Loss of cytosolic NADP-malic enzyme 2 in Arabidopsis thaliana is associated with enhanced susceptibility to Colletotrichum higginsianum


  • Lars M. Voll,

    1. Division of Biochemistry, Department of Biology, Friedrich-Alexander-University Erlangen-Nuremberg, Staudtstr. 5, 91058 Erlangen, Germany
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Martina B. Zell,

    1. Institute for Developmental and Molecular Biology of Plants, Heinrich-Heine-University, Universitätsstr. 1, 40225 Düsseldorf, Germany
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Timo Engelsdorf,

    1. Division of Biochemistry, Department of Biology, Friedrich-Alexander-University Erlangen-Nuremberg, Staudtstr. 5, 91058 Erlangen, Germany
    Search for more papers by this author
  • Alexandra Saur,

    1. Division of Biochemistry, Department of Biology, Friedrich-Alexander-University Erlangen-Nuremberg, Staudtstr. 5, 91058 Erlangen, Germany
    Search for more papers by this author
  • Mariel Gerrard Wheeler,

    1. Centro de Estudios Fotosintéticos y Bioquímicos (CEFOBI), Universidad Nacional de Rosario, Suipacha 531, Rosario, Argentina
    Search for more papers by this author
  • Maria F. Drincovich,

    1. Centro de Estudios Fotosintéticos y Bioquímicos (CEFOBI), Universidad Nacional de Rosario, Suipacha 531, Rosario, Argentina
    Search for more papers by this author
  • Andreas P. M. Weber,

    1. Institute of Biochemistry, Heinrich-Heine-University, Universitätsstr. 1, 40225 Düsseldorf, Germany
    Search for more papers by this author
  • Veronica G. Maurino

    1. Institute for Developmental and Molecular Biology of Plants, Heinrich-Heine-University, Universitätsstr. 1, 40225 Düsseldorf, Germany
    Search for more papers by this author

Author for correspondence:
Veronica G. Maurino
Tel: +49 211 8115446
Email: veronica.maurino@uni-duesseldorf.de


  • While photosynthetic NADP-malic enzyme (NADP-ME) has a prominent role in the C4 cycle, the biological function of nonphotosynthetic isoforms remains elusive. Here, we analysed the link between Arabidopsis thaliana cytosolic NADP-ME2 and the plant defence response.
  • Arabidopsis thaliana plants with wild-type and modified NADP-ME2 expression levels were analysed after elicitation with pathogen-associated molecular patterns (PAMPs) and during the interaction with the hemibiotrophic fungal pathogen Colletotrichum higginsianum.
  • Under normal growth conditions, the lack or gain of NADP-ME2 activity produced large changes in plant metabolite pool sizes without any effect on morphology or development. Total NADP-ME activity and NADP-ME2 transcript level were enhanced after PAMP treatment and pathogen infection. During infection with C. higginsianum, loss-of-function mutants of NADP-ME2 (nadp-me2) showed enhanced susceptibility. Transient apoplastic reactive oxygen species (ROS) production after elicitation and callose papilla formation after infection were dampened in nadp-me2. Late salicylic acid (SA)-dependent and SA-independent defence responses were not affected.
  • Taken together, our results indicate that NADP-ME2 is an important player in plant basal defence, where it appears to be involved in the generation of ROS. Moreover, NADP-ME2 was found to be dispensable for later defence responses.


NADP-malic enzyme (NADP-ME; E.C. catalyses the oxidative decarboxylation of l-malate using NADP+ as coenzyme, producing pyruvate, CO2, and NADPH. NADP-ME is well known for its prominent role in C4 plants (Drincovich et al., 2010). In these species, the most abundant NADP-ME isoform is localized to chloroplasts of bundle sheath cells, where it releases CO2 for further fixation by Rubisco (Maier et al., 2011). Apart from this role, nonphotosynthetic NADP-ME isoforms play less well-understood functions in plant metabolism. Some suggested functions are the provision of pyruvate and NADPH for plastidic lipid or amino acid biosynthesis, the control of malate concentrations during hypoxia, the generation of reducing power for anabolic processes in the cytosol, possibly assisting the oxidative pentose phosphate pathway, the provision of NADPH during the detoxification of cytotoxins and xenobiotics via glutathione conjugation in trichomes and the support of CO2 fixation in the mid-veins (Maurino et al., 2001; Gerrard Wheeler et al., 2005; Brown et al., 2010). It has also been long speculated that nonphotosynthetic NADP-MEs would be involved in defence responses. It was shown that transcription of an NADP-ME gene in bean (Phaseolus vulgaris) was rapidly and transiently activated by fungal elicitor treatment in cell-suspension cultures (Walter et al., 1988) and that wounding caused a marked induction of NADP-ME promoter activity, which was also strongly induced upon application of stimuli related to pathogen defence, for example, reduced glutathione, fungal elicitor, cellulase and NADPH (Schaaf et al., 1995). NADP-ME was also shown to respond to plant defence elicitors in bean (Walter et al., 1994), wheat (Triticum aestivum; Casati et al., 1997) and maize (Zea mays; Maurino et al., 2001). It was shown that the expression of a rice (Oryza sativa) cytosolic NADP-ME isoform was induced by NaCl and polyethylene glycol and its overexpression in Arabidopsis thaliana conferred high salt and osmotic stress tolerance (Liu et al., 2007). Recently, it was revealed that NADP-ME is induced more rapidly upon infection with Magnaporthe grisea in resistant Brachypodium distachyon ecotypes than in susceptible ones (Parker et al., 2009).

Colletotrichum higginsianum is a hemibiotrophic ascomycete fungus that was isolated from Brassica rapa (O’Connell et al., 2004). As most A. thaliana ecotypes are susceptible to C. higginsianum, it can be regarded as an adapted pathogen for A. thaliana (Shimada et al., 2006). At the beginning of its hemibiotrophic life cycle, C. higginsianum conidia land on the leaf surface and produce germ tubes, which then differentiate appressoria to penetrate the leaf surface (for reviews on the life style of Colletotrichum sp., see Mendgen & Hahn (2002) or Münch et al. (2008)). During maturation, appressorial cell walls melanize, while compatible solutes are accumulated inside. By diffusion of water into the appressorium, a high appressorial turgor pressure is built up that is translated into mechanical force upon the release of the penetration peg that breaches the host cell wall. In the penetrated host epidermis cells, C. higginsianum establishes itself as a biotroph within 36 h post inoculation by forming an infection vesicle that produces lobed biotrophic primary hyphae. Upon the subsequent colonization of neighbouring cells at c. 72 h post infection, a switch in both hyphal morphology and lifestyle occurs. Narrow-bore necrotrophic secondary hyphae grow rapidly and hyphal spread will eventually lead to necrotic lesions on the infected leaves that are visible as soon as 84 h post infection. In the macerated tissue, new conidiospores are formed mitotically at the base of pin-shaped setae in structures called acervuli.

