The role of NADPH oxidase and MAP kinase phosphatase in UV-B-dependent gene expression in Arabidopsis


Åke Strid. Fax: +46 19 303566; e-mail:


Plant responses to supplementary UV-B irradiation have been reported to include formation of reactive oxygen species (ROS), hydrogen peroxide, in particular, and regulation by mitogen-activated protein kinase (MAPK) cascades which in turn are fine-tuned by MAPK phosphatases (MKPs). Here we present direct genetic evidence for the involvement of plasma membrane NADPH oxidase, a source of superoxide and hydrogen peroxide in the apoplasts, in UV-B signalling in Arabidopsis thaliana, by analysis of gene expression of the UV-B molecular markers in NADPH oxidase (atrbohD, F and DF) and MAP kinase phosphatase 1 (MKP1) knockout mutants (mkp1). Whereas the NADPH oxidase mutants were affected in UV-B-dependent CHS, PYROA and MEB5.2 gene expression, the mkp1 mutant was affected in the general expression pattern of the pathogenesis-related (PR) and PDF1.2 genes. The results indicate involvement of MKP1 in repressive action on gene expression of more general stress response pathways, similar to those activated by pathogen attack, while NADPH oxidase is involved in quantitative (rather than absolute) regulation of more UV-B-specific genes. The expressions of the molecular markers in the knockout mutant mkp1 and in its complemented lines (lines 6 and 10) were similar, as opposed to the responses of the corresponding wild-type Wassilewskija-4 (Ws-4). Lines 6 and 10 showed much higher MKP1 mRNA than Ws-4 but did not complement the mutant. This suggests a complex dependency of the MAPK phosporylation level of the PR and PDF1.2 genes. Both NADPH oxidase mutants and the mkp1 mutant phenotypically responded to UV-B by growth retardation.


Depletion of the stratospheric ozone layer results in increased levels of the sun’s ultraviolet-B (UV-B) radiation (280–315 nm) at the Earth’s surface. This influx of short-wave photons with high energy implies serious effects for all living organisms. For instance, UV-B can damage DNA, proteins and membrane lipids, and induce defence reactions. Plants, being sessile and photosynthetic organisms, are constantly exposed to this harmful radiation. Numerous UV-B effects have been reported, affecting plants on both the molecular and the ecosystem level (Caldwell et al. 1998). The damage by UV-B irradiation can be the result of both direct action on the cellular components and indirect action through the induction of formation of chemical adducts such as reactive oxygen species (ROS; Brosché & Strid 2003; Barta et al. 2004). To resist the damaging UV-B radiation, plants have developed mechanisms of protection and repair that have been extensively studied over the years, such as accumulation of UV-B-absorbing compounds (Harborne & Williams 2000) and repair of UV-B-induced DNA damage (Sancar 2003). At the molecular level, natural and relatively low doses of UV-B radiation result in significant changes in gene expression (Broschéet al. 2002; Casati & Walbot 2003; Ulm et al. 2004). The diversity of the metabolic and signalling pathways affected by UV-B is evidence for their overlapping and close interactions. ROS and mitogen-activated protein kinases (MAPKs) could act as convergence points in these networks.

Different kinds of ROS were detected in plant tissues as a result of UV-B stress (Hideg & Vass 1996; Barta et al. 2004). In the context of the present work, we examined the role of superoxide anion O2• − as a signalling molecule during UV-B stress. Superoxide anions are short-lived radicals that readily undergo either spontaneous or enzymatic dismutation yielding hydrogen peroxide (H2O2). Therefore, at least partially, the detectable H2O2 has O2• − as the primary source, and the effects observed as the result of H2O2 production can be due to O2• − dependent signalling.

The enzyme NADPH oxidase is one of the potential sources of H2O2 in plants. The involvement of H2O2 and NADPH oxidase in UV-B signalling was suggested by both pharmacological studies (A.-H.-Mackerness et al. 2001) as well as direct measurements of NADPH-oxidase activity (Rao, Paliyath & Ormrod 1996) and NADPH-oxidase mRNA transcript levels (Casati & Walbot 2003). Genetically controlled production of O2• − through NADPH oxidase induction has been implicated as a signal in a wide range of biotic and abiotic stress responses, as well as in the regulation of cell expansion and plant development (Mittler 2002; Foreman et al. 2003). An NADPH oxidase complex homologous to that of activated mammalian phagocytes was suggested to be a likely source of O2• − and H2O2, and the apoplastic oxidative burst in plants (Torres et al. 1998). Ten NADPH oxidase genes were identified in Arabidopsis, of which two (AtRBOHD and AtRBOHF) are expressed in leaves. Should ROS act as a second messenger in signal transduction pathways, it would be able to transfer the perceived information into a change in gene expression. This could partially be achieved by interaction with MAPK cascades.

