The Arabidopsis outward K⁺ channel GORK is involved in regulation of stomatal movements and plant transpiration

The gork-1 knockout line was obtained by PCR screening of ≈40,000 Arabidopsis thaliana T-DNA insertion mutants (Wassilevskija ecotype; library constructed by the Station de Génétique et Amélioration des Plantes, Versailles, France; ref. 12), with primers corresponding to the GORK gene and to the T-DNA left and right borders. Selection on kanamycin revealed a single insertion locus. The exact position of the T-DNA insertion was determined by sequencing the T-DNA flanking sequences. Plants homozygous for the disruption were selected by PCR in the F₃ progeny of the positive line.

The sequence encoding the hallmark GlyTyrGlyAsp motif in the GORK pore domain (typical of K⁺-selective channels) was replaced by ArgArgGlyAsp (by site-directed mutagenesis) in the GORK gene. The mutated gene, named gork-dn (8.78 kb in total, with 2.714 kb upstream from the initiation codon) was cloned into the KpnI–SstI sites of the binary vector pBIB-HYGRO (13). The resulting plasmid was introduced into Agrobacterium tumefaciens GV3010 (pMP90) strain (14). A. thaliana (Wassilevskija ecotype) was transformed by using the floral dip method (15). Selection on hygromycin allowed us to identify transformed lines displaying a single insertion locus.

Arabidopsis plants were hydroponically grown in a growth chamber (22°C, 65% relative humidity, 8 h/16 h light/dark, 300 μmol⋅m⁻²⋅s⁻¹) for 4 weeks before being transferred (a single plant per experiment) to an experimental chamber allowing gas exchange measurements (dew-point hygrometers and infrared gas analyzer) as described (16). The root compartment (500 ml, 21°C) contained an aerated half-strength Hoagland solution. The shoot compartment (23°C, 8 h/16 h light/dark, 400 μmol⋅m⁻²⋅s⁻¹; HQI-TS NDL, Osram, Berlin) was attached to an open-flow gas circuit (air flow: 160 liters/h). At the inlet, the water-vapor pressure was held constant (1.5 kPa) and controlled with a dew-point hygrometer (Hygro, General Eastern Instruments, Woburn, MA). Humidity at the outlet was measured with a second dew-point hygrometer. The water-vapor pressure deficit in the leaf chamber was 0.9 ± 0.1 kPa (leaf and air temperatures measured with thermistors). Transpiration and photosynthesis were monitored for at least 8 days, including a 3-day adaptation period (the data corresponding to the adaptation period were not taken into account in the analyses). Photosynthesis was proportional to the leaf fresh weight and leaf area (data not shown).

Leaves from 4- to 6-week-old Arabidopsis plants were excised at the end of the night period, and epidermal strips were prepared as described (17). After peeling, epidermal strips were placed in Petri dishes containing 5 ml of the incubation solution (usual buffer unless otherwise noted: 30 mM KCl/10 mM Mes-iminodiacetic acid, pH 6.5). To standardize the initial state, epidermal strips were kept in the incubation solution for 30 min in darkness. Then, they were submitted to different treatments at 20°C in darkness or light (300 μmol⋅m⁻²⋅s⁻¹). Stomatal apertures were measured (pore width; at least 60 measurements per experiment in <5 min) with an optical microscope (Optiphot, Nikon) fitted with a camera lucida and a digitizing table (Houston Instruments) linked to a personal computer. Each experiment was performed at least in triplicate.

Plants were grown for ≈5 weeks in compost (individual containers) in a growth chamber (21°C, 70% relative humidity, 8 h/16 h light/dark, 300 μmol⋅m⁻²⋅s⁻¹). Guard cell protoplasts were isolated by digestion of leaf epidermal peels. The digestion solution contained 1 mM CaCl₂, 2 mM ascorbic acid, Onozuka RS cellulase (1% wt/vol, Yakult Pharmaceutical, Tokyo), Y-23 pectolyase (0.1% wt/vol, Seishin Pharmaceutical, Tokyo), and 1 mM Mes-KOH (pH 5.5). The osmolarity was adjusted to 420 mosM with D-mannitol. The epidermal peels were digested for 40 min at 27°C. Filtration through 50-μm mesh allowed recovery of protoplasts. The filtrate was mixed with three volumes of conservation buffer (10 mM potassium glutamate/1 mM CaCl₂/2 mM MgCl₂/10 mM Mes-HCl, pH 5.5, with osmolarity adjusted to 500 mosM with D-mannitol). The protoplast suspension (≈10% of guard cell protoplasts, based on microscopic observations) was kept on ice. Patch-clamp pipettes were pulled (P97, Sutter Instruments, Novato, CA) from borosilicate capillaries (Kimax-51, Kimble, Toledo, OH) and fire polished (L/M CPZ 101, List Medical, Darmstadt, Germany). The pipette solution contained 1 mM CaCl₂, 5 mM EGTA, 2 mM MgCl₂, 100 mM potassium glutamate, 2 mM MgATP, 10 mM Hepes-NaOH, pH 7.5, with osmolarity adjusted to 520 mosM with D-mannitol. The bath solution contained 20 mM CaCl₂, 2 mM MgCl₂, 100 mM potassium glutamate, 10 mM Mes-HCl, pH 5.5, with osmolarity adjusted to 500 mosM with mannitol. In these conditions, the pipette resistance was ≈10 MΩ. Seals with resistance >5 GΩ were used for electrophysiological analyses. Whole-cell recordings were obtained by using an Axon Instruments Axopatch 200A amplifier. PCLAMP 6.0.3 software (Axon Instruments, Foster City, CA) was used for voltage pulse stimulation, online data acquisition, and data analysis. The voltage protocol consisted in stepping the membrane potential to voltages from −200 to +80 mV in 20-mV increments from a holding potential of −60 mV.