The A. thaliana genome contains four genes encoding functional NADP-ME isoforms (Gerrard Wheeler et al., 2005). NADP-ME4 is localized to plastids, while NADP-ME1, -2 and -3 are cytosolic isoforms (Gerrard Wheeler et al., 2005, 2009). Although the identity among these isoforms is high, they differ in their expression patterns and biochemical properties (Gerrard Wheeler et al., 2005, 2008), which suggests that each isoform may participate in specific biological functions instead of representing functional redundancy (Maurino et al., 2009). Nonetheless, single and multiple loss-of-function mutants of the A. thaliana NADP-ME isoforms analysed under normal growth conditions did not give any hint of their possible biological functions (Gerrard Wheeler et al., 2005).

As reported here, the analysis of transcriptional co-response patterns indicated an association of A. thaliana NADP-ME2 with the plant defence response. To experimentally validate this hypothesis, we assessed the activity of NADP-ME in A. thaliana upon elicitation with pathogen-associated molecular pattern (PAMP) elicitors and after infection with the fungal pathogen C. higginsianum. The results indicated that the loss of A. thaliana cytosolic NADP-ME2 leads to enhanced susceptibility to C. higginsianum infection and biochemical data suggest that NADP-ME2 plays a role during the basal defence response.

Materials and Methods

Plant material, growth conditions and inoculation with PAMPs

Homozygous knock-out lines nadp-me2.1 (SALK-073818) and nadp-me2.2 (SALK-020607) were isolated through a PCR-based reverse genetic screen for T-DNA insertions in At5g11670 (NADP-ME2; Gerrard Wheeler et al., 2005). Both mutants are in the Columbia (Col) background. Arabidopsis thaliana (L.) Heynh plants were grown in pots containing three parts of soil (Gebr. Patzer KG, Sinntal-Jossa, Germany) and one part of vermiculite (Basalt Feuerfest, Linz, Germany) in a growth cabinet in a 16 h : 8 h light : dark cycle at 22°C day : 18°C night temperatures and at a photosynthetically active photon flux density of c. 100 μmol quanta m−2 s−1.

For elicitation with PAMPs, flagellin (1 μM in distilled water) or harpin (1 μM in distilled water) were syringe-infiltrated into leaves of 3-wk-old A. thaliana wild type. Control plants were infiltrated with the same amount of distilled water.

Colletotrichum higginsianum infection assays

Colletotrichum higginsianum isolate MAFF 305635 (kindly provided by the Ministry of Forestry and Fisheries, Japan) was grown on oatmeal agar plates (OMA: 5% (w/v) shredded oat meal and 1.2% (w/v) agar) for 7 d at 22°C under illumination to promote conidia formation. Colletotrichum higginsianum conidia were harvested by rinsing OMA plates containing a high density of acervuli with distilled water, the conidia titre was adjusted to 2 × 106 conidia ml−1 and the suspension was immediately used for infection experiments. To obtain a maximum level of homogeneity in the distribution of conidia on the leaf surface, infection experiments were performed by evenly spraying conidia suspension on 5-wk-old soil-grown plants that had been cultivated as described above. Spray inoculation was always conducted 30 min before the end of the light phase and the high humidity required for turgor build-up in fungal appressoria was immediately established by covering the tray bottom with a thin layer of water and tightly fitting a hood sprayed with water onto the trays with the plants. High humidity was terminated at the beginning of the light phase at 2.5 d post infection (dpi).

Construction of a binary vector to express NADP-ME2, plant transformation and selection of transformants

In order to complement NADP-ME2 loss-of-function mutants and produce NADP-ME2-overexpressing lines, the full-length cDNA encoding AtNADP-ME2 was cut out from the pCR-Blunt II-TOPO plasmid (Gerrard Wheeler et al., 2005) using the XbaI and KpnI sites. The 1.8-kb fragment was cloned into a modified version of the binary vector pGreen II bearing the BASTA resistance gene (pGreenII-35S-nosBAR; Fahnenstich et al., 2007). The resulting plasmid containing the full-length AtNADP-ME2 cDNA (35S::AtNADP-ME2) was introduced into A. thaliana (Col-0) and the knock-out mutants nadp-me2.1 and -2.2 by Agrobacterium tumefaciens (GV3101)-mediated transformation using the vacuum infiltration method (Bechtold et al., 1993). Transformants were selected by resistance to BASTA. DNA was extracted from leaf material collected from selected plants and used for PCR analyses. Plants containing the transgene were transplanted and allowed to self-pollinate. Seeds from the primary (T1) generation were sown, and resultant T2 plants were subjected to another round of BASTA selection and characterization by means of PCR and NADP-ME activity assay. The process was repeated to obtain nonsegregating T3 transgenic lines. All further analyses were performed with homozygous T3 transgene plants.

Extraction of leaf soluble protein and NADP-ME activity measurement

Leaf material of A. thaliana wild-type and mutant lines was ground in N2 and the resulting powder was suspended in 100 mM Tris-HCl, pH 7.5, 5 mM MgCl2, 2 mM EDTA, 10% (v/v) glycerol and 10 mM 2-mercaptoethanol, in the presence of a protease inhibitor cocktail (Sigma). The homogenates were clarified by centrifugation and the supernatants were separated for activity measurements. NADP-ME activity was determined spectrophotometrically using a standard reaction mixture containing 50 mM Tris-HCl, pH 7.0, 10 mM MgCl2, 0.5 mM NADP and 10 mM l-malate in a final volume of 0.5 ml. The reaction was started by the addition of l-malate. One unit (U) is defined as the amount of enzyme that catalyses the formation of 1 μmol of NADPH min−1 under the specified conditions.

Polyacrylamide gel electrophoresis

Native PAGE was performed using a 6% (w/v) acrylamide separating gel. Electrophoresis was performed at 150 V at 10°C. In gel NADP-ME activity analysis was performed as previously described (Gerrard Wheeler et al., 2005).

Quantitative RT-PCR

The relative expression levels of NADP-ME1, NADP-ME2 and NADP-ME3 were analysed by qRT-PCR. For this purpose, total RNA was isolated from 100 mg leaves using the TRIzol reagent (Gibco-BRL, Germany). RNA was converted into first strand cDNA using the SuperScriptII Reverse Transcriptase (Invitrogen, Germany). Subsequently, the cDNA was used as a template using the fluorescent dye SYBR Green (Applied Biosystems, Foster City, CA, USA) in a 7300 Real Time PCR System (Applied Biosystems). The primers used are listed in Supporting Information Table S1. The Ct, defined as the PCR cycle at which fluorescence levels significantly higher than the background fluorescence are detected, was used as a measure of the transcript level of the target genes. Relative quantification of expression levels was performed using the comparative ΔΔCt method using the A. thaliana ACTIN2 gene expression as a standard.