MAPK signalling pathways were reported to be actively involved in transducing oxidative signalling (Kovtun et al. 2000). Several reports have demonstrated the activation of MAPKs by H2O2 (Desikan et al. 1999; Grant, Yun and Loake 2000) in Arabidopsis. H2O2 activates AtMPK3 and AtMPK6 via Arabidopsis NPK1-related protein kinase 1 (ANP1; Kovtun et al. 2000) and strongly induces expression of nucleotide diphosphate kinase NDPK2 gene in Arabidopsis (Moon et al. 2003). Involvement of MAPKs in UV-B signalling and the convergence of different signalling pathways at the level of MAPKs were demonstrated by Holley et al. (2003) in suspension cell culture of the wild tomato species Lycopersicon peruvianum. Two highly homologous MAPKs, LeMPK1 and LeMPK2, were found to be activated in response to different stresses, including UV-B radiation, while an additional MAPK, LeMPK3, was activated by UV-B radiation only. Irradiation of evacuolated parsley protoplasts with UV stimulated a change in the phosphorylation pattern within seconds (Harter et al. 1994). A pharmacological approach was used by Christie & Jenkins (1996) to demonstrate the involvement of protein phosphorylation in regulation of CHS and PAL gene expression after UV-B and UV-A/blue irradiation. Addition of serine/threonine protein kinase inhibitors to Arabidopsis cell suspension cultures resulted in decreased CHS gene expression. At the same time, it was shown that protein phosphatase inhibitors lead to the opposite result.

Phosphatases are responsible for dephosphorylation of components of MAPK cascades and regulation of the magnitude and the duration of their activation and thereby the signal. Another way of modulation of MAPK activity is by direct interaction of phosphatases with H2O2 (Wu, Kwon & Rhee 1998). Oxidation of receptor-directed protein tyrosine phosphatases by ROS is one of the mechanisms of UV-B signal transduction in animal cells (Gross et al. 1999).

The first direct evidence of involvement of MKPs in genotoxic stress relief was presented by Ulm et al. (2001). The conditional mutant mkp1 was hypersensitive to UV-C and methyl methansulphonate, and insensitive to other abiotic stresses, such as osmotic stress or ROS (paraquat and H2O2). The disrupted gene appeared to be an Arabidopsis homologue of mitogen-activated protein kinase phosphatase (AtMKP1). It is interesting to note that MKP1 interacts with MPK3, 4 and 6 as was shown by using a yeast two-hybrid assay (Ulm et al. 2002).

Sequencing of the Arabidopsis genome revealed the existence of 20 genes encoding MAPKs. The number of known MKPs is five at the most (Ulm et al. 2002). The disproportion in the numbers of MAPKs and MKPs suggests the involvement of one MKP in regulation of several MAPKs. In this way, MKPs presumably can serve as integration points for several different signals.

The evidence found in the literature prompted us to study genetically the involvement of MKP1 and NADPH oxidase in UV-B signal transduction. In the present paper, we examine whether the Arabidopsis thaliana homologues of the human respiratory burst oxidase, and a homologue of MAPK phosphatase MKP1, are involved in UV-B signalling. We used the atrbohF, atrbohD and atrbohDF mutants to address the role of NADPH oxidase and the mkp1 mutant and its complemented lines to address the role of the MKP1. We characterize the effects of UV-B on regulation of expression of UV-B molecular markers in these mutants and in the corresponding wild-type plants. In addition, we investigate the morphologic phenotypes and the oxidative status of the plants as the consequence of exposure to UV-B radiation.


Plant material and growth conditions

The atrbohD and atrbohF are homozygous knockout mutants carrying a single dSpm transposon insertion in the corresponding genes (At5g47910 and At1g64060). AtrbohDF is a double mutant obtained by crossing the atrbohD and atrbohF single mutants. The background ecotype is Columbia-0 (Col-0). The mutants were kindly donated by Prof. Jonathan Jones and described in detail by Torres et al. (1998).

The mkp1 mutant is a null mutant carrying a single copy of T-DNA in the AtMKP1 gene (At3g55270). The background ecotype is Wassilewskija-4 (Ws-4). Lines 6 and 10 are complemented lines obtained by transformation of the mkp1 mutant with the wild-type copy of AtMKP1 driven by its own promoter. The mkp1 mutant and its complemented lines were obtained from and described by Ulm et al. (2001). To exclude that gene silencing had occurred, a phenomenon that sometimes is observed in transgenes after several generations, we examined the expression of MKP1 in the complemented lines. Indeed, the complemented lines 6 and 10 showed much higher levels of MKP1 mRNA transcripts than wild-type Ws-4 (Fig. 1).

Figure 1.

Northern blot analysis of RNA isolated from 4-week-old plants (Ws-4, mkp1, line 6 and line 10). 0C denotes control conditions at 0 h; 6B denotes treatment with UV-A+B during 6 h (UV-BBE,300 = 0.18 W m−2). The blot was probed with radioactively labelled AtMKP1 cDNA and with 18S rRNA cDNA as a loading and transfer control. Ws, Wassilewskija.

For UV-B-exposure experiments, the investigated plants were grown on fertilized soil: Perlite: Vermiculite mixture (Weibull Trädgård AB, Hammenhög, Sweden). Seeds were maintained for 5 d at 4 °C (darkness). They were then transferred to a growth chamber with 16 h light (22 °C), 8 h dark (22 °C) cycles at 70% humidity. The plants were additionally grown for 3 weeks at 100 µmol photons m−2 s−1 photosynthetically active radiation.

