• By comparing cDNA populations derived from chromium-stressed primary leaves of barley (Hordeum vulgare L.) with controls, differentially expressed cDNA fragments could be identified. The deduced amino acid sequence of one of these cDNAs [named ‘C2 domain 1’ (HvC2d1)] exhibits a motif that is similar to the known C2 domain and a nuclear localization signal (NLS).
• Expression of this member of a novel class of plant C2 domain-like proteins was studied using real-time PCR, and subcellular localization was investigated using green fluorescent protein (GFP) fusion constructs. Calcium binding was analysed using a 45Ca2+ overlay assay.
• HvC2d1 was transiently induced after exposure to different heavy metals and its mRNA accumulated during the phase of leaf senescence. HvC2d1 expression responded to changes in calcium levels caused by the calcium ionophore A23187 and to treatment with methylviologen resulting in the production of reactive oxygen species (ROS). Using overexpressed and purified HvC2d1, the binding of calcium could be confirmed. Chimeric HvC2d1-GFP protein was localized in onion epidermal cells at the plasma membrane, cytoplasm and the nucleus. After addition of calcium ionophore A23187 green fluorescence was only visible in the nucleus.
• The data suggest a calcium-dependent translocation of HvC2d1 to the nucleus. A possible role of HvC2d1 in stress- and development-dependent signalling in the nucleus is discussed.
Contamination of soil and water by toxic heavy metals such as chromium represents a major environmental problem. Plants growing on such soil can tolerate heavy metals to very different extents. Some plant species are severely damaged by low concentrations of heavy metals while others are not affected even by high concentrations because they have evolved adaptive mechanisms to cope with this stress (Van Assche & Clijsters, 1990; De Vos et al., 1991; Hall, 2002). Sometimes plants actually accumulate large amounts of these metals in their shoots and thus remove the toxic metals from the ground. This trait can be used in the process of phytoremediation to clean contaminated soil (Clemens et al., 2002). However, heavy metals do not only have toxic effects. Some metals are actually essential micronutrients fulfilling many crucial functions in plant metabolism. In order to resolve this dilemma that, on the one hand, some metals have to be taken up in various amounts, transported and finally assembled with their specific target apoproteins and that, on the other hand, surplus toxic metal species have to be detoxified, plants have evolved a complex regulatory network (Clemens et al., 2002).
We are still far from understanding the mechanisms underlying this network in plants, and to date only a few of the key players in the signalling processes, such as protein kinases, have been identified (Yeh et al., 2003). Interestingly, several genes induced in response to heavy metal stress are also induced during leaf senescence, which is the final stage of leaf development (Himelblau & Amasino, 2000). On the one hand, heavy metal stress often induces senescence-like degradation processes in plants (Gallego et al., 1996) and, on the other hand, heavy metals are liberated from their apoproteins during leaf senescence and have to be sequestered or transported to their new binding partners in other parts of the plant (Himelblau & Amasino, 2001). This indicates an overlap in the regulatory mechanisms underlying heavy metal homeostasis and leaf senescence. A common event during the exposure of plants to toxic concentrations of heavy metals and also during induction of leaf senescence seems to be the accumulation of ROS such as ·O2−, H2O2 and ·OH (Krupinska et al., 2003; Mithöfer et al., 2004). In addition to their harmful oxidative effects on biomolecules, these ROS have been suggested to function as signals in stress response and development (Mittler et al., 2004). Downstream signalling events associated with ROS also involve changes in Ca2+ concentrations, which serve as a general messenger in plant responses to biotic and abiotic stresses (Reddy, 2001; Sanders et al., 2002). The wide spectrum of stimuli evoking changes in calcium concentration also comprises abiotic stresses such as cold, drought, oxidative stress and aluminium exposure (Reddy, 2001). The complex spatio-temporal calcium signatures are decoded by a set of calcium-binding proteins including calmodulin, calmodulin-binding proteins and other calcium-binding proteins, for example C2-domain proteins (Reddy, 2001; Kim et al., 2003; Evans et al., 2004).
In this report, we describe the characterization of a cDNA induced by chromium and other heavy metals and also during natural senescence. The gene, named ‘C2 domain 1’ (HvC2d1), codes for a protein with 326 amino acids which includes a C2 domain-like motif at the N-terminal and a nuclear localization signal (NLS) at the C-terminal end. Our data indicate a possible calcium-dependent signal transduction pathway chain involving HvC2d1 during heavy metal stress and senescence.
