• catalase deficiency;
  • Arabidopsis thaliana;
  • microarray;
  • hydrogen peroxide;
  • high-light irradiation


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

In plants, hydrogen peroxide (H2O2) plays a major signaling role in triggering both a defense response and cell death. Increased cellular H2O2 levels and subsequent redox imbalances are managed at the production and scavenging levels. Because catalases are the major H2O2 scavengers that remove the bulk of cellular H2O2, altering their levels allows in planta modulation of H2O2 concentrations. Reduced peroxisomal catalase activity increased sensitivity toward both ozone and photorespiratory H2O2-induced cell death in transgenic catalase-deficient Arabidopsis thaliana. These plants were used as a model system to build a comprehensive inventory of transcriptomic variations, which were triggered by photorespiratory H2O2 induced by high-light (HL) irradiance. In addition to an H2O2-dependent and -independent type of transcriptional response during light stress, microarray analysis on both control and transgenic catalase-deficient plants, exposed to 0, 3, 8, and 23 h of HL, revealed several specific regulatory patterns of gene expression. Thus, photorespiratory H2O2 has a direct impact on transcriptional programs in plants.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

A fine-tuned balance between hydrogen peroxide (H2O2) production and its removal is crucial for plant survival. However, not only does perturbed H2O2 management lead to cellular damage, but also moderately elevated H2O2 levels induce a defensive response during both biotic and abiotic stresses (Dat et al., 2000). Such a defense response encompasses acclimation (Karpinski et al., 1999; Prasad et al., 1994), cross-tolerance (Bowler and Fluhr, 2000; Pastori and Foyer, 2002), and the induction of active cell death processes (Dat et al., 2003; Levine et al., 1994).

Like other eukaryotes, plants have an elaborate antioxidant system to control reactive oxygen species (ROS) concentrations during fluctuating environmental conditions. The different components of the enzymatic antioxidant system are diversified in their scavenging capacity not only in terms of specificity and affinity towards the different ROS species (e.g. superoxide, H2O2, singlet oxygen), but also by their appropriate subcellular localization in the vicinity of ROS production sites (Vranováet al., 2002). An important source of H2O2 during light stress is the photochemical quenching of excess light by the Mehler reaction and by photorespiration. In C3 plants, photorespiratory H2O2 production would account for the majority of total H2O2 formed (Karpinski et al., 2003; Noctor et al., 2002). This abundant H2O2 production is mainly counteracted by peroxisomal catalase, although other antioxidative enzymes are active in the leaf peroxisome (Corpas et al., 2001). Catalase is a tetrameric iron porphyrin that catalyzes the dismutation of H2O2 to water and oxygen. Peroxisomal catalase, perturbed by mutation or gene silencing, results in decreased H2O2 scavenging during high-light (HL) irradiation or low CO2 concentrations in different C3 plants. These catalase-deficient plants were not only more sensitive to a variety of environmental stress factors, such as ozone and salt stress (Kendall et al., 1983; Willekens et al., 1997), but also became very useful tools to study the signaling role of photorespiratory H2O2in planta. In catalase-deficient tobacco, elevated photorespiration-derived H2O2 is able to trigger both local and systemic defense responses and activate a cell death program that shares several features of programmed cell death (Chamnongpol et al., 1998; Dat et al., 2003). In a catalase-deficient barley mutant, Noctor et al. (2002) nicely demonstrated the ability of photorespiratory H2O2 to perturb antioxidant redox states in leaves. Such redox perturbations are inevitably potent drivers of signal transduction events at both the biochemical and the gene expression levels (Desikan et al., 2001; Vandenabeele et al., 2003).

In Arabidopsis thaliana, the catalase (cat) multigene family contains three genes (At4g35090, At1g20630, and At1g20620) with nucleotide and amino acid sequences 70–72% and 75–84% identity, respectively (Frugoli et al., 1996). All three genes are highly expressed in inflorescences, but only the cat2 and cat3 genes are highly expressed in leaves and are both under the control of the circadian clock. Six different enzyme isoforms have been described in Arabidopsis (Michael and McClung, 2002; Zhong and McClung, 1996). Here, we report the first production of transgenic Arabidopsis plants that are perturbed in their most predominant leaf catalase (CAT2). The importance of peroxisomal catalase was confirmed because these transgenic plants were more sensitive to ozone stress and, more importantly, these plants allowed us to study comprehensively the transcriptomic variances induced by photorespiratory H2O2.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

Transgenic Arabidopsis thaliana with different levels of residual catalase are more stress sensitive

Transgenic lines of Arabidopsis with various levels of decreased CAT2 activity were produced (for details, see Experimental procedures). Transcript, protein, and activity levels of CAT2, which are usually high in leaves, were reduced in four independent lines (CAT2AS, CAT2S, CAT2HP1, and CAT2HP2), retaining 65, 50, 20, and 7% of total residual catalase activity levels, respectively (Figure 1a,b). Under ambient growth conditions in vitro or in controlled growth chambers, no obvious phenotypic differences were noticeable between transgenic and control plants (wild-type or empty-vector control PTHW), except for CAT2HP2 plants, which showed significant growth retardation. These plants were on average 30% smaller than the control plants 24 days after germination (Figure 1c) and never reached a normal size at later developmental stages. DW and FW measurements confirmed the observed growth differences (see Experimental procedures and Supplementary Material).


Figure 1. Catalase levels and growth characteristics of transgenic plants.

(a) Measurements of catalase activity with a spectrophotometric assay. Catalase activity was lower in the catalase-deficient lines. Activity levels were in agreement with protein levels shown in (b). Results are average of measurements on three plants of each line. The control line (PTHW) was used as a reference.

(b) Immunodetection of CAT2 isozyme by using antiserum against a cottonseed catalase (Mullen et al., 1997). Catalase protein levels correlated with the residual catalase activity in catalase-deficient plants.

(c) Surface area measurements of leaves 1 and 2, indicating reduced size of CAT2HP2 plants. Leaf surface is presented as average leaf surface (n = 8–10) ± SE.

Download figure to PowerPoint

Although the molecular mechanisms remain unresolved, elevated levels of ROS are known to affect both growth and productivity of plants. Other mutants and transgenic Arabidopsis plants, deficient in their antioxidative potential, were also affected in growth. Pnueli et al. (2003) reported knockout ascorbate peroxidase (Apx)1 plants with a suppressed growth and development phenotype. Even though in the vitamin C-deficient mutant vtc1, decreased ascorbate pools, and not increased H2O2 levels, were responsible for growth deficiencies, severe perturbations within the antioxidant machinery had apparently serious consequences for cell growth and/or division (Pignocchi et al., 2003). CAT2HP2 was also able to undergo spontaneous cell death in individual plants. Spontaneous lesion formation in catalase-deficient plants has already been reported in a mutant barley line with 5–10% residual catalase activity (Smith et al., 1984). Because of this spontaneous lesion formation and its differential growth capacities, the CAT2HP2 line was omitted from further analyses. To assess whether cat1 and cat3 transcript levels were affected in the transgenic plants, we performed a quantitative real-time PCR experiment. With gene-specific primers, we determined cat1, cat2, and cat3 transcript levels in PTHW and in both CAT2AS and CAT2HP1. No significant differences were observed in cat1 and cat3 transcript levels within both the catalase-deficient lines and the control, whereas cat2 transcript levels were clearly reduced in the catalase-deficient lines. However, because all cat genes could be involved in formation of peroxisomal CAT isoforms, we cannot exclude a decrease in general catalase activity levels because of a specific silencing of cat2. Therefore, we will refer to catalase-deficient instead of CAT2-deficient lines (Frugoli et al., 1996).