UV-B irradiation

Three-week-old plants were transferred to UV chambers with the same environmental conditions and irradiated with either UV-A or UV-A+B radiation in addition to white light. UV-A+B was obtained from Philips TL40/12UV fluorescence tubes (Philips, Eindhoven, the Netherlands) by filtering through a cellulose acetate film with a cut-off at 292 nm. UV-A was obtained from the same tubes by filtering through a mylar film (Nordbergs Tekniska AB, Vallentuna, Sweden) with a cut-off at 315 nm. UV-irradiation was measured with an OL 754 UV-visible spectroradiometer (Optronic Laboratories, Orlando, FL, USA). In all experiments, plants were irradiated with UV-BBE,300 = 0.18 W m−2 where UV-BBE,300 was the biologically effective radiation, normalized to 300 nm, according to Caldwell (1971) and Green, Sawada & Shettle (1974). This corresponds to 1.24 µmol photons m−2 s−1. The control plants were exposed simultaneously to UV-A at the same level as in the UV-A+B experiments (0.38 W m−2; 1.09 µmol photons m−2 s−1). The whole rosettes were harvested after the indicated time of irradiation and quickly frozen in liquid nitrogen. One biological replicate contained a pool of six to seven rosettes.

RNA isolation and Northern blot analysis

RNA isolation was performed according to Strid, Chow & Anderson (1996). Samples containing 15 µg of total RNA were electrophoretically separated on a 1.2% agarose gel and transferred to a Hybond-N membrane (Amersham Biosciences, Uppsala, Sweden). Probes were labelled with 32P-dCTP using the random primers DNA labelling system (Invitrogen, Groningen, the Netherlands). Hybridization was performed in Church buffer (Church & Gilbert 1984). Membranes were sequentially washed 2 × 5 min with 2 × SSC/0.5% SDS and 4 × 5 min with 0.2 × SSC/0.1% SDS. Loading of RNA and Northern transfer was controlled by hybridization with a radiolabelled 18S ribosomal RNA cDNA probe (Kalbin et al. 1997) to the same membrane after stripping. The Arabidopsis Genome Initiative numbers for the molecular markers are CHS (At5g13930), PYROA (Atg5g01410), MEB5.2 (At3g17800), LHCB1*3 (At1g29930), PR-5 (At1g75040), PR-1 (At2g14610) and PDF1.2 (At5g44420).

cDNA preparation

In order to remove traces of genomic DNA from RNA preparations, on-column DNAase digestion was performed with RNAase-free DNAase (QIAGEN, Hilden, Germany) and RNeasy Mini Kit spin columns (QIAGEN) according to the manufacturer’s protocol. One microgram of total RNA in 20 µL was denatured (70 °C, 3 min) in the presence of 2 µL (2.5 pmol) of random oligonucleotide hexamers (Amersham Bioscienes). After chilling on ice, the following additions were made: 8 µL 5× RT buffer (Invitrogen), 2 µL 0.1 m dithiotreitol (DTT), 1 µL 25 mm dNTP, 1 µL RNAseOut (Invitrogen), 4.5 µL H2O and 1.5 µL SuperScript II reverse transcriptase (Invitrogen). The 40 µL reaction was incubated for 10 min at 25 °C, for 50 min at 42 °C and for 4 min at 94 °C.

Primer-probe design and real-time PCR

Quantification of gene-specific cDNA was performed by real-time PCR. Using the software tool Primer Express (Applied Biosystems, Foster City, CA, USA), primers and probes were selected for the use in a standardized TaqMan amplification protocol. All primers were commercially synthesized by Sigma–Genosys whereas all fluorogenic probes were commercially synthesized by Applera Sweden. The 5′-terminal reporter dye 6-carboxyfluorescein (FAM) and the 3′ quencher dye tetra-methylcarboxyrhodamine (TAMRA) were used. The sequences of the primers and probes used in this work are shown in Table 1. Amplification by real-time PCR was carried out with an ABI PRISM 7000 Sequence Detection System (Applied Biosystems).

Table 1.  The nucleotide sequences of the primers and probes used in the TaqMan real-time reverse transcriptase-PCR
Primer nameOrientationSequence (5′→3′)

The TaqMan PCR reactions were set up in 96-well reaction plates using PCR core reagents (Applied Biosystems). Samples of 3 µL of cDNA were amplified in 25 µL reaction volumes containing 1 µm of each primer, 250 nm FAM/TAMRA-labelled probe, 12.5 µL 2 × TaqMan Universal Master Mix (Applied Biosystems) and 7.75 µL H2O with the standard TaqMan temperature profile (2 min at 50 °C, 10 min at 95 °C and 40 cycles of 15 s at 95 °C, 1 min at 60 °C). Each PCR reaction was performed in triplicate and the absolute copy number of individual transcripts was analysed. The mRNA levels were normalized with respect to the level of 18S ribosomal RNA which was measured using eukaryotic 18S Assay-on-Demand (Applied Biosystems).

Histochemical staining for H2O2

H2O2 was analysed using 3,3′-diaminobenzidine (DAB) according to Schraudner et al. (1998). Arabidopsis rosettes were vacuum-infiltrated in 30 mL staining solution (0.1% (w/v) DAB, 10 mm 2-(N-morpholino)ethanesulphonic acid, pH 6.5) for 10 min in a dessicator. Irradiation of infiltrated plants was performed for 1 h. Rosettes were destained by boiling for 3 min in alcoholic lactophenol solution (ethanol : lactic acid : phenol : water was in the ratio 6 : 1 : 1 : 1) and were stored in 50% (v/v) ethanol. DAB staining was visualized as a brown precipitate. Staining for H2O2 was performed after 11 h of UV irradiation (UV-BBE,300 = 0.18 W m−2; UV-A fluence rate was 0.38 W m−2). Quantitation of the staining was performed by using the SigmaGel software (SPSS, Chicago, IL, USA).