Quantification of C. higginsianum in planta proliferation by qPCR

To obtain quantitative data for pathogen proliferation and thus host susceptibility, the amount of genomic C. higginsianum DNA in infected leaves was assessed at different time-points after infection as described by Horst et al. (2012) with minor modifications. A total of 2 cm2 leaf material was pooled from punches of at least three fully expanded leaves per replicate and tissue was homogenized with a mortar and pestle with liquid nitrogen. Total genomic DNA was extracted with the Qiagen DNeasy Kit (Qiagen, Hilden, Germany). Fungal DNA quantity was subsequently analysed with a Stratagene Mx3000P qPCR System using Brilliant II SYBR® Green QPCR Master Mix (Agilent Technologies, Waldbronn, Germany) and primers for the C. higginsianum TrpC (indole-3-glycerolphosphate synthase) gene and the A. thaliana RbcS gene ribulose-1,5-bisphosphate carboxylase/oxygenase small subunit (Table S1). The relative quantity of fungal DNA was then calculated as described by Rieu & Powers (2009) and was normalized either to the quantity of host DNA or to the sampled leaf area.

Metabolite analysis by gas chromatography–electron impact–time of flight mass spectrometry (GC-EI-TOF-MS)

For metabolite analysis in normal growth conditions, whole rosettes from 4-wk-old plants were harvested 4 h after the onset of the light period and immediately frozen in liquid nitrogen. For metabolite analysis during C. higginsianum infection, whole rosettes from 4-wk-old plants were collected 2, 2.5 and 3 dpi. Samples at 2 and 3 dpi were collected 1 h before the end of the day and those at 2.5 dpi were collected 1 h before the end of the night and were immediately frozen in liquid nitrogen. Three independent biological replicates were used. The tissues were ground in a mortar and a 50-mg fresh weight aliquot was extracted using the procedure described by Lee & Fiehn (2008). Ribitol was used as an internal standard for data normalization.

Determination of PR gene transcript accumulation

Transcripts of the PR (pathogenesis related) genes PR-1 (At2g19990), PR-2 (At3g57260), plant defensin1.2a (PDF1.2a) (At5g44420) and PR-3 (At3g12500) were detected by RNA gel blot analysis of 8 μg of total RNA per lane, while for the detection of flavin-dependent monooxygenase 1 (FMO1) (At1g19250) and respiratory burst oxidase homolog D or NADPH oxidase D (RBOH-D) (At5g47910) transcripts, 40 μg of total RNA was loaded per lane. RNA extraction, probe synthesis and blot hybridization were performed according to standard procedures.

Determination of free salicylic acid and camalexin

Free salicylic acid (SA) and camalexin were extracted and analysed as described by Nawrath & Métraux (1999) with minor modifications. A total of 2 cm2 of leaf material was pooled from at least two leaf punches, supplemented with 250 ng of o-anisic acid (Acros Organics, Geel, Belgium) as an internal standard, extracted once with 600 μl of 70% MeOH at 65°C for 1 h and once with 600 μl of 90% MeOH at 65°C for 1 h. The combined extracts were then evaporated with a vacuum concentrator, followed by precipitation of the analytes with 500 μl of 5% (w/v) trichloroacetic acid. Free phenols and camalexin were partitioned two times against 600 μl of cyclohexane/ethyl acetate (1 : 1). The combined organic phases were evaporated to near dryness to prevent sublimation of SA and the residue was resuspended in 400 μl of 20% acetonitrile in 25 mM KH2PO4 (pH 2.6).

HPLC separation of SA, o-anisic acid and camalexin was performed on a Dionex Summit system (P680, ASI-100, TCC-100, RF-2000; Dionex Sunnyvale, CA, USA) equipped with a Phenomenex Luna Security Guard C18 column (4.0 × 3.0 mm) followed by a 5-μm Luna C18(2) reversed-phase column (250 × 4.6 mm) (Phenomenex, Torrance, CA, USA). The columns were kept at 30°C. Elution began with 20% acetonitrile in 25 mM KH2PO4 (pH 2.6) and was kept isocratic for 5 min at a flow rate of 1 ml min−1. The concentration of acetonitrile was then increased to 50% within 23 min, further increased to 70% within 2 min and kept isocratic for 4 min. In the subsequent 2 min, the concentration of acetonitrile was reduced to 20% and the columns were equilibrated for 4 min before the injection of the next sample. The fluorescence detector was programmed appropriately to measure o-anisic acid (excitation 305 nm and emission 365 nm, between 0 and 14 min post injection), SA (excitation 305 nm and emission 407 nm, between 14 and 22 min post injection) and camalexin (excitation 318 nm and emission 370 nm, between 22 and 28 min post injection). Peak identification and quantification were performed by comparison to authentic standards. The authentic camalexin reference was extracted from leaf samples infected with C. higginsianum as described by Zhao & Last (1996), analysed by reversed-phase HPLC as described by Glawischnig et al. (2004), identified on the basis of the UV absorption spectrum using a Dionex PDA-3000 photodiode array detector and quantified using the molar extinction coefficient given in Tsuji et al. (1992).

Quantitative determination of reactive oxygen species (ROS) and staining of callose

The quantitative determination of transient production of ROS in response to flagellin (100 nM) and chitin was conducted using the luminiscence-coupled assay as described by Keppler et al. (1989). Chitin was dissolved in distilled water (1 : 1 w/v) and incubated for 30 min in a water bath, and after centrifugation the supernatant was used for the assay. Aniline blue staining and epifluorescence microscopy to visualize callose papillae after infection with Chigginsianum were performed exactly as described in Abbasi et al. (2009). Per leaf specimen, at least 140 appressoria at different positions across the leaf blade were scored for the presence of callose papillae to give the percentage of appressoria above papillae. Per genotype and time-point, four leaves were assessed.

Gene co-expression network analysis

Genes that are co-expressed with NADP-ME2 were identified using the ATTED-II database (http://atted.jp) and are summarized in Table S2. This database allows evaluation of genes that are co-expressed under five sets of experimental conditions (tissue, abiotic stress, biotic stress, hormones and light conditions) (Obayashi et al., 2009). The mutual rank (MR), which is based on weighted Pearson’s correlation coefficients, is used to assess the extent of gene co-expression. The MR is calculated as the geometric mean of the correlation rank of gene A to gene B and of gene B to gene A. The GO annotations were manually actualized and corrected.

Statistical analysis

Significance was determined using Student’s t-test with Excel software (Microsoft Corporation, Unterschleißheim, Germany) and the term ‘significant’ is used only in cases where the change observed had a value < 0.05. Principal component analysis (PCA) of log-transformed metabolite data was performed with the MarkerView software (Version; Applied Biosystems) using the Pareto algorithm after data normalization to sample means.


