ATR and MKP1 play distinct roles in response to UV-B stress in Arabidopsis

Authors

  • Marina A. González Besteiro,

    1. Department of Botany and Plant Biology, University of Geneva, Sciences III, Geneva 4, Switzerland
    2. Faculty of Biology, Institute of Biology II, University of Freiburg, Freiburg, Germany
    3. Spemann Graduate School of Biology and Medicine, University of Freiburg, Freiburg, Germany
    Search for more papers by this author
  • Roman Ulm

    Corresponding author
    • Department of Botany and Plant Biology, University of Geneva, Sciences III, Geneva 4, Switzerland
    Search for more papers by this author

For correspondence (e-mail roman.ulm@unige.ch).

Summary

Ultraviolet-B (UV-B) stress activates MAP kinases (MAPKs) MPK3 and MPK6 in Arabidopsis. MAPK activity must be tightly controlled in order to ensure an appropriate cellular outcome. MAPK phosphatases (MKPs) effectively control MAPKs by dephosphorylation of phosphothreonine and phosphotyrosine in their activation loops. Arabidopsis MKP1 is an important regulator of MPK3 and MPK6, and mkp1 knockout mutants are hypersensitive to UV-B stress, which is associated with reduced inactivation of MPK3 and MPK6. Here, we demonstrate that MPK3 and MPK6 are hyperactivated in response to UV-B in plants that are deficient in photorepair, suggesting that UV-damaged DNA is a trigger of MAPK signaling. This is not due to a block in replication, as, in contrast to atr, the mkp1 mutant is not hypersensitive to the replication-inhibiting drug hydroxyurea, hydroxyurea does not activate MPK3 and MPK6, and atr is not impaired in MPK3 and MPK6 activation in response to UV-B. We further show that mkp1 leaves and roots are UV-B hypersensitive, whereas atr is mainly affected at the root level. Tolerance to UV-B stress has been previously associated with stem cell removal and CYCB1;1 accumulation. Although UV-B-induced stem cell death and CYCB1;1 expression are not altered in mkp1 roots, CYCB1;1 expression is reduced in mkp1 leaves. We conclude that the MKP1 and ATR pathways operate in parallel, with primary roles for ATR in roots and MKP1 in leaves.

Introduction

Genome integrity is constantly challenged in all organisms by endogenous and environmental genotoxic stress. UV-B radiation (280–315 nm) is the most energetic part of sunlight reaching the earth, and may lead to the formation of pyrimidine dimers that inhibit DNA replication and transcription, and that are pre-mutagenic lesions (Britt, 2004). The sessile growth habit of plants and their direct dependence on the energy from sunlight requires that plants be particularly well-armed against UV-B stress. In addition to protection provided by the plant-specific UVR8 photoreceptor-dependent acclimation pathway (Favory et al., 2009; Heijde and Ulm, 2012), plants contain damage signaling pathways that are partially conserved in other organisms (Culligan et al., 2004; Cools and De Veylder, 2009; González Besteiro et al., 2011).

To cope with damaged DNA, cells have evolved efficient mechanisms to sense and repair a broad range of lesions. The protein kinases ATAXIA-TELANGIECTASIA MUTATED (ATM) and ATM AND RAD3-RELATED (ATR) are key regulators of the DNA damage response and are activated by double-strand breaks and replication stress, respectively. The DNA damage response initiated by ATM and ATR is conserved and well-studied in mammals. It controls cell-cycle transitions, transcription, DNA repair and apoptosis (Sancar et al., 2004). In Arabidopsis, atr mutants are hypersensitive to UV-B and other replication-blocking agents, including hydroxyurea (HU) (Culligan et al., 2004). In the absence of exogenously introduced replication stress, Arabidopsis atr mutants develop normally throughout their entire life cycle (Culligan et al., 2004), whereas mice lacking ATR die early in development (Brown and Baltimore, 2000; de Klein et al., 2000). Arabidopsis WEE1 KINASE HOMOLOG (WEE1) and SUPPRESSOR OF GAMMA RESPONSE 1 (SOG1) are downstream transducers of ATR (Preuss and Britt, 2003; De Schutter et al., 2007; Yoshiyama et al., 2009). The checkpoint kinase WEE1 acts as a G2/M checkpoint regulator in non-plant organisms, but Arabidopsis WEE1 times the G1/S transition in response to replication stress (Cools et al., 2011). WEE1 operates by directly modulating the activity of cyclin-dependent kinase (CDK) complexes, the effectors that control cell-cycle progression (De Schutter et al., 2007; Cools et al., 2011). SOG1 is a plant-specific NAC domain transcription factor that governs the transcriptional response to DNA damage (Yoshiyama et al., 2009). It is required for autolytic cell death in Arabidopsis root tips following UV-B or gamma radiation (Fulcher and Sablowski, 2009; Furukawa et al., 2010).

