SRK2C, a SNF1-related protein kinase 2, improves drought tolerance by controlling stress-responsive gene expression in Arabidopsis thaliana

Abstract

Protein phosphorylation/dephosphorylation are major signaling events induced by osmotic stress in higher plants. Here, we showed that a SNF1-related protein kinase 2 (SnRK2), SRK2C, is an osmotic-stress-activated protein kinase in Arabidopsis thaliana that can significantly impact drought tolerance of Arabidopsis plants. Knockout mutants of SRK2C exhibited drought hypersensitivity in their roots, suggesting that SRK2C is a positive regulator of drought tolerance in Arabidopsis roots. Additionally, transgenic plants with CaMV35S promoter::SRK2C-GFP displayed higher overall drought tolerance than control plants. Whereas stomatal regulation in 35S::SRK2C-GFP plants was not altered, microarray analysis revealed that their drought tolerance coincided with up-regulation of many stress-responsive genes, for example, RD29A, COR15A, and DREB1A/CBF3. From these results, we concluded that SRK2C is capable of mediating signals initiated during drought stress, resulting in appropriate gene expression. Our present study reveals new insights around signal output from osmotic-stress-activated SnRK2 protein kinase as well as supporting feasibility of manipulating SnRK2 toward improving plant osmotic-stress tolerance.

Drought or high osmolarity are critical environmental factors that limit agricultural production worldwide. Drought or osmotic stress causes a series of morphological, physiological, biochemical, and molecular changes in plants. Expression of a number of genes play a crucial role in plant stress responses by driving adaptive responses, such as accumulation of compatible solutes, like proline and sugars, and alteration of lipid composition (1, 2). Extensive studies have elucidated the regulatory mechanisms of stress-responsive gene expression (2) and that manipulation of stress-responsive genes for transcription factors, such as dehydration-responsive element-binding factor (DREB)/C-repeat-binding factor (CBF), should prove useful for improving plant stress tolerance (3–8). However, signal transduction factors that function upstream of stress-responsive gene expression are still under investigation.

In eukaryotes, protein phosphorylation/dephosphorylation plays essential roles in the adaptation to hyperosmotic stress. For instance, a yeast mitogen-activated protein kinase (MAPK) cascade, named the HOG1 pathway, transduces stress signals from hyperosmolarity sensors, leading to expression of appropriate genes (9). In plants, activation of protein phosphorylation events has also been observed when plants are exposed to water deficit (10). Several protein kinases have been described as signal transduction factors related to osmotic-stress responses in plants, for example plant MAPKs are activated by abiotic stress (11–15). In addition to MAPKs, another protein kinase family has been recently identified as osmotic-stress-activated protein kinases in plants: the SNF1-related protein kinase 2 (SnRK2) family, which is a relatively small plant-specific gene family (i.e., 10 members designated as SRK2A-J in the Arabidopsis genome) (16, 17). Originally, the SnRK2 protein kinase was identified as a central regulator of abscisic acid (ABA)-dependent stomatal closure in fava bean designated AAPK (for ABA-activated protein kinase) (18, 19). Recently, one Arabidopsis SnRK2, SRK2E/OST1/SnRK2.6, which is an ortholog of AAPK, was identified as an ABA-activated protein kinase and a key regulator of stomatal closure in Arabidopsis (17, 20). Also, several reports indicate that some SnRK2 are also activated by hyperosmotic stress, for instance, in tobacco cells (13) or soybean (21). Among 10 members of the rice SnRK2 family, three were activated by exogenous ABA and osmotic stress, whereas others were activated solely by osmotic stress (22). These data suggest that the SnRK2 family may play an important role in osmotic-stress signaling; however, knowledge of the specific function of osmotic-stress-activated SnRK2 is still fragmentary. Questions, such as whether activation of SnRK2 regulates a plant's stress response or where SnRK2 transduces stress signals, still remain to be answered.

In the functional analysis of the Arabidopsis SnRK2 family, we found that several SnRK2 were activated by osmotic stress. Here, we demonstrate that activation of SRK2C/SnRK2.8 mediates drought-stress signaling and improves the drought tolerance of Arabidopsis plants. Overexpression of SRK2C enhanced drought tolerance significantly in concert with up-regulation of stress-responsive gene expression involving DREB1A/CBF3. We discuss the in vivo function of SRK2C relative to osmotic-stress response, as well as the possibility for biotechnology applications of SRK2C to improve the stress tolerance of plants, which is another type of regulon biotechnology that uses signaling factors.

