INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Stomata close when guard cell turgor drops, a process initiated by plasma membrane depolarization. As soon as the membrane voltage becomes positive of the equilibrium potential (E_K), K⁺ leaves the cells accompanied by water due to osmotic forces. This depolarization is caused by anions (Cl and malate) rapidly moving down their electrochemical gradient, or by a "destimulation" of the plasma membrane proton pump. Although evidence is emerging that the activation of anion channels may be the preponderant event therein, it is clear that during stomatal closure, the activity of the pump must be reduced to be activated again upon stomatal opening.

Stomata close when the CO₂ concentration within the leaf increases (Willmer and Fricker, 1996). However, the way CO₂ is sensed by the leaf, and the signal transduced to the respective transporters, is poorly understood (Assmann, 1999). One reason for our ignorance with this problem may be the difficulty in studying the response of the functional entity (stomata plus substomatal cavity with adjacent cells) to well-controlled substomatal CO₂ concentration changes with high-time resolution in a quantitative manner under different physiological conditions. We have developed a technique to monitor ion activities in the guard cell apoplast within the intact leaf (Hanstein and Felle, 1999; Felle et al., 2000). To control the substomatal CO₂ concentration, this technique was combined with a minitube cuvette and a newly developed CO₂ microsensor that measures the CO₂ concentration within an individual substomatal cavity (Hanstein et al., 2001). With this system, it is possible to measure and to control the CO₂ concentration inside the leaf and thus any CO₂ concentration gradient that may be built up during tests. Because large Cl fluxes occur before stomata visually start to close, we use the online measurement of Cl efflux within the cavity (Felle et al., 2000) as an early and reliable indicator of stomatal closure.

Monitoring guard cell Cl efflux from within the substomatal cavity, whereas at the same time the substomatal CO₂ concentration can be controlled, offers access to two areas of scientific interest: In the field of guard cell signal transduction, we were able to address the problem of CO₂ sensing by guard cells. For instance, the question of whether guard cell anion channels are direct targets of CO₂ was addressed. In the field of integration of stomatal responses to different environmental stimuli, the interplay of light and CO₂ in the stomatal response to sudden shading was revisited. Sudden shading is accompanied by a rise in substomatal CO₂ within seconds (Laisk et al., 1984; Pearcy, 1990). Based on comparative analysis of the steady-state conductance of C₃ plants as a function of substomatal CO₂ and of the photon flux density in well-watered conditions, the influence of CO₂ on stomatal conductance is considered to be slight (Wong et al., 1978; Sharkey and Raschke, 1981; Ball et al., 1987). Differing from this steady-state approach, the current investigation focuses on the trigger potential of a fast CO₂ increase in the physiological range to accelerate the onset of stomatal closure.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Substomatal CO₂ Rapidly Responds to Light off and Light on

Light off is a stomatal closing signal. As CO₂ is incorporated into ribulose-1,5-bisphosphate, and the processing thereof is energy dependent, loss of light energy will disturb the dynamic equilibrium of CO₂ incorporation and CO₂ release (e.g. from respiration). Figure 1 shows a CO₂ measurement at a CO₂ concentration within the cuvette of 350 µL L¹. The sensor was first positioned in the air stream above the leaf surface and was then carefully moved toward the leaf surface and into a substomatal cavity. Within the cavity the CO₂ concentration was 90 ± 31 µL L¹ (SE, n = 7) higher compared with the cuvette air. Turning off the light quickly triggered a CO₂ increase of 120 ± 20 µL L¹ (SE, n = 6). Upon light on, the CO₂ level before light off was approximately restored. When the leaf surface was flushed with CO₂-free air, a substomatal CO₂ concentration of 149 ± 34 µL L¹ (SE, n = 9) was measured. Under these conditions, light off increased substomatal CO₂ by 95 ± 30 µL L¹ (SE, n = 4).

View larger version (12K): [in this window] [in a new window]   Figure 1.   Effect of light off on substomatal CO2 concentration measured directly with a CO2 microsensor. The leaf surface was enclosed in a minitube cuvette flushed with air containing 350 µL L1 CO2 (see text). The sensor tip was first placed 2 mm above the leaf surface, and was then moved toward a stoma, and was finally positioned within the substomatal cavity. Representative trace of six independent experiments.

