Arabidopsis OST1 Protein Kinase Mediates the Regulation of Stomatal Aperture by Abscisic Acid and Acts Upstream of Reactive Oxygen Species Production

Isolation of ost1 Mutants

The ost1 mutants were identified in a screen based on thermal imaging of drought-stressed plants (Merlot et al., 2002). Plantlets derived from an ethyl methanesulfonate–mutagenized population in the Arabidopsis accession Landsberg erecta (Ler) were subjected to drought stress. These plantlets then were analyzed by infrared thermography, and the individuals that displayed a lower leaf temperature than that of the wild type were selected. Among several mutants recovered during this screen, two corresponded to independent recessive alleles of a novel locus that was called OST1. Upon drought, the leaf temperature of ost1-1 and ost1-2 plantlets was ∼1°C lower than that of wild-type plantlets (Figure 1A). This temperature difference reflected the inability of ost1 leaves to limit their transpirational water loss in response to water stress (Figure 1B).

ost1 Mutations Impair the ABA Regulation of Stomatal Aperture

The responsiveness of ost1 stomata to various stimuli was analyzed in vitro using isolated epidermal peels. Preopened ost1 stomata failed to close in response to ABA (Figure 2A), which is consistent with the inability of ost1 plants to reduce water loss upon drought. By contrast, ost1 stomata closed to the same extent as the wild type in response to darkness (Figure 2B). Similarly, the ost1 mutations had no effect on the ability of preclosed stomata to open in response to light (Figure 2C) or CO₂-free air (Figure 2D). However, the ost1 mutants were impaired in the ABA inhibition of light-induced stomatal opening (Figure 2C). These data indicate that the ost1 mutations do not cause a general defect in stomatal functioning but specifically disrupt ABA signaling in guard cells.

When grown under well-watered conditions, the ost1 plants did not display any visible phenotypic alteration. Moreover, the ost1 mutations did not affect the degree of seed dormancy or the sensitivity of seed germination to inhibition by exogenous ABA (data not shown), indicating that OST1 does not control ABA responsiveness in seeds.

Map-Based Cloning of the OST1 Gene

To genetically map the OST1 gene, homozygous ost1-2 mutant plants in the Ler accession were crossed to wild-type plants in the Columbia accession. The resulting F1 plants were allowed to self. From the segregating F2 population, ost1 mutants were identified by thermal imaging as colder individuals under drought conditions. A mapping survey of polymorphism markers distributed throughout the five Arabidopsis chromosomes placed the OST1 locus on chromosome 4, between markers AG and nga1139. New simple sequence length polymorphism markers in this region were designed using the Arabidopsis polymorphism database generated by Cereon Genomics (Cambridge, MA; http://www.arabidopsis.org/cereon/). Mapping of 1024 chromosomes narrowed OST1 to the 119-kb region between markers CER449763 on BAC clone F17I5 and CER451680 on BAC clone F28A23 (Figure 3A).

We examined the sequence of this genomic region for genes predicted to encode signal transduction elements. Among them, the protein kinase gene At4g33950 on BAC clone F17I5 was considered a likely candidate. This gene was found by sequence analysis to be mutant in both ost1-1 and ost1-2 (Figure 3B). Compared with wild-type Ler, ost1-1 contains a G-to-A change 1486 bp downstream of the predicted translation initiation codon that disrupts the acceptor splice site of intron 6. ost1-2 contains a G-to-A change at nucleotide 97, which converts Gly-33 to Arg. Finally, transformation with a 4.5-kb genomic fragment encompassing At4g33950 restored the wild-type phenotype in both ost1-1 and ost1-2 (Figure 3C). These data demonstrate that At4g33950 corresponds to OST1.

OST1 cDNA was obtained by reverse transcriptase–PCR (RT- PCR) and sequenced. Comparison of the genomic and cDNA sequences showed that OST1 consists of 10 exons (Figure 3B). The wild-type OST1 sequence has been deposited in GenBank.

OST1 Kinase-Like Proteins

OST1 encodes a 362–amino acid protein with a predicted molecular mass of 41 kD. The N-terminal region of the OST1 protein (amino acids 1 to 297) presents all of the conserved features of the catalytic domains of protein Ser/Thr kinases (Figure 4A). In particular, the ost1-2 mutation affects an invariant residue required for ATP binding and thus is predicted to abolish OST1 kinase activity (Hanks and Quinn, 1991).

