by the oas-a1.1 mutant, whereas the co-infiltration of Pst DC3000 avrRpm1 with 5 mM glutathione did not produce any effect on the HR response in the wild-type or oas-a1.1 mutant plants (Fig. 4a). Furthermore, the addition of cysteine, but not of glutathione, seemed to increase the intensity of the total ion leakage reached by both wild-type and oas-a1.1 mutant plants (Fig. 4a). In order to exclude that the cysteine directly affected pathogen growth instead of functioning through the plant response to the pathogen, we performed growth tests of Pst DC3000 avrRpm1 in solid culture LB media in the absence and in the presence of 500 μM of cysteine. No differences were observed in any of two conditions (Fig. 4b), therefore excluding the possibility of a direct effect of cysteine in the pathogen’s growth rather than an effect in rescuing the oas-a1.1 phenotype.
Cytosolic cysteine is required for the transcriptional regulation of R proteins
One of the earliest steps during the ETI is the effector recognition, that is, a direct or indirect recognition of microbe avirulence effectors by plant resistance proteins (reviewed in Panstruga et al., 2009). In the A. thaliana–Pst DC3000 avrRpm1 system, the interaction between the effector avrRpm1 and the RPM1 CC-NB-LRR resistance protein initiates the plant response, driving an incompatible interaction (Grant et al., 1995). In wild-type plants, the RPM1 transcript is transiently induced during the Arabidopsis–Pst DC3000 avrRpm1 interaction (Fig. 5a). In oas-a1.1 and oas-a1.2 plants, the RPM1 transcription was reduced by 50% with respect to the wild type in control conditions without infection (Fig. 5b). RPM1 expression in noninfected plants was significantly unaltered in the des1-1 and cad2-1 mutants (Fig. 5b) and was slightly increased in the des1-2 mutant. These data suggest that the repression of RPM1 transcription observed in oas-a1.1 plants was a result of their decreased cytosolic cysteine content. A reduction in the RPM1 transcription would be reflected in reduced effector recognition and poor plant response when infected with Pst DC3000 avrRpm1, which was exactly the effect observed. To test whether this poor ETI response affected the development of bacteria in planta, we quantified the evolution of the infection by counting the number of bacteria developing in infected wild type and oas-a1.1 mutant plants at 1 dpi and 3 dpi. As already established (Tao et al., 2003), bacteria growth in wild-type plants is arrested 1 d after infection. As we expected, we observed a poor plant response to infection in the oas-a1.1 mutant, thus the incompatible Pst DC3000 avrRpm1 bacteria were able to grow as a compatible Pst DC3000 strain does (Fig. 5c) while des1-1, des1-2 and cad2-1 mutants showed a phenotype indistinguishable from wild-type plants.
Cysteine and glutathione accumulation patterns and transcript regulation during ETI
The results presented here indicate that cytosolic cysteine is essential to induce the HR response associated with the ETI triggered by the Arabidopsis–Pst DC3000 avrRpm1 interaction, so we determined the kinetics of accumulation of cysteine and glutathione during infection with an avirulent Pst DC3000 avrRpm1 strain. Samples were taken at 1, 3, 6, 9 and 24 h post-infection (hpi). Fig. 6 shows the accumulation of cysteine (Fig. 6a) and glutathione (Fig. 6b) during an incompatible Arabidopsis–Pst interaction. Interestingly, cysteine content decreased transiently, and at 1–3 hpi, its content was reduced to 50% of the level of cysteine in the absence of infection. After 3 hpi, cysteine content started increasing drastically reaching 150% of the noninoculated plant content at 9 hpi and 200% at 24 hpi. Glutathione concentrations, in turn, increased drastically during an incompatible Arabidopsis–Pst DC3000 avrRpm1 interaction, even in the earlier stages of infection, reaching up to 300% of the noninoculated plant level at 6 hpi.