An Arabidopsis mutant line, named gork-1, was identified by PCR screening of a collection of T-DNA-transformed plants (Fig. 1A; T-DNA insertion flanking sequences indicated in the figure). Growth tests on selection medium revealed a single insertion locus (data not shown). The T-DNA insertion was shown to result in a 3.7-kb deletion in the 5′ region of the gene (Fig. 1). RT-PCR experiments performed on total RNA extracted from aerial parts of 4-week-old plants grown in a greenhouse indicated that GORK transcripts were not accumulated in homozygous gork-1 plants (Fig. 1B).

Shaker channels are tetrameric proteins (18–20). This structural feature allows us to develop reverse genetics approaches by using dominant negative mutant subunits (obtained by site-directed mutagenesis) able to coassemble with wild-type subunits and to lead to formation of nonfunctional channels (11, 21). Mutations were introduced in the GORK coding sequence to replace the hallmark motif GlyTyrGlyAsp, expected to play a crucial role in the formation of the channel-conducting pathway (22), by ArgArgGlyAsp. Experiments in Xenopus oocytes revealed that the polypeptide encoded by the mutated sequence was electrically silent (not shown) and indeed endowed with dominant negative behavior; for instance, coinjection of mutant cRNA with wild-type cRNA (in vitro transcription) with a mutant/wild-type stoichiometry of 1:1 resulted in 82 ± 6% (n = 10) inhibition of the GORK current. The corresponding mutation was introduced in the GORK gene (8.7 kb in total, with a 2.7-kb promoter region), and the mutated gene was introduced into Arabidopsis plants. Single locus transformed lines (checked on hygromycin selection medium) were obtained. Patch-clamp experiments were performed on homozygous transgenic plants from the progeny (F₄ generation) of two of them, named gork-dn1 and gork-dn2 in this article.

Homozygous gork-1, gork-dn1, or gork-dn2 plants were grown in a greenhouse or growth chamber (20°C, 50% relative humidity, 8 h/16 h light/dark, 300 μmol⋅m⁻²⋅s⁻¹). Comparison with control wild-type plants grown in parallel in these conditions did not reveal any obvious phenotype. Because previous RT-PCR experiments had shown that GORK transcripts are also present in root hairs (23), root systems from plants grown in vitro on agar plates were examined under microscope. No obvious effect of the gork-1 mutation on root hair density or development could be detected (data not shown).

Patch-clamp experiments on guard cell protoplasts prepared from wild-type plants revealed both inward and outward currents (Fig. 2A), in agreement with previous analyses (6). The outward current was dominated by a slowly activating sigmoidal component, appearing beyond a threshold potential of ≈0 mV (with 100 mM K⁺ in both the bath and pipette solutions) and strongly reminiscent of that recorded in Xenopus oocytes expressing GORK cRNA (10). Similar experiments were performed on guard cell protoplasts prepared from homozygous gork-1 plants (Fig. 2B) or from gork-dn1 or gork-dn2 plants (Fig. 2C). They revealed inward K⁺ currents quite similar to those recorded in wild-type protoplasts but strongly reduced outward currents. The time course of the remaining outward K⁺ current observed in gork-dn1 (see the current trace at +80 mV magnified ×10 in Fig. 2C) and gork-dn2 (data not shown) protoplasts suggested some residual GORK activity. No GORK activity could be detected in gork-1 guard cells, as shown by the flat kinetics of the residual outward current (current trace at +80 mV magnified ×10 in Fig. 2B). It can therefore be concluded that gork-1 is actually a knockout mutation.