Materials and Methods
Barley (Hordeum vulgare L. cv. Steffi) was used in this study. Seedlings were grown hydroponically on Murashige and Skoog (MS) medium (Duchefa Biochemie bv, Haarlem, the Netherlands) for 7 d under controlled growth-chamber conditions [16 h at 21°C and a photosynthetic photon fluence rate (PPFR), 400–700 nm of 100 µmol m−2 s−1; 8 h at 16°C in the dark], and then treated either with 1 mm potassium dichromate (K2Cr2O7) for various times (1, 5, 10, 24, 48, 72 and 96 h) or with cadmium chloride (CdCl2) or copper chloride (CuCl2) (for 48, 72 and 96 h) or not treated (controls). For analysis of senescence-dependent expression of HvC2d1, encoding the C2 domain 1 protein, plants were grown for 9, 26 and 38 d in 16 h light (21°C with a PPFR (400–700 nm) of 100 µmol m−2 s−1) and 8 h darkness (16°C) on soil containing 4 g of fertilizer (Osmocote 5 m; Urania, Hamburg, Germany) per litre of soil. In order to analyse the effects of calcium on expression of HvC2d1, primary leaves of 7-d-old barley plants were cut and then immersed in water containing 200 µm of calcium ionophore A23187 for 5, 10, 24 and 48 h. For application of calcium ionophore A23187, a stock solution of 1 mm was prepared in dimethylsulphoxid (DMSO) (Roth, Karlsruhe, Germany) and then dissolved in distilled water by rapid mixing following the instructions of the manufacturer (Sigma-Aldrich, Taufkirchen, Germany). Controls were treated in the same way except for addition of calcium ionophore A23187. Samples of leaves were harvested at appropriate times, immediately frozen in liquid nitrogen and stored at −80°C until use. In order to analyse the effects of reactive oxygen species (ROS), barley plants were grown hydroponically as described. Then 50 µm methylviologen in 0.1%[volume : volume (v : v)] Tween 20 was sprayed onto the leaves. Control plants were treated with 0.1% (v : v) Tween 20 only. After an incubation of 1 h in the dark, for improved uptake of the herbicide, plants were exposed to a PPFR (400–700 nm) of 300 µmol m−2 s−1 to induce accumulation of ROS. The effect of methylviologen treatment was characterized by measuring the decrease in photosystem II (PSII) efficiency, as described below (data not shown).
Relative chlorophyll content per unit leaf area was determined in the middle region of intact leaves during 0–144 h of chromium treatment using a soil plant analysis development (SPAD) analyser (Minolta; Hydro Agri, Dülmen, Germany) which measures transmission of wavelengths absorbed by chlorophyll in the intact leaves. Each data-point represents the mean of 10 independent measurements.
Photosystem II efficiency
Chlorophyll fluorescence was measured in the middle region of intact leaves after dark adaptation, as described by Humbeck et al. (1996), using a chlorophyll fluorometer (Mini PAM; Walz, Effeltrich, Germany). Mean values of the ratio of variable fluorescence to maximal fluorescence (Fv/Fm), which is a measure of PSII efficiency, are based on 10 independent measurements.
Total RNA was extracted from leaf tissue using a method described by Chirgwin et al. (1979) and was quantified spectrophotometrically. To verify the quality of RNA, 10 µg of total RNA was fractionated on a 1%[weight : volume (w : v)] agarose gel containing 4% formaldehyde, stained with ethidium bromide and then visualized under UV light.
Restriction fragment differential display PCR (RFDD-PCR)
The RFDD-PCR was performed according to the instructions for the Display Profile Expression Profiling Kit (Qbiogene GmbH, Heidelberg, Germany). Poly(A)+ RNA was isolated from the total RNA sample using PolyATtract mRNA Isolation System IV (Promega, Madison, WI, USA). cDNA was synthesized and digested with TaqI enzyme, adaptors were ligated and finally PCR was performed with 33P-labelled primers.
The populations of cDNA fragments of each sample were loaded on a 8% (w/v) denaturing polyacrylamide gel and autoradiographed with Kodak Biomax MR-film (Eastman Kodak, Rochester, NY, USA). The cDNA fragments differentially expressed were eluted by boiling the gel pieces in 10 mm Tris and 1 mm EDTA (TE) for 10 min at 95°C, re-amplified by PCR with the same pair of primers as used for the first amplification, cloned using the pGEM-T® Vector System I (Promega) and sequenced.