Materials and Methods

Generation of Transgenic Plants or Cells. The coding region of SRK2C cDNA was amplified by RT-PCR by using primers 5′-TCTAGAATGGAGAGGTACGAAATAG-3′ (XbaI site is in italics) and 5′-GGATCCCAAAGGGGAAAGGAGATC-3′ (BamHI site is in italics) and cloned into pBE2113 vector (23) as a GFP-fused fragment downstream of the CaMV35S promoter (17). The Prosrk2c::GUS construct was obtained by amplifying 2 kb of the SRK2C promoter region by using oligonucleotides 5′-AAGCTTGCATTAGTTTTCTTGATG-3′ (HindIII site is in italics) and 5′-GGATCCGATTCTACAAACTGCAAC-3′ (BamHI site is in italics) and cloning the amplified DNA into the pBI101 vector. Arabidopsis thaliana (Columbia ecotype) or T87 cultured cells were transformed by using Agrobacterium tume-faciens C58 harboring a 35S::GFP, 35S::SRK2C-GFP, or Prosrk2c::GUS construct as described in ref. 17.

Isolation and Analysis of T-DNA Insertional Mutants. Two SRK2C insertion alleles were registered in the Salk Institute Genomic Analysis Laboratory (SIGnAL) T-DNA Express Database (http://signal.salk.edu), and we obtained those lines from the Arabidopsis Biological Resource Center (Columbus, OH). After isolation of plants homozygous for the T-DNA [a portion of the Ti (tumor-inducing) plasmid that is transferred to plant cells] insertion, the two lines were designated srk2c-1 and srk2c-2. One-week-old seedlings of WT, srk2c-1, and srk2c-2 plants were carefully pulled out from agar plates and exposed to drought stress on filter paper for 45 min. Seedlings were placed on germination medium agar plates (24), and grown on the plate standing vertically. After 6 days, root elongation and the number of lateral roots were measured. Water loss was measured as described in ref. 17. Experiments were performed in at least three replicates.

Cell Culture, Plant Growth, and Stress Treatments. T87 cultured cells were grown under conditions as described in ref. 17. Aliquots (1 ml) of T87 cells on day 10 in culture were exposed to 0.5 M NaCl, 0.8 M mannitol, 1 mM H₂O₂, or 5% glucose treatment or cold treatment on ice for 30 min and stored at –80°C. For time-course or dose-dependent analysis, aliquots of cell culture were treated with 0.2 M NaCl and harvested at indicated times. Arabidopsis plants were grown under conditions described in ref. 17. For drought treatment, seedlings were transferred onto paper towels and harvested after designated time points. Alternatively, plants were grown on soil until they were 4 weeks old and exposed to drought stress by withholding water supply. Drought-stressed plants were rewatered after 14 days, and the survival rate was calculated after 4 days from rewatering. Transpiration rate was measured with LI-6400 portable photosynthesis system (Li-Cor, Lincoln, NE) as described in ref. 25.

In-Gel Kinase Assay. Crude extracts from plants or cultured cells were prepared as described in ref. 11. Portions of the protein extracts (10–15 μg) were electrophoresed in 8–10% SDS-polyacrylamide gels that contained 0.5 mg/ml Histone III-S (Sigma) as a substrate. The in-gel kinase assay was performed as described in refs. 11 and 26.

Immunocomplex Kinase Assay. Anti-SRK2C polyclonal antibody was generated against a synthetic peptide corresponding to the C terminus of SRK2C (CDDLDTDFDDIDTADLLSPL). The synthetic peptide was conjugated with keyhole limpet hemocyanin carrier. Polyclonal antisera were raised in rabbits and purified by affinity chromatography (Medical and Biological Laboratories, Nagoya, Japan). Immunoprecipitation was performed as described in ref. 11. The precipitates with anti-SRK2C polyclonal antibody were separated by electrophoresis; then the gels were subjected to in-gel kinase assay or Western blot analysis according to standard protocol (27).

RNA Gel Blot Analysis. Total RNA was extracted from Arabidopsis seedlings or adult plants by using TRIzol Reagent (Invitrogen). RNA gel blot analysis was performed essentially as described in refs. 27 and 28.