Several adverse environmental conditions, such as heavy metal stress, drought, temperature stress, and air pollution, provoke oxidative stress, and consequently affect growth and productivity of plants (Van Breusegem and Inzé, 2002). To assess oxidative stress sensitivity of catalase-deficient lines, the lines were exposed for 5 days to elevated ozone concentrations (350 nl l−1, 8 h per day). Cell death on the middle-aged leaves in a population of 100–150 plants per line and the percentage of damaged leaf surface on 10–15 plants per line were scored. Increased ozone sensitivity was inversely correlated with residual catalase activity (Figure 2). Columbia (Col-0) accessions are relatively resistant to ozone stress (Wohlgemuth et al., 2002), and this resistance was clearly confirmed in this experiment because rarely cell death was scored in the control lines. Increased ozone sensitivity also correlated with increased diaminobenzidine (DAB) staining, which is indicative of increased H2O2 levels. No differences were observed in the nitro blue tetrazolium (NBT) staining patterns, showing no differential superoxide accumulation between transgenic and control lines (data not shown). These results indicate that catalase, like other antioxidants, is a determining factor in the ozone-resistance phenotype of Arabidopsis accession Col-0. Whether differences in catalase activity levels are also responsible for the variable sensitivity of different Arabidopsis accessions remains to be investigated.


Figure 2. Cell death triggered by ozone stress in catalase-deficient plants.

(a) Cell death phenology after 3 days of ozone exposure.

(b) H2O2 accumulation after a 5 h ozone treatment (350 nl l−1). Middle-aged leaves were stained with DAB.

(c) Quantification of cell death after 3 days of ozone (350 nl l−1). Dark gray bars represent the percentage of plants with visually scored cell death events. Data are presented as average (n = 100–150) ± SE. Light gray bars indicate the percentage of the leaf area that was damaged. Data are presented as average (n = 10–15) ± SE. Control and catalase-deficient plants were exposed to ozone (350 nl l−1) for 5 days.

Download figure to PowerPoint

HL-induced H2O2 and cell death in catalase-deficient Arabidopsis thaliana

Despite their extensive antioxidative capacities, peroxisomes are capable of generating and releasing signal molecules, such as H2O2, into the cytosol under specific conditions (Corpas et al., 2001). H2O2 production is elevated during photorespiration. HL irradiation induces photorespiration and consequently strongly increases H2O2 levels in catalase-deficient plants (Chamnongpol et al., 1998; Dat et al., 2003). We assessed the response of the catalase-deficient Arabidopsis plants under HL irradiation conditions (1600 µmol m−2 sec−1), which correspond to a bright, summer sunny day at mid-northern latitudes. Six-week-old control and catalase-deficient plants were exposed to HL for 23 h. Within such a long-term exposure, leaves from control plants colored brown-purple, probably because of the accumulation of anthocyanins (data not shown). In catalase-deficient plants, this coloration was impaired and within 8 h of HL, cell death was visible. Both the timing and the extent of cell death were inversely correlated with the residual catalase activity in the different transgenic lines. Nomarski optics revealed that cell death occurred both in the mesophyll and palisade parenchyma layer, whereas trypan blue staining of leaves exposed for 24 h indicated its spatial distribution. Cell death was detected predominantly at the top of the leaf and in the vicinity of the veins (data not shown). Both catalase-deficient and control plants were also infiltrated with DAB and NBT after 5 and 23 h of HL. Whereas with NBT staining, no difference in superoxide accumulation was seen between control and catalase-deficient plants, DAB staining showed that cell death correlated spatially with increased H2O2 levels (Figure 3).


Figure 3. H2O2 accumulation and cell death triggered by HL irradiation in catalase-deficient plants.

(a) Control and catalase-deficient plants were exposed to HL (1600 µmol m−2 sec−1) for 23 h. Cell death was visible after 8 h of HL only on the leaves of catalase-deficient lines. Middle-aged leaves of control plants (PTHW) and two catalase-deficient lines (CAT2AS and CAT2HP1) are shown.

(b) H2O2 accumulation. Control and catalase-deficient lines were illuminated at 1600 µmol m−2 sec−1 for 5 h, and middle-aged leaves were stained with DAB.

Download figure to PowerPoint

To validate the photorespiratory nature of the elevated H2O2 levels in HL-treated catalase-deficient plants, plants were exposed to HL under ambient air conditions (400 p.p.m., 21% O2) and under high CO2 levels (1500 p.p.m., 21% O2). As expected, high CO2 levels clearly impaired the induction of cell death in the catalase-deficient plants exposed to HL (Figure 4).


Figure 4. Suppression of cell death in catalase-deficient plants exposed to HL combined with high CO2 levels.

(a) Quantification of cell death. Bars represent the mean injury of middle-aged leaves of five plants with visually scored cell death events in control plants (Col-0 WT and PTHW) and two catalase-deficient lines (CAT2AS and CAT2HP1). High CO2 levels significantly reduce leaf injury.

(b) Representative middle-aged leaves of control and catalase-deficient line plants, were exposed to HL (1600 µmol m−2 sec−1) for 23 h under ambient air conditions (400 p.p.m., 21% O2) and under high CO2 levels (1500 p.p.m., 21% O2).

Download figure to PowerPoint

Thus, as shown before in catalase-deficient tobacco, a severe perturbation of catalase, combined with HL irradiance, allows H2O2 to accumulate, followed by the advent of cell death. Hence, these Arabidopsis lines could be used as tools to study the signaling role of H2O2.

Transcriptome analysis of HL-treated control and catalase-deficient plants

Previously reported H2O2 transcriptome studies relied on the use of cell suspensions and exogenously applied H2O2 and transcriptional changes were mostly assessed at only one time point during the stress treatment. Pnueli et al. (2003) described the transcriptome in Arabidopsis plants deficient in cytosolic Apx. The catalase-deficient transgenic lines allow us to follow transcriptional changes during a sustained H2O2 stress over time. By using two different transgenic lines (CAT2AS and CAT2HP1, each containing 65 and 20% residual catalase activity, respectively), we were also able to assess a potential relative responsiveness of the plant transcriptome toward increasing H2O2 levels. Twenty individual plants of the CAT2AS, CAT2HP1, and control lines were exposed continuously to a 1600 µmol m−2 sec−1 HL treatment for 0, 3, 8, and 23 h. Catalase-deficient plants developed visible cell death after 8 h of HL, while in the control plants, no cell death was evident even after 23 h of HL.

For the microarray study, a reference design was used that consisted of 12 dye-swap experiments in which pools of test samples were compared with a reference sample on a total of 24 cDNA microarrays. These microarrays had duplicated sets of 6008 cDNA clones of Arabidopsis. The 12 test samples hybridized on the microarrays were the pooled middle-aged leaves of CAT2AS, CAT2HP1, and the control line harvested at each of the four time points, while the reference sample represents a pool of all lines harvested at 0 and 23 h. None of the 6008 genes were uniformly below the signal threshold (see Experimental procedures). To select genes that are significantly differentially expressed over time between the three lines, we assessed for each gene, the significance of the interaction between genotype and time (the effect of primary interest) beyond that of the main effects genotype and time, which are of secondary interest. Therefore, we applied two sequentially mixed model analyses of variance (anovas; Wolfinger et al., 2001) to the base-2 logarithm of all the ‘lowess’-transformed spot measurements with the residual maximum likelihood method (see Experimental procedures). As a measure of variability in expression levels among the effects, we used Wald statistics, which were tested against the χ2 distribution. The P-value cut-off was set arbitrarily at the stringent level of 0.001. No further adjustments for multiple testing were performed. A significant difference in expression between catalase-deficient and control plants over time was found for 1552 genes. All expression data are available according to minimum information about a microarray experiment guidelines at

Hierarchical and adaptive quality-based clustering reveal specific patterns of H2O2-regulated gene expression

For the 1552 genes selected as differentially expressed over time between the three lines, we used hierarchical average linkage clustering (HALC) and adaptive quality-based clustering (AQBC) to group genes with similar expression profiles (De Smet et al., 2002; Eisen et al., 1998). HALC of all differentially expressed genes resulted in three main clusters. Cluster A comprises all genes that are downregulated in the two catalase-deficient lines (Figure 5). They were further grouped based on their expression profile in the control line: rapidly downregulated (cluster A1); downregulated at 3 h, but upregulated again after 8 h (cluster A2); and maximally expressed after 8 h in the control line (cluster A3). Cluster B groups genes that were induced by HL stress in control plants, but whose induction was retarded or completely abolished in the catalase-deficient lines. Finally, the most predominant cluster C includes genes that were not or only transiently induced in the control line, but were sustainably upregulated in the two catalase-deficient lines during the HL treatment.