Nucleotide sequences were determined by the dideoxy chain-termination method using the BigDye® Terminator v1.1 Cycler Sequencing Kit (Applied Biosystems, Forster City, CA, USA) with an ABI Prism™ 370 automatic DNA sequencer (Applied Biosystems), and analysed using Lasergene expert sequence analysis software (DNASTAR Inc., Madison, WI, USA). The sequencing primers pUC/M13 forward (24 bp) and pUC/M13 reverse (22 bp) (Promega) and the specific primer 5′-GGT ACG GGA CGA CGG GCG TGG C were used for sequencing of cDNA clone HV_Ceb0017E22f (BF064709). Primers NewT7 5′-GTA ATA CGA CTC ACT ATA GGG C and SP6 5′-AGC TAT TTA GGT GAC ACT ATA G were used for sequencing of cDNA fragments cloned into the pGEM-T® vector.
To produce a recombinant GST-HvC2d1 protein in Escherichia coli, the full-length coding region [open reading frame (ORF)] of HvC2d1 was amplified using cDNA clone HV_Ceb0017E22f (BF064709) and primers HvC2d1 (forward) 5′-GAG AAT TCA TGG GCT CGC GGT ACG AGG TGG AGG TGA C with an EcoRI site and HvC2d1 (reverse) 5′-GAC TCG AGC TAG TAG TCG TCG TCG CCG CCG TAG TC with an XhoI site. The PCR product was ligated into the EcoRI/XhoI site of a pGEX-2TK vector (Amersham Biosciences, Freiburg, Germany). The recombinant plasmid was introduced into E. coli Rosetta (DE 3) pLys S (Novagen, Darmstadt, Germany), and overexpression of the protein was induced by addition of 0.1 mm isopropyl β-D-thiogalactoside (IPTG; Duchefa Biochemie bv) to the culture medium. After incubation at 30°C for 2 h, bacterial cultures were harvested and resuspended in phosphate-buffered saline (PBS) [150 mm NaCl, 16 mm Na2HPO4, 4 mm KH2PO4 and 2% (v/v) Triton X-100] before lysis by sonification. The recombinant protein GST-HvC2d1 was purified in a Glutathione-Sepharose 4B affinity column (Amersham Biosciences).
Ca2+-binding assay for HvC2d1
The ability of HvC2d1 to bind calcium was confirmed by 45Ca2+ overlay assay using the method described by Maruyama et al. (1984). Different amounts (2, 4, 8 and 12 µg) of the purified HvC2d1 protein, of calmodulin (Sigma-Aldrich) and of bovin serum albumin (BSA; Sigma-Aldrich) were blotted onto nitrocellulose transfer membranes (Protran, Dassel, Germany) using a dot-blot apparatus (Minifold I; Schleicher & Schuell, Dassel, Germany). The membrane sheet was washed four times with 10 mm MES-KOH (pH 6.5), 5 mm MgCl2 and 60 mm KCl. Then the membrane was incubated in the same buffer supplemented with 37 KBq ml−1 45Ca2+ (45CaCl2; Amersham Biosciences) at 23°C for 10 min. The membrane was washed three times with 50% (v : v) ethanol and dried at room temperature, then exposed to an imaging plate for 3 d and analysed using a fluorescent image analyser (FLA-3000 Series; Fuji Photo Film Co., Tokyo, Japan).
Subcellular localization of HvC2d1-GFP
The full-length coding region (ORF) of HvC2d1 was amplified using cDNA clone HV_Ceb0017E22f (BF064709) and PCR primers HvC2d1 (forward) 5′-GAG AAT TCA TGG GCT CGC GGT ACG AGG TGG AGG TGA C with an EcoRI site at the 5′ end and HvC2d1 (reverse) 5′-GAA GAT CTG TAG TCG TCG TCG CCG CCG TAG TCG C with a Bgl II site at the 5′ end. The PCR product was digested with EcoRI and Bgl II enzymes and cloned into the EcoRI and BamHI sites of the smRSGFP-pKE4xtr-G construct to generate HvC2d1-smRSGFP (smRSGFP = solubility modified red shifted-green fluorescent protein) under the control of the cauliflower mosaic virus 35S promoter. The HvC2d1-GFP and control smRSGFP constructs were used for transformation of onion epidermal cells using a custom-made vacuum helium particle gun as described by Barth et al. (2004). The epidermal layers isolated from onions were placed on MS basal medium containing 2% (w : v) plant agar, 2.5 µg ml−1 amphotericin B and 5 µg ml−1 chloramphenicol (Duchefa Biochemie bv). The transformed onion cells were either treated with 30 µm calcium ionophore or not treated and were incubated in the dark at 28°C for 10–12 h. GFP fluorescence was determined by fluorescein isothiocyanate (FITC)-filtered visual inspection using a confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany).