Microarray Analysis and Data Mining. 35S::GFP and 35S::SRK2C-GFP plants were grown on soil for 4 weeks. Total RNA was isolated by using TRIzol reagent (Invitrogen) and used for preparation of Cy5-labeled and Cy3-labeled cDNA probes (29). An Arabidopsis 2 Oligo Microarray kit (Agilent Technologies, Palo Alto, CA) was used to compare transcription profiles of 35S::GFP and 35S::SRK2C-GFP plants. All microarray experiments, including the data analysis, were carried out as described in ref. 30. The data were compared with previous expression profiles collected from drought-, salt-, and cold-stressed Arabidopsis plants and DREB1A/CBF3 overexpressing plants (31).

Results

SRK2C Is an Osmotic-Stress-Activated Protein Kinase. SRK2C is a member of SnRK2 family in A. thaliana and categorized within the SnRK2a subfamily (Fig. 1A) (17). SRK2C is annotated as a putative serine/threonine protein kinase identical to At1g78290 according to the Munich Information Center for Protein Sequences A. thaliana Database (http://mips.gsf.de/proj/thal/db/index.html), and it is designated as SnRK2.8 or OSKL4 by other research groups (16, 20). The SRK2C cDNA contains an ORF of 1,032 bp and encodes a polypeptide of 344 amino acid residues with an estimated molecular mass of 38.2 kDa.

Fig. 1.

SRK2C is an osmotic-stress-activated protein kinase. (A) A phylogenic tree of the SnRK2 family from Arabidopsis and other plant species. Arabidopsis SnRK2 are drawn with solid lines and others are drawn with gray lines. SRK2C is enclosed in a box. (B) Activation of SRK2C-GFP in Arabidopsis T87 cultured cells by various stresses. T87 cells were transformed with pBE2113 vector harboring 35S::SRK2C-GFP and treated with 0.5 M NaCl, 0.8 M mannitol, 10 mM H₂O₂, or 5% glucose for 30 min or exposed to 4°C for 30 min. Protein kinase activities in cell extracts were monitored by an in-gel kinase assay with histone as the substrate. A solid arrow indicates the protein kinase activity of SRK2C-GFP, and open arrows indicate native protein kinases in T87 cells. (C) Time course of SRK2C activation. 35S::SRK2C-GFP cells were treated with 0.2 M NaCl, and aliquots of cell culture were taken at the indicated time points. Protein kinase activities were detected as described for B.(D) Dose response of SRK2C activation. 35S::SRK2C-GFP cells were treated with various concentrations of NaCl for 30 min, and kinase activities were detected as described for B. Molecular mass markers are indicated at the left and expressed in kilodaltons.

To evaluate the enzymatic activity of SRK2C, we performed an in-gel protein kinase assay employing SRK2C-GFP fusion protein overexpressed in Arabidopsis T87 cultured cells. SRK2C-GFP-overexpressing cells were exposed to 0.5 M NaCl, 0.8 M mannitol, 10 mM H₂O₂, or 5% glucose treatment or cold treatment on ice for 30 min, and crude extracts from stress-treated cells were subjected to an in-gel kinase assay. We found that the kinase activity of SRK2C-GFP was elevated by 0.5 M NaCl and 0.8 M mannitol treatment but not by 10 mM H₂O₂,5% glucose, or cold stress (Fig. 1B). Activation of SRK2C-GFP by NaCl was achieved within 2 min and reached the maximum level at 0.5–2 hr (Fig. 1C). SRK2C was activated by high NaCl concentrations (>100 mM) (Fig. 1D). These results indicate that SRK2C is an osmotic-stress-activated protein kinase in Arabidopsis.

In Arabidopsis plants, three bands of protein kinase activity ranging from 36 to 42 kDa were activated within 30 min by drought stress (Fig. 2A). Immunoprecipitates with anti-SRK2C polyclonal antibody showed that 36-kDa kinase activity was elevated in response to drought stress (Fig. 2 A). In addition, 36-kDa kinase activity was completely depleted in two independent T-DNA insertional mutant lines of SRK2C, designated srk2c-1 and srk2c-2 (Fig. 2 B and C). These data taken together suggest that SRK2C activity corresponded with the 36-kDa protein kinase activated by drought stress in Arabidopsis plants.

Fig. 2.