Apoplastic Voltage

The apoplastic voltage was monitored with a blunt microelectrode, routinely used as voltage reference to the Cl-selective microelectrode. Because there is a high electrical resistance between the site of measurement and the bath harboring the cut petiole, initial voltage changes qualitatively represent the inverse of the plasma membrane potential. Thus, the apoplastic voltage can be used as reliable, noninvasive indicator of the direction of membrane potential changes occurring immediately after imposing the test conditions. As shown in Figure 2A, the increase in the CO₂ concentration from 350 to 600 µL L¹ was responded to by an apoplastic depolarization by 8 ± 1 mV within 1.6 ± 0.3 min (SE, n = 6), which reflects a hyperpolarization of the membrane potential; the decrease in CO₂ had the inverse effect. The apoplastic voltage also responded with an initial depolarization of comparable extent and velocity to light off. To determine the contribution of the mesophyll to the observed apoplastic voltage changes, the epidermis was removed and the apoplastic voltage was measured at a mesophyll cell (Fig. 2B). It was almost insensitive to CO₂ changes, whereas the light off effect was comparable with that measured in the guard cell apoplast (Fig. 2A). Figure 2C shows that the guard cell apoplastic depolarization depends on the CO₂ concentration, saturating around 700 µL L¹.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Effect of CO2 changes and of light off on apoplastic voltage (the voltage between the guard cell apoplastic fluid and ground, see text). A, Response of the apoplastic voltage in the vicinity of guard cells to an increase in CO2 within the cuvette compared with the light off effect. Representative kinetics of six independent experiments. B, Apoplastic voltage recorded within the mesophyll without interference from the epidermis. Measurement was performed after locally removing the epidermis. C, Response of the apoplastic voltage in the vicinity of guard cells to different CO2 concentration changes in the cuvette air. The CO2 concentration was switched from 0 µL L1 to 100 µL L1 for 1 min, and was then switched back to 0 µL L1. After 2 min, a concentration of 300 µL L1 was applied for 1 min, again followed by a CO2-free period of 2 min, before the next CO2 concentration level was applied.

CO₂ Triggers Cl Efflux

In Figure 3 the apoplastic Cl activity is shown to respond continuously to changes in CO₂ within the cuvette from 150 to 800 µL L¹ and back. Short CO₂ pulses of 1 or 2 min in duration (measurement of substomatal CO₂ see inset) did not or only marginally changed Cl activity, did not initiate stomatal closure, but they clearly affected the membrane potential (kinetics of membrane potential changes see Fig. 2). Extended exposure to 800 µL L¹ CO₂ resulted in a rapid apoplastic increase in Cl activity that, after approximately 13 min, peaked around 16 mM, and over about 90 min, spontaneously returned to the starting level of about 2 mM. Stomata were completely closed 45 min upon the rise in CO₂. A subsequent decrease in CO₂ caused a rapid drop in Cl to below 1 mM. As indicated by the ovals (visual observation of stomatal aperture), the stomata were still open when Cl activity peaked; this test demonstrates that Cl efflux precedes stomatal closure and signals its onset.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Response of guard cell apoplastic Cl activity (pCl) to periods of elevated CO2 of different duration. The CO2 concentration within the cuvette ([CO2]cuv) is given in the bar on the top with shaded areas for periods of high [CO2]cuv. Inset shows the substomatal CO2 change (ci) during the 2-min CO2 pulse, which was directly recorded by a CO2 microsensor within the substomatal cavity. Stomatal apertures were microscopically observed (open ovals, stomata open; closed ovals, stomata closed). The experiment was conducted at a photon flux of 300 µmol m2 s1, a temperature of 22°C, and a relative air humidity of 80%. The figure gives an example of three experiments performed with different leaves.

The Physiological Relevance of the CO₂ Trigger

We were interested in the control potential (physiological impact) of the closing signal "elevated CO₂" in relation to the closing signal light off. In Figure 4, the substomatal CO₂ concentration and the apoplastic Cl activity are shown responding to different CO₂ concentrations in the light and after light off. The switch from 350 to 600 µL L¹ CO₂ within the cuvette is responded to by an increase in apoplastic Cl activity to a level of 13.8 ± 2.5 mM (SE, n = 5), which is comparable with that shown in Figure 3. It took 1.6 ± 0.9 min until the Cl activity had changed by more than 10%. A time of 3.8 ± 1.2 min was required until Cl had changed by more than 50%. When CO₂ was increased by 120 µL L¹, which is equivalent to the CO₂ increase that occurs upon light off, the Cl increase was about 50% of that induced by 250 µL L¹.