Isolation of a cDNA for a C2 domain-like protein by RFDD-PCR of chromium-stressed barley leaves
To identify genes that respond to chromium in barley leaves, we used RFDD-PCR. Hydroponically grown seedlings were treated with 1 mm potassium dichromate (K2Cr2O7) and the stress response of primary leaves was analysed using the photosynthesis-related parameters chlorophyll content and PSII efficiency (Fig. 1). Chlorophyll content during the first 24 h of chromium stress remained almost constant. Only after this time-point did the chlorophyll content start to decrease. Over a long period of exposure to potassium dichromate (144 h), the chlorophyll content was clearly reduced to c. 60% of that in controls not treated with chromium, indicating stress-dependent chlorophyll degradation in the later stages of the treatment. Another sensitive parameter reflecting stress-induced damage to photosynthetic activities is PSII efficiency measured after dark adaptation with a chlorophyll fluorometer (Humbeck et al., 1996). Only after 24 h of chromium treatment, PSII efficiency started to decrease.
The changes in photosynthetic parameters indicated that, under the stress conditions used in this experiment, leaves showed no response during the first 24 h, and there were slight responses in both parameters only 48 h after the onset of chromium stress. In the later stages of this chromium stress, leaves were severely damaged. In order to identify genes that are either up-regulated before photosynthetic functions are severely reduced or induced at the time of the very first effects of chromium stress on chlorophyll content and PSII efficiency, mRNAs from 48-h stressed leaves and of corresponding controls were extracted and cDNAs prepared. The cDNA populations were separated on 8% (w : v) polyacrylamide gels and compared using the differential display approach. This method yielded 48 cDNA fragments, presumably representing genes differentially expressed during the first phase of chromium treatment. After sequence analyses, these novel chromium-induced genes could be classified into the following groups: (i) genes possibly involved in signalling, pathogen responses and amino acid synthesis, and (ii) genes already known to be involved in the heavy metal response, such as that encoding the wheat aluminium-induced protein (wali3) (Snowden & Gardner, 1993). One cDNA clone (AJ630120; 81 bp) was investigated in more detail in this report. Using the software HarvEST Triticeae version 0.99 from the University of California (Los Angeles, CA, USA), 100% identity to a part of the barley cDNA clone HV_Ceb0017E22f (BF064709) was demonstrated. This clone was then sequenced. The full-length sequence was 1274 bp, including an ORF of 978 bp (nucleotides 84–1062; CAI58613) coding for 326 amino acids. By analysis of the deduced amino acid sequence, a C2 domain-like motif comprising amino acids 7–95 was identified showing homology to the known C2 domain sequence (pfam00168; National Center for Biotechnology Information [NCBI] Conserved Domain Search, shown in Fig. 2a). The gene was therefore named C2 domain 1 (HvC2d1).
C2 domains are known to interact with membranes in a Ca2+-dependent manner and are found in various types of protein, for example protein kinases, phospholipid-modifying enzymes such as phospholipase D, and so-called small C2-domain proteins with a single C2 domain (Kim et al., 2003). On the basis of its sequence, which shows only a single C2 domain-like motif and no other conserved protein domains, the novel barley C2-domain protein HvC2d1 investigated in this report is similar to the small C2-domain protein class, which has only been found in a limited number of plants, such as Arabidopsis (AtC2-1; AAG52148), maize (Zea mays) (ZmC2-1; U64437), pumpkin (Cucurbita maxima) (CmPp16-1 and 16-2; AF079170 and AF079171, respectively) and rice (Oryza sativa) (OsERG1a and OsERG1b; U95135 and U95136, respectively) (Kim et al., 2003). Alignment of the sequences of these proteins with that of the novel barley C2 domain-like protein revealed clear homology in the C2 domain. In addition to the C2 domain (at the N-terminus from Val7 to Arg95), at the C-terminal end a NLS (a K-rich region from Lys264 to Lys267) was identified (Fig. 2c).
The novel HvC2d1 has a much higher molecular weight than the known plant small C2-domain proteins, and shows the highest homology to genes with unknown function that have only recently been added to the NCBI database (OsBAD09616, AtC2-2, OsBAB84404 and At1g07310). As shown in the phylogenetic tree in Fig. 2(b), these genes cluster in their own novel C2 domain-like group, separately from the known small C2-domain proteins. Alignment of this novel class of C2 domain-like proteins shows high homology in conserved protein sequence areas, as presented in Fig. 3. Interestingly, the deduced amino acid sequence of another protein in this novel C2 domain-like class also exhibits a NLS-like motif (Fig. 3).