The nadp-me2 metabolic phenotype and its genetic complementation

In our previous work, independent T-DNA insertion lines in the AtNADP-ME2 gene were isolated (nadp-me2.1 and nadp-me2.2; Gerrard Wheeler et al., 2005). These plants have a total residual leaf NADP-ME activity of 1.9–4.4% of the wild type (Fig. 1). These results together with previous studies (Gerrard Wheeler et al., 2005) indicated that this isoform accounts for the major NADP-ME activity in leaves, roots, flowers and stems of A. thaliana. Under glasshouse conditions, no differences in growth rate and morphology were visible between the nadp-me2 and the wild-type plants at all stages of the life cycle. However, a particular metabolite fingerprint was found in nadp-me2 leaves harvested 4 h after the onset of the light period (Table 1). In both nadp-me2 lines, malate, fumarate, (iso-)citrate, glycerate, and glycerol accumulated significantly with respect to wild type. However, pyruvate, 2-oxoglutarate, glutamine, ornithine, serine, and glycine were significantly reduced in nadp-me2 plants with respect to wild type.

Figure 1.

NADP-malic enzyme (NADP-ME) activity in rosette leaves of Arabidopsis thaliana wild type (wt), loss-of-function (nadp-me2.1 and -2.2), complemented (nadp-me2.1/ME2-1 and -2) and overexpressing (ME2-1 and -2) lines. The values are means ± SE of three replicates of at least six plants each.

Table 1.   Metabolite levels (relative to the wild-type) in rosette leaves of loss-of-function (nadp-me2.1 and -2.2), complemented (nadp-me2.1/ME2-1 and -2) and overexpressing (ME2-1 and -2) lines assayed by gas chromatography–electron impact–time of flight mass spectrometry (GC-EI-TOF-MS)
  1. Samples were harvested 4 h after the onset of the light period from 4-wk-old Arabidopsis thaliana plants.

  2. The entire data set of metabolite levels relative to the internal standard ribitol for all genotypes is available in Supporting Information Table S3.

  3. Those values that are significantly different from the wild type as determined by Student’s t-test (< 0.05) are shown in bold type.

Nicotinic acid1.131.360.870.870.520.33

To demonstrate that the lack of NADP-ME2 activity is responsible for the chemotype observed, a genetic complementation of nadp-me2.1 and nadp-me2.2 with the NADP-ME2 full-length cDNA driven by the cauliflower mosaic virus (CaMV) 35S promoter was conducted. Moreover, overexpression lines were produced by the introduction of the same construct in wild-type plants (Col-0). Many independent lines were generated and analysed in the T3 generation. As shown in Fig. 1, leaf extracts of the selected complemented lines (nadp-me2.1/ME2-1, nadp-me2.1/ME2-2, nadp-me2.2/ME2-1, and nadp-me2.2/ME2-2) showed similar total NADP-ME activity to the wild type, while overexpressors (ME2-1 and ME2-2) exhibited 2.3- to 3.8-fold more NADP-ME than the wild type. As in the case of the loss-of-function mutants, no differences in growth and morphology were observed in the transgenic lines in comparison to the wild-type under glasshouse conditions at all stages of the life cycle. Independent complemented and overexpressing lines were further used for metabolite analysis (Table 1). The results indicated that expression of NADP-ME2 in the nadp-me2 background induced a shift in the opposite direction in the contents of the organic acids malate, fumarate, and (iso)-citrate, which accumulated in nadp-me2 (Table 1). Moreover, those metabolite concentrations that were lower in the loss-of-function mutants remained reduced in at least one complemented line, or reached wild-type concentrations in both of them, as in the case of serine and ornithine (Table 1). Similar trends in changes in the above-mentioned metabolite concentrations were observed in the overexpressing lines (Table 1). Other common changes found in complemented and overexpressing lines were reductions of the concentrations of aspartate, GABA, glycolate, maleate, and phenylalanine, which correlated with the level of NADP-ME activity (Table 1). In addition to the changes common to these two genotypes, both overexpressing lines showed a significant reduction in the concentrations of alanine, gluconate, glutamate, histidine, nicotinic acid, shikimate, succinate, fructose, glucose, and sucrose (Table 1).

In this way, either the lack or the gain of NADP-ME2 activity produced significant changes in primary metabolite pool sizes without any effect on morphology or development in normal growth conditions.

Analysis of transcriptional co-response patterns related NADP-ME2 to plant defence responses

In an attempt to better understand the biological function of AtNADP-ME2, we performed an analysis of transcriptional co-response patterns using the ATTED-II database (Obayashi et al., 2009). The analysis revealed a tight connection between NADP-ME2 and many genes involved in plant defence and stress responses and signalling. These genes include several glutathione transferases, mitogen-activated protein (MAP) kinases, avirulence-responsive proteins, hydrolases, penetration (PEN)2 and PEN3 (both components of preinvasion resistance), bcl-2 homologous antagonist/killer (BAK)1 (brassinosteroid-Insensitive (BRI)1-associated receptor kinase), peroxidases, powdery-mildew-resistance gene o (MLO) (powdery mildew resistance locus), pathogenesis-related proteins, disease resistance proteins, chitinases, chitin elicitor receptor kinase 1 (CERK1), cytochrome P450 momooxygenase 83B (CYP83B1) and cadmium resistance (CAD)1 (induced during the defence response by callose deposition in the cell wall), and others listed in Table S2.

Enhanced NADP-ME2 activity in wild-type leaves during elicitation of defence responses

To experimentally validate the results of the co-expression analysis, total NADP-ME activity was analysed after the treatment with PAMPs. Total NADP-ME activity increased after elicitation with flagellin (flg22) and harpin, showing a maximum at 24 h after infiltration, while in mock-inoculated leaves NADP-ME activity did not change (Fig. 2a).

Figure 2.

Effect of flagellin and harpin on the activity of NADP-malic enzyme (NADP-ME). (a) NADP-ME activity in leaves of wild-type Arabidopsis thaliana during the course of elicitation by flagellin (flg22; grey bars, upper panel) and harpin (hrp; grey bars, lower panel) Mock, black bars. The values are means ± SE of three replicates of at least six plants each. Significant differences to time-point 0 h with < 0.05 are indicated by asterisks. (b) In gel NADP-ME activity assay after native PAGE of the same samples as in (a). As controls, samples of nadp-me2.1 and ME2-1 were processed in parallel.

As it had been shown that NADP-ME activity was increased to a much greater extent in B. distachyon ecotypes resistant to the hemibiotrophic ascomycete fungus M. grisea than in susceptible ecotypes (Parker et al., 2009), we chose to investigate the induction of NADP-ME during the interaction of A. thaliana with the hemibiotrophic ascomycete C. higginisianum. Upon infection of wild-type plants, total NADP-ME activity increased up to twofold at 4 dpi (Fig. 3a). Total NADP-ME activity did not increase in C. higginsianum-infected nadp-me2.1 (Fig. 3a), which indicates that the increase in total NADP-ME activity in wild-type infected leaves is connected to higher transcript levels of NADP-ME2; and that fungal NADP-ME activity does not contribute substantially to the total measured NADP-ME activity in infected leaves.