View larger version (21K): [in this window] [in a new window]   Figure 4.   Response of guard cell apoplastic Cl activity (pCl) to CO2 changes within the cuvette (bar on the top) in the light (at 300 µmol m2 s1) and in darkness. Substomatal CO2 concentration ([CO2]sub) was recorded by a CO2 microsensor directly within the substomatal cavity. Measurements of pCl and of substomatal CO2 were sequentially performed within the same cavity. The data basis is two experiments with different leaves.

Figure 4 shows that Cl activity dropped to levels below 1 mM following the reduction in CO₂ to 0 µL L¹. Visual control proved that stomata were open at that point. When light was turned off, substomatal CO₂ increased to 300 µL L¹, but apoplastic Cl activity responded with a minor transient increase only. Under these conditions, stomata remained open for the next 1 h. Even an increase in CO₂ within the cuvette to 250 µL L¹ (=550 µL L¹ in cavity) could not trigger the typical Cl efflux to a level above 10 mM. Only when the CO₂ within the cuvette was increased to 500 µL L¹ did the apoplastic Cl rise to about 16 mM and the stomata close.

The kinetics in Figure 5 show that the effectiveness of the light off signal in inducing Cl efflux increases at a higher CO₂ level within the cavity. The protocol shows that, in agreement with Figure 4, the changes in the CO₂ concentration within the cuvette from 350 to 600 µL L¹ and back (control) yielded the expected Cl response. However, when in contrast to the test shown in Figure 4, the cavity CO₂ concentration was experimentally kept constant at 400 µL L¹ (CO₂ clamp), light off alone (without concomitant CO₂ increase) triggered Cl efflux to a level of 23.4 mM (n = 2; 22.8 and 24 mM) within 25 min and caused stomatal closure.

View larger version (21K): [in this window] [in a new window]   Figure 5.   Response of guard cell apoplastic Cl activity (pCl) to a CO2 increase by 250 µL L1 compared with the CO2-independent light off effect. The CO2 concentration within the cuvette is given in the bar on the top. Upon light off, the substomatal CO2 concentration ([CO2]sub) was held constant by simultaneously lowering the CO2 concentration in the cuvette (CO2 clamp). The data basis is two experiments with different leaves.

Adaptation

The Cl responses to light off and to elevated CO₂ differed according to the duration of exposure to low CO₂ concentrations. Whereas light off shortly after switching to 0 µL L¹ (CO₂ concentration within the cuvette) only caused a minor rise in apoplastic Cl to a level of 3.2 mM (n = 3: 1.3, 2.5, and 5.8 mM), and within the following 1 h did not increase apoplastic Cl activity or closed the stomata (Fig. 4), a prolonged adaptation to 0 µL L¹ (150 µL L¹ cavity CO₂) yielded a different response. As shown in Figure 6, light off still did not trigger an immediate Cl efflux after being exposed to low CO₂ for about 100 min, but did so after being about 20 min total darkness. The apoplastic Cl activity increased to a level of 6.2 mM (n = 3; 4.5, 6.9, and 7.2 mM). Similar results were obtained for elevated CO₂ (Fig. 7): After a 30-min adaptation to 0 µL L¹, the apoplastic Cl activity remained indifferent to increases from 0 to 225 µL L¹ or to 450 µL L¹, and only 900 µL L¹ triggered the typical Cl transient (Fig. 7A; n = 3). However, after prolonged adaptation (70 min) to 0 µL L¹, the increase from 0 to 225 µL L¹ triggered the Cl transient in five out of six leaves (Fig. 7B). These experiments demonstrate that for both stimuli, light off and elevated CO₂, adaptation shifts the sensing threshold necessary to initiate stomatal closure. A remarkably delayed CO₂ response was observed in the case of temporary darkness during the adaptation period to 0 µL L¹ CO₂ within the cuvette (Fig. 6): Approximately 10 min after the end of the intermittent dark phase, the CO₂ concentration was raised to 175 µL L¹, and the Cl response was delayed by about 30 min.