In addition to the ATR-mediated pathway, MAP kinase (MAPK) signaling constitutes another UV response that is conserved among organisms (Herrlich et al., 1997; Ulm, 2003). In Arabidopsis, MPK3 and MPK6 are activated by DNA-damaging agents, and map kinase phosphatase 1 (mkp1) mutants are hypersensitive to UV-B stress (Ulm et al., 2001, 2002; Kalbina and Strid, 2006; González Besteiro et al., 2011). MKP1 negatively regulates MAPK signaling by dephosphorylating stress-activated MPK3 and MPK6 (Bartels et al., 2009, 2010). Concomitantly, mkp1 hypersensitivity to UV-B stress is associated with hyperactive MPK3 and MPK6 (González Besteiro et al., 2011).

It has been speculated that MKP1 may function as a trigger of cell-cycle checkpoints (Cools and De Veylder, 2009; González Besteiro et al., 2011). However, the role of MKP1 in the DNA damage response and its relationship with the ATR-mediated pathway remain poorly understood. Here we provide evidence suggesting that MPK3 and MPK6 are activated by UV-B-induced DNA damage. Although MKP1 does not affect the G2/M-specific B-type cyclin CYCB1;1 in roots, it positively regulates CYCB1;1 expression in response to UV-B radiation in leaves. As the leaves of the atr mutants are not detectably UV-B hypersensitive (in contrast to mkp1), our data suggest that the MKP1-regulated, DNA damage-activated MPK3/MPK6 cascade is a major pathway contributing to UV-B-induced DNA damage tolerance in the above-ground tissues of Arabidopsis.

Results

MPK3 and MPK6 activation in response to UV-B stress is enhanced in the absence of photorepair

The UV-B hypersensitivity of mkp1 is associated with reduced and delayed MPK3 and MPK6 inactivation (González Besteiro et al., 2011), and is reminiscent of the UV-B phenotype of plants that are deficient in the photolyase PHR1/UVR2 (Landry et al., 1997). The Arabidopsis PHR1 and UVR3 photolyases are light-dependent DNA repair enzymes that perform error-free repair of UV-B-induced cyclobutane pyrimidine dimers and (6–4) photoproducts, respectively (Britt, 2004). The photolyase mutants phr1 and uvr3 thus accumulate elevated levels of pyrimidine dimers and are UV-B-hypersensitive (Jiang et al., 1997; Landry et al., 1997; Castells et al., 2010). To determine whether DNA damage induces MPK3 and MPK6 activation in response to UV-B, we first compared their activities in UV-B-treated plants that were incubated after treatment in darkness (no photorepair) or in the light (allowing photorepair by photolyases). MPK3 and MPK6 activation was clearly sustained when photorepair was inhibited in the dark (Figure 1a). Consistently, Arabidopsis phr1 and uvr3 mutants showed sustained MPK3 and MPK6 activation when recovery from UV-B was allowed under white light (under which conditions the two photolyases are active in wild-type; Figure 1b). The double mutant phr1 uvr3 showed even higher MPK3 and MPK6 activities in comparison with either single mutant (Figure 1b). Collectively, these results suggest that UV-B-dependent MPK3 and MPK6 activation is initiated by pyrimidine dimers in UV-B-damaged DNA.

Figure 1.

Sustained MPK3 and MPK6 activation upon UV-B stress in the absence of photorepair by the photolyases PHR1 and UVR3. (a) Immunoblot analysis using anti-phospho-p44/42 antibodies for protein extracts from 7-day-old wild-type (Col) seedlings that were irradiated for 10 or 20 min with broadband UV-B and subsequently kept in the dark (D) or under continuous light (L) at 21°C for the indicated times (1, 8 or 24 h). Control seedlings were irradiated for 20 min under broadband UV-B but under a WG345 cut-off that filters out the UV-B portion (−UV-B). The anti-phospho-p44/42 antibodies detect the active, phosphorylated form of Arabidopsis MPK3 (p-MPK3) and MPK6 (p-MPK6). The immunoblot was re-probed using anti-actin antibody (loading control). (b) Immunoblot analysis using anti-phospho-p44/42 antibodies for protein extracts from 7-day-old wild-type (Col), phr1, uvr3 and phr1 uvr3 seedlings that were irradiated for 20 min with broadband UV-B and immediately harvested or kept under continuous light (L) at 21°C for 8 h, as indicated. The immunoblot was re-probed using anti-actin antibody (loading control).