Figure 5. Hierarchical average linkage clustering of 1552 significantly differentially expressed genes between the catalase-deficient A. thaliana lines and the control line during the HL time course.

Each column represents the time point of sampling during HL treatment (0, 3, 8, and 23 h) in control (PTHW) and catalase-deficient plants (CAT2AS and CAT2HP1). Each row represents the expression profile of a transcript. Expression data were processed as described (see Experimental procedures). Three main clusters of expression (A, B, and C) are indicated on the right. Cluster 1 was divided into three subclusters (A1, A2, and A3).

Download figure to PowerPoint

AQBC of the differentially expressed genes resulted in 31 different clusters (required probability, 0.97; minimal number of genes in a cluster, 5) containing a total of 1248 genes (Figure 6; Clusters 1 and 4 included transcripts induced exclusively in the catalase-deficient lines and corresponded largely to cluster C of the HALC. Clusters 2 and 3 contain transcripts that are downregulated and/or expressed at a lower level from the start in the catalase-deficient lines (approximately cluster A of HALC). The induction of 64 genes belonging to cluster 5 was delayed or prevented in catalase-deficient lines (approximately cluster B of HALC). These five main clusters together contain 929 genes (75% of the genes that were included in an AQBC cluster with the above mentioned parameters).


Figure 6. Adaptive quality-based clustering.

The 1552 differentially expressed genes were clustered with the aqbc software (De Smet et al., 2002). For each cluster, the number of genes and the trend line, representing the average profile, is shown: dotted, control (PTHW); dashed, CAT2AS (65% catalase activity); and solid, CAT2HP1 (20% catalase activity). Each trend line connects, for each plant line, the four time points during the HL treatment at which the samples were collected. Five main clusters are shown; the remaining smaller clusters are available at

Download figure to PowerPoint

The perturbation of H2O2-scavenging capacities has a significant effect on nuclear gene expression when plants are exposed to HL irradiation. Therefore, we wanted to define specifically in which of the 1552 differentially expressed genes, HL irradiation played an important role in combination to their H2O2 responsiveness. Exposure of Arabidopsis to HL intensities for 1 h led to a differential expression of at least 185 genes (Rossel et al., 2002), and cyanobacterial acclimation to HL intensities was accompanied by changes in gene expression of at least 160 genes (Hihara et al., 2001). Here, screening for genes differentially expressed by HL irradiation only in the control plant revealed 1561 genes responsive to HL (see Of these 1561 genes, 910 genes were affected in their responsiveness to HL in catalase-deficient plants. Consequently, the expression of 642 genes was only modulated by H2O2 in the catalase-deficient background, suggesting that these genes responded uniquely to H2O2. This responsiveness correlated with increasing levels of photorespiratory H2O2 because kinetics and amplitude of the response were stronger in the CAT2HP1 line, containing only 20% of residual catalase activity, than in the CAT2AS line with 65% of residual catalase activity (Figure 5). Table 1 presents the top 30 significantly upregulated transcripts with a minimal 10-fold upregulation in the catalase-deficient lines during HL exposure. The complete list of differentially expressed genes is available at

Table 1.  Top 30 of significantly (P < 0.001) upregulated genes in catalase-deficient plants (CAT2HP1) compared with control plants (PTHW) from HALC cluster C ranked according to their maximal fold change observed during HL exposure
EST accession numberGene annotationAGI codeFold changeTime point (h)
  1. Fold change refers to the maximal induction in catalase-deficient plants relative to the control plants observed at the time point at which the maximal fold change induction was observed. Complete tables of microarray results of all lines can be found at

AI99890417.6-kDa class I small heat shock proteinAt1g5354085.98
AI992977Expressed proteinAt1g0534044.38
AI998968Heat shock protein 101At1g7431030.98
AW004348Dihydroflavonol 4-reductaseAt5g4280028.03
AI99840817.7-kDa class II heat shock protein 17.6 AAt5g1203027.78
AI993358Legume lectin family proteinAt3g1653025.98
AW004568Putative mannitol dehydrogenase ELI3-2At4g3799022.223
AI996814β-Fructosidase BFRUCT1/cell wall invertaseAt3g1379021.78
AI996408Arabinogalactan protein AGP2At2g2247020.78
AI996535Plastocyanin-like domain-containing proteinAt5g2023020.68
AI99553118.1-kDa class I heat shock proteinAt5g5972019.98
AI993740Expressed proteinAt2g4000019.88
AW004579Mutator-like transposaseAt1g3386017.68
AI993868Putative cytochrome P450At2g4557017.223
AI994052Putative heat shock protein 70At3g1258016.68
AI993534Heat shock protein 81-1At5g5264016.68
AI994585Expressed proteinAt4g3967016.38
AI997508Expressed proteinAt2g1869015.28
AI992420Putative cytochrome P450At2g3077015.023
AW004360Ribonuclease 1At2g0299015.023
AI996816Arabinogalactan protein AGP1At5g6431014.68
AI998304Trypsin and protease inhibitor family proteinAt1g7326013.68
AI994827Glycosyl hydrolase family 17 proteinAt4g1626013.38
AI992663Expressed proteinAt2g2786013.38
AI997082Anthranilate synthase, α subunit, component I-1 ASA1At5g0573013.023
AI996685Lateral organ boundaries domain protein 41At3g0255012.98
AI993083Hevein-like proteinAt3g0472011.68
AI997494Mitogen-activated protein kinase MPK17At2g0145011.18
AI998638Protease inhibitor/seed storage/lipid transfer protein LTP family proteinAt3g2260010.723
AI993906Class IV chitinaseAt3g5442010.58

Our results confirm the presence of both an H2O2-dependent and -independent HL-triggered regulatory pathway (Kimura et al., 2001). Interestingly, we can also note a modest transient induction or repression for most of the 642 strictly H2O2-modulated genes (clusters A2 and C; Figure 5) after 3 h of HL in control plants. This probably reflects responsiveness toward oxidative stress that rises transiently in control plants within the first hour of HL exposure and is able to affect nuclear gene expression. Because these genes are clearly either continuously up- or downregulated in catalase-deficient plants, we can assume that in the control plants also, this transient H2O2 responsiveness is provoked by photorespiratory H2O2, but that it is probably efficiently counteracted by the antioxidant system. High levels of photorespiratory H2O2 can explain the abolishment of an adaptive response in the two catalase-deficient lines. In addition, cell death events at 8 h of HL in catalase-deficient plants will be able to amplify the signal, thereby sustaining the transcriptional response in these lines even more. Although we cannot discriminate between a direct responsiveness toward elevated photorespiratory H2O2 and a more integrated response via additional H2O2-steered oxidative stress at other subcellular locations, such as chloroplasts (i.e. photoinhibition) or mitochondria, our results clearly consolidate earlier reports on a potent early signaling role of peroxisomal H2O2 (Corpas et al., 2001; Noctor et al., 2002).