Transient expression pattern of HvC2d1 during chromium, copper and cadmium treatment
To analyse the heavy metal stress-dependent expression of HvC2d1, 7-d-old barley seedlings were treated with 1 mm potassium dichromate or were not treated (controls). The changes in mRNA levels were investigated via quantitative real-time PCR by comparison of mRNA levels of the HvC2d1 gene to levels of the reference 18S ribosomal RNA from treated samples and controls at different time-points. In Fig. 4 this expression rate for HvC2d1 is referred to that of the untreated control, which at each time-point is set as 1. The HvC2d1 mRNA had already started to accumulate after 10 h and reached a maximum transcript level 24 h after onset of stress (6.4 times higher than in the control; Fig. 4). In the later stages of the treatment, mRNA levels decreased again, reaching a basal transcript level after c. 96 h. These data show a transient expression pattern of HvC2d1 during the initial phase of chromium stress.
In order to determine whether HvC2d1 expression is also induced in response to heavy metals other than chromium, 7-d-old barley seedlings were either treated with 1 mm cadmium chloride or copper chloride for 48, 72 or 96 h or were not treated (controls). The changes in expression levels of HvC2d1 in the barley leaves exposed to cadmium chloride or copper chloride measured by quantitative real-time PCR (compared with the controls) are shown in Fig. 5. The mRNA levels of HvC2d1 exhibited similar changes during cadmium or copper treatment. In both treatments, HvC2d1 was significantly induced after 48 h. At this time, the HvC2d1 transcript level was 2.8 times higher in the copper-treated seedlings and 15.2 times higher in the cadmium-treated seedlings compared with the untreated control (Fig. 5). Similar to the chromium treatment, in the later stages the relative mRNA levels of HvC2d1 declined again, also showing a transient expression pattern after exposure to essential (copper) and nonessential (cadmium) heavy metals.
HvC2d1 is also induced during leaf senescence but not by drought stress
It is known that some heavy metal-induced genes, such as methallothionein 1 (MT1) and blue copper-binding protein (BCB), are also up-regulated during leaf senescence (Himelblau & Amasino, 2000). In order to determine whether this is also true for HvC2d1, barley plants were grown for 9, 26 and 38 d in 16 h light (21°C with a PPFR [400–700 nm] of 100 µmol m−2 s−1) and 8 h darkness (16°C) on soil as described in Materials and Methods. As characterized by measurements of chlorophyll content and PSII efficiency (data not shown), primary leaves of 9-d-old seedlings were in the late developmental stage and those of 26- and 38-d-old seedlings were in the early and late senescence stages, respectively, with decreased chlorophyll content and photosynthetic activities. Our data indicate that the mRNA level of HvC2d1 was significantly increased during senescence (Fig. 6a).
We further examined the effect of drought stress on transcript levels of HvC2d1, as the VrPLC3 (phospholipid C) gene in Vigna radiata L. containing a C2 domain is induced under drought stress (Kim et al., 2004). Barley plants were grown hydroponically on MS medium for 7 d and then removed from the medium to induce drought stress. The impact of drought stress on barley plants was characterized by measuring the decreasing water content in the leaves at different times (data not shown). Our data indicated that HvC2d1 transcripts do not accumulate but rather decrease during drought stress (Fig. 6b).
Expression of HvC2d1 responds to changes in cytosolic calcium and is affected by abscisic acid and methylviologen
In plants, calcium is a possible mediator of extracellular signals such as hormones and biotic and abiotic stressors (Takezawa, 1999). It is known that the C2 domain is a Ca2+-activated membrane-targeting motif present in a wide range of Ca2+-regulated proteins. In order to investigate whether our newly identified C2 domain-like HvC2d1 also responds to changes in the Ca2+ concentration, we investigated the effects of calcium ionophore A23187, which induces an increase in cytosolic calcium concentration (Takezawa, 1999; Kim et al., 2003). Primary leaves from 7-d-old barley plants were cut and then treated with 200 µm of calcium ionophore A23187 or not treated (controls) for 5, 10, 24 and 48 h as described in Materials and Methods. Figure 7(a) shows that addition of calcium ionophore A23187 resulted in a clear increase in the level of HvC2d1 mRNA, which started to accumulate significantly 5 h after addition of the calcium ionophore. This result suggests that calcium signalling is involved in the regulation of HvC2d1.