Figure 3.

Effect of the infection with Colletotrichum higginsianum on the activity of NADP-malic enzyme (NADP-ME). (a) NADP-ME activity in leaves of Arabidopsis thaliana during the course of infection with C. higginsianum. The values are means ± SE of three measurements. Significant differences to time-point 1 d post infection (dpi) with < 0.05 are indicated by asterisks. (b) In gel NADP-ME activity assay after native PAGE of the same samples as in (a) taken at 2 and 3 dpi.

We next aimed to gather additional evidence that NADP-ME2 can account for the observed increase in total NADP-ME activity in infected leaves. In leaves of A. thaliana only two NADP-ME isoforms are expressed, the cytosolic NADP-ME2 and the plastidic NADP-ME4 (Gerrard Wheeler et al., 2005). As a consequence of different electrophoretic mobilities, it is possible to distinguish the two isoforms by native PAGE assayed for NADP-ME activity. While NADP-ME2 is detected as a band of c. 440 kDa, NADP-ME4 is detected as two bands of c. 230 and 140 kDa (Gerrard Wheeler et al., 2005). We employed these in gel NADP-ME activity assays to determine which NADP-ME isoform is responsible for the increased ME activity in response to PAMP or pathogen challenge. As shown in Figs 2(b) and 3(b), the observed increase of total NADP-ME activity in wild-type leaves correlated with an increased activity of a band at 400 kDa, while no activity was found associated with bands of 230 and 140 kDa in any of the conditions tested. However, the two other cytosolic NADP-ME isoforms, NADP-ME1 and NADP-ME-3, also run as bands at an apparent size of c. 440 kDa in in gel activity assays (Gerrard Wheeler et al., 2005) and the possibility cannot be ruled out that expression of these isoforms could be induced upon infection. To clearly ascertain which gene confers the observed increase in NADP-ME activity, transcript levels of all cytosolic NADP-ME isoforms were determined by qRT-PCR. While the levels of the transcripts of NADP-ME1 and -3 were at the detection limit both before and after infection with C. higginsianum, NADP-ME2 transcript was present before infection and showed a significant increase upon infection (Table 2). Together with the absence of NADP-ME induction in nadp-me2 mutants, these results strongly indicate that total NADP-ME activity is enhanced after pathogen infection as a result of a transcriptional induction of the NADP-ME2 gene and/or post-transcriptional processes resulting in higher NADP-ME2 transcript stability.

Table 2.   Relative expression level of NADP-malic enzyme 1 (NADP-ME1), -2 and -3 in leaves of Arabidopsis thaliana wild type (WT) and nadp-me2.1 directly after infection (0 d post infection (dpi)) and 4 dpi determined by qRT-PCR
0 dpi4 dpi0 dpi4 dpi
  1. Values at 4 dpi that are significantly different from the respective values at 0 dpi as determined by Student’s t-test (< 0.05) are shown in bold. ND, not detected.

NADP-ME1ND8.0 ± 0.82.0 ± 0.34.0 ± 1.1
NADP-ME22054 ± 21016088 ± 25097 ± 1751 ± 14
NADP-ME355 ± 0.8ND40 ± 6.3ND

Enhanced susceptibility to infection by C. higginsianum in nadp-me2

The fact that NADP-ME2 accounts for the increase in NADP-ME activity upon PAMP or pathogen challenge prompted us to examine whether NADP-ME2 is important for defence against microbial pathogens. To this end, we tested how NADP-ME2 dosage affected susceptibility to the hemibiotrophic ascomycete C. higginsianum, employing nadp-me2.1 mutants, complemented mutants (nadp-me2.1/ME2-1) and overexpressing (ME2-1) lines as well as Col-0 wild type.

After 4 d of infection, while most of the old leaves of all genotypes were fully collapsed, fully expanded leaves of nadp-me2 showed large necrotic areas more often than those of the other genotypes (as indicated by white asterisks in Fig. 4a). As a quantitative and more objective measure for susceptibility to fungal colonization, the accumulation of fungal genomic DNA in the host tissue was quantified by qPCR at different time-points after infection (Fig. 4b). During the early biotrophic infection phase, fungal DNA content was not significantly different from that of control leaves harvested immediately after spray inoculation, which was used to normalize the relative quantification data. With the onset of the necrotrophic colonization phase between 3 and 4 dpi, substantial fungal proliferation could be observed. At 4 dpi, nadp-me2 mutants contained approx. fourfold more C. higginsianum DNA than wild type (Fig. 4b). The nadp-me2.1/ME2-1 and the ME2-1 transgenic lines behaved like wild type (Fig. 4b).

Figure 4.

Susceptibility to Colletotrichum higginsianum infection. (a) Representative Arabidopsis thaliana plants at 4 d post infection (dpi) with 2 × 106 C. higginisanum conidiospores ml−1. The plant genotypes are indicated above the images. White asterisks indicate fully expanded leaves that are covered with extensive necrotic lesions. (b) Quantitation of C. higginsianum genomic DNA in infected, fully expanded A. thaliana leaves by qPCR at the indicated time points after infection given as NRQ (normalized relative quantity) of the C. higginsianum TrpC signal relative to the A. thaliana RbcS signal. One representative out of four experiments is shown and NRQ data are normalized to the 0 dpi control. In each experiment, at least three biological replicates per genotype and time point with three technical replicates each were analysed. Data represent mean values ± SE.

ROS production and callose papilla formation are dampened in nadp-me2

As susceptibility was increased in nadp-me2 mutants, we aimed to assess the output of basal and SA-triggered defence based on various readouts. During plant basal defence, ROS and callose papillae are formed at the site of attempted penetration. To analyse the involvement of NADP-ME2 in these early defence reactions, we first quantified the transient production of ROS with a luminescence-based assay. To better assess the capacity of the examined genotypes to produce ROS, the bacterial elicitor flg22 and the fungal PAMP chitin were used in the assays. As shown in Fig. 5(a,b), nadp-me2.1 plants produced significantly less ROS production after flg22 and chitin elicitation than the wild type. Complemented plants produced ROS at wild-type levels after flg22 treatment (Fig. 5a), while much higher ROS production than in the wild type was observed after chitin elicitation in these transgenic lines (Fig. 5b). NADP-ME2 overexpressors showed elevated production of ROS compared with the wild type after flg22 and chitin treatment (Fig. 5a,b). As expected, ROS production was absent in all genotypes in the absence of flg22 induction and in the flagellin receptor mutant flagellin sensitive2 (fls2) after induction. Taken together, these results indicate a strong link between the PAMP-triggered production of ROS and NADP-ME2.

Figure 5.