Database searches revealed that OST1 shares between 67 and 80% amino acid identity with nine other predicted protein kinases from Arabidopsis. We named the members of this kinase subfamily OSKL for OST1 kinase-like proteins (Figure 4A). All of these protein sequences but OSKL5 and OSKL10 are supported by full-length cDNA sequences present in the databases. OSKL7 and OSKL8 were previously designated ASK1 and ASK2, respectively, by Park et al. (1993). OSKL2 and OSKL3 were previously designated ATHPROKINB and ATHPROKINA, respectively. The OSKL proteins contain an N-terminal kinase domain similar to that of SNF1/AMPK (Hardie et al., 1998) and a novel C-terminal domain that presumably is regulatory. This group of plant protein kinases also has been designated subfamily 2 of SNF1-related protein kinases (SnRK2) by Hardie (1999). To date, no biological function has been assigned to any of the Arabidopsis OSKL2 to OSKL10 genes.

OST1 also is similar to kinases from various other plant species, including the V. faba AAPK (Li et al., 2000) and barley PKABA (Gomez-Cadenas et al., 1999) proteins, which have been proposed to act in ABA signaling (Figure 4). In particular, the AAPK protein kinase has been shown by in-gel assays to be activated by ABA in V. faba guard cell protoplasts (Li and Assmann, 1996; Mori and Muto, 1997).

Expression Pattern of OST1

Using RT-PCR analysis, the OST1 mRNA was detected in leaves, flowers, stems, and roots (Figure 5A). The specificity of the oligonucleotides used for RT-PCR was tested on OSKL2 and OSKL3 cDNAs (Figure 5B). Characterization of transgenic plants harboring a transcriptional fusion between the β-glucuronidase (GUS) reporter gene and the OST1 promoter (pOST1::GUS) further indicated that in leaves, OST1 expression is confined to guard cells and the vascular system (Figures 5C and 5D). In agreement with this observation, the OST1 mRNA was detected by RT-PCR in guard cell protoplasts but not in mesophyll cell protoplasts (Figure 5A). pOST1::GUS expression also was detected in the guard cells present in stems and flower sepals (data not shown) and in the root vascular system (Figure 5E).

ABA treatment had no detectable effect on OST1 mRNA abundance (Figure 5A) or on the level of pOST1::GUS activity (data not shown), indicating that OST1 expression is not regulated by ABA.

OST1 Kinase Is Activated by ABA

A recombinant OST1 protein was produced in Escherichia coli and migrated at the expected molecular mass upon gel electrophoresis. However, this protein displayed no detectable kinase activity in vitro (data not shown), suggesting that OST1 might be active only upon stimulation of the ABA signaling cascade in planta. A similar result in vitro also was observed in V. faba AAPK (Li et al., 2000). Based on OST1 expression patterns, protein extracts were prepared from Arabidopsis root and guard cell protoplasts and were analyzed by in-gel protein kinase assays. In both roots and guard cell protoplasts from wild-type Ler plants, a 42-kD band with protein kinase activity was enhanced markedly in response to ABA (Figure 6A). This band also was observed when calcium was replaced by EGTA in the assay (Figure 6B). The apparent molecular mass of this kinase closely correlates with the theoretical molecular mass of OST1. Furthermore, the ABA-activated 42-kD band was absent from extracts of the ost1-2 mutant (Figure 6A). These data indicate that, like the V. faba AAPK, OST1 is an ABA-activated Ca²⁺-independent protein kinase (Li et al., 2000). Furthermore, we observed a band with an apparent molecular mass of <42 kD in root extracts, which seems to be downregulated by ABA and calcium dependent. Although this band seems not to have a direct correlation with the ost1 phenotype, it needs further investigation.

OST1 Acts Upstream of ROS Production

ROS emerged recently as an essential second messenger in ABA signaling (Guan et al., 2000; Pei et al., 2000; Zhang et al., 2001). In particular, ABA-induced stomatal closure involves the production of ROS that activate plasma membrane Ca²⁺ channels that mediate Ca²⁺ influx (Pei et al., 2000; Murata et al., 2001). ROS levels were analyzed in guard cells using the fluorescent dye 2′,7′-dichlorofluorescein diacetate. In agreement with previous studies (Pei et al., 2000; Murata et al., 2001), the relative fluorescence emission increased after treatment of wild-type guard cells with ABA (Figure 7A). By contrast, ABA did not increase ROS levels in ost1 guard cells (Figure 7A).

To investigate the physiological relevance of this impairment of ABA-induced ROS production in ost1, we determined whether this mutation could be bypassed by an experimental increase of H₂O₂. As shown in Figure 7B, externally applied H₂O₂ elicited the same degree of stomatal closure in ost1 as in the wild type. Likewise, increased extracellular calcium induced similar stomatal closing in the wild type and in ost1 (Figure 7C). Similar results have been obtained for the inhibition of stomatal opening mediated by H₂O₂ and calcium (data not shown). These data support the idea that the ost1 mutation disrupts ABA signaling upstream of ROS production.