Multiple protein kinases, including SRK2C, are activated by drought stress in Arabidopsis plants. (A) In-gel kinase assay of crude extracts and immunoprecipitates with anti-SRK2C polyclonal antibody. Two-week-old seedlings were exposed to drought stress for 0, 0.5, and 1 hr. Crude extracts (Left) and immunoprecipitates (IP) (Center) with anti-SRK2C polyclonal antibody were subjected to an in-gel kinase assay with histone as the substrate. (Right) Amounts of SRK2C in immunoprecipitates were confirmed by Western blot analysis by using an anti-SRK2C polyclonal antibody. (B) Genomic structure of Arabidopsis SRK2C gene. SRK2C gene consists of six exons (closed boxes). The positions of T-DNA insertion in srk2c-1 and srk2c-2 are indicated (not to scale). (C) Drought stress-activated protein kinases in WT (Col-0) and srk2c mutants. Two-week-old seedlings of WT and srk2c mutants were exposed to drought stress for the indicated times, and protein kinase activities were detected by in-gel kinase assay, with histone as the substrate. An arrow indicates the position of kinase activity that is depleted in srk2c mutants. Molecular mass is indicated at the left and expressed in kilodaltons.

SRK2C Positively Regulates Drought Tolerance in Arabidopsis Roots. To determine tissue-specificity of SRK2C expression in Arabidopsis plants, gel blots of total RNA from roots, leaves, cauline leaves, flowers, siliques, and seeds were hybridized to a cDNA probe of SRK2C. SRK2C was expressed abundantly in roots and weakly in leaves and siliques (Fig. 3A). Tissue-specific expression of SRK2C was also analyzed by using transgenic Arabidopsis plants expressing the GUS reporter gene driven by the SRK2C promoter. Histochemical analysis detected significant GUS activity in root tips (Fig. 3B). These results suggest that SRK2C is an osmotic-stress-activated protein kinase that functions mainly in root tips.

Fig. 3.

SRK2C is mainly expressed in root tips. (A) RNA gel blot analysis was performed to analyze SRK2C expression in different tissues. Total RNA was extracted from roots (R), stems (St), rosette leaves (L), cauline leaves (CL), flowers (F), siliques (Si), and dry seeds (Sd). Ten micrograms of total RNA was blotted and hybridized with a ³²P-labeled SRK2C cDNA probe. Ribosomal RNAs stained with ethidium bromide (EtBr) are shown as a loading control. (B) Localization of SRK2C expression in root tips. Two-week-old transgenic plants expressing GUS reporter gene under SRK2C promoter were assayed for GUS activity. Each panel shows whole plant (a), root tip (b), and root stem (c) after GUS staining. Five independent lines were analyzed; of these, one representative line was photographed.

The effects of SRK2C disruption were evaluated by using two srk2c T-DNA-tagged mutants (Fig. 2 B and C). These mutants were exposed to drought stress for 45 min and then placed on agar plates to monitor their root growth. Root elongation and newly emerged lateral roots were measured and counted (respectively) after 6 days on agar. Although the growth of srk2c-1 and srk2c-2 was completely normal under nonstress conditions (Fig. 4 A and B), the root growth of srk2c mutants after drought treatment was inhibited significantly compared with controls. The root elongation and lateral root number of srk2c mutants were decreased to 21–24% (Fig. 4B) and 29–33% (Fig. 4C) relative to WT controls. By contrast, water-loss levels of srk2c mutants were similar to that of WT (Fig. 4D), suggesting that drought hypersensitivity of srk2c mutants was largely dependent on sensitivity of roots.

Fig. 4.

srk2c-1 and srk2c-2 mutants are sensitive to drought stress in roots. (A) Root elongation assay of WT and srk2c mutants after drought stress. (Upper) Twelve-day-old seedlings without stress treatment. (Lower) Eight-day-old seedlings were exposed to drought stress for 45 min then placed on agar plates for 4 days. (B) Root elongation of WT (Col, open bar) and srk2c mutants (solid bars) under normal conditions or after drought stress. Plants were exposed to drought stress as described above. Root elongation was measured at 6 days after stress treatment. Vertical bars indicate SE (n = 5). This result was confirmed in three independent experiments. (C) Lateral root number of WT and srk2c mutants after drought stress. Newly emerged lateral roots were counted at 6 days after stress treatment as described above. Vertical bars indicate SE (n = 5). This result was confirmed in three independent experiments. (D) Water loss in WT and srk2c mutants (mean ± SE, n = 4).