The role of GORK in stomatal movements was investigated in vitro by submitting epidermal peels from wild-type, gork-1, or gork-dn1 plants to treatments inducing either stomatal closure (azobenzenearsonate or darkness; Fig. 3 A and B, respectively) or opening (light; Fig. 3C). Stomatal closure was strongly altered in the gork-1 and, to a lesser extent, in the gork-dn1 plants (Fig. 3 A and B), indicating that GORK activity was required for efficient stomatal closure. On the other hand, the gork-1 mutation weakly increased light-induced stomatal opening (Fig. 3C). It is worth noting, however, that gork-1 peels consistently displayed slightly larger apertures, e.g., by 10–15% at the end of the 3-h light pretreatment in Fig. 3 A and B (t = 0).

As a first step in investigating the role of GORK in the control of leaf transpiration, we measured water loss (decrease in weight) of rosettes excised from wild-type, gork-1, or gork-dn1 plants. The gork-1 rosettes displayed greater water loss than the wild-type ones, by ≈35% during the first hour after the excision (Fig. 4). The gork-dn1 rosettes displayed an intermediate phenotype. This provided the first experimental evidence that GORK activity participates in leaf transpiration control.

In a second set of experiments, the effect of the gork-1 mutation on intact plant transpiration was investigated by using an experimental setup that allowed us to control the water-vapor pressure deficit (0.9 kPa) and to continuously record the plant transpiration and CO₂ fixation rates (open-flow gas circuit). Switching light on and off induced an increase and decrease, respectively, in transpiration rate, leading to a new steady state (Fig. 5A). Analysis of the steady-state data (Fig. 5B) revealed that the gork-1 plants displayed higher transpiration rates than the wild-type controls, by ≈25% in darkness and ≈7% in light (the latter difference, however, being not statistically significant). The kinetics of the transitions between the successive steady states were fitted (least-squares adjustment) with first order exponential functions to derive the time constants, τ (Fig. 5C), assumed to reflect the rate of stomatal movements. The gork-1 mutation was without any significant effect on the rate of stomatal opening but resulted in a strong decrease in the rate of stomatal closure, the mean τ value being shifted from ≈12 to 19 min. The former value is consistent with previous analyses of dark-induced stomatal closing in Arabidopsis (24) and of dark-induced Rb⁺ efflux kinetics in Commelina (25).

In another experiment, 6-week-old wild-type and gork-1 plants grown on compost were submitted to a progressive drought stress over 7 days, leading to loss of leaf turgor. The gork-1 plants displayed higher transpirational water loss (weight loss) than the wild-type plants, during both the night and the light periods (Fig. 6). The relative difference was clearly more important at the end than at the beginning of the stress treatment (≈30–50% and 5–15%, respectively), suggesting that the relative contribution of GORK to transpiration control is more important in conditions of water shortage.

In planta (electro)physiological analyses have led to the conclusion that K⁺ release from the guard cell, leading to stomatal closure, involves the activity of K⁺-selective slowly activating voltage-gated outwardly rectifying channels (6, 26). The functional features of these channels characterized in vivo in guard cells of Arabidopsis and of a number of other species are very similar to those displayed by the Arabidopsis SKOR and GORK channels expressed in heterologous systems (10, 27). These two channels belong to the so-called Shaker family, which comprises nine members in Arabidopsis (7). SKOR is expressed in root stelar tissues, where it plays a role in K⁺ secretion into the xylem sap (27). RT-PCR experiments have revealed GORK transcripts in guard cells and in root hairs (9, 23). This information supported the hypothesis that the outward K⁺ channel active in guard cells was encoded by the GORK gene (in Arabidopsis; ref. 10). Here, we show that expressing a dominant negative mutant allele of the GORK gene in Arabidopsis leads to a strong decrease in the outward K⁺ conductance of the guard cell membrane and that disruption of the GORK gene results in full suppression of this conductance (Fig. 2). Thus, the present data demonstrate that GORK and SKOR are not redundantly expressed in guard cells and that functional expression of the GORK gene is required for voltage-gated outwardly rectifying K⁺ channel activity in the guard cell plasma membrane. The simplest hypothesis is that a single gene, GORK, encodes the major voltage-gated outwardly rectifying K⁺ channel characterized in this membrane (6).

