Regulation of Abscisic Acid-Induced Stomatal Closure and Anion Channels by Guard Cell AAPK Kinase

See allHide authors and affiliations

Science  14 Jan 2000:
Vol. 287, Issue 5451, pp. 300-303
DOI: 10.1126/science.287.5451.300


Abscisic acid (ABA) stimulates stomatal closure and thus supports water conservation by plants during drought. Mass spectrometry–generated peptide sequence information was used to clone a Vicia faba complementary DNA, AAPK, encoding a guard cell–specific ABA-activated serine-threonine protein kinase (AAPK). Expression in transformed guard cells of AAPK altered by one amino acid (lysine 43 to alanine 43) renders stomata insensitive to ABA-induced closure by eliminating ABA activation of plasma membrane anion channels. This information should allow cell-specific, targeted biotechnological manipulation of crop water status.

The hormone ABA regulates various processes in plants including responses to stressors such as drought, cold, and salinity (1). During drought, ABA alteration of guard cell ion transport promotes stomatal closure and prevents stomatal opening, thus reducing transpirational water loss. That this is a fundamental component of plant desiccation tolerance is indicated by the wilty phenotype of some ABA-insensitive mutants ofArabidopsis thaliana [dominant mutations abi1-1and abi2-1 (2)]. Conversely, the ABA supersensitive mutant, era1, shows enhanced drought tolerance (3). However, the abi1-1,abi2-1, and era1 phenotypes are pleiotropic, showing altered seed dormancy (2, 3), for example, which indicates that these genes would not be ideal targets for biotechnological manipulations seeking specifically to regulate stomatal responses.

Guard cells express an AAPK, which has Ca2+-independent and ABA-activated phosphorylation activities (4). AAPK activity is detected in guard cells but not in leaf epidermal or mesophyll cells (4) or in roots (5). AAPK is activated by ABA but not by darkness or elevated CO2concentrations (Fig. 1), conditions that also engender stomatal closure (6). We thus hypothesized that AAPK could be a guard cell–specific ABA response regulator. Here we report cloning of the AAPK cDNA, AAPK function, and manipulation of that function in planta.

Figure 1

ABA, but not other signals that cause stomatal closure, activate AAPK. AAPK autophosphorylation activity was assayed by the in-gel kinase method (4, 7). Closing signals applied to the guard cell protoplasts were darkness [(A), lane 1], ABA (10 μM [±]-cis,trans-ABA) [(A), lane 2], and elevated CO2 concentrations (700 ppm CO2), versus the normal concentration of 350 ppm Co2[(B), lane 3]. The asterisk indicates a previously identified Ca2+-dependent protein kinase (28). The arrowhead indicates AAPK.

Guard cell protoplasts (4.8 × 107; 99.6% pure) were prepared (4) from Vicia faba.Protoplast proteins were extracted and subjected to two-dimensional (2D) gel electrophoresis. AAPK was identified as a 48-kD ABA-dependent and Ca2+-independent autophosphorylation spot with the in-gel kinase assay (4, 7). The AAPK spot was excised and subjected to peptide sequencing by tandem mass spectrometry (8). Two sequenced AAPK peptides had similarity to the PKABA1 (protein kinase ABA1) subfamily (9) of protein kinases in subdomains I and VIb. PKABA1 is transcriptionally up-regulated by ABA (9) and PKABA1 may suppress gene induction by gibberellic acid during cereal grain germination (10).

Degenerate primers, whose design was based on conserved sequences in subdomain II of the PKABA1 subfamily and on the AAPK peptide sequence corresponding to protein kinase subdomain VIb, were used for reverse transcription–polymerase chain reaction (RT-PCR) with guard cell total RNA as template (11). The 310–base pair (bp) product that was generated encoded the previously determined AAPK peptide sequences as well as an amino acid sequence similar to that of the PKABA1 subfamily from subdomains II to VIb. This product was used to screen aV. faba guard cell cDNA library. A full-length cDNA of the appropriate size and sequence to encode AAPK was obtained.