Both mkp1(Col) and atr are hypersensitive to UV-B at the root level, but at above-ground level, it is primarily mkp1(Col) that is hypersensitive to UV-B

The ATR protein kinase is activated in response to various types of DNA damage, including cross-links and replication stress (Cimprich and Cortez, 2008). In particular, human ATR was shown to bind to UV-damaged DNA, and this was able to stimulate ATR activity (Unsal-Kacmaz et al., 2002; Choi et al., 2007). The activation of mammalian ATR by UV-damaged DNA suggests that Arabidopsis ATR may constitute a link between pyrimidine dimers and MAPK activation. To address this possibility, we analyzed MPK3 and MPK6 activities in UV-B stressed atr seedlings in comparison with wild-type, mkp1 and mkp1 atr double mutants. While mkp1 and mkp1 atr mutants displayed MPK3 and MPK6 hyperactivation, wild-type and atr seedlings displayed similar activity levels (Figure 2). These results do not support a major function of ATR upstream of MAPK activation.

Figure 2.

atr shows wild-type levels of UV-B-induced MAPK activation. Immunoblot analysis using anti-phospho-p44/42 antibodies for protein extracts from 7-day-old wild-type (Col), mkp1(Col), atr-2 and mkp1(Col) atr-2 seedlings. The seedlings were irradiated for 30 min with broadband UV-B (+UV-B) or under a WG345 cut-off filter (−UV-B) and harvested directly after UV-B irradiation or after an additional 1 h under standard growth conditions under white light (WL). The immunoblot was re-probed using anti-actin antibody (loading control).

To further understand the relationship between ATR and MKP1, we analyzed the UV-B stress phenotype of the mkp1 atr double mutant in comparison with wild-type and the respective single mutants. First, we confirmed the previously characterized UV-B hypersensitive root phenotype of atr (Figure 3a; Culligan et al., 2004). Moreover, mkp1(Col) displayed root growth inhibition by UV-B that was comparable to that in the atr mutant, and mkp1 atr double mutant roots displayed enhanced UV-B sensitivity compared with the single mutants (Figure 3a,b). Second, we found that mkp1(Col) leaves bleached and subsequently died in response to UV-B stress much more readily but much less so than the leaves of wild-type and atr, with no detectable difference between the latter two (Figure 3c). The UV-B hypersensitivity of mkp1 atr double mutant leaves was comparable to that of the mkp1(Col) single mutant (Figure 3c). Altogether, the mkp1 atr UV-B phenotype indicates that MKP1 and ATR regulate separate pathways, with approximately equal importance in roots but a major role for MKP1 in above-ground tissues.

Figure 3.

MKP1 and ATR contribute to UV-B tolerance in roots, and MKP1 also contributes to UV-B tolerance in above-ground tissues. (a) Six-day-old wild-type (Col), mkp1(Col), atr-2 and mkp1(Col) atr-2 seedlings were irradiated for 15 min with 1.3 mW/cm2 UV-B or kept under standard conditions (−UV-B), and allowed to recover for 7 days. Representative plants are shown. (b) Cumulative root length of plants grown and treated as described in (a). The length of the primary root was measured 2 days before the treatment, just before UV-B radiation, and every 1 or 2 days afterwards. Values are means ± SD (n = 12-15). (c) Seven-day-old wild-type (Col), mkp1(Col), atr-2 and mkp1(Col) atr-2 seedlings were irradiated for 2.5 h with broadband UV-B under a WG305 cut-off filter (+UV-B) or a WG360 cut-off filter (−UV-B), and allowed to recover for 9 days.

mkp1(Col) mutants are not hypersensitive to hydroxyurea-mediated replication stress

In addition to UV-B, atr root growth is also hypersensitive to the replication-inhibiting drug hydroxyurea (HU), supporting the importance of ATR in coping with diverse replication stresses (Culligan et al., 2004). We thus determined whether mkp1(Col) is similarly affected by replication-blocking agents other than UV-B. When grown in 1 mm HU, the mkp1(Col) mutant and wild-type roots grew similarly, but atr root growth was strongly affected (Figure 4a,b). The HU hypersensitivity of mkp1 atr double mutant roots was comparable to that of the atr single mutant (Figure 4a,b).

Figure 4.

mkp1 is not hypersensitive to hydroxyurea (HU). (a) Five-day-old wild-type (Col), mkp1(Col), atr-2 and mkp1(Col) atr-2 seedlings were transferred for 7 days to control medium or medium supplemented with 1 mm HU. Representative plants grown in the presence of HU are shown. A representative control plate is shown in Figure 2(a). (b) Quantification of root length of plants grown and treated as described in (a). Values are means ± SD (n = 8-15). (c) Cell-death phenotype of mkp1(Col) compared with wild-type (Col) and atr-2. Roots were stained with PI 24 h after transfer to control medium (−HU) or medium containing 1 mm HU (+HU). Scale bar = 100 μm. (d) Immunoblot analysis using anti-phospho-p44/42 MAPK for protein extracts from 7-day-old seedlings. Sixteen hours after transfer to liquid medium without sucrose, the seedlings were irradiated for 10 min with unfiltered broadband UV-B (+UV-B), or treated with 0 (−HU), 5 or 40 mm HU for the indicated times. The immunoblot was re-probed using anti-actin antibody (loading control).