Therefore, cysteine content showed a ‘down and up’ regulation during an incompatible Arabidopsis–Pst DC3000 avrRpm1 interaction, which decreased at the beginning of the interaction and increased later. To determine whether the cysteine content increase is caused by protein degradation during the HR response or de novo cysteine synthesis, we analysed regulation of the transcripts of genes involved in cytosolic cysteine synthesis and degradation during the infection of A. thaliana with Pst DC3000 avrRpm1. OAS-A1- and DES1-coding genes were regulated in an opposite way in response to the infection with an avirulent strain of Pst DC3000 (Fig. 7). While the OAS-A1 transcript transiently decreased, with a minimum of 20% of the uninfected level at 3 hpi, the DES1 transcript showed significantly different kinetics of expression, remaining high in the early hours of infection and declining after 3 hpi. These results were consistent with the pattern of cysteine accumulation observed during an incompatible interaction. Therefore, the variation in cysteine content observed during an incompatible interaction would be produced by modulation of the enzymatic activities related to the synthesis and degradation of this thiol, rather than from protein degradation associated with the HR.
Cysteine and glutathione accumulation and degree of oxidation during ETI in oas-a1 and des1 mutants
The accumulation of cysteine and glutathione, as well as their degree of oxidation, was also measured in oas-a1 and des1 mutants in order to further distinguish the roles of these two thiols in the establishment of an incompatible interaction. As the basal cysteine and glutathione content (at time 0) of each plant is different (Fig. 3b), Fig. 8 shows the evolution of the glutathione and cysteine content of both Col-0 and No-0 wild types and des1 and oas-a1 mutants relative to the basal content of each plant. Glutathione content showed, in all cases, a very similar curve, as observed in Fig. 6(b), with a slight decrease in glutathione content at 1 hpi, an increase after 3 hpi and a maximum at 6 hpi, decreasing after this time to reach at 24 hpi a content of c. 120–160% of the plant at time 0. By contrast, cysteine content is evolved in a different way in each plant line, as wild-type plants and des1 mutants showed an early decrease, followed by a drastic increase at 6, 9 and 24 hpi, as shown in Fig. 8(a), whereas the oas-a1 mutants clearly differed in cysteine content evolution from wild types during ETI, showing a rapid and maintained increase in cysteine content from 1 hpi.
The basal oxidation degree of cysteine and glutathione is also different for every plant line used in this work (Lopez-Martin et al., 2008a; Table S2). Therefore, the progression of the degree of oxidation during ETI was calculated as percentage of the basal oxidation degree of each plant (Fig. 9). In wild types (Col-0 and No-0), the degree of cysteine and glutathione oxidation decreased early (1 hpi) and showed an increase at 3–6 hpi, diminishing slowly to reach values similar to the basal values at 24 hpi (100%). The des1 mutants did not show an early diminution in the degree of glutathione oxidation and the increase in both thiol oxidation degrees at 1–3 hpi was much greater than their respective wild types, although the shapes of the curves displayed by des1 mutants were similar to those of the respective wild types. In the case of oas-a1 mutants, the degree of glutathione and cysteine oxidation remained constant and was unchanged during infection by PstDC3000 avrRpm1, which is consistent with the absence of HR during their interaction with the bacteria (Fig. 5a).