Overexpression of SRK2C-GFP Increases Drought Tolerance of Arabidopsis Plants. SRK2C-GFP fusion protein was overexpressed under a constitutive CaMV 35S promoter in Arabidopsis plants for a gain-of-function analysis of SRK2C. Mainly three independent transgenic lines (nos. 1–3) with different levels of the transgene were analyzed. Results of RNA gel-blot analysis indicated that SRK2C-GFP was expressed to various degrees in lines 1–3 (Fig. 5A). In-gel kinase assay revealed that the SRK2C-GFP protein was actually activated by drought stress in transgenic plants (Fig. 5B). These kinase activation data for SRK2C-GFP correlated well with expression levels determined by the RNA gel-blot analysis.

Fig. 5.

Elevated drought tolerance in SRK2C-GFP-overexpressing plants. (A) Expression analysis of SRK2C-GFP fusion gene. Total RNA was isolated from two-week-old seedlings of 35S::GFP and 35S::SRK2C-GFP (lines 1–3), and 10 μg of RNA was blotted and hybridized with a ³²P-labeled SRK2C cDNA probe. (B) Activation of SRK2C-GFP in 35S::SRK2C-GFP (lines 1–3) plants. Two-week-old seedlings of 35S::SRK2C-GFP plants were exposed to drought stress for 0, 0.5, and 1 hr. Protein kinase activities were detected by in-gel kinase assay with histone as the substrate. Molecular mass markers are given at the left in kilodaltons. (C) Drought tolerance of 35S::GFP (VC) and 35S::SRK2C-GFP (lines 1–3) plants. We withheld from watering 3-week-old plants to impose drought treatment. After 14 days, each plant was photographed. (D) Survival rate of 35S::GFP (VC) and 35S::SRK2C-GFP (lines 1–3) plants after 14 days of drought stress. The survival rate and SE were calculated based on three independent experiments. (E) Transpiration rates of 35S::GFP (VC) and 35S::SRK2C-GFP (lines 1–3) plants under normal (open bars) and drought (solid bars) conditions. Each data set is represented as mean ± SE (n = 5).

Although 35S::SRK2C-GFP plants exhibited weak late-flowering phenotypes (see Fig. 7 and Tables 1 and 2, which are published as supporting information on the PNAS web site), their growth and size were almost similar to those of 35S::GFP or WT plants (as shown in Fig. 5C). Seed production of transgenics was also the same as vector controls (supporting information). To assess the effect of SRK2C overexpression on stress tolerance, soil-grown 35S::SRK2C-GFP plants were exposed to drought stress for 2 weeks. During the treatment course, the majority of 35S::GFP control plants died, whereas ≈70–98% of the 35S::SRK2C-GFP plants survived (Fig. 5 C and D). This result should not be due to any differences between each pot, because we reconfirmed the elevated drought tolerance of 35S::SRK2C-GFP plants interspersed among 35S::GFP control plants in the same pot (data not shown). Stress tolerance of 35S::SRK2C-GFP plants correlated with the expression and activation levels of introduced SRK2C-GFP (Fig. 5 A–C). The transpiration rate of the 35S::SRK2C-GFP transgenic plants under drought conditions was similar to that of the 35S::GFP control plants. Thus, overexpression of SRK2C did not affect stomatal regulation (Fig. 5E).

Overexpression of SRK2C-GFP Affects Stress-Responsive Gene Expression. Microarray analysis of soil-grown 35S::GFP and 35S::SRK2C-GFP transgenic plants revealed that expression profiles of a number of genes were changed in the 35S::SRK2C-GFP plants. A total of 18 and 14 genes were up-regulated and down-regulated, respectively, in the 35S::SRK2C-GFP plants under nonstress conditions (supporting information). The 18 up-regulated genes were compared with the expression profiles of drought-stressed plants or DREB1A-overexpressing plants (29, 31). Results of this comparison indicated that 12 or 6 genes overlapped with drought-inducible or DREB1A-target genes, respectively (Fig. 6A). Although the down-regulated genes in the 35S::SRK2C-GFP plants were not significantly impacted by drought stress, the majority of them were also down-regulated in the 35S::DREB1A plants (Fig. 6A). We partially confirmed the microarray results by RNA gel blot analysis, which showed, in concert with our microarray data, that transcripts of RD29A, COR15A, AtGolS3, DREB1A, and PKS18 were up-regulated in soil-grown 35S::SRK2C-GFP plants (Fig. 6B). CBL1 and RD29B, the expression levels of which were not altered according to the microarray analysis, were again demonstrated as negative controls.