The root phenotype of atr under HU-induced replication stress is associated with massive cell death, both in the prevascular tissue and the stem cell zone (Cools et al., 2011). To analyze the role of MKP1 in HU-induced cell death, we stained HU-treated roots with propidium iodide (PI), which stains the walls of living cells but penetrates dead cells. Similar to wild-type and in stark contrast to atr, mkp1(Col) showed no detectable cell death after a 24 h treatment with 1 mm HU (Figure 4c). Moreover, in contrast to UV-B, HU replication stress did not result in detectable MPK3 and MPK6 activation (Figure 4d). We conclude that, whereas ATR is important in response to various forms of replication stress, MKP1 and MPK3/MPK6 activation are more specifically involved in the response to UV-B stress.

mkp1(Col) mutant roots show normal levels of UV-B-induced cell death

One strategy to avoid propagation of genetic defects involves selective killing of irreparably damaged stem cells. In Arabidopsis root tips, UV-B-induced stem cell death is controlled by ATR (Furukawa et al., 2010). UV-B-induced root growth inhibition of mkp1(Col) is similar to that of atr mutants, which exhibit enhanced stem cell death upon UV-B treatment (Furukawa et al., 2010). We therefore hypothesized that mkp1 mutants may show enhanced root stem cell death in response to UV-B. However, mkp1(Col) shows wild-type levels of stem cell and vascular tissue death under a variety of UV-B conditions (Figure 5). We conclude that the UV-B hypersensitive root phenotype of mkp1 manifests without altered cell death.

Figure 5.

mkp1 shows wild-type levels of UV-B-induced stem cell death. (a) Four-day-old wild-type (Col), mkp1(Col) and atr-2 seedlings were irradiated for 2.75 min with broadband UV-B under a WG305 (+UV-B) or WG345 cut-off filter (−UV-B), and stained with PI after 24 h incubation in the dark. Scale bar = 100 μm. (b) Three-day-old wild-type (Col) and mkp1(Col) seedlings were transferred to fresh medium without sucrose, irradiated the following day for 4 min with broadband UV-B under a WG305 (+UV-B) cut-off filter, and stained with PI after 24 h incubation in the dark. Scale bar = 100 μm. (c) Quantification of dead cells in root tips lacking MKP1 (mkp1) in comparison with wild-type (Col). Seedlings were irradiated for 2 min with broadband UV-B under a WG305 (+UV-B) cut-off filter, and stained with PI after 24 h incubation in the dark. The number of dead cells per root per genotype were counted, and the data are shown as the percentage of roots (= 63) with no dead cells (0), 1, 2–3 or 4–6 dead cells per root. (d) Roots of 4-day-old wild-type (Col), mkp1(Col), atr-2 and mkp1(Col) atr-2 seedlings were irradiated for 2.5 h with unfiltered narrowband UV-B (+UV-B) and stained with PI after 24 h incubation in the dark. Scale bar = 100 μm.

mkp1(Col) displays altered CYCB1;1 expression in young leaves, but not in root tips

The mitotic cyclin B1;1 (CYCB1;1) is induced by DNA damage, and its mis-regulation has been associated with hypersensitivity to ionizing radiation and replication stress (Culligan et al., 2004, 2006; Hefner et al., 2006; De Schutter et al., 2007). As mkp1 and atr mutant roots respond similarly to UV-B radiation (Figure 3a,b) and the atr phenotype correlates with reduced CYCB1;1 expression (Culligan et al., 2004), we determined whether MKP1 also affects the behavior of this cyclin. For this purpose, we used the pCYCB1;1::dbCYCB1;1-GUS construct containing a mitotic destruction box (dbCYCB1;1) between the CYCB1;1 promoter and the GUS reporter, allowing determination of the amount of CYCB1;1–GUS protein (Colon-Carmona et al., 1999). We observed similar accumulation of dbCYCB1;1–GUS in mkp1(Col) compared with wild-type roots (Figure 6a). Consistently, endogenous CYCB1;1 transcript levels in whole roots did not differ between mkp1 and the wild-type (Figure 6b). We therefore conclude that ATR and MKP1 affect the UV-B tolerance of roots via different mechanisms.

Figure 6.