Functional categories of genes with H2O2-driven transcriptional changes

A complete overview of the H2O2-responsive genes is available at Based on the automatically derived functional categories as described by the Munich Information Center for Protein Sequences (MIPS;; Schoof et al., 2002), the 1552 H2O2-responsive transcripts could be classified into 16 functional groups. A potential over-representation was assessed by comparing the number of genes of a specific functional category within a cluster with the total amount of genes present on the microarray classified within this category. These functional categories within the six largest AQBC clusters are presented in Table S1. In the AQBC cluster 1, genes that are involved in protein synthesis, protein fate, and energy are over-represented, whereas in AQBC cluster 4 (Figure 6), the group of cell rescue, defense, and virulence is dominant; AQBC clusters 5 and 6 contain mainly metabolic genes. Because these functional categories are automatically derived from the MIPS database, the use of this information is limited for the biological interpretation of genome-wide expression data. Therefore, we have additionally assessed specific functional categories, allowing us to analyze and interpret the microarray data into more detail.

H2O2 drives upregulation of the translational and proteolytic machinery

Together with an upregulation of several splicing-related components, the translational machinery is clearly transcriptionally upregulated because several cytosolic ribosomal proteins (60S components L30, L18A, and L6), and at least eight cytosolic translation initiation factors (EF1A, EF1Bγ, eIF3a, eIF3b, eIF3c, eIF3e, eIF5, eIF6, and eRF1) and a translation release factor are present in the AQBC cluster 1. Although active repression of protein synthesis protects cells against protein misfolding during oxidative stress, a long-term adaptation could involve an increase of proteins necessary for the translation of genes involved in this adaptation (Novoa et al., 2003). In yeast and plants, Rausell et al. (2003) have recently shown that the translational factor eIF1A is an important determinator in salt stress tolerance. Proteolysis is also clearly upregulated by H2O2. Four (poly)ubiquitin genes, together with a ubiquitin-conjugating enzyme-like protein, are upregulated (Bachmair et al., 2001). Within the AQBC cluster 1, four proteins contain an F-box motif, three a WD40 repeat, and six a RING-finger domain. Additionally, five different components of the 26S proteasome and three proteasome regulatory subunits are upregulated by H2O2. Whether H2O2-increased proteolysis within our system solely reflects the increased demand for removal of misfolded proteins or it plays a signaling role by targeted degradation of regulatory components within the defense response of programmed cell death remains to be elucidated (Estelle, 2001).

H2O2 reverses or mitigates HL-driven expression of antioxidant, hypersensitive response-related and anthocyanin biosynthetic genes

Whereas in the AQBC clusters 1 and 2, the expression of genes in catalase-deficient plants was mainly induced or repressed while their profile hardly changed in control plants, expression of genes in clusters 3–5 changed in the control line, where H2O2 exerted an additionally suppressive or inducing effect in the catalase-deficient lines. Cluster 3 contains genes that are induced in the control line and whose induction is delayed in the catalase-deficient lines. Surprisingly, four antioxidant genes belong to this cluster: one coding for a Cu/Zn superoxide dismutase-like protein and three for peroxidase. Cluster 4 groups genes that are downregulated by HL in the control plants, but are induced in the catalase-deficient plants by HL. Here, several genes are present that are classically linked with the hypersensitive response, including the oxidative burst (A. thaliana respiratory burst oxidase homolog (atrbohC); Torres et al., 2002). The exact function of AtrbohC is unknown, but because of its H2O2-triggered upregulation, it probably plays a role in the induction of cell death and defense response. Interestingly, a hevein-like protein, which is an allergenic lectin from rubber latex that is able to activate oxidative burst of human neutrophils (Rojas et al., 2001); four resistance (R) genes, of which two containing an NB-ARC domain (van der Biezen and Jones, 1998); and two β-1,3-glucanases are upregulated. Within the hypersensitive cell death response, Ca2+ fluxes are an important early signal (Blume et al., 2000). In cluster 4, an anion channel protein, a Ca2+-transporting ATPase, and a calmodulin domain protein kinase (Hrabak et al., 1996) are present. Four other genes contain an EF hand motif (Day et al., 2002). A concomitant downregulation of specific antioxidant enzymes and the upregulation of hypersensitive response-related genes can contribute to the observed cell death in catalase-deficient plants (Clarke et al., 2002; Fath et al., 2001; Király et al., 2002). Cluster 5 contains at least three genes involved mainly in phenylpropanoid biosynthesis and, in particular, in anthocyanin biosynthesis (phenylalanine ammonia lyase, chalcone synthase, chalcone isomerase, and dihydroflavonol 4-reductase). The induction of these genes is delayed and reduced in the catalase-deficient plants. Leaves of control plants colored mostly brown-purple during long-term HL exposure. This coloration was not observed in catalase-deficient plants (data not shown).

H2O2-dependent masterswitch controls part of the HL-induced gene expression

Within the experimental set-up, we followed nuclear gene expression changes provoked by two primary perturbations: exposure to HL irradiation and peroxisomal catalase deficiency. The combination of these perturbations led to a third primary perturbation in the transgenic lines: elevated photorespiratory H2O2, which allowed us to assess transcriptomal changes provoked by HL and H2O2 only. We were able to identify not only a signaling role for H2O2 within the HL response, but also the negative impinging of elevated levels of photorespiratory H2O2 at the HL-modulated gene expression. Our results suggest that a masterswitch transcriptionally controls the HL response of plants, as already described for the nuclear chloroplast transcriptome (Richly et al., 2003), and that the activity of this masterswitch can be reversed by elevated H2O2 levels originally produced in the peroxisomes.

H2O2 regulates nuclear-encoded chloroplastic and mitochondrial proteins

We addressed the potential interplay between different ROS-producing subcellular sites via the expression behavior of nuclear-encoded chloroplastic and mitochondrial proteins (Heazlewood et al., 2003; Kruft et al., 2001; Leister, 2003). The most abundant functional group of genes in the AQBC cluster 2 consists of nuclear-encoded genes involved in photosynthesis. In control plants, at least 17 different nuclear genes coding for photosystem I and II components drop to a basal expression level within 3 h of HL, while in the transgenic plants, this downregulation is delayed but finally more pronounced at 8 and 23 h of HL. This downregulation in control plants probably reflects an adaptation response to the sudden shift to HL intensities, thereby avoiding additional damage caused by the excessive energy influx (Allen and Pfannschmidt, 2000). A similar transcriptional regulation of photosynthetic activities during HL treatment was reported by Rossel et al. (2002). The modulation of photosynthetic components is a common reflex during stress, such as the combination of drought and heat shock (Rizhsky et al., 2002). Nuclear-encoded plastid proteins are controlled by a transcriptional switch that functions in a binary mode, either by inducing or by repressing the same large set of genes (Richly et al., 2003). The signals for this downregulation could originate from the chloroplasts themselves because chloroplasts not only provide energy but also are sensors for environmental information and, hence, are clearly involved in interorganellar signaling (Karpinski et al., 1999; Pfannschmidt, 2003; Pfannschmidt et al., 1999). Apparently, catalase perturbation initially impairs the downregulation in the transgenic lines. Possibly, these sustained high expression levels of photosynthetic components promote additional chloroplastic oxidative stress, fastening the induction of cell death, which, in turn, could release signals to decrease the expression of these genes. To obtain a more global view on the expression behavior of nuclear-encoded chloroplastic genes, we analyzed separately all the genes on the array that had been assigned to code for chloroplastic proteins (Leister, 2003). Of the 985 such genes, 23% present on the array had a differential expression pattern. The majority was downregulated or completely repressed and, hence, was present in the AQBC clusters 2 and 3.

An integrative role for mitochondria in oxidative stress signaling and, more specifically, cell death is not only consolidated in mammalian systems (Desagher and Martinou, 2000; Kaufmann and Hengartner, 2001), but is emerging in plants as well (Beers and McDowell, 2001; Dat et al., 2003; Dutilleul et al., 2003; Tiwari et al., 2002). Here, we show that an increase in mitochondrial metabolism is also reflected at the transcriptional level: cytochrome c, a cytochrome c-like protein, a mitochondrial elongation factor G, chaperonin of heat shock protein 60, two mitochondrial processing peptidases, an uncoupling protein BCS1, and several genes involved in the citrate cycle were upregulated in catalase-deficient plants during HL stress. Our results demonstrate that photorespiratory H2O2 influences the expression of both photosynthetic and mitochondrial genes, providing additional evidence for the importance of intraorganellar communication within the plant's defense response.