Phenotype investigations

Three-week-old plants were grown as described earlier and exposed for 6 h (centred around midday each day) to either UV-A (control) or UV-A+B radiation. The exposure was performed for 7 d. The level of UV-BBE,300 was 0.12 W m−2 and the UV-A fluence rate was 0.26 W m−2. After the plants had been UV-B-exposed, they were left for further growth in climate chambers at standard conditions (without supplementary UV radiation). One week later, the plants were examined for visible alterations. The fresh and dry weights, leaf area and rosette size were investigated.


Molecular marker expression in NADPH oxidase mutants

The expression of the following molecular markers was studied: CHS (encoding the enzyme catalysing the first committed step of the flavonoid biosynthesis pathway; Fuglevand, Jackson & Jenkins 1996), MEB5.2 (encoding a protein with unknown function but being most strongly up-regulated by UV-B and other stresses; Broschéet al. 2002), PYROA[encoding an enzyme involved in formation of pyridoxine (vitamin B6) and participating in inactivation of ROS and protection of cells from oxidative damage; Osmani, May & Osmani 1999], PR-1, PR-5, PDF1.2[encoding pathogenesis-related (PR) proteins and defencin, respectively, the expression of which increases in response to UV-B exposure; A.-H.-Mackerness et al. 2001] and LHCB1*3 (corresponding to the chlorophyll a/b binding protein of the photosystem II light-harvesting antenna complex; Jordan et al. 1991; Jackowski, Kacprzak & Jansson 2001).

According to Brosché & Strid (2003), the genes used as molecular markers in this work belong to different groups depending on the sensitivity of their regulation by UV-B radiation. All up-regulated genes, except PR genes and PDF1.2, are induced by low or very low levels of UV-B, whereas PR-5, for instance, is induced by comparatively higher level of UV-B. The transcripts of CHS, PYROA and MEB5.2 were measured after 0, 0.5 and 3 h of irradiation, whereas PR-5 and LHCB1*3 transcript abundance was studied after 0, 3 and 11 h of irradiation.

Figure 2a–e shows the expression of the different molecular markers in the Col-0 wild type and three different atrboh mutants (atrbohD, atrbohF and atrbohDF) as studied by real-time reverse transcriptase RT-PCR. At short exposures to UV-B (0.5 h), no or very little CHS expression was seen and no difference between mutants was noticeable (Fig. 2a). After 3 h of UV-B exposure, the transcript levels for CHS were the same both in the wild type and in the two different single mutants. On the other hand, the CHS mRNA levels in the double mutant atrbohDF were about half of that of the wild type and single mutants.

Figure 2.

mRNA levels of (a) CHS, (b) PYROA, (c) MEB5.2, (d) PR-5 and (e) LHCB1*3 were quantified by real-time reverse transcriptase -PCR. The relative amounts of transcripts were normalized with respect to 18S rRNA. A white column denotes Columbia-0 (Col-0) plants exposed to ultraviolet-A (UV-A); a black column denotes Col-0 plants exposed to UV-A+B; a red striped column denotes atrbohD plants exposed to UV-A; a red closed column denotes atrbohD plants exposed to UV-A+B; a blue striped column denotes atrbohF plants exposed to UV-A; a blue closed column denotes atrbohF plants exposed to UV-A+B; a green striped column denotes atrbohDF plants exposed to UV-A; a green full column denotes atrbohDF plants exposed to UV-A+B. The values are the means of three independent biological replicates. The error bars show the SD of the means. a–c: UV-B exposure was carried out for 0, 0.5 and 3 h; d–e: UV-B exposure was carried out for 0, 3 and 11 h.

For PYROA, induction of expression by UV-B was not clearly evident until the 3 h time point, where the mRNA levels in all three mutants were about half of that of the wild type, indicating the quantitative involvement of NADPH oxidase in UV-B-dependent up-regulation of PYROA.

An expression pattern similar to that for CHS was observed for the MEB5.2 gene. Col-0, atrbohD and atrhohF showed a similar increase of the MEB5.2 transcripts after 3 h of UV-B exposure, whereas the double mutant atrbohDF had half the MEB5.2 message levels of Col-0 and the single mutants.

No consistent differences in expression of PR-5 and LHCB1*3 were observed between the three different atrboh mutants and the wild type. The results suggest that production of ROS by NADPH oxidase does quantitatively regulate CHS, PYROA and MEB5.2 during UV-B stress but not LHCB1*3 or PR-5.

We were unable to detect any significant levels of PR-1 and PDF1.2 transcripts, neither in wild type, nor in the atrboh mutants at the chosen fluence rates of UV.

Molecular marker expression in the mkp1 mutant

The expression levels of the molecular markers in the mkp1 Arabidopsis mutant and two complemented overexpressing lines (6 and 10) compared with the background wild-type strain Ws-4, were determined by Northern blot analysis. The experiments were repeated three times obtaining the same results, and representative autoradiographs are shown in Fig. 3. It is important to note that complementation of the mkp1 mutant with the wild-type gene resulted in significantly higher levels of MKP1 mRNA compared with Ws-4 (Fig. 1 and Ulm et al. 2001). For this reason we call lines 6 and 10 complemented ‘overexpressing’ lines. We observed no significant differences in the expression pattern of the CHS, LHCB1*3, PYROA and MEB5.2 genes between Ws-4, mkp1 and its complemented overexpressing lines although a slight reduction in gene expression of MEB5.2 and PYROA was noticeable in the complemented lines 6 and 10 as compared with Ws-4 and the mkp1 mutant.