Calcium acts as an intracellular messenger in many phytohormone signalling processes, including that involving abscisic acid (ABA), which plays a major role in stress responses. In these processes, the Ca2+ signal is often triggered by secondary messengers such as cyclic ADP ribose (cADPR), inositol 1,4,5 trisphosphate (InsP3) and myo-inositol hexakiphosphate (InsP6) or ROS such as H2O2 (Himmelbach et al., 2003). It is also known that the small C2-domain containing protein OsERG1 is induced in response to H2O2 (Kim et al., 2004). In order to clarify whether ABA and/or ROS is also involved in the regulation of HvC2d1, two experiments were performed with ABA and methylviologen, which generates ROS. For the first experiment, primary leaves from 7-d-old barley plants were cut, then exposed in beakers to 50 µm ABA with 1% (v : v) ethanol for 8, 12 and 48 h or not exposed to ABA (controls). Figure 7(b) shows that the expression level of HvC2d1 was more than doubled in response to ABA treatment.
For methylviologen application, 7-d-old barley plants cultivated hydroponically as described in Materials and Methods were sprayed with 50 µm methylviologen in 0.1% (v : v) Tween 20, then incubated for 1 h in the dark for improved uptake of methylviologen. Plants were then exposed to a PPFR (400–700 nm) of 300 µmol m−2 s−1 to induce oxidative stress for 1, 3, 5 and 6 h. The control was treated only with 0.1% (v : v) Tween 20. As shown in Fig. 7(c), the treatment of plants with methylviologen clearly causes the accumulation of HvC2d1 mRNA.
Confirmation of calcium binding of HvC2d1 by 45Ca2+ overlay analysis
The ability of HvC2d1 to bind Ca2+ was tested in a 45Ca2+ overlay analysis. Recombinant GST-HvC2d1 protein was overexpressed in E. coli and purified via glutathione-sepharose affinity chromatography. Different amounts of the purified protein, of BSA and of calmodulin were blotted to nitrocellulose membranes, exposed to 45Ca2+ and then washed to remove nonspecifically bound calcium. As shown in Fig. 8, HvC2d1 and the known calcium-binding protein calmodulin clearly bound 45Ca2+ whereas the negative control BSA did not.
Ca2+-dependent subcellular localization of HvC2d1-GFP constructs
It is known that C2-domain proteins play a role in accurate Ca2+-dependent spatio-temporal targeting in different regulatory signal transduction chains (Evans et al., 2004). Furthermore, it has been shown that the small rice C2-domain protein OsERG1 is translocated to the plasma membrane of plant cells in a Ca2+-dependent manner (Kim et al., 2004). In order to determine whether the newly identified C2-domain protein HvC2d1 also exhibits calcium-dependent subcellular localization, HvC2d1-smRSGFP constructs were, after particle bombardment, overexpressed in onion epidermal cells for 12 h either with or without the addition of calcium ionophore A23187. As an additional control, smRSGFP alone was also overexpressed. Subcellular localization was then analysed using a confocal laser scanning microscope (Fig. 9). SmRSGFP alone was localized at the plasma membrane, the cytoplasm and the nucleus (Fig. 9a). Treatment of these cells overexpressing smRSGFP with the calcium ionophore A23178 did not change its distribution (data not shown). All cells transformed with HvC2d1-smRSGFP showed a similar pattern of localization at the plasma membrane, the cytosole and the nucleus to smRSGFP alone. In contrast, after treatment with calcium ionophore A23178, green fluorescence was only seen in the nucleus, with higher green fluorescence intensity in the nucleoli of the examined cells (an example is shown in Fig. 9c). In all the cells analysed, we never observed green fluorescence at the plasma membrane, as was always seen after transformation with either smRSGFP alone or the HvC2d1-smRSGFP construct without calcium ionophore A23178 treatment. These results indicate a calcium-dependent translocation of the HvC2d1 protein to the nucleus, as expected from the presence of a NLS motif in the deduced amino acid sequence (shown in Fig. 2c).
In order to identify factors involved in the response of plants to chromium, transcript populations of 48-h chromium-treated barley leaves were compared with controls by RFDD-PCR. This experimental approach yielded 48 cDNA fragments representing genes up-regulated during the early phase of chromium treatment when leaves were not yet severely damaged, as shown by the photosynthesis parameters chlorophyll content and PSII efficiency (Fig. 1). In the present report, we focused on one of these genes which, after full-length sequencing, was found to exhibit a NLS and a conserved calcium binding C2 domain-like motif (Fig. 2c). The expression studies (Fig. 4) showed that the mRNA of this gene transiently accumulated in the early phase of chromium treatment, and sequence alignment revealed that the novel chromium-induced gene is similar to a subgroup of C2-domain genes coding for the so-called small C2-domain proteins which, in contrast to other known C2-domain proteins, contain only a single C2 domain and do not have further conserved domain motifs such as, for example, protein kinase domains (Kim et al., 2003). Only a few small C2-domain proteins are known in plants and none have been found in animals (Kim et al., 2003). The functions of the already described small C2-domain proteins are not yet clear. In pumpkin, a small C2-domain protein has been reported to increase the size of mesophyll plasmodesmata to enable transport of cellular materials, including RNA molecules, from cell to cell (Xoconostle-Cazares et al., 1999). In rice, the isolation and characterization of two small C2-domain rice proteins, named OsERG1a and OsERG1b, which are significantly induced by a fungal elicitor have been reported (Kim et al., 2003). Kim et al. (2003) proposed that small C2-domain proteins play a functional role in defence signalling systems in plant cells. There are no other reports on the function of small C2-domain proteins in plants. The novel protein HvC2d1 has a much higher molecular weight than the known small C2-domain proteins. Four other sequences recently added to the NCBI databases show these characteristic features of HvC2d1 (AtC2-2, At1g07310, OsBAD09616 and OsBAB84404 in Fig. 3), and cluster in their own group (Fig. 2b). Interestingly, one of these sequences also has a NLS-like motif (OsBAD09616).