Signal transduction

Of the 340 transcription factors (Riechmann et al., 2000) present on our microarray, 85 were significantly differentially expressed in the catalase-deficient plants compared to control plants during the HL treatment. Most of them were affected within 3 h of HL, making them strong candidates to be part of the regulatory circuit that governs downstream H2O2 processes. As shown in Figure 7, these transcription factors are spread over the different clusters. Predominant categories are AP-2-related transcription factors (12), WRKY (8), Myb (9), and NAC (7) domain-containing factors and, interestingly, five CONSTANS-like transcription factors are also differentially expressed. Among the other transcription factors, we identified factors assigned to the bZIP (4), Zinc finger (4), bHLH (3), a MADS box, two NF-Y (CCA box binding), two SCARECROW (GRAS), two GARP, and one HSF4 family. The complexity of the H2O2-driven regulatory network is illustrated by the observation that transcription factors of the same family respond inversely to the same stress situations. AP-2/EREBP factors are not only up- and downregulated, but also present in the HALC cluster B, in which H2O2 mitigates the HL responsiveness. In addition to the transcription factors, potential upstream signal transducers are transcriptionally regulated. At least 36 different kinases, including LRR kinases, receptor-like protein kinases, wall-associated kinases, SNF1-like kinases, five members of the known MAPK cascades, calcium-dependent kinases, and several phosphatases are differentially regulated. Now that we have a comprehensive inventory of transcription factors and other signal transducers responsive to an initial H2O2 stress, a more detailed analysis will be necessary to clarify how these different factors interconnect with each other and how specific their regulatory role is within the management of the downstream response. Systematic perturbations of these factors and a more detailed kinetic study of their responsiveness toward H2O2 will be the initial steps to be taken to unravel the hierarchical structure of the genetic network that governs the H2O2 response in plants.


Figure 7. Hierarchical average linkage clustering.

HALC resulted in 127 differentially expressed genes involved in signal transduction. Transcription factors, kinases, and phosphatases were selected for HALC. Each line, representing a differentially expressed gene, contains the unique array spot number of the gene and its description as based on the MIPS database.

Download figure to PowerPoint

Regulatory elements in promoters

We used different approaches and algorithms to detect conserved motifs in the promoters of co-regulated genes. The clusters deduced from the combination of AQBC and HALC defined the potentially co-regulated genes. To reduce the unavoidable noise that might spoil the promoter analysis, we subclustered the larger clusters into smaller groups with gradually more stringent parameters. We analyzed 800 bp of sequence upstream of the ATG and screened for common motifs with the motifsampler, ymf, and the rsa tools (see Experimental procedures). Only one motif (GAAGnnGAwG) was found to be significantly over-represented in the promoters of one set of co-regulated genes, whose expression was repressed by H2O2 (AQBC cluster 2; Figure 6). This motif was not found in PlantCARE (Lescot et al., 2002) and, to our knowledge, has not been described before, although, the motif GAAGAA might be recognized as part of motifs known to be involved in stress responses in Brassica oleracea. Because of the combination of several perturbations and late sampling, the analyzed data set represents a complex network of regulatory circuits and the presence of secondary effects within each response is not unrealistic. This combined response resulted probably in the introduction of a considerable amount of noise in the data sets and might explain why no other regulatory elements were significantly over-represented within other sets of co-regulated genes. Because different classes of transcription factors are spread over the different clusters, we have to be cautious when drawing conclusions from the identified regulatory elements. Also, despite significant progress, the current algorithms that look for regulatory elements have still significant drawbacks to unambiguously predict regulatory elements within linear promoter fragments. The available software is able to identify potentially good candidate motifs, but currently these predictions can only be confirmed by extensive experiments, which is the subject of a separate investigation (Rombauts et al., 2003).

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

Production of transgenic Arabidopsis thaliana with decreased CAT2 levels

Sense and antisense cassettes were constructed by cloning the PCR-amplified (5′-GCGCCATGGTTCCTTACAAGTATCGTCCAG-3′; 5′-CGGGATCCTTAGATGCCTGGTCTCACGTTC-3′) open-reading frames of cat2 in both sense and antisense directions into the NcoI–BamHI sites of the vector pH35S, which resulted in pH35SCATS and pH35SCATAS (Hemerly et al., 1995). Subsequently, the EcoRI–XbaI fragment of the pH35SCAT(A)S, encompassing the CaMV 35S promoter cat2 and the 3′ nopaline synthase terminator, was cloned into the binary pBINPLUS vector containing the neomycin phosphotransferase II (NPTII) gene as a selectable marker (van Engelen et al., 1995). For the hairpin construct, a 515-bp PCR product of the 5′ coding sequence of CAT2 (5′-CGCGGATCCTTACAAGTATCG-3′; 5′-CGCGGATCCTCCAGTTCTCTTGGATGTGAG-3′) was inserted as a BamHI fragment into pH35SCATS and selected for inverted direction by PCR. A PstI fragment was ligated into the PstI-digested pthw142 vector, a binary vector containing the bialaphos resistance gene (BAR), which is a resistance gene for tranformant selection. Binary constructs were mobilized into Agrobacterium tumefaciens (pMP90) with the Escherichia coli helper strain HB101. A. thaliana (L) Heyhn. ecotype Col-0 plants were transformed by floral dip (Clough and Bent, 1998). For the hairpin construct, Col-4 plants were used. Residual catalase expression levels in T1 and T2 transformants were monitored by Northern analysis with the 3′ untranslated region of cat2 as a probe and by Western blot analysis with rabbit antiserum against a cotton catalase (Mullen et al., 1997).

Growth measurements, catalase activity measurements, Western blot analysis, and real-time PCR experiments

Seedlings were grown in vitro on Murashige and Skoog medium (2% sucrose) for 24 days under 16 h light/8 h dark conditions. Surface area of leaves 1 and 2 (8–10 plants per line) was determined with imagej software (publically available at 18, 20, 22, and 24 days after germination, knowing that leaves 1 and 2 are fully grown 20–22 days after germination (De Veylder et al., 2001).

FW and DW measurements were performed on 2-week-old plants. The plants were grown under a 16 h light/8 h dark photoperiod, with a relative humidity of 70% and temperature of 21°C. Fifteen individual plants of the CAT2AS, CAT2HP1, CAT2HP2, and control lines were weighed just after harvesting (FW) and after 48 h at 70°C (DW). Water content was calculated as follows: (FW − DW)/FW. Statistically significant differences were found in both FW and DW of the CAT2HP2 line, and in FW and nearly in DW of the CAT2HP1 line. No significant differences could be observed in the water content (see Table S2).

Leaf tissue was ground in extraction buffer (60 mm Tris (pH 6.9), 1 mm phenylmethylsulfonylfluoride, 10 mm DTT) on ice with 1 ml extraction buffer for 0.15 g tissue. The homogenate was centrifuged at 18 000 g for 15 min at 4°C. The supernatant was used for spectrophotometric catalase analysis and SDS–PAGE protein separation. Catalase activity assays were performed according to Clare et al. (1984). For Western blot analysis, 50 µg of leaf proteins were separated on a 12% SDS–PAGE gel and hybridized with a 1 : 1000 dilution of rabbit antiserum against cotton seed catalase (Mullen et al., 1997). One sense line (CAT2S), one antisense line (CAT2AS), and two hairpin construct-containing lines (CAT2HP1 and CAT2HP2) were withdrawn for further analysis.