Fig. 6.

Gene expression profiles of 35S::SRK2C-GFP plants. (A) Classification of up-regulated and down-regulated genes in 35S::SRK2C-GFP plants. Gene expression analysis with an Agilent Arabidopsis 2 Oligo Microarray kit revealed that a total of 18 and 14 genes were up-regulated (Left) and down-regulated (Right), respectively, in 35S::SRK2C-GFP plants. The microarray data were compared with the previous data from drought-stressed WT plants (drought-responsive genes) and DREB1A-overexpressing plants (DREB1A-target genes). Gray areas indicate the overlap between SRK2C-regulated genes and drought-responsive genes (Upper) or DREB1A-target genes (Lower). (B) RNA gel blot analysis of 35S::GFP (VC) and 35S::SRK2C-GFP (lines 1–3) plants. Total RNA was extracted from 4-week-old plants grown on soil. Ten micrograms of total RNA was blotted and hybridized with ³²P-labeled RD29A, COR15A, AtGolS3, DREB1A, PKS18, CBL1, and RD29B cDNA probes. Ribosomal RNAs stained with ethidium bromide (EtBr) were used as loading control. (C) Induction of COR15A by drought stress in 35S::GFP and 35S::SRK2C-GFP seedlings. Two-week old seedlings grown on agar plates were exposed to drought stress for indicated times. Ten micrograms of total RNA was blotted and hybridized with ³²P-labeled COR15A cDNA probes. Ribosomal RNAs stained with ethidium bromide were used as loading control.

The expression of stress-responsive genes seemed to be constitutive in soil-grown 35S::SRK2C-GFP plants (Fig. 6B). However, there are no COR15A transcripts in 35S::SRK2C-GFP seedlings grown on agar plates under nonstress condition (Fig. 6C). COR15A expression increased under drought stress in 35S::SRK2C-GFP plants to a much higher degree than in the 35S::GFP plants (Fig. 6C), suggesting that up-regulation of stress-responsive genes require SRK2C-GFP activation. Constitutive expression of those particular genes in soil-grown 35S::SRK2C-GFP plants may be due to low humidity in a growth room in contrast to saturated humidity in Petri dishes.

Discussion

In plants, protein phosphorylation is suggested as one of the major signaling events occurring in response to environmental stress (32). In this study, we identified an osmotic-stress-activated protein kinase, SRK2C/SnRK2.8, which is a member of SnRK2 protein kinases in Arabidopsis. Osmotic-stress-dependent activation of SRK2C was demonstrated in T87 cultured cells and Arabidopsis plants as follows: (i) Activation of SRK2C-GFP was detected in transgenic T87 cultured cells treated with hyperosmotic stress (Fig. 1B), (ii) results of an immunocomplex kinase assay using anti-SRK2C antibody indicated that a 36-kDa kinase corresponded to SRK2C (Fig. 2 A), and (iii) the 36-kDa protein kinase activity was depleted in srk2c T-DNA-tagged mutants (Fig. 2B). The activation pattern of SRK2C differs significantly from other osmotic-stress-activated protein kinases in plants. SRK2C was rapidly activated within 2 min and showed maximal activity from 0.5 to 1 hr after osmotic shock was imposed (Fig. 1C), in contrast to tobacco SnRK2 kinase, which exhibited maximal activity within 1 min after osmotic shock (13). Furthermore, although MAPKs are also activated by osmotic stress in plants, they can be affected by hypotonic conditions as well (11, 33), whereas SRK2C was not (data not shown). Accordingly, SRK2C can be classified as a different type of osmotic-stress-activated protein kinase in plants.

Osmotic-stress-dependent activation of SRK2C indicated that SRK2C could mediate stress signaling in plants. In fact, loss-of-function analysis using T-DNA-tagged srk2c mutants demonstrated that SRK2C acts as a positive regulator of drought tolerance in Arabidopsis roots (Fig. 4). Although the drought sensitivity of srk2c mutants was not so apparent, the initial recovery from drought stress was clearly delayed in srk2c-1 and srk2c-2 roots. Such limited phenotype of srk2c mutants may be due to tissue-specific localization of SRK2C in root tips (Fig. 3B) or functional redundancy of SnRK2. Because water loss of srk2c mutants was not altered, the drought sensitivity of srk2c was attributed largely to roots. In maize, a Ca²⁺-independent protein kinase was also detected in the elongation zone of primary roots (10). Generally, the root tip is a dynamic and specialized tissue that plays a crucial role in sensing water and nutrients and responds by transmitting appropriate signals (34). Root-specific protein kinases, such as SRK2C, may have some important roles in root tissues as a sensor of water and nutrient status in soil. In addition, we detected that other SnRK2 were expressed in vegetative tissues that may function in osmotic-stress response in vegetative tissues of Arabidopsis. One of the SnRK2, SRK2E/OST1/SnRK2.6, plays an important role in stomatal closure in Arabidopsis leaves (17, 20).