Measurement of reactive oxygen species (ROS) production after elicitation with pathogen-associated molecular patterns (PAMPs) and callose papilla formation. Flagellin (a) and chitin (b) induced ROS formation determined by the luminiscence-coupled assay. In (a) flagellin sensitive2 (fls2) after flg22 treatment and all genotypes after mock treatment showed similar responses. Error bars represent SE of three independent replicates of four plants each. (c) Callose papillae formation in response to Colletotrichum higginsianum infection. At 2 d post infection (dpi) (left) and 3 dpi (right) after C. higginsianum infection of Arabidopsis thaliana leaves with 2 × 106C. higginisanum conidiospores ml−1, four leaves per genotype were stained with aniline blue as described in the Materials and Methods section. From each leaf, at least 140 appressoria from different parts of the leaf blade were scored for the presence of a callose papilla underneath and the fraction of appressoria exhibiting callose papilla was calculated. Data are thus means of four biological replicates ± SE. Significant differences from the wild type sample sampled at the same time-point with < 0.05 are indicated by an asterisk.

In response to microbial infection, plant cells swiftly deposit callose papillae at the site of pathogen recognition to prevent penetration into the underlying apoplastic compartment. As an additional readout for the strength of basal defence, we stained callose with aniline blue before scoring the occurrence of callose papillae underneath fungal appressoria by fluorescence microscopy. Callose deposition in response to C. higginsianum infection was delayed in nadp-me2 plants compared with the wild type, while NADP-ME2 overexpressors responded like wild type (Fig. 5c).

PR gene expression and SA and camalexin accumulation

It is well established that the strength of the basal defence response influences the amount of the defence messenger SA that is produced (Durrant & Dong, 2004). Therefore, we tested whether the dampened basal defence response in nadp-me2 mutants resulted in diminished SA production and PR gene induction.

At 3 dpi, a time-point at which fungal proliferation was similar among the genotypes (Fig. 4b), free SA contents were similar among all genotypes tested (Fig. 6a). In addition, the accumulation of both the SA-dependent PR genes PR-1 and PR-2 and the jasmonic acid (JA)-dependent PR genes PDF1.2 and PR-3 remained unaltered at 3 dpi (Fig. 6b). To assess defence reactions that are either controlled upstream or largely independent of SA or JA, we looked at the induction of FMO1 (Fig. 6b) and at the accumulation of the major phytoalexin camalexin (Fig. 6c) in response to C. higginsianum infection. Both remained similar among the genotypes. Taken together, our results indicate that late SA-dependent and SA-independent defence responses are not substantially affected by the loss of NADP-ME2 activity.

Figure 6.

Free salicylic acid (SA) (a), Pathogenesis-Related (PR) gene transcript accumulation (b) and camalexin (c) accumulation in response to Colletotrichum higginsianum infection. At 0 dpi (control), 2 d post infection (dpi), 3 and 4 dpi after C. higginsianum spray infection of Arabidopsis thaliana leaves with 2 × 106 conidiospores ml−1 (from left to right), leaf punches from three leaves were pooled per biological replicate for the determination of free SA (a) and camalexin contents (c). Data are means of four biological replicates ± SE. (b) At 0 dpi (control, left panel), 2 dpi (middle panel) and 3 dpi (right panel) the indicated PR gene transcripts were detected by RNA blot analysis of fully expanded A. thaliana leaves pooled from four plants. Either 8 μg of total RNA (PR-1, PR-2, PR-3, and plant defensin1.2a (PDF1.2)) or 40 μg of total RNA (FMO1 and RBOH-D) were loaded and hybridized to the appropriate DNA probes. Plant genotypes are indicated above each lane.

Metabolite profiling during C. higginsianum infection

Once nadp-me2 mutants proved to be more susceptible to C. higginsianum infection, we were interested as to whether metabolite profiles changed upon infection. To this end, three replicate samples were taken from mock control and infected leaves of Col-0, nadp-me2.1 and ME2-1 at 2 dpi (initial biotrophic phase), 2.5 dpi (biotrophic phase) and 3 dpi (initial necrotrophic phase) (see raw data in Table S4). PCA of all samples revealed that principal component 1 (PC1) reflected the progress of infection, explaining > 40% of the variance in the data set (Fig. 7). In addition, PC2 separated the different mock control time-points from each other. Interestingly, infected and control samples were not separated for the 2-dpi time-point. However, a PCA of the 2-dpi samples revealed that samples from control and infected leaves could be separated (Fig. 8a), with infection representing the strongest determinant in this data subset (PC1 with > 30% of the total variance). This indicates that detectable changes in metabolite steady-state pools occur already during the biotrophic phase. Among the 2-dpi samples, nadp-me2 and wild-type profiles could be clearly distinguished by PC2, irrespective of whether infected or mock control samples were considered. The metabolite profile of ME2-15 exhibited higher divergence in comparison with the two other genotypes. Interestingly, the described differences among the three genotypes were absent in samples of infected leaves taken at 2.5 and 3 dpi, indicating that the progress of C. higginsianum infection had a much greater impact on the metabolite profile than genotype (Fig. 8b). Furthermore, the two infection time-points, representing the biotrophic and initial necrotrophic infection stage, could be distinguished quite well (PC1 explaining > 40% of the variance). Nevertheless, malate contents remained largely unaltered in nadp-me2 mutants upon C. higginsianum infection at 3 dpi (Table 3), while they increased twofold in infected leaves of wild type (Table 3). At 3 dpi, infected wild-type leaves also accumulated other carboxylic acids, such as fumarate, succinate and 2-oxoglutarate, soluble sugars, and most free amino acids, including the nonproteinogenic amino acid GABA, to concentrations more than twofold higher than in mock control plants (Table 3). In nadp-me2 mutants, the accumulation of soluble sugars and free amino acids was already evident in noninfected control leaves relative to the wild type (see RMA in Table 3). Thus, the increase in the pool sizes of these metabolites was reduced upon C. higginsianum infection, impairing the changes of primary carbon and nitrogen metabolism in response to C. higginsianum colonization in the mutants.

Figure 7.

Principal component analysis with all metabolite profiling data (principal component 1 (PC1) vs PC2). GC-MS metabolite profiling was conducted with infected and mock control Arabidopsis thaliana plants at 2, 2.5 and 3 d post infection (dpi) from Columbia (Col-0), nadp-me2.1 and ME2-1 with three biological replicates each. The left panel depicts PC1 vs PC2, while the right panel displays the loading plot for the individual metabolites. Circles, control; squares, infected. Black symbols, Col-0; open symbols, nadp-me2.1 (me2.1); grey symbols, ME2-1 (ME-OX); c, control; i, infected; 2, 2 dpi; 2.5, 2.5 dpi; 3, 3 dpi.

Figure 8.