CYCB1;1 regulation in the mkp1 mutant is affected in response to UV-B in leaves but not in roots. (a) Accumulation of db CYCB1;1-GUS in roots of wild-type Col, mkp1 and the complemented mkp1/Pro35S::Pyo-MKP1 line under UV-B (+UV-B) or without UV-B (−UV-B). (b) Quantitative RT-PCR analysis of CYCB1;1 expression levels in roots of mkp1(Col) and mkp1(Col) atr-2 compared with atr-2 and wild-type (Col). (c) Accumulation of db CYCB1;1-GUS in young leaves of wild-type (Col), mkp1(Col) and the complemented mkp1/Pro35S::Pyo-MKP1 line under UV-B (+UV-B) or without UV-B (−UV-B). (d) Quantitative RT-PCR analysis of CYCB1;1 expression levels in above-ground tissues of mkp1(Col) and mkp1(Col) atr-2 compared with atr-2 and the wild-type (Col). For (a) and (c), 6-day-old transgenic seedlings (harboring the pCYCB1;1::dbCYCB1;1-GUS construct) were fixed and stained 24 h after a 15 min irradiation with unfiltered broadband UV-B (+UV-B) or under a WG345 cut-off filter (−UV-B). Scale bar = 100 μm. For (b) and (d), 8-day-old seedlings were acclimated to weak light for 24 h, treated for 10 min with broadband UV-B (+UV-B) or under a WG345 cut-off filter (−UV-B), and the roots (b) or above-ground tissues (d) were harvested for RNA preparation 24 h later. Error bars represent standard deviations for three independent experiments.

Despite the similarity of mkp1(Col) and atr root phenotypes, we did not find any evidence for the involvement of MKP1 in ATR-regulated responses in root tips. However, as our data suggest that MAPKs are activated by cyclobutane pyrimidine dimers and (6–4) photoproducts in seedlings, and mkp1(Col) aerial tissues are highly UV-B hypersensitive, we analyzed the possibility of a defective DNA damage response in the above-ground tissues of mkp1 mutant plants. We stained wild-type, mkp1(Col) and the complementation line mkp1/Pro35S:Pyo-MKP1 carrying the pCYCB1;1::dbCYCB1;1-GUS construct 24 h after a 15 min UV-B exposure. In contrast to both controls, mkp1 showed reduced accumulation of dbCYCB1;1–GUS in young leaves in response to UV-B (Figure 6c). In support of this observation, after a similar treatment (10 min UV-B plus 24 h recovery), CYCB1;1 transcripts increased by up to 4.5-fold in the above-ground tissues of wild-type plants, but only 2.5-fold in mkp1 and mkp1 atr mutants (Figure 6d and Figure S1). In accordance with previous results (Culligan et al., 2006), the atr mutant displayed levels of CYCB1;1 mRNA similar to those of wild-type plants (Figure S1). It is important to note that, in root tips, atr mutants regulate CYCB1;1 at the transcriptional level (Culligan et al., 2004). We conclude that MKP1 controls the DNA damage response marker CYCB1;1 specifically in above-ground tissues, but not in roots.

Discussion

Plants must be able to resist and/or tolerate the effects of UV-B radiation. This ability involves a combination of reducing exposure to UV-B and repairing damaged molecules (Britt, 2004). Damage prevention (e.g. production of ‘sunscreen’ flavonols) and repair (e.g. induction of genes encoding photolyases) are positively regulated during UV-B acclimation, a process that is mediated by the UV-B photoreceptor UVR8 (Rizzini et al., 2011; Heijde and Ulm, 2012). If damage occurs due to insufficient UV-B protection, plants use the ATR kinase that relays DNA damage signals to regulate the cell cycle (Culligan et al., 2004; Cools and De Veylder, 2009), and activate a UVR8-independent stress-response MAPK pathway (González Besteiro et al., 2011). Arabidopsis MKP1 regulates UV-B-activated MPK3 and MPK6 and is required to cope with UV-B stress (González Besteiro et al., 2011). However, it was previously not known which signal activates the MAPK pathway in response to UV-B stress, how the ATR- and MAPK-regulated pathways relate to each other, and which downstream processes associated with UV-B stress are affected by the MKP1-regulated MAPK pathway. Here we provide evidence that (i) the UV-B-activated MAPK cascade responds to UV-B-induced DNA damage (pyrimidine dimers), (ii) the MKP1-regulated and ATR-mediated DNA damage pathways operate independently of each other, and (iii) MKP1 promotes expression of the DNA-damage-associated mitotic CYCB1;1 in UV-B-exposed aerial tissues.