Figure 3.

Autoradiographs of LHCB1*3, MEB5.2, PYROA, CHS, PDF1.2, PR-5, PR-1 and 18S rRNA transcript levels in Ws-4, mkp1, lines 6 and 10. Northern blots were hybridized with [32P]-labelled cDNA probes. Hybridization with 18S rRNA was employed as a control for even loading of nucleic acids. A: Control plants exposed to ultraviolet-A (UV-A) at the same level of UV-A as the UV-A+B exposed plants (0.38 W m−2); B: Plants exposed to UV-A+B at UV-BBE,300 = 0.18 W m−2. The exposure was carried out for 1, 3 or 6 h in all cases. Ws, Wassilewskija.

Instead, the most significant differences were found in the expression of the PR-1, PR-5 and PDF1.2 genes. No or extremely low levels of mRNA were found in Ws-4 for these genes, whereas mkp1 and its complemented overexpressing lines 6 and 10 displayed constitutive expression of these genes independently of UV-B irradiation. The overexpressing lines 6 and 10 therefore did not complement the mkp1 mutant with respect to the regulation of PR-1, PR-5 and PDF1.2 expression. Instead, these results suggest a rather unusual involvement of MKP1 in transcriptional regulation of these genes (see the Discussion section).

H2O2 accumulation

The oxidative status of experimental plants was monitored by DAB staining for H2O2. Plants exposed to UV-A at the same level as in the UV-A+B experiments were used as controls. The staining was performed after 11 h of exposure (UV-BBE,300 = 0.18 W m−2). The mutant mkp1, and lines 6 and 11 showed no differences in the steady-state concentration of H2O2 compared with the Ws-4 wild type (data not shown). This indicates that MKP1 is not involved in regulation of H2O2 levels in Arabidopsis at low fluence rates of UV-B.

No significant differences were observed in H2O2 steady-state concentrations between Col-0 and the atrboh mutants under UV-A exposure (Fig. 4a). However, UV-A+B exposure led to an increase in H2O2 concentration in Col-0 whereas the oxidative status of the atrboh mutants remained unchanged (Table 2). These results indicate the contribution of plasma membrane-bound NADPH oxidase in the regulation of the H2O2 pool in plants. It is also important to note that the UV-B fluence rate used (0.18 W m−2) was not high enough to cause any significant non-specific changes in H2O2 accumulation as a result of UV-B exposure in any of the Arabidopsis lines, which in turn would have led to general oxidative damage.

Figure 4.

(a) Accumulation of hydrogen peroxide in atrboh mutants and the Columbia-0 (Col-0) wild type. Plants were either exposed to ultraviolet UV-BBE,300 = 0.18 W m−2 supplemented with UV-A radiation (UV-A+B) for 11 h or to UV-A = 0.38 W m−2 alone (UV-A) for 11 h. Control plants did not receive any UV irradiation. The details of the experiment are as described in the Materials and Methods; (b) Growth retardation after UV-B irradiation of mkp1 mutant compared with the Ws-4 wild type and lines 6 and 10 (overexpressing lines complemented by ectopic expression of MKP1). Five-day-old seedlings were exposed to UV-A+B (0.07 W m−2) or UV-A (0.15 W m−2) for 14 d. Ws, Wassilewskija.

Table 2.  H2O2 accumulation in Columbia-0 (Col-0) and atrboh mutants after 11 h of ultraviolet-A+B (UV-A+B) irradiation (UV-BBE,300 = 0.18 W m−2) and UV-A irradiation (0.38 W m−2) compared with untreated plants (control)
Wild type/mutantMean (SD)UV-A – controlUV-A+B – UV-A
ControlUV-AUV-A+BMean95% limitsMean95% limits
  1. The values are presented as arbitrary units of the intensity of DAB-stained leaves (n = 16–23). Quantitation of the intensity of staining was performed using the SigmaGel software (SPSS, Chicago, IL, USA). DAB, 3,3 ′-diaminobenzidine.

Col-05448 (426)5728 (468)6750 (574) 280−52 to 6121022 644 to 1400
atrbohD5820 (420)5614 (563)5537 (481)−206−562 to 150 −77−462 to 308
atrbohF5586 (478)5614 (586)5490 (674)  28−353 to 409−124−559 to 308
atrbohDF5612 (676)5550 (523)5563 (628) −62−487 to 363  13−390 to 416

UV-B-dependent phenotypes of the atrboh and mkp1 mutants

The UV-BBE,300 levels used in the phenotype experiments were 2.6 kJ m−2 per day. For comparison, the daily exposure in Lund, Sweden, at midsummer is 4.8 kJ m−2 (Yu & Björn 1997). After 7 d of UV-B exposure (see Materials and Methods) and additional 7 d of unexposed growth, we observed no statistically significant changes in fresh and dry weight of above-ground tissue, leaf area and diameter of rosettes in the wild-type strain Col-0. The same treatment resulted in decrease of all investigated traits in the atrboh mutants (Table 3). No significant differences between atrbohDF and the corresponding single mutants in the UV-B-induced decrease of fresh and dry vegetative mass and diameter of rosettes were found (approximately 25% decrease compared with Col-0 in all three cases). The UV-B effect on the leaf area, on the other hand, was stronger for the double mutant (approximately 50% decrease compared with Col-0) than for the single mutants (33% decrease compared with Col-0).