In contrast to the srk2c loss-of-function mutants with drought-stress-sensitive phenotype, 35S::SRK2C-GFP transgenic plants exhibited higher drought tolerance (Fig. 5 C and D), suggesting that excessive SRK2C-GFP protein resulted in enhanced transduction of drought response signals. Because the transpiration rate of 35S::SRK2C-GFP plants was not altered, SRK2C may not be involved in stomatal response, suggesting that SRK2C-mediated signaling should be quite different from SRK2E/OST1/SnRK2.6, a central regulator of ABA-signaling in guard cells (17, 20). However, the elevated drought tolerance of 35S::SRK2C-GFP plants coincided with up-regulation of stress-responsive genes, including RD29A, COR15A, and DREB1A/CBF3 (Fig. 6). DREB1A/CBF3 is regarded as an important transcription factor that broadly regulates stress-responsive genes (6, 31, 35). We propose, then, that introduced SRK2C-GFP can lead to DREB1A/CBF3 expression and subsequent up-regulation of its target genes, resulting in enhanced drought tolerance of transgenic plants. However, DREB1A-target genes are merely a part of SRK2C-regulated genes (Fig. 6A), suggesting that some other stress pathways may be involved in SRK2C-mediated signal-transduction pathways. It is noteworthy that stress-responsive genes are not induced constitutively in 35S::SRK2C-GFP plants, suggesting that SRK2C-GFP does not function under nonstress condition (Fig. 6C). This feature should be preferable for avoiding growth retardation as observed in DREB1A-overexpressing plants (6).

Although DREB1A/CBF3 is thought to be a transcriptional activator functioning mainly during cold stress, SRK2C was not activated by cold stress in T87 cells (Fig. 1B). Our microarray analysis revealed that drought stress leads to a 6.7-fold induction of DREB1A/CBF3 and a >50-fold induction by cold stress, suggesting that DREB1A/CBF3 is mainly involved in cold-responsive gene expression, but it may also function under drought stress in specific tissues. Furthermore, salt-stress- or ABA-dependent induction of DREB1A/CBF3 was observed in Arabidopsis plants (36, 37), suggesting the complex regulation of DREB1/CBF expression in addition to major cold response. At present, it has remained to be elusive whether SRK2C and DREB1A/CBF3 actually cooperate in Arabidopsis roots under osmotic stress. To fully understand SRK2C-mediated signal transduction, factors upstream or downstream of SRK2C must be focused on in other studies. In addition, microarray analysis indicated that some other pathways may exist in SRK2C-mediated signaling as well as the DREB1A/CBF3 pathway (Fig. 6A). Further functional dissection of the SRK2C, its target proteins, and their interplay in signaling pathways will enrich the understanding of osmotic-stress-signaling networks in plants.

The present study indicates that SRK2C is a functional and osmotic-stress-activated protein kinase in Arabidopsis. The enhanced drought tolerance of 35S::SRK2C-GFP plants revealed that SRK2C could mediate signal transduction, regulating a series of drought stress-responsive genes. The SnRK2 family, of which SRK2C is a member, is a relatively small, plant-specific gene family, i.e., 10 members within the Arabidopsis or rice genomes, all of which are activated by osmotic stress in rice (22). The osmotic-stress-activated SnRK2s have also been identified in tobacco (13) or soybean (38). Therefore, the signal transduction system of osmotic-stress-activated SnRK2s may be highly conserved among higher plants. Our present approach of manipulating SRK2C expression in Arabidopsis may be developed as a valuable tool for improving the osmotic-stress tolerance of agronomically important crop plants and ornamental grasses or trees. We hope that this report provides beneficial information that supports biotechnology applications and molecular breeding that leads to improved drought tolerance in plants.