UV-B-induced pyrimidine dimers may serve as a signal for MAPK pathway activation

UV radiation activates MAPKs in various organisms. However, the identity of the upstream activating signals remains controversial, including both nuclear and non-nuclear signals (Herrlich et al., 2008). For example, UV activation of mammalian MAPKs may take place in the cytoplasm, as enucleated cells showed close to maximal activation (Devary et al., 1993). More recently, however, MAPK activation was reported to rely on UV-dependent accumulation of cyclobutane pyrimidine dimers (Hamdi et al., 2005). To the best of our knowledge, this issue has not previously been addressed in plants. Using photolyase mutants as a genetic tool, we provide evidence that pyrimidine dimers may be a trigger for MAPK activation in Arabidopsis. Arabidopsis encodes two photolyases, the cyclobutane pyrimidine dimer photolyase PHR1/UVR2 and the (6–4) photoproduct photolyase UVR3, that repair pyrimidine dimers by error-free direct reversal via a process that requires UV-A/blue photons (Ahmad et al., 1997; Jiang et al., 1997; Landry et al., 1997). We found that UV-B-mediated MPK3 and MPK6 activation is higher in photolyase mutants (particularly phr1 uvr3 double mutants) and under dark conditions that are non-permissive for photorepair. Thus, we conclude that pyrimidine dimers probably play an important role as a signal of UV-B stress in the MAPK pathway, but this is probably not the only UV-B-associated stress signal. It is known that UV-B damages a broad variety of cellular components, providing a range of putative signals for MAPK pathway activation. However, the existence of a sensor for pyrimidine dimers upstream of the MAPK pathway in plants is supported by our findings but its identity remains to be determined . Similarly, the MAPK kinase kinase and MAPK kinase components of the UV-B-activated MAPK cascade that affect MPK3 and MPK6 activation remain to be determined. Several good candidates for these missing components in UV-B signaling may be deduced from other stress signaling pathways involving MPK3 and/or MPK6 (Andreasson and Ellis, 2010).

MKP1-regulated and ATR-mediated DNA damage pathways operate independently of each other

ATR activation constitutes an early step in replication stress sensing, and may therefore link DNA damage to MAPK activation. Indeed, mammalian p38 MAPKs may be activated by ATR (Raman et al., 2007; Reinhardt et al., 2007). However, we showed that UV-B-induced MAPK activation was not altered in Arabidopsis atr seedlings. Moreover, our genetic evidence showing an additive effect for UV-B-mediated root growth inhibition of mkp1 and atr in the double mutants does not support a link between ATR- and MKP1-related signaling. In contrast to ATR, MKP1 appears not to be involved in HU-mediated replication stress signaling: HU does not activate MPK3 and MPK6, and mkp1 null mutants tolerate HU to a similar level as wild-type. We thus conclude that UV-B-induced DNA damage rather than associated replication stress activates MPK3 and MPK6, and that this pathway operates parallel to the ATR-dependent pathway. Whereas ATR has a role in response to replication stress (UV, HU, aphidicolin), the role MKP1 appears to be restricted to DNA-damage stress (e.g. UV).

Stem cells undergo cell death upon exposure to exogenous DNA-damaging agents. Previous studies have identified ATR as a crucial regulator of this response in Arabidopsis (Fulcher and Sablowski, 2009; Furukawa et al., 2010). Elevated stem cell death in atr mutants is associated with root growth delay (Furukawa et al., 2010). Moreover, a positive correlation between growth and stem cell death was reported previously when analyzing mutants in DNA translesion polymerases η and ζ (Curtis and Hays, 2007). Therefore, it was inferred that stem cell death is a crucial mechanism to prevent perpetuation of mutations and restore growth after replication-blocking lesions have occurred (Curtis and Hays, 2011). However, despite a similar root growth inhibition phenotype for mkp1 and atr, we did not observe altered stem cell death in mkp1 mutants, demonstrating that MKP1 and ATR differ in the downstream physiological response that they invoke.

Despite this, we cannot at present discard the possibility that the UV-B-activated MAPK cascade participates in DNA damage-induced cell death. Previously, UV-C-induced bleaching of Arabidopsis seedlings was found to be associated with apoptosis-like cell death (Danon and Gallois, 1998; Danon et al., 2004). In animals, the stress-activated MAPKs JNK and p38 function as pro-apoptotic factors in UV signaling (Batista et al., 2009). Although JNK and p38 may play a cytoprotective role under certain conditions (Liu and Lin, 2005; Jinlian et al., 2007), there is no evidence for such a mechanism for stress-activated MAPKs in plants. However, it was recently shown that MPK6 promotes apoptosis-like death in roots challenged by cadmium stress (Ye et al., 2013). Whether MKP1-regulated UV-B-activated plant MAPKs promote apoptosis-like cell death in UV-B-exposed aerial tissues remains to be investigated.

MKP1 facilitates expression of the DNA damage-associated mitotic CYCB1;1 in UV-B-exposed aerial tissues