For the quantification of cat1, cat2, and cat3 transcript levels by real-time PCR, plants were grown under the same conditions as above. Total RNA of 5-week-old plants was prepared with TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). For each sample, 1 µg of total RNA was reverse-transcribed with Superscript II RNaseH Reverse Transcriptase (Invitrogen). First-strand cDNA was generated with an oligo(dT)12–18 primer. One microliter of the 1 : 25 diluted first-strand cDNA was used as a template in a subsequent PCR. The transcripts were amplified with gene-specific primers: 5′-TGCTCCTCTTACGGTTTGGTTTC-3′ and 5′-ACCCTTCTTTAAGCGTTTCATTGTC-3′ for cat1; 5′-GCTGGCAAGCCGTCTGAAC-3′ and 5′-AGCACAGAAGATCCACATGATGAAG-3′ for cat2; 5′-GGCCAATCTCCATATAAGCTCAGT-3′ and 5′-GGATTTAACGACCAAGCGATGATAG-3′ for cat3, and 5′-AACTTGTGCTCATCTGCCATTAGG-3′ and 5′-TGATTCTGCGGAAACACCACTTTAG-3′ for actin (At3g60830).

Plant growth conditions and stress treatments

Unless mentioned otherwise, the plants were grown in exposure chambers, which had been specially designed for plant studies (Thiel et al., 1996). The light regime was 12 h/12 h (from 6:00 to 18:00) at 140 µmol m−2 sec−1, the climate was adjusted to relative humidity of 70% and 22°C day/18°C night temperatures. Six-week-old plants were exposed for 6 days to ozone (350 nl l−1) for 8 h (8:00–16:00) during the light period. The remainder of the time, ozone-free air was applied. For HL treatments, 6-week-old plants were transferred to a sun-simulator with identical growth conditions and exposed to continuous HL irradiation (photosynthetically active radiation 400–700 nm at approximately 1600 µmol m−2 sec−1). For the transcriptome analysis, the middle-aged leaves of five plants per line were pooled after 0, 3, 8, and 23 h of HL irradiation.

ROS stainings

Leaves or whole plants were infiltrated with 0.1% DAB (Sigma-Aldrich, St Louis, MO, USA) in 10 mm 2-(N-morpholino)ethanesulfonic acid (pH 6.5) for H2O2 staining. After a 45-min incubation, leaves were de-stained in ethanol. Leaves were infiltrated with 0.1% NBT (Sigma-Aldrich) in 50 mm potassium phosphate buffer (pH 6.4) for superoxide staining. After 10–30 min of incubation in the dark, leaves were de-stained in 95% ethanol.

Microarray preparation, target labeling, hybridizations, and data analysis

The Arabidopsis 6K microarray used in our study contained 6008 cDNA fragments from the Incyte Unigene collection (Arabidopsis Gem I, Incyte, Palo Alto, CA, USA) together with 520 positive and negative controls. For details, see For each sample, 5 µg total RNA was amplified and hybridized as described previously by Puskás et al. (2002). Arrays were scanned with a Generation III scanner (Amersham BioSciences, Little Chalfont, UK) at 532 and 635 nm wavelengths for the two fluorescence dyes Cy3 and Cy5 (Amersham Biosciences), respectively. Image analysis was performed with ArrayVision (Imaging Research, St. Catharines, Ont., Canada). Spot intensities were measured as artifact-removed total intensities without correction for background. Variation in gene expression was assessed by using a two-step procedure essentially as outlined by Wolfinger et al. (2001). In a first step, a linear normalization anova model accounts for experiment-wide systematic effects (i.e. array effects) that could bias inferences made on the data from the individual genes. The residuals from this model represent normalized values and serve as input data for the gene models. In a second step, the gene models were fit separately to the normalized data from each gene, allowing inferences to be made by using separate estimates of variability. Here, estimates of primary interest are genotype × time effects. So finally, as a measure of variability in expression levels among the effects, Wald statistics were calculated and significance was assigned to these effects for each gene (for detailed description of the statistical analysis, see Supplementary Material).

Promoter analysis

On the normalized data, both HALC and AQBC were used. To avoid noisy big clusters that would spoil the promoter analyses, each cluster was independently subclustered by using the same algorithms but with gradually more stringent parameters. The promoter analysis was performed on the 800 bp upstream of the start codon. As software ymf (Sinha and Tompa, 2000) and the rsa tools (van Helden, 2003) were used for word-counting methods and the motifsampler (Thijs et al., 2002) as Gibbs sampler. The default parameters from the web site were used for the rsa tools. For the ymf software, version 2.0 was used locally and a specific A. thaliana background model was built by using curated intergenic sequences bordered at both sides by experimentally verified genes. On each cluster, different sets of parameters were run that mainly varied in the size of the gap (0–2, 3–5, 6–8) and the length of the words (up to 15). With the top 100 words returned by ymf 2.0, the software ‘find explanators’ was used to extract the top five words. The motifsampler was run 100 times for each cluster and each combination of parameters. The background orders alternated between 2 and 3 while looking for three motifs of lengths 7–17 bp. To extract the relevant motifs from these 100 runs, MotifRanking was used to present the top five words. Results from all three methods were then compared, expecting motifs to be found by at least two methods. Alternatively, motifs with distinctively high scores were also considered.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

The authors thank Dr Paul Van Hummelen, Anouk Brackenier, Roel Sterken, Dr Dominique Vercammen, and Dr Philippe Alard for helpful contribution; Lucia Gössl and Brigitte van de Cotte for excellent expert technical assistance; Dr Richard Trelease for the cottonseed catalase antibody; and Dr Martine De Cock for help in preparing the manuscript. This work was supported by a grant from the Research Fund of the Ghent University (Geconcerteerde Onderzoeksacties no. 12051403). S.V. is indebted to the Instituut voor de aanmoediging van Innovatie door Wetenschap en Technologie in Vlaanderen for a predoctoral fellowship.

Supplementary Material

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

Statistical microarray data analysis

Table S1 Over- or under-represented functional categories of genes for the six largest AQBC clusters