View larger version (11K): [in this window] [in a new window]   Figure 6.   Effect of light off (L. off) on guard cell apoplastic Cl activity (pCl) after adaptation to different CO2 levels (given in the bar on the top). The data basis is three experiments with different leaves.

View larger version (21K): [in this window] [in a new window]   Figure 7.   Adaptation to low CO2 concentrations. Apoplastic Cl activity (pCl) responding to a CO2 increase following a prolonged CO2-free period. The CO2 concentration within the cuvette is given in the bar on the top. Leaves were exposed to CO2-free cuvette air for 30 min (A) or for 70 min (B) before the CO2 concentration was stepwise increased, as indicated. Data are representative of six leaves investigated after 70 min of adaptation and three leaves investigated after 30 min of adaptation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

A Sensitive Parameter to Monitor the Onset of Stomatal Closure

Upon a change in ambient CO₂, there is a general pattern of short-term stomatal response: Decreasing the CO₂ concentration stimulates stomatal opening, whereas increasing the CO₂ concentration triggers closure (Morison, 1987). Because stomatal closure only occurs after ions (anions and K⁺) have started to be released from the guard cells, any method to directly observe the actual movement must be inferior to the measurement of the preceding channel activation, monitored as ion fluxes. The measurement of a pronounced Cl efflux from guard cells within the intact leaf has been demonstrated following the closing signals light off and abscisic acid (Felle et al., 2000). In the current investigation, we adopted the measurement of apoplastic Cl activity to monitor the onset of stomatal closure upon a rise in ambient CO₂. As expected, the increase in the apoplastic Cl activity can be measured before the progressive release of osmotica from guard cells visibly affects stomatal aperture (Fig. 3). We do not infer that Cl efflux was the primary event because there is evidence that the activation of the anion channels is triggered by cytosolic Ca²⁺ elevation (Ward et al., 1995; Webb et al., 1996; Webb and Hetherington, 1997; Allen et al., 1999). Ca²⁺ was not chosen as the measuring parameter because Ca²⁺ as a signaling element is not directly involved in the osmotic or charge balance during stomatal closure, whereas Cl is and thus reliably signals the onset of stomatal closure. The validity of this notion is underscored by the observation that short CO₂ pulses from 150 to 800 µL L¹ (Fig. 3) did not trigger Cl efflux or induce stomatal closure. As a consequence, measuring apoplastic Cl activity provides information not only on the kinetics of the upstream signal transduction that activates Cl efflux, but also on the effectiveness of a given CO₂ increase to bring about an aperture change.

Over extended periods, the composition of osmotica within the guard cell may shift between ionic species on the one hand and Suc on the other hand (Talbott and Zeiger, 1996). Of course, whenever such shifts occur, apoplastic Cl activity is only of limited use in comparing the effectiveness of environmental signals to initiate stomatal closure. The same holds true for long-term shifts between chloride and malate as the major counter ion of potassium. This investigation reports remarkable lag phases upon a rise in CO₂ that precede Cl efflux (Fig. 6: CO₂ increase from 0 to 175 µL L¹) and even insensitivity of the "sensitive parameter" (apoplastic Cl activity) toward an increase in CO₂: When the CO₂ concentration in the cuvette drops from 350 µL L¹ to 0 µL L¹ (substomatal CO₂ concentration in the light about 150 µL L¹; see Fig. 4), a subsequent rise in CO₂ by 225 µL L¹ does not trigger an increase in apoplastic Cl activity (Fig. 7A). By further stepwise increase in CO₂, it was possible to activate the anion channels, but it is interesting that between the last CO₂ step and the onset of Cl release, several minutes passed (Fig. 7A). Because the apoplastic Cl activity was in the range of 100 µM, any earlier Cl release from guard cells would have been easily detected. The insensitivity occurring upon CO₂ steps from 0 to about 200 µL L¹ raises the question whether the lack of the Cl response could be attributed to a replacement of Cl as a major osmoticum by organic compounds. However, a 30-min treatment with CO₂-free air is very unlikely to cause a significant build-up of carbon assimilates like Suc or malate. Moreover, after extended exposure to CO₂-free air, the influence of CO₂ on the apoplastic Cl activity did not further decrease, but it increased again (Fig. 7B). As a consequence, we have no argument to underscore that Cl was replaced by other osmotica during the exposure to CO₂-free air. Thus, the data strongly indicate that the CO₂ signal transduction chain does not convey the rise in CO₂ to guard cell anion channels or that the signal transduction requires a longer time, thereby preventing or delaying stomatal closure. The resulting hints for the mechanism of CO₂ signal transduction will be discussed below.