The deduced AAPK sequence shows greatest homology to the PKABA1 subfamily (Fig. 2). However, the predicted protein also has unique regions, and none of the other PKABA1 family members has been implicated in stomatal function. Northern analysis (12) shows that AAPK is expressed in guard cell protoplasts, but not in mesophyll cell protoplasts, flowers, leaves, or seeds (Fig. 3), paralleling the guard cell specificity previously observed for AAPK activity (4). Southern analysis (5) implies that AAPK is a single copy gene.

Figure 2

Alignment of the deduced AAPK amino acid sequence with those of homologous protein kinases. GenBank accession numbers are: AAPK (AF186020), Arabidopsis Atpk (L05562), tobacco WAPK (AF032465), soybean SPK-4 (L38855), rice REK (AB002109), ice plant MK9 (Z26846), and wheat PKABA1 (M94726). Amino acids are highlighted when there are at least four identical residues among the seven sequences. Conserved subdomains of the protein kinase family are indicated by roman numerals. Peptide sequences obtained by tandem mass spectrometry are marked by lines. Peptide regions used for designing degenerate PCR primers are indicated by arrows. Sequences were aligned by the Clustal method in MegAlign (DNASTAR, Madison, WI). Numbers indicate amino acid positions (16).

Figure 3

Specific expression of AAPK in guard cells. (A) Northern blot analysis with 10 μg total RNA per lane (12). The AAPK transcript is approximately 1.6 kb. (B) Ethidium bromide staining of the gel used in (A), confirming equal sample loading. GCP, guard cell protoplasts; MCP, mesophyll cell protoplasts.

Functional analysis of the AAPK gene product was complicated because ABA activation of AAPK does not occur in vitro. ABA activation is evident only when AAPK is extracted from intact guard cells previously treated with ABA (4), which presumably reflects a requirement for an intact cellular signaling cascade. Therefore, a green fluorescent protein (GFP)–tagged construct of AAPK (pAAPK-GFP) was made (13), expressed in guard cells (14), and shown to produce a kinase whose activity was up-regulated by ABA treatment of the cells (14). The conserved lysine residue in subdomain II of protein kinases is critical for adenosine triphosphate (ATP) binding. Mutation of this residue yields kinases with reduced or absent catalytic activity (15). To create a comparable AAPK mutant, Lys43in AAPK was mutagenized to an alanine and a pAAPK(K43A)-GFP construct was created (13, 16). The kinase encoded by this construct had reduced activity (14), as predicted.

Next, V. faba leaves were biolistically transformed with pGFP, pAAPK-GFP, or pAAPK(K43A)-GFP (17–19). Abaxial epidermal peels were isolated, and transformed guard cells, indicated by their green fluorescence, were assayed for ABA-prevention of stomatal opening or for stomatal closure stimulated by ABA, CO2, or darkness (19). The “half-aperture” of each transformed guard cell was compared with the half-aperture of the other, untransformed guard cell in the pair. Transformation with pAAPK(K43A)-GFP eliminated ABA-induced stomatal closure (Fig. 4) (Table 1), but had no effect on stomatal closure induced by CO2 or darkness. Transformation with wild-type AAPK (pAAPK-GFP) had no effect on any of the responses assayed (Table 1), suggesting either that native AAPK was not rate-limiting or that recombinant AAPK was not sufficiently overexpressed to alter function.

Figure 4

Block of ABA-induced stomatal closure by AAPK(K43A) mutant kinase. (A) Brightfield image showing that treatment of V. faba epidermal peels with 25 μM [±]-cis,trans-ABA causes closure of “half-stomates” associated with untransformed guard cells (left cell), while half-stomates associated with transformed guard cells (right cell) remain open. Scale bar, 10 μm. (B) Fluorescence image corresponding to the brightfield image of (A), showing the green fluorescence resulting from expression of pAAPK(K43A)-GFP in the right guard cell. Scale bar, 10 μm.

Table 1

Overexpression of AAPK(K43A) in guard cells inhibits ABA-induced stomatal closure. V. faba leaves were transformed and stomatal responses measured as described in (19). ABA treatment was 25 μM (for closure) or 50 μM (for opening) (±)-cis,trans-ABA, elevated CO2treatment was 700 ppm CO2. Except for the darkness treatment, peels were illuminated (0.20 mmol m−2s−1 white light) for the duration of each treatment. All numbers represent the change in half aperture of stomata as measured in micrometers. ND, not determined. Numbers in parentheses indicate sample sizes.