In this article, we present evidence consistent with a regulatory role for cysteine during plant–pathogen interaction. Previous work using plant mutants with increased (des1-1 and des1-2) or decreased (oas-a1.1 and oas-a1.2) cytosolic cysteine content has demonstrated that cysteine is very precisely regulated to maintain cellular homeostasis (Lopez-Martin et al., 2008b; Alvarez et al., 2010). Despite this maintenance, the transcriptomes of the oas-a1.1 and des1-1 mutants and the transcriptomes of plants challenged with pathogens show a high co-regulation. In fact, these mutants show altered basal resistance to pathogens, so that increased cytosolic cysteine content is associated with enhanced resistance to pathogens, whereas decreased cytosolic cysteine content is associated with decreased resistance to pathogens. The expression of genes associated with the plant immune response, as shown by analysis of the transcriptome of the oas-a1.1 and des1-1 mutants or by quantitative real-time RT-PCR, agrees to the phenotype observed in both kinds of mutants, particularly in the case of the des1 mutants, where the levels of SA are increased, as well as the transcription of WRKY54 and PR1. Moreover, the transcriptomic analysis of the des1-1 mutant shows that four PR proteins are induced in the mutant, the defensins PDF1-2a and PDF1-2b among them. The oas-a1.1 and oas-a1.2 mutants have SA levels indistinguishable of those of wild-type plants, as well as PR1 transcription. However, the WRKY54 transcription is decreased in the oas-a1.1 and oas-a1.2 mutants and transcriptomic analysis of the oas-a1.1 mutant shows that WRKY54, WRKY38, WRKY46, WRKY33 and WRKY26 are also repressed in the mutant more than twofold (Lopez-Martin et al., 2008a). Moreover, several PR genes and other signaling components of the plant immune response are altered in the oas-a1.1 mutant (Lopez-Martin et al., 2008a; see Fig. 1). In addition to their altered cysteine levels, oas-a1 and des1 mutants showed respectively a 24–28% decrease and a 14–15% increase in glutathione content. Pathogen attack and application of SA trigger an increase in total glutathione content (Fodor et al., 1997; Vanacker et al., 2001; Mou et al., 2003; Mateo et al., 2006). Other glutathione mutants deficient in the γ-glutamylcysteine synthetase gene (GSH1) and displaying a 70–80% decrease in glutathione content and no cysteine decrease are more susceptible to several pathogens (Ball et al., 2004; Parisy et al., 2007; Schlaeppi et al., 2008). Therefore, we cannot exclude the possibility that glutathione may have a function in basal resistance, and, in fact, it has been proposed that adequate concentrations of glutathione are important in Arabidopsis for limiting the spread of virulent P. syringae, probably by regulating the accumulation of resistance-related compounds (Glazebrook & Ausubel, 1994; Roetschi et al., 2001). However, the results in this report demonstrate a specific function for cytosolic cysteine in plant–pathogen interactions. This function could arise from cysteine per se, or from its role as a generator of elemental sulfur involved in fungal pathogen defense (Cooper et al., 1996; Williams et al., 2002). Another possible role of cysteine is as a determinant of the oxidative capacity of the cytosol (Lopez-Martin et al., 2008a,b; Alvarez et al., 2010). Apart from decreased cysteine and glutathione levels, the oas-a1.1 and oas-a1.2 mutants showed an increased degree of cysteine and glutathione oxidation, which was decreased in des1-1 and des1-2 plants (Lopez-Martin et al., 2008a; Alvarez et al., 2010) (Table S2). Thus, it is also possible that maintaining an adequate thiol redox state is essential for a proper basal response to plant pathogens. In accord with this line of reasoning, des1-1 and des1-2 plants possess a nonoxidizing cytosol and displayed enhanced resistance to plant pathogens, whereas the oas-a1.1 and oas-a1.2 plants have an oxidizing cytosol and showed enhanced sensitivity to plant pathogens.
