Trafficking of the plant potassium inward rectifier KAT1 in guard cell protoplasts of Vicia faba


For correspondence (fax +49 6151 164630; e-mail


Trafficking of K+ inward (Kin+) rectifying channels was analyzed in guard cells of Vicia faba transfected with the Kin+ rectifier from Arabidopsis thaliana KAT1 fused to the green fluorescent protein (GFP). Confocal images and whole-cell patch-clamp measurements confirmed the incorporation of active KAT1 channels into the plasma membrane of transfected guard cell protoplasts. The Kin+ rectifier current density of the plasma membrane was much larger in transfected protoplasts than in wild-type (wt) protoplasts. This shows a coupling between K+ channel synthesis and incorporation of the channel into the plasma membrane. Pressure-driven increase and decrease in surface area led to the incorporation and removal of vesicular membrane carrying active Kin+ rectifier in wt and transfected protoplasts. These vesicular membranes revealed a higher channel density than the plasma membrane, suggesting that Kin+ rectifier remains in clusters during trafficking to and from the plasma membrane. The observed results can be explained by a model illustrating that vesicles of a pre-plasma membrane pool carry K+ channels preferentially in clusters during constitutive and pressure-driven exo- and endocytosis.


The movement of stomatal pores is mediated by swelling and shrinking of guard cells. Under accumulation of K+ salts and subsequent water influx these cells swell for pore opening. In the reverse process, the pores close. Central proteins in this process are two K+ channels. One, a K+ inward (Kin+) rectifier conducts K+ uptake, the other, a K+ outward (Kout+) rectifier, mediates K+ discharge from guard cells.

Open and closed stomata differ in volume by up to 40% (Raschke, 1979). To accomplish these large changes in volume, also the surface area of the cells has to change significantly. As the elastic extensibility of biological membranes is limited to about 2%, (Wolfe ., 1986) the plasma membrane cannot accomplish this extension in surface area by stretching. Alterations in surface area therefore require addition and removal of plasma membrane material. Previous studies on guard cell protoplasts have shown that osmotically induced changes in surface area are indeed associated with fusion and fission of vesicular membrane ( Homann, 1998; Homann and Thiel, 1999).

A recent investigation examined the ion channel composition of the plasma membrane during pressure-stimulated changes in the guard cell surface area. It was found that an increase in plasma membrane area was coupled to a proportional rise in the conductance of both the Kin+ and the Kout+ rectifiers (Homann and Thiel, 2002). The absence of any kinetic effects on K+ channel activity during pressure stimulation fostered the hypothesis that during swelling, vesicular membrane with active K+ channels is inserted into the plasma membrane of guard cells.

The present study further investigates the mechanisms underlying the trafficking of K+ channels and the control of K+ channel density in the plasma membrane. We therefore transiently expressed a chimera of the green fluorescent protein (GFP) and KAT1, in guard cells of Vicia faba. Using confocal microscopy, we obtained information on the spatial distribution of KAT1 channels in guard cell protoplasts. To examine trafficking of these channels, we have carried out parallel recordings of membrane capacitance and conductance. The electrophysiological data argue for a model in which vesicles deliver clusters of K+ channels to the plasma membrane. The channels remain as clusters in the plasma membrane and are again retrieved in this form via endocytosis.


Localization of KAT1::GFP in guard cell protoplasts

Guard cells were transfected with KAT1::GFP using ballistic bombardment of leaves of V. faba. Fifteen hours after transfection about 30% of the guard cells in the area of the bombardment were fluorescently labelled. Guard cell protoplasts showing fluorescent staining could be enzymatically isolated from epidermal peels of bombarded leaves. For localization of the fusion protein, fluorescent images of the transfected protoplasts were obtained by confocal laser scanning microscopy (CLSM). These protoplasts showed a pronounced fluorescent line at the periphery, which corresponds to the plasma membrane (Figure 1a). GFP fluorescence was also detected in the cytosol in association with numerous small fluorescent structures with a diameter below 250 nm (Figure 1a,b). As shown in a more detailed view (Figure 1c–f), some of these small fluorescent structures were associated with the labeled plasma membrane.

Figure 1.

Confocal images of a guard cell protoplast expressing KAT1::GFP.

(a) Single equatorial optical slice of a confocal image stack of a guard cell protoplast expressing KAT1::GFP.