Table S2 DW and FW measurements


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information
  • Allen, J.F. and Pfannschmidt, T. (2000) Balancing the two photosystems: photosynthetic electron transfer governs transcription of reaction centre genes in chloroplasts. Philos. Trans. R. Soc. Lond. B, 355, 13511359.
  • Bachmair, A., Novatchkova, M., Potuschak, T. and Eisenhaber, F. (2001) Ubiquitylation in plants: a post-genomic look at a post-translational modification. Trends Plant Sci. 6, 463470.
  • Beers, R.P. and McDowell, J.M. (2001) Regulation and execution of programmed cell death in response to pathogens, stress and developmental cues. Curr. Opin. Plant Biol. 4, 561567.
  • Van Der Biezen, E.A. and Jones, J.D.G. (1998) The NB-ARC domain: a novel signalling motif shared by plant resistance gene products and regulators of cell death in animals. Curr. Biol. 8, R226R227.
  • Blume, B., Nürnberger, T., Nass, N. and Scheel, D. (2000) Receptor-mediated increase in cytoplasmic free calcium required for activation of pathogen defense in parsley. Plant Cell, 12, 14251440.
  • Bowler, C. and Fluhr, R. (2000) The role of calcium and activated oxygen as signals for controlling cross-tolerance. Trends Plant Sci. 5, 241246.
  • Chamnongpol, S., Willekens, H., Moeder, W., Langebartels, C., Sandermann, H., Jr, Van Montagu, M., Inzé, D. and Van Camp, W. (1998) Defense activation and enhanced pathogen tolerance induced by H2O2 in transgenic plants. Proc. Natl. Acad. Sci. USA, 95, 58185823.
  • Clare, D.A., Duong, M.N., Darr, D., Archibald, F. and Fridovich, I. (1984) Effects of molecular oxygen on detection of superoxide radical with nitroblue tetrazolium and on activity stains for catalase. Anal. Biochem. 140, 532537.
  • Clarke, S.F., Guy, P.L., Burritt, D.J. and Jameson, P.E. (2002) Changes in the activities of antioxidant enzymes in response to virus infection and hormone treatment. Physiol. Plant. 114, 157164.
  • Clough, S.J. and Bent, A.F. (1998) Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, 735743.
  • Corpas, F.J., Barroso, J.B. and Del Río, L.A. (2001) Peroxisomes as a source of reactive oxygen species and nitric oxide signal molecules in plant cells. Trends Plant Sci. 6, 145150.
  • Dat, J., Vandenabeele, S., Vranová, E., Van Montagu, M., Inzé, D. and Van Breusegem, F. (2000) Dual action of the active oxygen species during plant stress responses. Cell. Mol. Life Sci. 57, 779795.
  • Dat, J.F., Pellinen, R., Beeckman, T., Van De Cotte, B., Langebartels, C., Kangasjärvi, J., Inzé, D. and Van Breusegem, F. (2003) Changes in hydrogen peroxide homeostasis trigger an active cell death process in tobacco. Plant J. 33, 621632.
  • Day, I.S., Reddy, V.S., Shad Ali, G. and Reddy, A.S.N. (2002) Analysis of EF-hand-containing proteins in Arabidopsis. Genome Biol. 3, 0056.10056.24.
  • De Smet, F., Mathys, J., Marchal, K., Thijs, G., De Moor, B. and Moreau, Y. (2002) Adaptive quality-based clustering of gene expression profiles. Bioinformatics, 18, 735746.
  • De Veylder, L., Beemster, G.T.S., Beeckman, T. and Inzé, D. (2001) CKS1At overexpression in Arabidopsis thaliana inhibits growth by reducing meristem size and inhibiting cell-cycle progression. Plant J. 25, 617626.
  • Desagher, S. and Martinou, J.-C. (2000) Mitochondria as the central control point of apoptosis. Trends Cell Biol. 10, 369377.
  • Desikan, R., Mackerness, S.A.-H., Hancock, J.T. and Neill, S.J. (2001) Regulation of the Arabidopsis transcriptome by oxidative stress. Plant Physiol. 127, 159172.
  • Dutilleul, C., Garmier, M., Noctor, G., Mathieu, C., Chétrit, P., Foyer, C.H. and De Paepe, R. (2003) Leaf mitochondria modulate whole cell redox homeostasis, set antioxidant capacity, and determine stress resistance through altered signaling and diurnal regulation. Plant Cell, 15, 12121226.
  • Eisen, M.B., Spellman, P.T., Brown, P.O. and Botstein, D. (1998) Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA, 95, 1486314868.
  • Van Engelen, F.A., Molthoff, J.W., Conner, A.J., Nap, J.-P., Pereira, A. and Stiekema, W.J. (1995) pBINPLUS: an improved plant transformation vector based on pBIN19. Transgenic Res. 4, 288290.
  • Estelle, M. (2001) Proteases and cellular regulation in plants. Curr. Opin. Plant Biol. 4, 254260.
  • Fath, A., Bethke, P.C. and Jones, R.L. (2001) Enzymes that scavenge reactive oxygen species are down-regulated prior to gibberellic acid-induced programmed cell death in barley aleurone. Plant Physiol. 126, 156166.
  • Frugoli, J.A., Zhong, H.H., Nuccio, M.L., McCourt, P., McPeek, M.A., Thomas, T.L. and McClung, C.R. (1996) Catalase is encoded by a multigene family in Arabidopsis thaliana (L.) Heynh. Plant Physiol. 112, 327336.
  • Heazlewood, J.L., Millar, A.H., Day, D.A. and Whelan, J. (2003) What makes a mitochondrion? Genome Biol. 4, 218.1218.4.
  • Van Helden, J. (2003) Regulatory sequence analysis tools. Nucleic Acids Res. 31, 35933596.
  • Hemerly, A., De Almeida Engler, J., Bergounioux, C., Van Montagu, M., Engler, G., Inzé, D. and Ferreira, P. (1995) Dominant negative mutants of the Cdc2 kinase uncouple cell division from iterative plant development. EMBO J. 14, 39253936.
  • Hihara, Y., Kamei, A., Kanehisa, M., Kaplan, A. and Ikeuchi, M. (2001) DNA microarray analysis of cyanobacterial gene expression during acclimation to high light. Plant Cell, 13, 793806.
  • Hrabak, E.M., Dickmann, L.J., Satterlee, J.S. and Sussman, M.R. (1996) Characterization of eight new members of the calmodulin-like domain protein kinase gene family from Arabidopsis thaliana. Plant Mol. Biol. 31, 405412.
  • Karpinski, S., Reynolds, H., Karpinska, B., Wingsle, G., Creissen, G. and Mullineaux, P. (1999) Systemic signaling and acclimation in response to excess excitation energy in Arabidopsis. Science, 284, 654657.
  • Karpinski, S., Gabrys, H., Mateo, A., Karpinska, B. and Mullineaux, P.M. (2003) Light perception in plant disease defence signalling. Curr. Opin. Plant Biol. 6, 390396.
  • Kaufmann, S.H. and Hengartner, M.O. (2001) Programmed cell death: alive and well in the new millennium. Trends Cell Biol. 11, 526534.
  • Kendall, A.C., Keys, A.J., Turner, J.C., Lea, P.J. and Miflin, B.J. (1983) The isolation and characterization of a catalase-deficient mutant of barley (Hordeum vulgare). Planta, 159, 505511.
  • Kimura, M., Yoshizumi, T., Manabe, K., Yamamoto, Y.Y. and Matsui, M. (2001) Arabidopsis transcriptional regulation by light stress via hydrogen peroxide-dependent and -independent pathways. Genes Cells, 6, 697617.
  • Király, Z., Barna, B., Kecskés, A. and Fodor, J. (2002) Down-regulation of antioxidative capacity in a transgenic tobacco which fails to develop acquired resistance to necrotization caused by TMV. Free Rad. Res. 36, 981991.
  • Kruft, V., Eubel, H., Jänsch, L., Werhahn, W. and Braun, H.-P. (2001) Proteomic approach to identify novel mitochondrial proteins in Arabidopsis. Plant Physiol. 127, 16941710.
  • Leister, D. (2003) Chloroplast research in the genomic age. Trends Genet. 19, 4756.
  • Lescot, M., Déhais, P., Thijs, G., Marchal, K., Moreau, Y., Van de Peer, Y., Rouzé, P. and Rombauts, S. (2002) PlantCare: a database of plant cis-acting regulatory elements and a portal to tools for in silico analysis of promoter sequences. Nucleic Acids Res. 30, 325327.
  • Levine, A., Tenhaken, R., Dixon, R. and Lamb, C. (1994) H2O2 from the oxidative burst orchestrates the plant hypersensitive disease resistance response. Cell, 79, 583593.
  • Michael, T.P. and McClung, C.R. (2002) Phase-specific circadian clock regulatory elements in Arabidopsis. Plant Physiol. 130, 627638.