Table 3.  The effect of ultraviolet-B (UV-B) radiation on morphologic traits of three atrboh mutants and the background ecotype Columbia-0 (Col-0)
Morphologic traitWild type/mutantMean (SD)Ratio (UV-A+B)/(UV-A)
UV-AUV-A+BMean95% limits
  • a

    Each of these observations was an average using three plants.

Fresh vegetative mass, mg n = 30aCol-0202 (15)197 (23)0.970.91–1.05
atrbohDF110 (10) 87 (12)0.750.67–0.85
atrbohD194 (15)141 (11)0.740.68–0.81
atrbohF127 (11)104 (7)0.830.77–0.90
Dry vegetative mass, mg n = 30aCol-0 14 (2) 13 (2)0.960.86–1.08
atrbohDF  9 (1)  7 (1)0.780.71–0.85
atrbohD 14 (2) 11 (1)0.780.72–0.85
atrbohF 10 (1)  8 (1)0.790.74–0.85
Leaf area, mm2n = 25Col-0166 (23)160 (19)0.980.94–1.01
atrbohDF101 (11) 55 (13)0.520.49–0.56
atrbohD177 (24)121 (18)0.670.61–0.75
atrbohF119 (12) 81 (12)0.660.61–0.72
Diameter of rosette, mm n = 25Col-0 68 (5) 66 (5)0.950.91–1.00
atrbohDF 48 (5) 37 (6)0.760.70–0.83
atrbohD 72 (5) 51 (5)0.710.67–0.75
atrbohF 58 (4) 47 (3)0.810.78–0.85

The same morphologic traits after the same treatments were investigated with Ws-4, mkp1 and the overexpressing complemented line 10. Neither line 10 nor Ws-4 showed any statistically significant changes in any of the investigated traits as the result of UV-B treatment (Table 4). However, statistically significant decreases in leaf area (by approximately 30% compared with Ws-4) and diameter of the rosettes (by approximately 10%) were detected in the mkp1 mutant when compared with Ws-4 and line 10.

Table 4.  The effect of ultraviolet-B (UV-B) radiation on morphologic traits of the mkp1 mutant, the complemented line 10 and the background ecotype Wassilewskija-4 (Ws-4)
Morphologic traitWild type/mutantMean (SD)Ratio (UV-A)/(UV-A+B)
UV-AUV-A+BMean95% limits
  • a

    Each of these observations was an average using three plants.

Fresh vegetative mass, mg n = 30aWs-4 86 (12) 83 (14)0.940.78–1.11
mkp1 71 (17) 65 (13)0.920.81–1.04
Line 10 87 (10) 90 (15)0.960.81–1.12
Dry vegetative mass, mg n = 30aWs-4  7 (1)  7 (2)0.940.83–1.08
mkp1  7 (2)  6 (2)0.800.68–1.02
Line 10  9 (2)  9 (2)0.950.81–1.14
Leaf area, mm2n = 25Ws-4102 (11) 98 (14)0.950.88–1.04
mkp1113 (12) 86 (17)0.670.61–0.74
Line 10102 (18)103 (23)0.970.84–1.12
Diameter of rosette, mm n = 25Ws-4 45 (3) 43 (3)0.980.93–1.02
mkp1 42 (4) 37 (4)0.870.83–0.91
Line 10 45 (5) 45 (4)1.000.93–1.08

The 14-day exposure of 5-day-old mkp1 seedlings to 0.07 W m−2 of UV-BBE,300 revealed not only significant growth retardation but also some bleaching as well (Fig. 4b), which we did not observe in mature plants. This suggests different functional roles of MKP1 at different stages of plant development.


Molecular marker expression in NADPH oxidase mutants

ROS, in particular superoxide and H2O2, are considered to be important signalling molecules in both animal and plant cells. The involvement of ROS in UV-B signalling has been studied using pharmacological approaches (A.-H.-Mackerness et al. 2001). It was argued that O2• − and H2O2 act as second messengers in up-regulation of PR and defensin (PDF1.2) genes and down-regulation of photosynthetic genes (LHCB). The sources for the ROS involved in regulation of different molecular markers were not the same, according to the same study. PDF1.2 expression was probably regulated through the activity of peroxidases and involve O2• − directly, whereas regulation of PR-1 and LHCB transcription was suggested to be mediated by ROS of clearly distinct origin, namely H2O2 derived from O2• −.

Plasma membrane NADPH oxidase is an important source of O2• − and H2O2. Knockout experiments demonstrated that the AtRBOHF and AtRBOHD genes (encoding two out of 10 NADPH oxidase genes) were required for H2O2 generation during bacterial and pathogen challenge in Arabidopsis (Torres, Dangl & Jones 2002). Using a reverse genetics approach and a number of molecular markers (CHS, PYROA and MEB5.2 in addition to those mentioned earlier), we here investigated the involvement of NADPH oxidase in ROS-mediated UV-B signalling. Compared with the high UV-B levels used by A.-H.-Mackerness et al. (2001), the levels of UV irradiation used in our experiments were comparable with ambient levels. At such low fluence rates of UV irradiation, we did not observe any significant expression of PDF1.2 and PR-1 genes, either in the atrboh mutants or wild type.