(b) Maximum projection of a confocal image stack of the protoplast shown in (a).

(c) Magnification of the area highlighted in (a).

(d–f) Further magnifications of the confocal image stack shown in (b).

Kin+ conductance is greatly enhanced in protoplasts expressing KAT1::GFP

To analyze the K+ conductance of guard cell protoplasts expressing KAT1::GFP membrane, currents were recorded with the patch-clamp technique in the whole-cell mode. Figure 2(a) shows a representative example of the current response to a voltage step protocol. At voltages positive of about 0 mV, the guard cell typical time- and voltage-dependent currents through the Kout+ rectifier were elicited. Stepping the membrane voltage more negative than − 80 mV resulted in the slow activation of a large time-dependent inward conductance. Typically, the Kout+ current recorded in transfected guard cells was similar to the outward current in wild-type (wt) protoplasts (Figure 2b). In contrast, the time-dependent inward conductance of protoplasts expressing KAT1::GFP generally exceeded those of wt protoplasts by an order of magnitude (Figure 2b). This large inward conductance in transfected protoplasts most likely results from the activity of both a small number of endogenous Kin+ rectifier channels and a much larger number of KAT1 channels. Therefore, the time dependent inward conductance recorded in transfected protoplasts is hereafter referred to as . The average current density of in protoplasts expressing KAT1::GFP was 159 ± 28 pA pF−1 (n = 10) at −160 mV (Figure 2c). This is about 15 times the Kin+ current density found in wt protoplasts under the same conditions (10.3 ± 0.6 pA pF−1, n = 9).

Figure 2.

Comparison of K+ currents recorded in KAT1::GFP transfected and wt guard cell protoplasts.

(a) Current traces (button) recorded in response to voltage protocol (top) in a guard cell protoplast expressing KAT1::GFP (●) and in a wt guard cell protoplast (○).

(b) Current–voltage relation of transfected (●) and wt protoplasts (○).

(c) Average current density of transfected (solid bar, n = 10) and wt (open bar, n = 8) guard cell protoplasts measured before onset of changes in surface area.

Reversible changes in K+ conductance during changes in surface area

To analyze changes in K+ conductance during changes in surface area, we recorded membrane capacitance and membrane conductance in parallel in the same protoplast. While the former parameter is proportional to the cell surface area and monitors excursions in the plasma membrane area, the latter parameter provides information on the activity of the prominent K+ channels. Swelling and shrinking of protoplasts was induced by pulses of hydrostatic pressure applied via the patch pipette. Figure 3(a) shows an example of the recording of the membrane capacitance and the current in a KAT1::GFP-expressing protoplast. Swelling and shrinking of the protoplast led to a parallel increase and decrease, respectively, of membrane capacitance and the current (Figure 3a). In the present example, the current density increased during swelling from 273 to 423 pA pF−1 (Figure 3a). Corresponding measurements on wt protoplasts revealed equivalent results. For the recording presented in Figure 4(a), swelling of the protoplasts led to an increase in the current density from 9 to 26 pA pF−1. The rise in the current density implies that the channel density of the added membrane is higher than that of the plasma membrane. Figures 3(a) and 4(a) also demonstrate that the change in the current density was reversed upon subsequent decrease in surface area. The channel density of membrane retrieved by endocytosis is therefore also higher than the overall channel density of the plasma membrane. This is a strong indication for clustering of channels in the plasma membrane.

Figure 3.

Whole-cell K+ currents of a transfected guard cell protoplast recorded during increase and decrease in surface area.

(a) Membrane capacitance (Cm, ●), Kin+ currents (IKin*, ○), and current density (IKin*/Cm, inline image) recorded from a guard cell protoplasts transfected with KAT1::GFP. Currents were recorded in response to square-wave voltage pulses from a holding voltage of −65 mV to test voltages of −160 mV.

(b) Correlation between the time constant of the activation of the Kin+ current (τ) and changes in Cm from the measurement shown in (a).

Figure 4.

Whole-cell K+ currents of a wt guard cell protoplast recorded during increase and decrease in surface area.

(a) Membrane capacitance (Cm, ●), Kin+ currents (IKin, ○), and current density (IKin/Cm, inline image) recorded from a wt guard cell protoplasts. Currents were recorded in response to square-wave voltage pulses from a holding voltage of −65 mV to test voltages of −160 mV.