Controlling the Substomatal CO₂ Concentration by a Minicuvette

The use of microsensors for observing the onset of stomatal closure has implications for the experimental system. A system was required that did not only allow a controlled CO₂ concentration and fast concentration changes, but also access of microelectrodes into substomatal cavities. Fast cuvette systems for measuring transient gas exchange have been designed for studying the impact of transient light (sunflecks in forest understories) on photosynthesis (Pearcy, 1990). The use of 2-mm thin, single-sided minicuvettes allowed for response times below 1 s for measuring transient gas exchange (1990). In our approach, a single-sided "leaky" minicuvette had to be used containing a window for electrode access. The employed tube-shaped minicuvette offered the advantage of efficiently suppressing CO₂ exchange between the leaf surface and the electrode window on the opposite side at a low gas supply rate due to the well-defined air flow through the tube (Hanstein et al., 2001). Because the leaf surface enclosed by the leaf window was small, substomatal CO₂ could not be measured by conventional infrared gas analyzers (Ball, 1987). Thus, a CO₂ microsensor was used to measure CO₂ directly within the cuvette or within a single substomatal cavity (Hanstein et al., 2001).

It is surprising that measurements with this sensor within the illuminated leaf revealed substomatal levels higher than the CO₂ concentration in the cuvette at any concentration tested between 0 and 800 µL L¹ (Figs. 1, 4, and 5). As this was not due to concentration differences between inside and outside, it must have metabolic sources representing dynamic equilibria between CO₂ consuming and producing processes, where high respiration rates obviously preponderate. The validity of this observation is underscored by the demonstration that any change in CO₂ within the cuvette is closely tracked by substomatal CO₂ (e.g. Fig. 4). We consider the elevated substomatal CO₂ concentration to be a consequence of the pressure exerted by the sealing cuvette edge, which obviously triggers respiratory processes in the underlying cells and increases CO₂ also in the neighboring tissue where the measurement is performed. In fact, after 1 d of acclimation, a substomatal CO₂ concentration below 350 µL L¹ was measured at a CO₂ concentration of 350 µL L¹ in the cuvette (not shown).

Hints for CO₂ Signal Transduction Mechanisms

Although the input signal of CO₂-dependent regulation of guard cell movements is considered to be the substomatal CO₂ (Mott, 1990), the mechanism by which guard cells sense and respond to CO₂ signals remains poorly understood. The combination of Cl and CO₂ measurements is a versatile tool in testing current hypotheses on CO₂-dependent stomatal control. Upon a fast rise in substomatal CO₂ by 650 µL L¹, it took several minutes until the apoplastic Cl activity in the direct vicinity of guard cells had increased by 1 mM (calculated from the pCl values in Fig. 3). This slow initial stage of Cl release indicates that upon contact with increased CO₂ concentrations, Cl channels are not activated immediately, suggesting an indirect influence of CO₂ on these channels.

Hedrich and Marten (1993) have proposed a positive feedback model in which CO₂-stimulated and photosynthetically driven changes in apoplastic malate activate guard cell anion channels (R-type), resulting in anion loss. From the moment of a CO₂ concentration change to the arrival of malate in the apoplast at concentrations high enough to stimulate the anion channels (Hedrich et al., 1994), time elapses that would depend on the amount of CO₂ added, but also on the CO₂ concentration to which the stomata were adapted. This model agrees with our observations of a considerable delay in Cl efflux that occurs after a switch to CO₂-free cuvette air (unless an adaptation process has taken place). The dependence on substomatal CO₂ is shown by the data in Figures 4 and 5 where the change from 350 to 600 µL L¹ CO₂ (cuvette) triggered Cl efflux to an apoplastic level above 10 mM within 10 min in comparison with CO₂ changes from 0 to 250 µL L¹ (Fig. 4), 0 to 175 µL L¹ (Fig. 6), and 0 to 225 µL L¹ or 225 to 450 µL L¹ (Fig. 7A), which did not immediately trigger Cl fluxes. Brearley et al. (1997) have reported that in epidermal strips activation of anion currents occurs within 2 min upon a rise in CO₂ from 350 to 1,000 µL L¹. This fast response of Cl currents in the absence of the mesophyll seems to argue against a role for malate in the activation of guard cell anion channels. However, it does not contradict the more general hypothesis that some CO₂ product (not necessarily malate) is required for channel activation, because the rate of formation of a CO₂ product is expected to be a function of the CO₂ concentration, and the CO₂ concentration applied by Brearley et al. (1997) was rather high. Moreover, activation thresholds of guard cell anion channels could be affected by removing the mesophyll tissue from the epidermis due to differences in the apoplastic ionic conditions, a different pressure exerted by the cell wall corset or due to the wounding response of the epidermis itself.