Many but not all known C2-domain proteins have conserved aspartate residues implicated in Ca2+ coordination (amino acids 15, 21, 71, 73, 78 and 79 in the consensus C2-domain pfam00168). As shown in Fig. 2(a), the novel protein HvC2d1, the Arabidopsis proteins AtC2-2 (AAV85706) and At1g07310, and also the rice proteins OsBAD09616 and OsBAB84404 do not exactly follow this pattern but have either aspartates nearby or other amino acids which might also be capable of Ca2+ coordination (threonine, glutamine, serine and asparagine). In spite of this variability, HvC2d1 is clearly able to bind calcium, as shown by the 45Ca2+ overlay analysis (Fig. 8), indicating the presence of a functional C2 domain. Such variability in the Ca2+-binding sites of different C2-domain proteins is also found in other C2-domain proteins and seems to be important for the specialization of these different C2 domains, presumably to provide optimized Ca2+-binding parameters, to allow specific changes in conformation upon Ca2+ binding or docking interactions for different biological functions (Nalefski & Falke, 1996). From work with animal systems it is known that, in general, the C2-domain is a Ca2+-dependent membrane-targeting module found in many cellular proteins involved in signal transduction or membrane trafficking (Nalefski & Falke, 1996; Cho, 2001). This membrane-targeting domain shows a wide range of lipid selectivity for the major components of cell membranes (Stahelin et al., 2003). The majority of C2 domains bind the membrane in a Ca2+-dependent manner and thereby play an important role in Ca2+-dependent membrane targeting of peripheral proteins (Stahlen & Cho, 2001). Recent investigations of Ca2+-activated membrane docking of the C2-domain protein phospholipase A2 have revealed an ordered binding of two Ca2+ ions with positive cooperativity (Malmberg et al., 2004).
The present study of the novel barley gene HvC2d1 has revealed a stress- and development-dependent expression pattern. We have shown, for the first time, induction of a gene encoding a C2-domain protein during heavy metal stress and leaf senescence. Some senescence-associated genes (SAGs) are also induced after exposure of the plant to heavy metals (Buchanan-Wollaston & Ainsworth, 1997; Himelblau & Amasino, 2000). The reason for these overlapping expression patterns might be that during leaf senescence proteins, including those containing metals, are degraded. The liberated metals have to be sequestered and a certain amount of these metals is transported to the growing tissues of the plant (Himelblau & Amasino, 2001). Therefore, the same regulatory factors might be involved in both the response of plants to heavy metals and leaf senescence.
Interestingly, HvC2d1 is induced by ABA but not by drought, indicating a signalling pathway different from that of many other genes regulated by ABA. A similar expression pattern was shown for a pathogen-related gene that was induced by high salinity, ABA and wounding, but not by drought and cold stress (Jung et al., 2004). Not much is known about the signalling pathways underlying the induction of heavy metal-induced genes or the onset of leaf senescence. One common signal could be ROS, which accumulate in response to abiotic and biotic stressors (Neill et al., 2002; Mithöfer et al., 2004) and have also been suggested to play a role in the induction of leaf senescence (Ye et al., 2000; Krupinska et al., 2003). In order to determine whether accumulation of ROS also induces the novel HvC2d1 gene, plants were treated with methylviologen to generate superoxid anion radicals by uptake of an electron from photosystem I (Donahue et al., 1997). Our data showed a clear induction of HvC2d1 during methylviologen treatment, indicating that ROS are involved in the signalling pathway.