Principal component analysis of metabolite profiling data. (a) Principal component 1 (PC1) vs PC2 with samples of the 2 d post infection (dpi) time-point only. (b) PC1 vs PC2 with data for infected leaves harvested at 2.5 and 3 dpi. GC-MS metabolite profiling was conducted at 2, 2.5 and 3 dpi with infected and mock control Arabidopsis thaliana plants from Columbia (Col-0), nadp-me2.1 and ME2-1 lines with three biological replicates each. Left panels in (a) and (b) depict PC1 vs PC2 for the indicated comparisons, while the right panels display the loadings plots for the individual metabolites in the respective analyses. Circles, control; squares, infected. Black symbols, Col-0; open symbols, nadp-me2.1 (me2.1); grey symbols, ME2-1 (ME-OX); c, control; i, infected; 2, 2 dpi; 2.5, 2.5 dpi; 3, 3 dpi.

Table 3.   Levels (relative to the internal standard ribitol) of selected metabolites in Arabidopsis thaliana wild type (WT) and nadp-me2.1 at 3 d post infection (dpi) measured by gas chromatography–electron impact–time of flight mass spectrometry (GC-EI-TOF-MS)
3 dpiWTnadp-me2.1RMA upon infection
Control (C)Infected (I)Control (C)Infected (I)
  1. Samples were taken at the end of the light period. RMA, relative metabolite accumulation upon infection, which is given by the ratio nadp-me2.1 I/nadp-me2.1 C divided by WT I/ WT C. Values indicate the mean ± SE and those shown in bold case indicate significant differences compared with the corresponding control values calculated by Student’s t-test (< 0.05).

  2. Asterisks indicate significant differences of nadp-me2.1 C and I from WT C and I, respectively, calculated by Student’s t-test (< 0.05).

2-oxoglutarate0.02 ± 0.000.08 ± 0.020.04 ± 0.00*0.14 ± 0.010.88
Alanine9.59 ± 1.0548.05 ± 6.0421.40 ± 3.6348.45 ± 3.810.45
Asparagine0.22 ± 0.030.70 ± 0.070.64 ± 0.230.74 ± 0.060.36
Aspartate1.81 ± 0.561.76 ± 0.193.32 ± 0.862.13 ± 0.440.66
(Iso)Citrate2.94 ± 0.083.61 ± 0.104.39 ± 0.692.88 ± 0.490.53
Fructose1.11 ± 0.1710.91 ± 2.216.24 ± 1.628.28 ± 0.650.14
Fumarate54.96 ± 4.67109.10 ± 22.59140.82 ± 16.19*71.19 ± 7.490.25
GABA0.02 ± 0.000.39 ± 0.130.04 ± 0.010.39 ± 0.000.50
Gluconate0.01 ± 0.000.09 ± 0.010.11 ± 0.060.08 ± 0.010.08
Glucose9.88 ± 2.0946.41 ± 13.6534.16 ± 5.95*63.76 ± 7.040.40
Glutamate11.39 ± 2.2425.80 ± 1.8827.88 ± 7.6622.55 ± 0.970.36
Glycerate0.96 ± 0.071.05 ± 0.141.98 ± 0.351.15 ± 0.120.53
Glycerol1.69 ± 0.182.27 ± 0.402.61 ± 0.311.71 ± 0.080.49
Glycine11.57 ± 1.6749.95 ± 6.1928.11 ± 5.7451.50 ± 4.050.42
Glycolate0.61 ± 0.100.71 ± 0.150.94 ± 0.080.90 ± 0.050.82
Isoleucine0.07 ± 0.010.37 ± 0.090.17 ± 0.030.31 ± 0.020.34
Lactate3.81 ± 0.403.39 ± 0.596.69 ± 0.893.65 ± 1.400.61
Lactose0.005 ± 0.0010.11 ± 0.070.01 ± 0.000.23 ± 0.111.05
Leucine0.03 ± 0.010.31 ± 0.090.10 ± 0.020.27 ± 0.020.26
Lysine0.06 ± 0.010.14 ± 0.030.12 ± 0.040.18 ± 0.010.64
Malate10.69 ± 0.9816.33 ± 2.2419.37 ± 2.35*17.82 ± 0.800.60
Maleate0.36 ± 0.080.416 ± 0.060.61 ± 0.110.52 ± 0.040.74
Maltose0.37 ± 0.1011.11 ± 5.360.81 ± 0.1810.27 ± 1.960.42
Mannitol0.54 ± 0.140.96 ± 0.260.96 ± 0.250.88 ± 0.030.52
Mannose0.02 ± 0.000.08 ± 0.020.07 ± 0.010.09 ± 0.010.32
Methionine0.12 ± 0.030.16 ± 0.010.22 ± 0.070.22 ± 0.070.75
(Myo-)Inositol5.86 ± 1.124.06 ± 0.2913.37 ± 2.504.37 ± 0.390.47
Ornithine0.02 ± 0.000.05 ± 0.010.06 ± 0.010.06 ± 0.010.40
Phenylalanine0.23 ± 0.043.68 ± 1.170.54 ± 0.133.00 ± 0.600.35
Proline0.04 ± 0.010.58 ± 0.120.13 ± 0.050.32 ± 0.040.17
Raffinose0.10 ± 0.010.04 ± 0.000.18 ± 0.040.04 ± 0.000.56
Serine3.26 ± 0.368.52 ± 0.075.73 ± 0.839.60 ± 0.23*0.64
Shikimate1.04 ± 0.232.08 ± 0.382.01 ± 0.341.90 ± 0.120.47
Sorbitol0.06 ± 0.010.09 ± 0.020.26 ± 0.070.11 ± 0.030.28
Succinate1.30 ± 0.184.44 ± 0.782.20 ± 0.464.49 ± 0.360.60
Sucrose57.24 ± 7.54123.01 ± 18.59110.69 ± 2.52*144.79 ± 11.950.61
Threonine1.74 ± 0.253.56 ± 0.313.19 ± 0.10*3.35 ± 0.270.51
Tyrosine0.03 ± 0.010.11 ± 0.040.13 ± 0.00*0.11 ± 0.010.23
Valine0.40 ± 0.091.65 ± 0.350.96 ± 0.221.68 ± 0.230.42


The main goal of this work was to investigate the long-discussed potential link between A. thaliana cytosolic NADP-ME2 and plant defence.

NADP-ME2 is involved in basal defence in A. thaliana

In previous studies, A. thaliana NADP-ME2 was shown to be almost constitutively expressed in A. thaliana organs and it was found to be responsible for the major part of the total NADP-ME activity in different organs (Gerrard Wheeler et al., 2005). In this report, we show that in normal growth conditions the lack or gain of NADP-ME2 activity produces large changes in plant metabolite pool sizes without any effect on morphology or development. Furthermore, we have shown that NADP-ME2 was the only NADP-ME isoform that presented an enhanced transcript level and increased enzymatic activity in A. thaliana after both PAMP and pathogen challenge.