We found that the sensitivity of the Cl efflux to CO₂ was approximately the same in the light and in darkness. Due to the lack of reliable malate concentration data from the substomatal cavity, it is difficult to judge whether the malate model would hold in light and in darkness. However, it is clear that because of the long lag phases, the anion channel cannot be the primary CO₂ sensor. Minor Cl responses to CO₂ increase that are observed here and there may be related to changes in apoplastic pH (data not shown) or in voltage (see below).

Zeiger and Zhu (1998) proposed a model in which an increase in the CO₂ level leads, via CO₂ fixation, to changes in ATP and NADPH status and thus to an increase in the pH of the chloroplast lumen. Because of the pH sensitivity of the enzymes involved in zeaxanthin formation within the chloroplast lumen, this would lead to a decrease in zeaxanthin levels, which is in fact demonstrated when leaves or epidermal peels were exposed to elevated CO₂. Our observation that CO₂ sensitivity of Cl efflux was the same in light and darkness would not favor the idea of different CO₂ sensors for light and dark, as suggested by Zeiger and Zhu (1998). Zeaxanthin is supposed to be the photoreceptor for blue light that triggers blue light-dependent stomatal opening (Shrivastava and Zeiger, 1995; Assmann and Shimazaki, 1999). Our data obtained with white light indicate that zeaxanthin does not participate in the transduction of the light off signal: At a low ambient CO₂ concentration, which has been shown to increase the zeaxanthin concentration (Zeiger and Zhu, 1998; Zhu et al., 1998), triggering of Cl efflux by light off was less efficient compared with normal ambient CO₂ (Figs. 4 and 6).

Is there a role for cytosolic pH? Cl efflux was shown to be independent of initial changes in the simultaneously recorded apoplastic voltage. Figure 2 shows that the initial response of the apoplastic voltage (qualitatively the inverse of the membrane potential) to an increase in CO₂ and to light off is a rapid depolarization (i.e. membrane hyperpolarization) of comparable magnitude. In fact, these voltage changes depend on the CO₂ concentration supplied and they saturate at around 700 µL L¹ (Fig. 2C). Using pH-sensitive microelectrodes, Felle and Bertl (1986a, 1986b) demonstrated in a variety of green cells, that light off always resulted in an initial cytosolic acidification of about 0.3 pH units. As protons are transport substrate of the plasma membrane H⁺ pump, this doubling of the cytosolic H⁺ activity was responded to by a transient hyperpolarization (equivalent to a depolarization in the apoplast). Here, it is suggested that the apoplastic depolarization (= plasma membrane hyperpolarization) due to CO₂ elevation may also be due to cytosolic acidification (proposed by Raschke, 1975). Because the short CO₂ pulses (Fig. 3) did not trigger Cl fluxes, but caused apoplastic depolarizations (and thus presumably also cytosolic acidification), it is unlikely that these early pH changes are involved in CO₂ sensing in a causal manner. Because this notion concerns only the primary responses to CO₂, it does not collide with Blatt and Armstrong's (1993) report on the activation of the K⁺_out rectifier by cytosolic alkalization. Likewise, apoplastic pH is also not a primary factor in CO₂ sensing because it is the result of ion fluxes and not their cause (Felle et al., 2000).