Another secondary signal involved in stress and developmental regulation is Ca2+ (Knight, 2000). It has been shown that calcium-dependent kinases play a role in different stress reactions in plants (Cheng et al., 2002). Furthermore, there are several reports indicating that signalling via Ca2+ and via ROS is connected (Himmelbach et al., 2003; Olmos et al., 2003). It is also known that the small C2-domain containing protein OsERG1 is induced in response to H2O2 (Kim et al., 2003). Our data concerning the novel heavy metal- and senescence-induced gene HvC2d1, which has a calcium-related C2 domain and is induced during accumulation of ROS, also indicate that calcium and ROS might be involved in the underlying signalling pathways. Such C2-domain proteins are known to be expressed in a Ca2+-dependent manner (Kim et al., 2003). We also clearly showed that the novel gene HvC2d1 is induced in response to calcium ionophore A23187, which is known to enhance calcium influx (Torrecilla et al., 2001). Known C2-domain proteins are involved in Ca2+ signalling or membrane trafficking processes mediated by Ca2+-dependent binding of C2 domains to membrane phospholipids (Kim et al., 2003). By using GFP constructs, Kim et al. (2003) showed that the small C2-domain protein OsERG1 is translocated to the plasma membrane of plant cells by treatment with a Ca2+ ionophore and also by a fungal elicitor. Immunocytochemical analyses with the other known small C2-domain protein from pumpkin (CmPP16-1) also suggested an association with the plasma membrane (Xoconostle-Cazares et al., 1999). Analyses of mammalian C2 proteins, for example phospholipases, synaptogamin I and protein kinase C, also showed that these proteins migrate after binding of Ca2+ from the cytosol to the plasma membrane and thus are able to transduce foreign signals into the cell (Pepio & Sossin, 2001; Ananthanarayanan et al., 2002; Teruel & Meyer, 2002). In contrast to these other C2-domain proteins, our novel HvC2d1 protein contains a NLS, indicating nuclear localization. Consistent with the presence of this motif, our localization studies using HvC2d1-GFP constructs showed a calcium-dependent nuclear localization. This is the first time that a calcium-dependent translocation of a C2-domain protein to the nucleus has been shown. It is well known that cytosolic Ca2+ fluctuations act as intracellular secondary messengers in the responses of plants to a number of stimuli, including ABA, osmotic stress, ionic stress and oxidative stress (Knight, 2000; Sanders et al., 2002). In these signalling processes the kinetics, amplitude and duration of Ca2+ transients are important for the transmission of specific information (Allen et al., 2000; Rudd & Franklin-Tong, 2001; Sanders et al., 2002). Recent data have shown that Ca2+ concentrations in response to different external stimuli vary not only in the cytosol but also in other cell compartments such as chloroplasts (Johnson et al., 1995) and the nucleus (Pauly et al., 2000; Xiong et al., 2004), indicating complex spatio-temporal characteristics of Ca2+ signatures. In the Ca2+-mediated signalling networks in plants, the calcium signatures are decoded by calcium sensors such as calmodulin and other calcium-binding proteins (Yang & Poovaiah, 2003). Interestingly, there are several reports showing that such calcium-binding proteins are not only localized in the cytoplasm but also in the nucleus. Among these proteins with nuclear localization are calcium-dependent protein kinases (Dammann et al., 2003; Chehab et al., 2004), a novel calmodulin-binding protein (Perruc et al., 2004) and an activator of a H+/Ca2+ antiporter (Cheng et al., 2004). In addition, the calcium-dependent protein kinase McCPK1 from the ice plant was shown to undergo a reversible change in subcellular localization from the plasma membrane to the nucleus, endoplasmatic reticulum and actin filaments of the cytoskeleton in response to environmental stimuli (Chehab et al., 2004). In the present study we identified another novel protein with a calcium-binding C2 domain-like motif which also shows, in a calcium-dependent manner, localization in the nucleus. From our results we cannot definitely conclude that the nuclear localization is a direct response of the binding of calcium to the C2 domain. Other calcium-dependent factors could also be involved in this process. However, despite this uncertainty, our results strongly indicate that HvC2d1, in the context of other calcium-dependent factors, plays a role in nuclear localized calcium signalling processes both in response to external stressors such as heavy metals and also in the specific developmental phase of senescence. Further studies are needed to elucidate the function of this novel C2-domain protein in the calcium signalling network.
We thank Nicole Sommer for providing RNA samples from senescent primary leaves, Wiebke Zschiesche for providing RNA samples from barley plants treated with cadmium and copper, Kathleen Clauss for providing RNA from methylviologen treatment and Olaf Barth for technical help with sequence alignments. The German Academic Exchange Service (DAAD) and the German Research Foundation (DFG) are thanked for financial support.