It has been known for quite a while that NADP-ME is induced in various abiotic stress conditions (Maurino et al., 2001; Liu et al., 2007) and during elicitor treatment and plant–pathogen interactions (Walter et al., 1988; Schaaf et al., 1995; Casati et al., 1997, 1999; Maurino et al., 2001; Parker et al., 2009) and recently it was shown that NADP-ME induction proceeds more rapidly in incompatible compared with compatible interactions with B. distachyon (Parker et al., 2009). Consequently, the sites of NADP-ME induction and ROS production coincided during early infection with this hemibiotrophic fungus (Parker et al., 2009). In the present study, we demonstrated that, in A. thaliana leaves infected with the hemibiotrophic fungus C. higginsianum, only one NADP-ME cytosolic isoform, NADP-ME2, responded exclusively to pathogen challenge. We also demonstrated that NADP-ME2 was required for effective basal resistance in A. thaliana, as callose papilla formation was delayed upon challenge with C. higginsianum and ROS burst was dampened in response to PAMPs in nadp-me2 mutants compared with the wild type. As papillary callose formation does not seem to affect penetration resistance of A. thaliana to adapted and nonadapted Colletotrichum species (Narusaka et al., 2004; Shimada et al., 2006), it appears quite unlikely that the retarded formation of callose papillae in nadp-me2 mutants represents the cause of the increased susceptibility of the nadp-me2 mutant to C. higginsianum infection. Rather, the reduced potential of the entire basal defence, comprising a variety of cellular and molecular response, might determine enhanced susceptibility of the nadp-me2 mutant.

In contrast to the basal defence response, free SA accumulation and PR gene induction in response to C. higginsianum infection remained largely similar between wild type and nadp-me2, demonstrating that host defence is not hampered in nadp-me2 at later stages. This suggests that NADP-ME2 is an important player in PAMP-triggered immunity, but is dispensable for later defence responses.

As PAMPs, we have employed chitin, a polymer of N-acetyl-d-glucosamine that is a major component of fungal cell walls, harpin, a cell-envelope-associated protein, and flg22, a 22-amino acid peptide representing the elicitor-active epitope of flagellin (Felix et al., 1999). Perception of flg22 depends on the leucine-rich repeat (LRR)-receptor kinase flagellin sensing 2 (FLS2), which rapidly associates with another LRR-receptor-like kinase, BAK1, initiating a downstream mitogen-activated protein kinase pathway (as reviewed by Boller & Felix, 2009). BAK1, which is structurally similar to FLS2, mediates brassinosteroid signalling and is required for the immune responses triggered by multiple microbe-associated molecular patterns (Chinchilla et al., 2007). Interestingly, a proteomic study indicated that NADP-ME2 protein is induced by brassinosteroid treatment in A. thaliana (Deng et al., 2007) and gene co-expression analysis showed that NADP-ME2 is in the same network as BAK1 (Table S2). All these findings together indicate that NADP-ME2 may be a target of a convergent BAK1 downstream signalling.

What could be the role of NADP-ME2 during host defence?

During pathogen attack, the ROS produced orchestrate the establishment of plant defence responses, which include the stimulation of cell wall cross-linking, the activation of downstream redox signalling and the induction of programmed cell death, and aid molecular attack against the pathogens (Torres et al., 2006; Mur et al., 2008). Some sources of ROS in plants, for example, NADPH oxidases, rely on NADPH as an electron donor (Scharte et al., 2009). Cytosolic NADPH may be produced via several enzymes, which include glucose-6-phosphate dehydrogenase, proline dehydrogenase, glyceraldehyde-3-phosphate dehydrogenase and cytosolic NADP-isocitrate dehydrogenase (Noctor, 2006; Cecchini et al., 2012; Mhandi et al., 2010). Like the above-mentioned enzymes, NADP-ME may also be involved in the provision of NADPH for the production of H2O2 and superoxide, as suggested by the results of Parker et al. (2009). Corroborating the findings of Parker et al. (2009), our results clearly show that cytosolic NADP-ME2 is involved in the generation of ROS during basal defence in A. thaliana.

However, we can assume that the demand for NADPH for anabolic processes, such as the biosynthesis of PR proteins, the biosynthesis of plant phytoalexins and the deposition of lignin, is higher in infected cells (Casati et al., 1999). NADP-ME2 could thus make a significant contribution to meeting the overall cellular demand for reducing power. However, we did not determine the reduction state of the NADP pool in our study.

As indicated by similar clustering patterns of infected relative to control samples for both wild type and the nadp-me2 mutants in the PCA, similar changes in the metabolome occurred in both genotypes during C. higginsianum infection (Figs 7, 8). However, the accumulation of most metabolites of central carbon and nitrogen metabolism was substantially less pronounced in infected nadp-me2 mutants at 3 dpi, because most metabolite pool sizes were already increased relative to the wild type in nonchallenged mutant leaves (Table 3). This indicates that altering malate metabolism by knocking out NADP-ME2 has profound consequences for the response of the entire primary metabolism upon C. higginsianum infection. In particular, the diminished accumulation of some organic acids in infected nadp-me2 mutant leaves in comparison to wild type may indicate reduced metabolic flux through the tricarboxylic acid (TCA) cycle into respiration. In nadp-me2 mutants, NADPH-dependent ROS production and the subsequent activation of the basal defence response by ROS (Torres et al., 2006) could be hampered in comparison to controls. Similarly, B. distachyon leaves infected with M. grisea exhibited a sustained fourfold increase in the content of malate, but not of other organic acids, irrespective of whether a compatible or an incompatible interaction was being observed (Parker et al., 2009), indicating that the in vivo flux through cytosolic NADP-ME relative to the total carboxylic acid pool might be different in challenged rice leaves compared with A. thaliana. Nevertheless, a substantial accumulation of proline and transient increases in ornithine and GABA contents were consistently observed in both pathosystems, indicating that the proline-P5C cycle, which is thought to represent another route to providing reducing equivalents for the production of ROS during plant defence (Cecchini et al., 2011), operates in both pathosystems. Interestingly, proline accumulation was fivefold reduced in infected nadp-me2 mutants (2.5-fold increase with respect to noninfected plants) at 3 dpi as compared with infected wild-type leaves (14.5-fold increase with respect to noninfected plants) (Table 3), suggesting that the provision of reducing equivalents for ROS production by both NADP-ME and the proline-P5C cycle is dampened in the mutants, which will consequently impair the generation of ROS in the mutants. Using PAMP elicitors, we provided experimental evidence that ROS production is indeed decreased in the mutants compared with the wild type (Fig. 2a).

Taken together, our results strongly support a role of NADP-ME2 in plant basal defence, where it appears to be required for the production of ROS following pathogen recognition. Although the exact mechanism of action is still unknown, NADP-ME2 could also be involved in meeting the demand of reducing power for the increased synthesis of building blocks and for the production of superoxide and H2O2 through NADPH oxidase and/or cell wall peroxidases.


We thank Ulrike Hebbeker and Katrin Weber for excellent technical assistance. Financial support was provided by the Deutsche Forschungsgemeinschaft (DFG) through grants MA2379/8-1 and 8-2 to V.G.M. and by funding in the framework of the DFG-FOR666 to L.M.V.