The Potential of CO₂ to Initiate Closing Responses upon Light off

The physiological role of substomatal CO₂ changes for the stomatal adjustment to varying irradiance has been a matter of extensive debate (Mansfield and Meidner, 1966; Raschke, 1975; Wong et al., 1979; Sharkey and Raschke, 1981; Morison, 1987). The observation that over a wide range of photon flux densities, the substomatal CO₂ concentration is stabilized by stomatal adjustment caused Raschke (1975) to assume that stomatal conductance is regulated by the substomatal CO₂ concentration in a feedback loop. The assumed feedback loop should also restore substomatal CO₂ when the ambient CO₂ concentration is changed. However, in many plant species, after stomatal adjustment to increased ambient CO₂, substomatal CO₂ is at a higher level (Wong et al., 1979; Jarvis and Morison, 1981). After Morison (1987) had raised the question of whether there is a direct CO₂ effect on stomata at all, investigations in guard cells of epidermal strips have demonstrated that a rise in CO₂ by about 700 µL L¹ increases cytosolic Ca²⁺ activity (Ward et al., 1995; Webb et al., 1996) and regulates the conductance of plasma membrane ion channels so that the efflux of K⁺ and Cl is facilitated (Brearley et al., 1997). Our investigation provides data for a comparison between smaller CO₂ steps and light off as triggers of guard cell Cl release in intact leaves. We have demonstrated that stepping up substomatal CO₂ within physiologically relevant margins triggers guard cell Cl efflux. For instance, an increase of 250 µL L¹ CO₂ leads to a Cl peak within 10 to 12 min (Figs. 4 and 5) and induces stomatal closure. Keeping substomatal CO₂ constant by CO₂ clamp, light off yields a comparable increase in Cl activity and a closing response (Fig. 5). On the other hand, when substomatal CO₂ is not kept constant, light off causes an increase in CO₂ of 120 µL L¹ (physiological CO₂ increase; see Fig. 1), which is roughly one-half of the CO₂ step-up shown in Figure 5. From this, it follows that light off and a 2-fold physiological CO₂ increase have approximately the same potential to trigger Cl efflux, i.e. to effectively initiate stomatal closure.

The pronounced Cl efflux upon light off at constant substomatal CO₂ (Fig. 5) demonstrates a strong potential of guard cells to initiate stomatal closure independent of CO₂. However, the existence of an additional CO₂-dependent closing trigger may be required to accelerate stomatal closure under certain environmental conditions: Based on analysis of air trapped in ice cores (Delmas et al., 1980), the atmospheric CO₂ level of the past 165,000 years was found to shift between 190 and 280 µL L¹ (Tolbert, 1997). At a substomatal CO₂ concentration of 150 µL L¹ compared with 440 µL L¹, we have clearly detected that light off becomes less efficient in triggering Cl release from guard cells. In most cases, the response was quickly initiated, but the Cl amplitude remained small (Fig. 4) or the response was delayed (Fig. 6). Under such conditions, selection for accelerated stomatal closing after shading stops photosynthetic carbon acquisition may have driven the evolution of supplementary closing mechanisms.

Fluctuation in irradiance is common within plant canopies where leaves are shaded by other leaves of the same or of other plants, and where the shade moves with the sun or as a consequence of wind. In a forest or a crop canopy on clear days, 20% to 80% of the irradiance is received as light flecks with intermittent shading (Pearcy, 1990). It has been reported for trees of a tropical rainforest that stomatal conductance remains high in shade leaves subjected to periodic sunflecks, owing to the fact that the opening response of stomata after the shade-light transition shows a hysteretic nature that predominates the closing response after the subsequent light-shade transition (Pearcy and Calkin, 1983; Assmann et al., 1985; Pearcy, 1990). The CO₂-dependent closing trigger may play a major role in terminating this light-induced hysteretic opening response.

Essentially two points emerge from these data. First, we have demonstrated in the intact leaf that upon light off a light-independent CO₂ trigger acts in concert with a CO₂-independent light off stimulus in promoting Cl release. The physiological role of the CO₂ trigger at low substomatal CO₂ concentrations and in fluctuating light deserves further investigation. Second, although some direct influence of CO₂ on anion channels cannot be ruled out, the occurrence of pronounced lag phases before the onset of Cl efflux under conditions of low CO₂ supply (Fig. 6: CO₂ increase from 0 to 175 µL L¹) indicates that full channel activation requires a strongly time-dependent effector, possibly a product of carboxylation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plants and Gas Exposure