The functional importance of CYCB1;1 up-regulation observed upon genotoxic stress in Arabidopsis remains unclear (Culligan et al., 2006; Cools and De Veylder, 2009). CYCB1;1 accumulation is regarded as an indicator of cell-cycle arrest at the G2/M boundary (Culligan et al., 2004; Hefner et al., 2006; De Schutter et al., 2007), or is considered a marker of disorganized cortical microtubules and root swelling (Serralbo et al., 2006; Wu et al., 2010). Defective CYCB1;1 induction and CYCB1;1–GUS accumulation in mkp1 mutants may be therefore a reflection of either process. The role of CYCB1;1 in morphogenesis is regulated by controlling the stability of the CYCB1;1 protein, whereas regulation at the transcriptional level is considered to be a marker of checkpoint signaling (Wu et al., 2010). In the above-ground tissues of the mkp1 mutant, CYCB1;1 transcript accumulation is defective, but is normal in the atr mutant. This correlates with enhanced UV-B-induced leaf bleaching in mkp1, but not in atr. In contrast to ATR, MKP1 may thus regulate a UV-B-induced G2 checkpoint in the above-ground tissues of Arabidopsis to promote survival. It is of note that other tested DNA repair and checkpoint genes (namely PARP1 and WEE1) were similarly mis-regulated in mkp1(Col) compared to wild-type (Figure S2). However, CYCB1;1 may also play an as yet unknown role, unrelated to checkpoint activation, but required for genotoxic stress tolerance.

In summary, we provide evidence that pyrimidine dimers are a signal for MAPK pathway activation in plants in response to UV-B stress, that ATR and MKP1 function independently of each other in response to UV-B, with equal importance in root tips and a primary role for MKP1 in aerial tissues, and that the mkp1 hypersensitive phenotype is associated with reduced expression of the CYCB1;1 marker gene. Together, these results add significantly to our understanding of the UV-B stress response in plants. Further research is required to identify the damage sensor and components of the downstream MAPK cascade, as well as the underlying physiological reason for the UV-B hypersensitivity of mkp1.

Experimental Procedures

Plant material and growth conditions

The mkp1(Col), atr-2 (SALK_032841), phr1 (WiscDsLox466C12), uvr3 (WiscDsLox334H05) and uvr3 phr1 mutants, and the mkp1/Pro35S:Pyo-MKP1 and pCYCB1;1::dbCYCB1;1-GUS lines are in the Columbia accession (Col) (Colon-Carmona et al., 1999; Garcia et al., 2003; Culligan et al., 2004; Bartels et al., 2009; Castells et al., 2010). Plants were grown under aseptic conditions or on soil as described previously (Bartels et al., 2009).

To generate the mkp1 atr-2 double mutant, we used the mkp1(Col) introgression line for crossing (Bartels et al., 2009). The mkp1(Col) line exhibits growth defects associated with a constitutive defense response when grown on soil (Bartels et al., 2009); however, the UV-B stress phenotype of mkp1(Col) is clearly independent of the constitutive defense response (González Besteiro et al., 2011). The growth defects of the mkp1 atr double mutant on soil were indistinguishable from those of the mkp1(Col) single mutant (Figure S3).

Plant treatments

For UV-B sensitivity assays of above-ground tissue, plants were treated exactly as described previously (González Besteiro et al., 2011). For root growth assays, 4-day-old seedlings were moved to fresh plates supplemented with 1.5% Phytagel (Sigma–Aldrich, Buchs, Switzerland; www.sigmaaldrich.com) and allowed to grow vertically. Two days later, they were either transferred to medium supplemented with 1 mm HU (Sigma–Aldrich) or treated for the indicated times with 1.3 mW/cm2 broadband UV-B. The plates were then returned to standard growth conditions for 6–9 days before taking photographs. The position of the root tip was marked daily, starting the day prior to irradiation, and the length of the primary root was analyzed using ImageJ (http://rsb.info.nih.gov/ij/). For cell death assays, seeds were germinated on vertical plates, treated 4 days later with broadband or narrowband UV-B, as indicated, and transferred to darkness for 24 h before analysis. For analysis of the GUS reporter lines, 5-day-old seedlings were transferred to liquid medium under a weak light field for at least 16 h. The seedlings were harvested after 15 min treatment with UV-B followed by 24 h incubation under the weak light field. For dtermination of MPK3 and MPK6 activation in response to HU and UV-B, 7-day-old seedlings were transferred to liquid medium without sucrose for 16 h before the treatment was applied. Broadband UV-B lamps (Philips TL40W/12RS, Philips AG Lighting, Zürich, Switzerland; www.philips.com) were used exactly as described previously (Ulm et al., 2004). The conditions for weak light (Osram L18W/30, OSRAM AG, Winterthur, Switzerland; www.osram.com) and narrowband UV-B irradiation (Philips TL20W/01RS) were as described previously (Favory et al., 2009). UV-B was filtered through 3-mm transmission cut-off filters of the WG series with half-maximal transmission at the indicated wavelength (WG305, WG345 and WG360; Schott, Mainz, Germany; www.schott.com), or unfiltered through a 3-mm quartz plate, as described before (Ulm et al., 2004).