As can be seen from Fig. 2d and e, there are no significant differences in regulation of the LHCB1*3 and PR-5 genes by UV-B between Col-0, and atrbohD and atrbohF mutants. This means that low-level UV-B regulation of these genes is not accomplished through NADPH oxidase. Either this regulation is totally ROS-independent, or another distinct pool of ROS not produced by NADPH oxidase participates in the regulation. Compared with Surplus et al. (1998) and A.-H.-Mackerness et al. (2001), the conclusions at first seem contrasting. However, our view that LHCB1*3 gene expression is not influenced by specific H2O2 production through NADPH oxidase is based on short-term low-level experiments, whereas the data of A.-H.-Mackerness et al. (2001) were produced in experiments carried out for extended periods of time (4–16-fold longer than ours) at a 2.6-fold higher UV-B level. The long exposure time at the higher level most likely resulted in ROS non-specifically produced through interaction of the UV-B with cellular oxygen-containing constituents, which in turn, as a secondary effect, depressed mRNA levels for LHCB. For a discussion on specific and non-specific regulation of genes during UV-B exposure, compare with Brosché & Strid (2003).

The levels of CHS mRNA transcripts in Col-0 and both atrboh single mutants were the same after 3 h of UV-B exposure, whereas the double mutant atrbohDF showed a CHS expression half of that in Col-0 and the single mutants (Fig. 2a). Our results suggest a quantitative role of ROS produced by NADPH oxidase in the UV-B regulation of CHS, fine-tuning the expression level of this gene. However, qualitatively, CHS expression can be induced in the absence of functional NADPH oxidase. Jenkins et al. (2001) have previously argued that ROS are not involved in regulation of CHS gene expression in Arabidopsis. The basis for this is primarily that addition of H2O2 itself to Arabidopsis cell cultures or the inhibition of catalase during normal metabolism in these cultures with the inhibitor 3-amino-1,2,4-triazole did not induce CHS expression. These data do not necessarily contradict our results because we have shown that the existence of a functional NADPH oxidase is not a prerequisite for CHS expression but instead enhances the accumulation of CHS mRNA, that is, H2O2 may not be the trigger of UV-B-induced CHS transcription but functions to amplify the expression. In addition, the ROS scavengers pyrrolidine carbamate and N-acetylcysteine, added to the Arabidopsis cell cultures, did not alter UV-dependent CHS expression (Jenkins et al. 2001), which would be anticipated keeping our data in mind. There are a number of plausible explanations for this discrepancy. One evident reason could be that the uptake of these pharmacological substances was small in the cell cultures, or that they did not scavenge the particular H2O2 pool that is responsible for UV-signalling. However, the most evident explanation would instead be that the difference in tissue differentiation state, cell cultures (Jenkins et al. 2001) and leaves of plants (our study), respectively, influenced the regulation of the CHS gene. Indeed, CHS regulation is known to be strongly influenced by, for instance, the developmental stage of an Arabidopsis plant. Phytochrome induces CHS induction in very young seedlings only, whereas UV-B and blue light regulation occurs also in mature plants (Jenkins et al. 2001). Therefore, we believe that using whole Arabidopsis mutant plants in our studies mimics a situation that is more closely resembling the physiological state than using cell cultures.

For the MEB5.2 gene, the expression pattern was very similar to the one for CHS (Fig. 2c), and the same points of discussion also hold for this gene.

The expression of the PYROA gene decreased in all three mutants (single mutants atrbohD, atrbohF and double mutant atrbohDF) compared with Col-0 (Fig. 2b). The expression pattern of this gene, as the result of UV-B exposure, clearly indicates the involvement of NADPH oxidase-dependent regulation of the levels of PYROA expression, although, again, substantial PYROA expression is found also in the absence of functional NADPH oxidase. However, for regulation of this molecular marker, the redundancy of the two mutant AtRBOH genes was not as clear as for the CHS and MEB5.2 markers.

Molecular marker expression in the mkp1 mutant

It has been shown that MAPK cascades can be activated by both UV-B irradiation and H2O2 (Harter et al. 1994; Christie & Jenkins 1996; Kovtun et al. 2000). Because H2O2 generation occurs in response to a number of different stresses and because certain defence responses to such stresses are shared, MAPK cascades could serve as a branching point in a complex signalling network. Given the importance of MAPK signalling cascades, it is clear that their activity must be tightly regulated. Evidence for the involvement of MKP1 in the integration and fine-tuning of plant responses to osmotic and genotoxic stresses was obtained by Ulm et al. (2002). In the present study, we tested whether MKP1 is also involved in regulation of UV-B-dependent gene expression of molecular markers.

The responses of different molecular markers to supplementary UV-B in the mkp1 mutant, in the Ws-4 wild type, and in lines 6 and 10 (mkp1 complemented by ectopic expression of MKP1) are presented in Fig. 3. Under our experimental conditions, we did not observe any differences between these lines with respect to the expression patterns for LHCB1*3 and CHS. These results show that regulation of CHS and LHCB1*3 is independent of MKP1 at low UV-B levels.

A slight decrease in PYROA and MEB5.2 mRNA transcript levels in the complemented lines compared with wild type and knockout mutant was observed. However, the most significant differences were found in expression of PDF1.2, PR-1 and PR-5. We observed constitutive expression of these markers in mkp1, lines 6 and 10, but no expression in the wild-type strain. The expression was not affected by UV-B at the low fluence rates used in this study.