(b) Correlation between the time constant of the activation of the Kin+ current (τ) and changes in Cm from the measurement shown in (a).

In a previous study (Homann and Thiel, 2002), we presented evidence that a rise in wt K+ currents during increase in surface area was most likely the consequence of an insertion of active K+ channels into the plasma membrane. To examine the reason for the large changes in the current during area excursions in the transfected cells, we analyzed the time constant (τ) of the activation of the current. The data presented in Figure 3(b), which are representative for 10 similar experiments, show that the time constant of the current was not related to changes in membrane capacitance (Figure 3b). This absence of a perceivable effect on channel kinetics suggests that changes in currents do not result from changes in the open probability of channels pre-existing in the plasma membrane. Instead, the data favor the view that during changes in capacitance, new active channels are added and retrieved via exo- and endocytosis, respectively. Similar results were found for wt protoplasts (Figure 4b).

Figure 5 summarizes the changes in the current as a function of changes in surface area from 10 KAT1::GFP-expressing guard cell protoplasts. In all cases, swelling and shrinking resulted in a parallel change in membrane capacitance and the current. The apparent linear relationship between the current and membrane capacitance has a slope of 831 pA pF−1. This is around seven times higher than what was found for the Kin+ in wt cells (Homann and Thiel, 2002). On the other hand, there was no apparent difference in the behavior of the Kout+ rectifier between transfected and wt cells (data not shown).

Figure 5.

Changes in Kin+ current are correlated with changes in membrane capacitance.

Correlation between changes in membrane capacitance (ΔCm) and Kin+ currents (ΔIKin*) recorded from 10 guard cell protoplasts (indicated by different symbols) expressing KAT1::GFP. Changes in current were calculated from the steady-state current amplitude recorded in response to square-wave voltage pulses from a holding voltage of −65 mV to test voltages of −160 mV. Line represents regression line fitted to the data points shown.

The slope value reported in Figure 5 is a measure for the current density of the membrane incorporated and removed by exo- and endocytosis, respectively. The current density of the vesicular membrane is thus more than five times that of the resting value of the plasma membrane. Accordingly, an increase in surface area by 2 pF would nearly double the current density of an average sized protoplast of 8.6 pF.


A previous study demonstrated that an increase in surface area in guard cell protoplasts is paralleled by a proportional rise in the number of the two dominant K+ channels, the Kin+ and Kout+ (Homann and Thiel, 2002). This study argued for a model in which vesicular membrane, including active K+ channels, is available in the form of a vesicular pool in the cytoplasm. This membrane can be incorporated on demand into the plasma membrane via exocytotic vesicles to accommodate for surface area extension. The present results on guard cells, which overexpress a GFP chimera of the Kin+ rectifier KAT1, add further details to this model.

A comparison of wt and transfected cells shows that the current density of the plasma membrane of the in the latter cells is more than 15 times as much as that in wt cells. The view that this additional current is because of a high number of KAT1 channels in the plasma membrane is consistent with the electrical features of this current, which closely resemble those of KAT1 (Schachtman ., 1992). Also, the pronounced concentration of KAT1::GFP fluorescence in the plasma membrane of the transfected cells supports the view that the additional current is not an upregulation of endogenous conductance but the result of incorporation of active KAT1 channels into the plasma membrane. In this sense, the present data corroborate a previous study showing that KAT1 from Arabidopsis can be heterologously expressed in other plant cells (Bei and Luan, 1998).

In agreement with our previous studies (Homann and Thiel, 2002), the present experiments also show that the channels are incorporated under pressure stimulation together with exocytotic membrane into the plasma membrane. The current density of the vesicular membrane incorporated during swelling was about eight times larger in transfected protoplasts than in wt protoplasts. This shows that not only the concentration of K+ channels in the plasma membrane but also the channel density of the vesicular pool is in a dynamic exchange with the sites of channel synthesis.

It is worth noting that the pressure-stimulated incorporation of outward rectifying channels in transfected cells was not different from those in wt cells. Assuming that vesicles always carry both Kin+ and Kout+ channels, these data suggest that the pool size, i.e. the number of vesicles, is not different between wt and transfected cells. It is more likely that the available vesicles in the pool of transfected cells have an eightfold higher concentration of KAT1 channels in their membrane.