Immunoblot analysis

Protein extraction and immunoblot analysis were performed exactly as described previously (González Besteiro et al., 2011). We used polyclonal primary antibodies against actin (Sigma–Aldrich) and phospho-p44/42 MAP kinase (Cell Signaling Technology, Danvers, MA, USA; www.cellsignal.com/) with horseradish peroxidase-conjugated anti-mouse and anti-rabbit immunoglobulins (Dako Schweiz AG, Baar, Switzerland; www.dako.com) as secondary antibodies, respectively. The anti-phospho-p44/42 antibodies detect the active phosphorylated form of Arabidopsis MPK3 and MPK6 (González Besteiro et al., 2011).

Quantitative real-time PCR

RNA was isolated using an RNeasy Plant Mini Kit (Qiagen AG, Hombrechtikon, Switzerland; www.qiagen.com) and treated with DNase I according to the RNeasy kit instructions. cDNA synthesis and SYBR Green quantitative RT-PCR were performed as previously described (Bartels et al., 2009) using a 7900HT real-time PCR system (Life Technologies Europe BV, Zug, Switzerland; www.appliedbiosystems.com). The gene-specific primers were CYCB1;1_fw (5′-AGACCAACAAGCTGCAGTTCT-3′) and CYCB1;1_rv (5′-GGAGTGGTTGATTGATGTGCAT-3′)for CYCB1;1 (AT4G37490), PARP1_fw (5′-ACCCATCAGAGGCTCAAACACT-3′) and PARP1_rv (5′-CCCCTTGGAGCATGAACGT-3′) for PARP1 (AT4G02390), and WEE1_fw (5′-CGTATGTGTTGAATCAGTCTGA-3′) and WEE1_rv (5′-GACTGAAATGTCCAGCACCAA-3′) for WEE1 (AT1G02970). Expression of PHR1/UVR2 and UVR3 was analyzed using TaqMan quantitative RT-PCR, as previously described (Favory et al., 2009). The gene-specific probes and primers were as follows: probe 6-carboxyfluorescein (FAM)-AACCAACGCGCCGGTGGC-6-carboxytetramethylrhodamine (TAMRA) with UVR2_for (5′-GCCGTTGATCTGGCAAACA-3′) and UVR2_rev (5′-TTGATCGAACAGATTGAAAACGA-3′) for PHR1/UVR2 (AT1G12370) , and probe FAM-TTGGATGCACCATCTAGCGCGTCA-TAMRA with UVR3_for (5′-GCCATCATGGTTCAGCTTCTG-3′) and UVR3_rev (5′-CACGAGTAAGAAAACAGGCTACACA-3′) for UVR3 (AT3G15620). cDNA concentrations were normalized to 18S rRNA transcript levels using the Applied Biosystems Eukaryotic 18S rRNA endogenous control assay (Life Technologies).

Staining and microscopy analysis

For propidium iodide (PI) staining, roots were submerged in a 10 μg/ml PI solution (Sigma–Aldrich) for 1 min before imaging with an LSM 510 Meta confocal microscope (Carl Zeiss AG, Feldbach, Switzerland; www.zeiss.ch). PI was excited using a 543 nm HeNe laser and emission light was collected through a long-pass (LP) 560 nm filter. For GUS staining, whole seedlings were fixed in ice-cold 90% acetone for 20 min. After washing in 100 mm phosphate buffer (pH 7.4), seedlings were vacuum infiltrated with staining buffer [0.5 mg/ml 5-bromo-4-chromo-3-indolyl-β-d-glucuronide (X-Glc), 10 mm EDTA, 0.5 mm ferricyanide, 0.5 mm ferrocyanide and 0.1% Triton X-100 in phosphate buffer]. After an overnight incubation at 37°C and clearing with 75% ethanol, samples were mounted in water and analyzed using a stereomicroscope (Leica MZ16, Leica Microsystems AG, Heerbrugg, Switzerland; www.leica-microsystems.com) or a differential interference contrast (DIC) microscope (Zeiss Axioscope II, Carl Zeiss AG, Feldbach, Switzerland, or Nikon Eclipse 80i, Nikon AG, Egg, Switzerland; www.nikon.com).

Acknowledgments

We would like to thank Melanie Binkert (Department of Botany and Plant Biology, University of Geneva, Switzerland), for helpful comments on the manuscript, Peter Doerner (University of Edinburgh, UK), Dorothy Shippen (Department of Biochemistry, Texas A&M University, College Station, TX), Ales Pecinka (Department of Plant Breeding and Genetics, Max Planck Institute for Plant Breeding Research, Cologne, Germany) and Frédy Barneche (Département de Biologie, Ecole Normale Supérieure, Paris, France) for seed material. We are also grateful to Christophe Bauer and Jérôme Bosset from the National Center of Competence in Research (NCCR) Frontiers in Genetics Bioimaging Platform for help with confocal microscopy, and Richard Chappuis for excellent technical assistance. This work was supported by the Excellence Initiative of the Deutsche Forschungsgemeinschaft (GSC-4, Spemann Graduate School), the Swiss National Science Foundation, and the University of Geneva.

Ancillary