Circumstantial evidence supported the hypothesis that, by mediating K⁺ release, the outward rectifier of the guard cell membrane could play a major role in stomatal closure and thereby in transpiration control. The reverse genetics approach developed in our study provides direct support to this hypothesis. The whole set of data indicates that transpiration is more important in gork-1 than in wild-type plants, from 10% (Fig. 5) up to ≈50% (Fig. 6), depending on the environmental conditions (water availability in these experiments). Thus, whereas disruption of the guard cell inward K⁺ channel gene KAT1 has been found to be without any effect on the regulation of stomatal aperture (9), that of the outward channel GORK results in both impaired stomatal movements and impaired transpiration control. This difference is likely to result from the fact that expression of inward K⁺ channel activity in guard cells would involve at least two different genes, KAT1 and KAT2 (8), whereas the outwardly rectifying K⁺ channel activity relies mainly on a single gene, GORK.

In vitro analyses using epidermal peels are likely to give a distorted view of in planta stomatal control because of the lack of functional interactions with the surrounding epidermal cells. The bioassays shown in Fig. 3 would suggest larger differences in transpiration between wild-type and gork-1 plants than those observed in Fig. 5. However, this discrepancy could result, at least in part, from the fact that hydroponic conditions (Fig. 5) lessen the importance of stomata, and thus of GORK activity, in the control of transpiration. Indeed, larger differences in transpirational water loss between the wild-type and gork-1 genotypes were observed when the plants were grown in compost and submitted to water shortage (Fig. 6).

Larger steady-state stomatal apertures were found in gork-1 than in wild-type epidermal peels not only at the end of closure-inducing treatments (azobenzenearsonate, darkness; Fig. 3 A and B) but also at the end of 3-h light-induced opening pretreatments/treatments (experimental points corresponding to time 0 in Fig. 3). Consistent with these observations, in planta analyses revealed that gork-1 plants displayed higher transpiration rates than wild-type plants not only in darkness but also in light (Figs. 5 and 6). These results suggest that GORK could act as a negative regulator of stomatal opening in light, in addition to playing a role in stomatal closure. It has been shown that the membrane voltage in guard cells can undergo large (>100 mV) and rapid (≈10-s period) oscillations between hyperpolarized values allowing K⁺ uptake and depolarized values allowing K⁺ release (28). This has led to the hypothesis that control of stomatal movements and steady-state aperture depends on variations in the pattern of these oscillations, enabling the net guard cell K⁺ content to increase or to decrease as a result of changes in the balance of K⁺ uptake through hyperpolarization-activated channels and K⁺ release through depolarization-activated channels (28). The proposal that GORK might act as a negative regulator of stomatal opening stands coherently within the framework of this hypothesis.

The available information indicates that the dominant solutes involved in guard cell osmoregulation and control of stomatal aperture are K⁺ salts and sugars (mainly sucrose), depending on the environmental conditions and the time of the day (4, 29). Genetic tools allowing the assessment of the relative contribution of these solutes and of the mechanisms responsible for their transport and accumulation are highly needed. The present data provide direct genetic evidence that GORK encodes the major voltage-gated outwardly rectifying K⁺ channel expressed in the guard cell plasma membrane, its disruption resulting in a dramatic decrease in the membrane outward K⁺ conductance. However, when grown in standard controlled conditions or even in the greenhouse, gork-1 plants do not display any obvious phenotype. Furthermore, the gork-1 stomata can undergo large movements (Fig. 3). This indicates that, in the absence of the GORK channel, different processes, likely to involve activity of other K⁺ efflux systems, can efficiently contribute to the decrease in turgor pressure leading to stomatal closure. Based on the present data, it cannot be said whether these processes/systems significantly contribute to stomatal closure also in the presence of GORK activity or whether they correspond to compensation mechanisms resulting from the loss of GORK activity. It is worthy to note that the very low level of GORK current remaining in the plants expressing the dominant negative mutant channel resulted in a stomatal phenotype closer to that of the wild-type plants than to that of the gork-1 plants (Fig. 3). The dominant negative mutant lines could be valuable genetic tools to further investigate why stomatal physiology requires the guard cell membrane to be fitted with a large voltage-gated outwardly rectifying K⁺ conductance.

In hydroponic conditions, the absence of GORK activity resulted in an increase in steady-state transpiration, by ≈7% in light and 25% in darkness (Fig. 5B). During the light-to-dark transition, the increase in water loss reached higher percentages (up to 40–50%; see the peak displayed by the gray curve in Fig. 5A) because of the slower closure kinetics during the transition from the open to the closed steady state once light was turned off. Thus, the contribution of GORK to water saving would be more important in natural conditions with fluctuating environmental conditions requiring rapid adaptation of stomatal opening (e.g., sudden changes in water-vapor pressure deficit and/or light intensity). Also, the data shown in Fig. 6 highlight that drought conditions amplify the relative contribution of GORK to water saving. Reduced water consumption by ≈10–20% can allow the plant to postpone dehydration by several days. Thus, the GORK gene is likely to play an important role in drought adaptation and to be under high selection pressure in natural fluctuating environments.