The constitutive expression of PR-1, PR-5 and PDF1.2 genes in mkp1 suggests the involvement of MKP1 in repression of the expression of these genes and the release of transcription at higher phosphorylation levels of MAPKs. Surprisingly, we observed the same expression pattern of these genes in both the knockout mutant and the complemented lines overexpressing AtMKP1, that is, both knockout and overexpression of MKP1 resulted in constitutive activation of PR and PDF genes. The observed action of the MKP1 is to some extent similar to a ‘double jeopardy’ scenario described by Samuel & Ellis (2002). They demonstrated that both overexpression and suppression of a redox-activated MAPK (SIPK) in tobacco plants gave rise to ozone sensitivity. Furthermore, transgenes overexpressing a non-phosphorylatable version of SIPK were not sensitive. Therefore, our data suggest that a balanced phosphorylation level of MAPKs is necessary for tight regulation of PDF1.2, PR-1 and PR-5 transcription, whereas both highly phosphorylated MAPKs (mkp1 mutant) and non-phosphorylated MAPKs (MKP1 overexpressing lines 6 and 10) result in derepression of expression of these genes (Fig. 5).

Figure 5.

A scheme showing the suggested relationship between genotype, phosphorylation level and the phenotype with respect to the transcriptional control on PR and PDF genes. MAP, mitogen-activated protein.

Samuel & Ellis (2002) also performed an analysis of the MAPK activation profiles that revealed an interplay between SIPK and wound-induced protein kinase (WIPK). Prolonged activation of SIPK was seen in the SIPK-overexpressing genotype, without WIPK activation, whereas strong activation of WIPK was observed in the SIPK-suppressed genotype. Therefore, an extension of our study would be a detailed study of the protein levels and the phosphrylation levels of each different Arabidopsis MAPK ortholog in the wild type, the mkp1 mutant and the two MKP1 overexpressing mutant lines 6 and 10. Such a large effort would decipher the gene expression pattern observed by us and reveal important mechanisms behind gene expression regulation by phosphorylated proteins.

Taken together, our results show the involvement of MKP1 in more general stress responses similar to those induced by pathogen attack as judged by the results on PR-1, PR-5 and PDF1.2 expression, and the absence of regulation (or very insignificant regulation) of more UV-B-specific molecular markers (CHS, PYROA and MEB5.2; Brosché & Strid 2003).

H2O2 accumulation and UV-B-dependent phenotypes

Staining for H2O2 in atrboh mutants and Col-0 wild type showed an increase in H2O2 accumulation in the wild type as a result of UV-B (Fig. 4a and Table 2). None of the three atrboh mutants showed any evident differences in H2O2 levels in leaves between UV-A and UV-A+B-exposed plants. The absolute H2O2 levels were comparable in all strains tested in the absence of UV-B. These results confirm the suggestion that NADPH oxidase contributes to H2O2 accumulation under UV-B exposure.

The phenotype studies of atrboh mutants and the Col-0 background ecotype clearly revealed the involvement of the AtRBOHD and AtRBOHF genes in UV-B responses. All tested atrboh mutants showed statistically significant decreases in fresh and dry weights, leaf area and rosette diameter as the result of UV-B exposure, whereas the same morphological traits remained unaffected in Col-0 (Table 3). These results strongly suggest NADPH oxidase as an important enzyme for protection against UV-B stress in plants.

Knockout of the MKP1 gene also resulted in growth retardation. Statistically significant decreases in leaf area and rosette size (Table 4) were observed in the mkp1 mutant but neither in the complemented line, nor in the wild type. Furthermore, we observed such an additional phenotypical response in young seedlings of the mkp1 mutant in response to UV-B that was not observed in mature plants. One-week-old seedlings exposed to very low fluence rate (UVBE,300 = 0.07 W m−2) revealed visible signs of bleaching (Fig. 4a). This implies that the involvement of MKP1 in stress relief is dependent on the developmental stage of the plants.


Our results show involvement of plasma membrane-bound NADPH oxidase and MKP1 in UV-B-dependent stress responses in Arabidopsis plants. NADPH oxidase is involved in regulating the extent of expression of the CHS, PYROA and MEB5.2 molecular markers, although expression to some extent was still seen in mutants lacking the functional enzyme. The differences in mRNA levels of the molecular markers were also reflected at the morphological level. atrboh mutants displayed growth retardation after UV-B irradiation when compared with the corresponding wild type. Involvement of MKP1 in UV-B tolerance was also evident on the morphological level because mkp1 was clearly UV-B-sensitive. However, this sensitivity is not a result of alteration in gene expression of our molecular markers. Instead, MKP1 is involved in repression of more general stress response pathways, which are normally activated by pathogen attack. This regulation results in constitutive expression of PR and PDF1.2 genes in the mkp1 mutant. Gene expression analyses thus suggest that MKP1 acts upon a MAPK that is not regulated by UV-B. Furthermore, MKP1 action was dependent on the developmental stage, as shown by differential UV-B effects on mutant seedlings and mature mutant plants.


We would like to thank Dr Roman Ulm for providing us with seeds for the mkp1 and complemented lines 6 and 10 seeds, and Prof. Jonathan Jones for the atrboh seeds. This work was supported by a grant to Irina Kalbina from Helge Ax:son Jonsons foundation, and by grants to Åke Strid from the Magn Bergvall foundation and from Örebro University’s Faculty for Medicine, Natural Science and Technology.