Fava beans (Vicia faba cultivar Witkiem Major; Nunhems Zaden bv, Haelen, The Netherlands) were grown in a glasshouse at a temperature of 20°C to 30°C in the day, 20°C at night, and at a relative air humidity of 60% ± 5%; during the photoperiod of 13 h d¹, natural light was supplemented with 100 µmol m² s¹ (SONT-Assimilationsleuchte, Philips, The Netherlands). Plants were grown in 250-mL pots containing a mixture of peat, compost, and sand at a volume ratio of 2:2:1, fertilized with 250 mg of Triabon (BASF, Ludwigshafen, Germany). For the experiments, the 3rd or the 4th fully developed leaf from 4-week-old plants was cut from the stem with a razor blade and was placed immediately in the standard test solution (1 mM KCl and 0.1 mM NaCl and CaCl₂ each) on a Plexiglas holder.

To control the CO₂ concentration at the leaf surface, a small flow-through cuvette was placed on one side of the leaf. Single-sided leaf cuvettes with a small volume have been useful for measuring fast transients in photosynthetic CO₂ uptake (Pearcy, 1990). In the current investigation, a small cuvette with a tube-like shape (described by Hanstein et al., 2001) served to perform fast CO₂ changes at the leaf surface. The transparent tube was mounted on the leaf with one oval hole (window) enclosing about 12 mm² of the leaf surface. To achieve a tight fit between the window edge and the leaf, the leaf support, a flexible adhesive compound (plastic-fermit; Nissen and Volk, Hamburg, Germany) had been preformed to match the shape of the window edge. Sealing between the window edge and the leaf was accomplished by a weight fixed to the tube, which exerted a pressure of 4.5 kPa onto the leaf surface lying under the window edge (approximately 40 mm²). The sensors were positioned within the cuvette through a second cuvette window above the enclosed leaf surface. Air with preset CO₂ concentrations passed the tube at a rate of 1.5 L min¹. Experiments were started 4 h after positioning the cuvette. During this period, the leaves were gradually adapted to a photon flux of 300 µmol m² leaf surface s¹ (white light). The light was supplied through a fiber cold-light source (KL 1500; Leica, Wetzlar, Germany), and photon flux was measured by the quantum sensor of an LCA 4 photosynthesis measurement system (ADC, Hoddesdon, Herts, UK).

Sensors

The fabrication and measuring principle of the CO₂ sensor has been described in detail previously (Hanstein and Felle, 2001). In brief, the sensor is built of two concentrically arranged capillaries, the inner one being a pH-sensitive microelectrode. The tip of this electrode is placed closely behind the tip of the outer capillary (diameter 2 µm), the opening of which is plugged with silicone (Dow Corning 3140 RTV Coating; Sasco Semiconductor, Frankfurt, Germany). CO₂ diffuses through this plug, reacts with water, and acidifies the tip solution, which is detected by the pH electrode. To accelerate this reaction, carbonic anhydrase is added to the solution behind the plug. A Teflon-coated silver wire (AG-3T; Clark Electromedical Instruments, Reading, UK) leading from the carbonic anhydrase solution to ground serves as reference.

Cl-selective microelectrodes were built and used as described recently (Felle et al., 2000). In brief, blunt, heat-polished microcapillaries with a tip diameter of 2 µm were internally silanized with 0.2% (w/v) tributylchlorosilane (Fluka Chemical, Ronkonkoma, NY) dissolved in chloroform. Capillaries were backfilled first with a mixture of Cl sensor cocktail (24902; Fluka), tetrahydrofuran, and polyvinylchloride. After evaporation of the tetrahydrofuran, the remaining gel was topped up with the undiluted sensor cocktail followed by a 0.1 mM KCl buffer solution. CO₂ sensor and Cl electrode were connected to high-impedance amplifiers (FD223; WP Instruments, Sarasota, FL), which simultaneously measured and subtracted the signals coming from the ion-selective electrodes and the voltage reference. The sensors were positioned with micromanipulators from Marzhauser (DC-3K; Wetzlar, Germany).

Measuring Conditions

Unless stated otherwise, measurements were performed at a photon flux of 300 µmol m² s¹, a relative air humidity of 95%, and at a temperature of 22°C. Stomatal aperture in leaves adapted to a CO₂ concentration within the cuvette of 350 µL L¹ was approximately 6 µm (measured by a micrometer eyepiece).