The view of a large reservoir of vesicles carrying active KAT1 channels agrees well with the fluorescent images of transfected protoplasts. There, it is possible to detect numerous small fluorescent structures with a diameter of about 250 nm. These structures are appreciably smaller than the dictyosomes of these cells. These dictyosomes are not round shaped; they have dimensions of about 150 nm × 550 nm (Fig. 2 by Diekmann ., 1993). The GFP-labeled structures are also unlikely to be because of a staining of the endoplasmic reticulum, because an analysis of their three-dimensional architecture reveals that they are vesicular and never reticular in structure. Hence, the small fluorescent structures most likely correspond to vesicles containing KAT1::GFP. It is interesting to note that in some cases even an association of these small structures with the plasma membrane can be detected. This may show the fusion or fission of vesicles with the plasma membrane.

It has already been noted previously that the density of K+ channels in the vesicular membrane is higher than that in the plasma membrane (Homann and Thiel, 2002). This imbalance in channel density is also visible in the transfected cells. In these cells, the vesicular membrane has a channel density, which is about five times higher than that of the plasma membrane. This means that the concentration of channels in the plasma membrane is not in direct equilibrium with the upstream vesicle pool. Stimulation of exocytosis hence leads inevitably to a rise in the channel density of the plasma membrane. Such an exocytosis-mediated increase in channel density appears not to be unique to guard cells. Also in kidney cells it was found that insulin-stimulated exocytosis resulted in doubling of the number of open Na+ channels in the plasma membrane, while at the same time, the membrane capacitance only increased by about 15% (Erlij ., 1994).

The following model (Figure 6) would give a plausible explanation for this phenomenon. We may assume that channel clusters occur during pressure-driven and constitutive exo- and endocytosis. We may further speculate that the ratio of vesicles with channel clusters over those without channels is low during exocytosis and high during endocytosis. This would lead to a higher channel density of the vesicular membrane over the plasma membrane.

Figure 6.

Minimal model for trafficking of K+ channels in guard cells.

Channel proteins are synthesized and assembled as functional tetramers in the endoplasmic reticulum and Golgi (Q1) and transported (k1) into a preplasma membrane pool (Q2). This pool comprises vesicles with and without K+ channels in the membrane. Constitutive exocytosis (k2) results in an insertion of K+ channels into the plasma membrane via exocytotic vesicles. The high density of K+ channels in the vesicular membrane remains as channel clusters in the plasma membrane. The process of exocytosis (k2) can be augmented by positive pressure (+P). A constitutive endocytotic process (k−1) retrieves membrane from the plasma membrane with a preference for membrane with channel clusters. This is the reason why the density of channels in the plasma membrane is lower than in the vesicular membrane. Negative pressure (−P) augments the endocytic process (k−2). The fate of the retrieved vesicles is not yet known. They could be retrieved back into the preplasma membrane pool or into another compartment for recycling.

The notion of channel clustering is supported by the fact that GFP-stained vesicles are visible in guard cell protoplasts. The resolution limit of the CLSM requires that at least 200 chromophores are in close proximity in order to give a detectable fluorescent signal (Kubitscheck ., 1996). Considering that KAT1 is a tetramer (Dreyer ., 1997) and presumably carries four GFP chromophores, we can predict that the detected vesicles must contain at least 50 or more KAT1 channels. Hence, when incorporated into the plasma membrane, KAT1 channels are in clusters. Indeed, molecular analysis has in the past identified a C-terminal domain, which appears to be unique to plant K+ channels, including KAT1. This domain favors a self interaction of channel proteins and leads to formation of K+ channel clusters (Ehrhardt ., 1997). In this sense, K+ channel clusters can be seen as self-organizing units, which do not require peripheral structures such as the cytoskeleton.

A further strong argument in support of channel clusters in the plasma membrane is the observation that the membrane retrieved during shrinking has a higher concentration in than the plasma membrane. If the Kin+* channels were not clustered, this would require a process in which channels are first concentrated in the plasma membrane prior to endocytosis. However, as endocytosis of membrane with high K+ channel density occurs quasi immediately with onset of pressure stimulation, it is unlikely that such a process is relevant. It is therefore more plausible to assume that channels are in stable clusters, and that during pressure-stimulated endocytosis, membrane areas with these clusters are retrieved preferentially. The fact that the reversible change in the current density was also found in wt protoplasts demonstrates that channel clustering does not result from over-expression of KAT1::GFP. The observation that the exo- and endocytotic membranes, which account for changes in surface area, are essentially of the same ‘quality’ suggests that the same membrane domains, which are inserted during exocytosis, are again retrieved by endocytosis.