View this table:

To determine the cellular basis for the mutant AAPK effect, guard cell protoplasts were isolated from biolistically transformed leaves and subjected to patch-clamp analyses (20). Slow anion channel activation is implicated in stomatal closure (6, 21–23). Anion loss presumably depolarizes the membrane potential, thus driving an efflux of K+, followed by H2O efflux and stomatal closure. We show here that in V. faba, ABA activates slow anion channels (Fig. 5) (24), not only reversing the typical decay in anion current magnitude over time in the whole cell configuration (25) but also increasing anion current magnitude. Guard cells transformed with pGFP or pAAPK-GFP show normal ABA regulation of anion currents (Fig. 5). However, in guard cells transformed with pAAPK(K43A)-GFP, ABA activation of anion channels is eliminated (Fig. 5).

Figure 5

Mutant AAPK blocks ABA-activation of slow anion channels in V. faba guard cell protoplasts. (A) In untransformed cells, or in cells transformed with GFP (pGFP) or AAPK (pAAPK-GFP) constructs, the typical decay in anion current magnitude over time in the whole cell configuration is prevented and anion current magnitude is enhanced by 50 μM [±]-cis,trans-ABA. By contrast, in cells transformed with mutant AAPK [pAAPK(K43A)-GFP], ABA has no effect, and current rundown proceeds. Currents at −85 mV are shown 15 and 30 min after achieving a stable whole-cell configuration. ABA was applied just after recording the 15-min traces. Note the “reversal” of the 15 and 30 min traces in the pAAPK(K43A)-GFP transformed guard cell treated with ABA. Currents were identified as slow anion currents by their time dependence, reversal potential, and sensitivity to the anion channel blocker 5-nitro-2-(3-phenylpropylamino)benzoic acid (29). (B) Summary of effects of AAPK and the kinase inhibitor K252A (1 μM) on ABA regulation of anion currents. Steady-state current magnitude at −85 mV (21, 22) at 15 min after ABA application relative to steady-state current magnitude just before ABA application is plotted using the formula: –[(I 15minI 30min)/I 15min] × 100. ABA was applied 15 min after attaining a stable whole-cell configuration. From left to right, bars represent mean ± SE of 9, 12, 5, 6, 5, 5, 9, 7, and 8 cells. Open bar, −ABA; striped bar, +ABA; solid bar, +ABA + K252a).

We suspect that the K43A mutant kinase competes with the activity of native AAPK in a dominant negative fashion. First, the kinase inhibitor K-252a inhibits (i) native AAPK activity (4,5), (ii) ABA-induced stomatal closure (22), and (iii) ABA regulation of anion channels in untransformed cells (Fig. 5B), implying that the channels are indeed normally regulated by AAPK. Second, although dominant abi1-1 and abi2-1mutations in ABI and ABI2 phosphatases confer ABA insensitivity to both anion channel activation and stomatal closure (2,3), recently identified recessive loss-of-function mutations inABI1 confer hypersensitivity to ABA (26). Thus, in wild-type plants an AAPK may mediate ABA-induced anion channel activation and stomatal closure through a phosphorylation event, while ABI1 opposes ABA action through a dephosphorylation event.

Neither wild-type nor mutant versions of recombinant AAPK affected ABA inhibition of stomatal opening (Table 1). ABA inhibition of stomatal opening and ABA promotion of stomatal closure may, therefore, employ different signaling cascades. Alternatively, and in contrast to current theory (21), ABA activation of anion channels may not be required for ABA inhibition of stomatal opening.

Agronomically, loss of ABA-stimulated stomatal closure in plants transformed with mutant AAPK under control of an inducible promoter might allow accelerated and controlled desiccation of crops that are dried before harvest or distribution. Basal levels of ABA remain even in irrigated crops (27); under these conditions, inhibition of AAPK activity might alleviate stomatal limitation of CO2 uptake, and thus accelerate growth or increase yield.

  • * To whom correspondence should be addressed. E-mail: sma3{at}


Stay Connected to Science

Navigate This Article