Plants induce the ETI when they recognize a pathogen effector (or avirulence factor, avr) through resistance (R) proteins. The avr–R interaction initiates a rapid response including an oxidative burst, a HR, SA accumulation and PR gene induction (Panstruga et al., 2009). In A. thaliana, variations in cysteine concentrations during the ETI induced by Pst DC3000 avrRpm1 result from de novo synthesis of this thiol because it is accompanied by specific regulation of genes involved in cytosolic cysteine synthesis and degradation. In the Arabidopsis–Pst DC3000 avrRpm1 system, the RPM1 resistance protein is essential for the initiation of the HR response after challenge by Pst DC3000 avrRpm1 because it mediates the AvrRpm1 recognition (Holt et al., 2000). In the oas-a1.1 mutant, RPM1 is 50% repressed in noninfected plants and it does not reach maximum levels upon infection. Thus, this level of RPM1 seems to be insufficient to initiate the HR in response to Pst DC3000 avrRpm1, and the infection is successful, although oas-a1 mutants displayed spontaneous cell-death lesions (Lopez-Martin et al., 2008a) and basal ion leakage was greater in these mutants than in wild-type plants. In this case, we have shown that the initiation of the HR is independent of glutathione levels in the cell because the lack of an HR response is rescued in the oas-a1.1 knockout mutants by cysteine, but not by reduced glutathione, and the cad2-1 mutant is unaltered in the HR response to Pst avrRpm1. Therefore, in the oas-a1.1 mutant, either the reduced amount of cytosolic cysteine or the oxidizing cytosol, or both, seem to be responsible for the poor transcription of RPM1 and the lack of an HR response after Pst avrRpm1 infection. Under sulfate deprivation, Arabidopsis plants transcribe 0.76 times the RPM1 gene (Genevestigator, https://www.genevestigator.com/), which suggests that sulfur-containing compounds, like cysteine, could be essential for full RPM1 transcription. Observations that supplementary sulfur-fertilization reduced the incidence of fungal pathogens on crops have been reported, and the terms ‘Sulfur-Induced Resistance’ (SIR) or ‘S-enhanced defense’ (SED) have been proposed to explain this phenomenon (Rausch & Wachter, 2005). It has been recently reported that SIR/SED is induced during a compatible plant–virus interaction in tobacco plants, linked to the activation of cysteine and glutathione metabolism (Holler et al., 2010).
NPR1 is a key regulatory protein involved in SA signaling, which has a redox-regulated cellular localization. When NPR1 is oxidized it is an oligomer and is located in the cytosol. The reduction of thiol groups on NPR1 leads to the monomerization of the protein and its localization in the nucleus, where it promotes the transcriptional activation of PR genes (Tada et al., 2008). The oxidized state of the cytosol of the oas-a1.1 mutants may prevent the pathogen-induced NPR1 nuclear localization and, consequently, NPR1 would not induce the cellular response to pathogens, like PR or WRKY induction. This lack of induction would lead to an enhanced sensibility to pathogens because of the lack of PR induction. Therefore, a deficiency in cysteine-mediated redox homeostasis could impair redox signaling such as the NPR1 redox modifications.
During the early steps of the interaction with Pst DC avrRpm1, the evolution of cysteine content in oas-a1 and des1 mutants is completely different, while the evolution of glutathione content is very similar in all plants. This is consistent with a regulatory role of cysteine, and not of glutathione, in the establishment of the ETI. Also, the oxidation state of cysteine and glutathione does not change considerably during the challenge of oas-a1 mutants with Pst DC avrRpm1 whether it is transiently augmented with either wild type or des1 mutants in the early steps of interaction with the avirulent bacteria. This suggests that a correct cytosolic cysteine content is essential for the initiation of the ETI, probably by regulating the redox state of the cell.
The SA content in the des1-1 and des1-2 mutants (with elevated cysteine levels) is higher than in wild-type plants and they act like SAR mutants because they accumulate SA and are more resistant to biotrophic and necrotrophic pathogens than wild-type plants. Consistent with this, PR-1 is constitutively expressed in these mutants.
In conclusion, we have demonstrated that cytosolic cysteine plays a (direct or indirect) role in the establishment and signaling of the plant response to pathogens. First, in light of our observations, we propose that cysteine could have a protective role during PTI, which cannot be separated from the already proposed protective role of glutathione. In addition, cytosolic cysteine is essential to the initiation of the HR response during ETI. Our work suggests that accurate regulation of cytosolic cysteine homeostasis is critical for orchestrating the plant response to pathogens.
This work was funded in part by the European Regional Development Fund (ERDF) through the Ministerio de Ciencia e Innovación (grant no. BIO2010-15201) and the Junta de Andalucía (grant no. BIO-273). This work was also funded by the CONSOLIDER CSD2007–00057, Spain, and by JAE program (CSIC) to CA for fellowship support. We thank Inmaculada Moreno for technical help with this research and M. Carmen López-Martín for performing the meta-analysis of the oas-a1.1 transcriptome. We would like to acknowledge Dr Olga Del Pozo for providing the bacterial strains used in this work.