Experimental procedures

Construction/design of the KAT1::GFP fusion protein

The cDNA for KAT1 was amplified by PCR for directed cloning. A BSPLU11I restriction site (in bold) was introduced 5′ of the start codon with a forward primer (GGAAAACATGTCTATCTCTTGGACTCGAAAT) and a restriction site for the isoshizomer NcoI (in bold) was introduced 3′ with a reverse primer (TTTCCCATGGCATTTGATGAAAAATACAAATGATCACC), depleting the KAT1 stop-codon from the cDNA sequence. After digesting the restriction sites, the PCR product was ligated into the plant expression vector pAVA393/NcoI restriction site in frame with the mGFP5 (Haseloff ., 1997) for expression of the fusion protein KAT1::GFP under the control of two strong 35S promotors. The plasmid was then cloned in Escherichia coli/DH5α, followed by preparation of plasmid DNA (Qiagen high speed Midi-Kit, Qiagen, Germany). The purified vector was used for ballistic bombardment of intact guard cells.

Transfection of intact guard cells via particle delivery and isolation of protoplasts

Whole leaves of V. faba L. cv. Bunyan were placed upside down on solid Murashige Skoog medium and bombarded with 2 mg of gold (1 µm particle diameter) coated with 10 µg of DNA according to manufacturers instructions, at a pressure of 600 psi, a distance of 6 cm, and a vacuum of 25 inches Hg. This resulted in a transfection efficiency of 30% of intact guard cells. Guard cell protoplasts were prepared from bombarded or wt leaves as described previously by Homann (1998).

Electrophysiological measurements

For measurements, protoplasts were bathed in 10 mm KCl, 10 mm CaCl2, 5 mm Mes/KOH, pH 5.6, with osmolarities adjusted to 530 mosmol kg−1 with sorbitol. Patch pipettes were filled with 150 mm K+-gluconate, 10 mm KCl, 2 mm MgCl2, 2 mm MgATP, 2 mm EGTA, 10 mm HEPES/KOH, pH 7.8, with osmolarities adjusted to 560 mosmol kg−1 with sorbitol.

For the measurement of membrane capacitance and membrane current in standard whole-cell patch-clamp experiments (Lindau and Neher, 1988), a two-phase lock-in amplifier (SWAM IIC, Ljubljana, Slovenia) and an A/D converter (DigiData 1200, Axon Instruments, USA) were used. Patch pipettes were prepared as described previously by Homann (1998). The reconstitution of membrane capacitance (Cm), access conductance, and the parallel combination of leak and membrane conductance (Gm) were carried out by the computer software ‘cap 3’ (J. Dempster, University of Strathclyde, Glasgow, UK). For reading out pulse protocol and membrane currents, we used the program winwcp 3.2.7 (J. Dempster, University of Strathclyde, Glasgow, UK).

For the swelling and shrinking of guard cell protoplasts, hydrostatic pressure was applied via the patch pipette, measured with a pressure monitor (PM01D, WPI, Berlin, Germany) and recorded on the computer.

Data are presented as mean ± SE.

Confocal microscopy

Confocal microscopic analysis on guard cell protoplasts was carried out using a confocal laser scanning microscope (Leica TCS SP, Leica Microsystems GmbH, Heidelberg, Germany), equipped with a 63x water immersion objective (plan apo, N.A. 1.2). For excitation of mGFP5, the 488 nm line of a 25 mW Ar/Kr-Ion-Laser was used, emission was detected at 505–535 nm. The confocal aperture was adjusted to give optical sections of around 0.68 µm. Guard cell autofluorescence was subtracted in all images which were processed using the Leica Confocal Software 2.00 (lcs, Leica Microsystems GmbH, Heidelberg, Germany).


We thank Prof. J. Dainty for helpful comments on the manuscript and Dr Guido Kriete for providing the plant expression vector. This work was supported by a grant to U.H. and G.T. from the Deutsche Forschungsgemeinschaft.