Like ost1, the ABA-insensitive abi1-1 mutation impairs ABA-induced ROS production and stomatal closure (Murata et al., 2001). In-gel kinase assays showed that ABA activation of the 42-kD OST1 kinase was inhibited in abi1-1 (Figure 7D). These data indicate that OST1 kinase acts downstream of ABI1 in the signaling cascade linking ABA to ROS production.

Plant Material and Culture Conditions

The ost1 (Merlot et al., 2002) and abi1-1 (Koornneef et al., 1984) mutants are derived from the Arabidopsis thaliana accession Landsberg erecta (Ler). Plants were grown routinely in a greenhouse (22°C with a 16-h photoperiod) on soil irrigated with mineral nutrients. For aseptic growth, seeds were surface-sterilized and plated on a medium containing the inorganic salts of Murashige and Skoog (1962) at half concentrations, 100 mg/L myoinositol, 1 mg/L thiamine-HCl, 0.5 mg/L nicotinic acid, 0.5 mg/L pyridoxine-HCl, 1% Suc, 0.8% agar, and 2.5 mM Mes-KOH, pH 5.7. For abscisic acid (ABA) induction, 2-week-old seedlings were treated with 10 μM ABA (mixed isomers; Sigma) diluted from 10 mM stock solutions prepared in ethanol; equivalent volumes of ethanol were included in the ABA-free controls.

Thermal Imaging

Thermal imaging of drought-stressed plantlets was performed as described previously (Merlot et al., 2002). In brief, plantlets were first grown under well-watered conditions (21°C, 60 to 70% RH, 16-h photoperiod) for approximately 1 week. Drought stress then was initiated by withholding watering and transferring the pots to a drier atmosphere (24°C, 50% RH, 16-h photoperiod). Thermal images were obtained using a Thermacam PM250 infrared camera (Inframetrics, North Billerica, MA). Images were saved on a Personal Computer Memory Card International Association card and were analyzed subsequently on a Macintosh computer using version 1.56 of the public domain image-analysis program NIH Image (http://rsb.info.nih.gov/nih-image/).

Stomatal Aperture Bioassays

Leaves from 4- to 5-week old plants (grown in 8 h of light at 22°C and 16 h of darkness at 20°C; 70% RH) were harvested in darkness at the end of the night. Paradermal sections of abaxial epidermis obtained in dim green light were incubated in 10 mM KCl, 7.5 mM potassium iminodiacetate, and 10 mM Mes-KOH, pH 6.15, at 20°C. In Figures 2 and 7, ABA, Ca²⁺, EGTA, and H₂O₂ were added to the solution or CO₂-free air was bubbled through the solution. Stomatal apertures were measured with an optical microscope (Optiphot-2; Nikon, Tokyo, Japan) fitted with a camera lucida and a digitizing table (TG 1017; Houston Instrument, Austin, TX) linked to a personal computer.

Protoplast Preparation

Guard cell protoplasts (GCPs) were prepared essentially as reported by Pandey et al. (2002). GCPs then were resuspended twice in 0.35 M mannitol, 10 mM KCl, and 1 mM Mes-NaOH, pH 6.1, and incubated in the same solution with ABA or ethanol solvent control. Then, GCPs were pelleted and stored at −80°C.

To prepare Arabidopsis mesophyll cell protoplasts (MCPs), rosette leaves from 4-week-old plants were excised and portions of the abaxial epidermis were peeled off. The peeled areas were floated peeled-side down on basic medium containing 0.07% cellulysin, 1.5% cellulase Onozuka RS, 0.02% pectolyase Y23, 0.1% polyvinylpyrrolidone 40, 0.25% BSA, and 0.5 mM ascorbic acid. After 2 h at 28°C with agitation, undigested material was removed by filtration through a 25-μm nylon net, and MCPs were collected by centrifugation at 150g for 20 min. MCPs were washed with fresh basic medium, pelleted again, and stored at −80°C.

The purity of the GCP and MCP preparations was 70 to 90% on a cell number basis.

DNA and RNA Analysis

The OST1 gene was amplified by PCR from genomic DNA using oligonucleotides 5′-CTCTGATGTCTTGGTGTC-3′ and 5′-GTGTACTCA-TTGAGACAA-3′. The ost1 cDNA was obtained by reverse transcription of total RNA using the Titan One Tube RT-PCR System (Roche Applied Science, Indianapolis, IN) and the same oligonucleotides as described above. The amplified DNA was cloned into PCR-TOPO2.1 (Invitrogen, La Jolla, CA).