  • Mullen, R.T., Lee, M.S. and Trelease, R.N. (1997) Identification of the peroxisomal targeting signal for cottonseed catalase. Plant J. 12, 313322.
  • Noctor, G., Veljovic-Jovanovic, S., Driscoll, S., Novitskaya, L. and Foyer, C.H. (2002) Drought and oxidative load in the leaves of C3 plants: a predominant role for photorespiration? Ann. Bot. 89, 841850.
  • Novoa, I., Zhang, Y., Zeng, H., Jungreis, R., Harding, H.P. and Ron, D. (2003) Stress-induced gene expression requires programmed recovery from translational repression. EMBO J. 22, 11801187.
  • Pastori, G.M. and Foyer, C.H. (2002) Common components, networks, and pathways of cross-tolerance to stress. The central role of ‘redox’ and abscisic acid-mediated controls. Plant Physiol. 129, 460468.
  • Pfannschmidt, T. (2003) Chloroplast redox signals: how photosynthesis controls its own genes. Trends Plant Sci. 8, 3341.
  • Pfannschmidt, T., Nilsson, A. and Allen, J.F. (1999) Photosynthetic control of chloroplast gene expression. Nature, 397, 625628.
  • Pignocchi, C., Fletcher, J.M., Wilkinson, J.E., Barnes, J.D. and Foyer, C.H. (2003) The function of ascorbate oxidase in tobacco. Plant Physiol. 132, 16311641.
  • Pnueli, L., Liang, H., Rozenberg, M. and Mittler, R. (2003) Growth suppression, altered stomatal responses, and augmented induction of heat shock proteins in cytosolic ascorbate peroxidase (Apx1)-deficient Arabidopsis plants. Plant J. 34, 187203.
  • Prasad, T.K., Anderson, M.D. and Stewart, C.R. (1994) Acclimation, hydrogen peroxide, and abscisic acid protect mitochondria against irreversible chilling injury in maize seedlings. Plant Physiol. 105, 619627.
  • Puskás, L.G., Zvara, Á., Hackler, L., Jr and Van Hummelen, P. (2002) RNA amplification results in reproducible microarray data with slight ratio bias. Biotechniques, 32, 13301340.
  • Rausell, A., Kanhonou, R., Yenush, L., Serrano, R. and Ros, R. (2003) The translation initiation factor eIF1A is an important determinant in the tolerance to NaCl stress in yeast and plants. Plant J. 34, 257267.
  • Richly, E., Dietzmann, A., Biehl, A., Kurth, J., Laloi, C., Apel, K., Salamini, F. and Leister, D. (2003) Covariations in the nuclear chloroplast transcriptome reveal a regulatory master-switch. EMBO Rep. 4, 491498.
  • Riechmann, J.L., Heard, J., Martin, G. et al. (2000) Arabidopsis transcription factors: genome-wide comparative analysis among eukaryotes. Science, 290, 21052110.
  • Rizhsky, L., Liang, H. and Mittler, R. (2002) The combined effect of drought stress and heat shock on gene expression in tobacco. Plant Physiol. 130, 11431151.
  • Rojas, E., Llinas, P., Rodríguez-Romero, A., Hernández, C., Linares, M., Zenteno, E. and Lascurain, R. (2001) Hevein, an allergenic lectin from rubber latex, activates human neutrophil's oxidative burst. Glycoconj. J. 18, 339345.
  • Rombauts, S., Florquin, K., Lescot, M., Marchal, K., Rouzé, P. and Van de Peer, Y. (2003) Computational approaches to identify promoters and cis-regulatory elements in plant genomes. Plant Physiol. 132, 11621176.
  • Rossel, J.B., Wilson, I.W. and Pogson, B.J. (2002) Global changes in gene expression in response to high light in Arabidopsis. Plant Physiol. 130, 11091120.
  • Schoof, H., Zaccaria, P., Gundlach, H., Lemcke, K., Rudd, S., Kolesov, G., Arnold, R., Mewes, H.W. and Mayer, K.F.X. (2002) MIPS Arabidopsis thaliana database (MAtDB): an integrated biological knowledge resource based on the first complete plant genome. Nucleic Acids Res. 30, 9193.
  • Sinha, S. and Tompa, M. (2000) A statistical method for finding transcription factor binding sites. In Proceedings of the Eight Intelligent Systems for Molecular Biology (Bourne, P., Gribskov, M., Altman, R., Jensen, N., Hope, D., Lengauer, T., Mitchell, J., Scheeff, E., Smith, C., Strande, S. and Weissig, H., eds). Menlo Park: AAAI Press, pp. 344354.
  • Smith, I.K., Kendall, A.C., Keys, A.J., Turner, J.C. and Lea, P.J. (1984) Increased levels of glutathione in a catalase-deficient mutant of barley (Hordeum vulgare L.). Plant Sci. Lett. 37, 2933.
  • Thiel, S., Döhring, T., Köfferlein, M., Kosak, A., Martin, R. and Seidlitz, H.K. (1996) A phytotron for plant stress research: how far can artificial lightning compare with natural sunlight? J. Plant Physiol. 148, 456463.
  • Thijs, G., Marchal, K., Lescot, M., Rombauts, S., De Moor, B., Rouzé, P. and Moreau, Y. (2002) A Gibbs sampling method to detect overrepresented motifs in the upstream regions of coexpressed genes. J. Comput. Biol. 9, 447464.
  • Tiwari, B.S., Belenghi, B. and Levine, A. (2002) Oxidative stress increased respiration and generation of reactive oxygen species, resulting in ATP depletion, opening of mitochondrial permeability transition, and programmed cell death. Plant Physiol. 128, 12711281.
  • Torres, M.A., Dangl, J.L. and Jones, J.D.G. (2002) Arabidopsis gp91phox homologues AtrbohD and AtrbohF are required for accumulation of reactive oxygen intermediates in the plant defense response. Proc. Natl. Acad. Sci. USA, 99, 517522.
  • Van Breusegem, F. and Inzé, D. (2002) Transgenic plants expressing tolerance towards oxidative stress. In Plant Biotechnology and Transgenic Plants, Books in Soils, Plants, and the Environment Series, Vol. 92 (Oksman-Caldentey, S. and Barz, W., eds). New York: Marcel Dekker, pp. 497516.
  • Vandenabeele, S., Van Der Kelen, K., Dat, J. et al. (2003) A comprehensive analysis of hydrogen peroxide-induced gene expression in tobacco. Proc. Natl. Acad. Sci. USA, 100, 1611316118.
  • Vranová, E., Atichartpongkul, S., Villarroel, R., Van Montagu, M., Inzé, D. and Van Camp, W. (2002) Comprehensive analysis of gene expression in Nicotiana tabacum leaves acclimated to oxidative stress. Proc. Natl. Acad. Sci. USA, 99, 1087010875.
  • Willekens, H., Chamnongpol, S., Davey, M., Schraudner, M., Langebartels, C., Van Montagu, M., Inzé, D. and Van Camp, W. (1997) Catalase is a sink for H2O2 and is indispensable for stress defence in C3 plants. EMBO J. 16, 48064816.
  • Wohlgemuth, H., Mittelstrass, K., Kschieschan, S., Bender, J., Weigel, H.-J., Overmyer, K., Kangasjärvi, J., Sandermann, H. and Langebartels, C. (2002) Activation of an oxidative burst in a general feature of sensitive plants exposed to the air pollutant ozone. Plant Cell Environ. 25, 717726.
  • Wolfinger, R.D., Gibson, G., Wolfinger, E.D., Bennett, L., Hamadeh, H., Bushel, P., Afshari, C. and Paules, R.S. (2001) Assessing gene significance from cDNA microarray expression data via mixed models. J. Comput. Biol. 8, 625637.
  • Zhong, H.H. and McClung, C.R. (1996) The circadian clock gates expression of two Arabidopsis catalase genes to distinct and opposite circadian phases. Mol. Gen. Genet. 251, 196203.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Supplementary Material
  8. References
  9. Supporting Information

Statistical microarray data analysis.

Table S1  Over or underrepresented functional categories of genes for the six largest AQBC clusters. The number of genes from a specific functional category within an AQBC cluster (CLU) was compared with that present on the microarray belonging to this functional category (Exp, expected). Functional groups over or underrepresented (?2 test statistically relevant) per cluster are underlined.

Table S2  Dry and fresh weight measurements.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

TPJ_2105_sm_Statisticanalysis.doc29KSupporting info item
TPJ_2105_sm_TableS1.doc14KSupporting info item
TPJ_2105_sm_TableS2.doc25KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.