For reverse transcription PCR (RT-PCR) analysis, total RNA was extracted, treated with DNase, and purified on RNeasy columns (Qiagen, Valencia, CA). Five micrograms of total RNA was reverse transcribed with SuperScript (Gibco BRL, Life Technologies). The cDNAs obtained were diluted 10-fold, and 2 μL was used in 50-μL PCR steps. A 590-bp fragment of OST1 cDNA was amplified with oligonucleotides 5′-CTTCCAGCAACTCATTTCAGGAGT-3′ and 5′-ACA-GTTGCTTCTGCAATGATCTGC-3′. As an internal standard for cDNA amounts, a 700-bp fragment of EF1α cDNA was amplified with oligonucleotides 5′-ATGCCCCAGGACATCGTGATTTCA-3′ and 5′-TTGGCGGCACCCTTAGCTGGATCA-3′. For RT-PCR control, OSKL2 and OSKL3 cDNAs were amplified by PCR from total cDNA with the following oligonucleotides: 5′-CCACAGGTCACTAAGGCATCCTA-3′ and 5′-GTTAAAAAAGGCTTTTTATTAGAGAGCGT-3′ for OSKL2, and 5′-TCATAGATCATTGAGACATCCTA-3′ and 5′-CAGACTTTGTCC-TCAGGAATCTC-3′ for OSKL3. PCR products were confirmed by sequence and used as templates for amplification with OST1 oligonucleotides used for RT-PCR.

Transgenic Plants

Transgenic Arabidopsis plants were generated by Agrobacterium tumefaciens–mediated transformation (Clough and Bent, 1998). For functional complementation of the ost1 mutants, an XbaI-SalI fragment of BAC F17I5 containing At4g33950 (nucleotides 55,462 to 59,974) was subcloned into the T-DNA vector pCAMBIA1390. The pOST1::GUS construct was obtained by amplifying 1.4 kb of the OST1 promoter region from BAC F17I5 using oligonucleotides 5′-GCCGGATCCCTCTAGAAGAGTGTATTGGG-3′ and 5′-GCCGGA-TCCTCTTTCCTCTGTCTCTTTGC-3′ and cloning the amplified DNA into the BamHI site of the T-DNA vector pCAMBIA1381Z. Four independent transgenic lines were analyzed by β-glucuronidase histochemical staining (Blazquez et al., 1997).

In-Gel Kinase Assays

Protein extraction and in-gel protein kinase assays were performed as described (Droillard et al., 2000). Protein extracts (30 μg) were electrophoresed on 10% SDS–polyacrylamide gels embedded with 0.5 mg/mL Histone III-S (Sigma) as a substrate for the kinases. After renaturation of the proteins, kinase activity was assayed in 40 mM Hepes, pH 7.5, 2 mM DTT, 20 mM MgCl₂, 0.1 mM orthovanadate, 1 mM CaCl₂, and 25 μM γ-³³P-ATP. Gels were dried on Whatman 3MM paper and analyzed using a Storm imager (Molecular Dynamics, Sunnyvale, CA). To test the Ca²⁺ independence of OST1 activity, CaCl₂ was replaced by EGTA in the kinase activity buffer.

Reactive Oxygen Species Detection

Reactive oxygen species production in guard cells was analyzed using 2′,7′-dichlorofluorescein diacetate essentially as described (Murata et al., 2001). Epidermal leaf peels were prepared from 5-week- old plants (grown under an 8-h photoperiod) and mounted on a microscope slide with medical adhesive (Hollister, Libertyville, IL). Epidermal tissues were incubated for 3 h in 30 mM KCl and 10 mM Mes-KOH, pH 6.15. A concentration of 30 μM 2′,7′-dichlorofluorescein diacetate was added to the incubation medium. After 20 min, the excess of dye was eliminated by two to three washes with distilled water. Epidermal tissues then were incubated for 20 to 30 min with 50 μM ABA or with an equal volume of ethanol as a control. Guard cells were observed with an epifluorescence microscope equipped with a charge-coupled device camera. Images were captured and the relative fluorescence emission of guard cells was analyzed using Adobe Photoshop 5.0 (Mountain View, CA) (Murata et al., 2001).

Upon request, all novel materials described in this article will be made available in a timely manner for noncommercial research purposes.

Accession Numbers

The GenBank accession numbers for the sequences described in this article are AJ316009 (wild-type OST1), L05562 and L05561 (OSKL2 and OSKL3, previously designated ATHPROKINB and ATHPROKINA), AL031032 (At4g33950), AAF27340 (AAPK), and BAB61735 (PKABA).