Red light activates a chloroplast-dependent ion uptake mechanism for stomatal opening under reduced CO2 concentrations in Vicia spp.

Authors


Author for correspondence: Gary Tallman Tel: +1 503 370 6611 Fax: +1 503 375 5425 Email: gtallman@willamette.edu

Summary

  •  Under red light in ambient CO2 guard cells of faba bean (Vicia faba) fix CO2 and accumulate sucrose, causing stomata to open. We examined whether at [CO2] low enough to limit guard cell photosynthesis stomata would open when illuminated with red (R) or far-red (FR) light.
  •  After illumination with R or FR in buffered KCl solutions, net stomatal opening was c. 3 µm (R and FR) in air containing 210–225 µl l−1 CO2 and was 5 µm (R) or 6.5 µm (FR) in air containing 40–50 µl l−1 CO2. Opening was fully inhibited by 3-(3,4-dichlorophenyl)-1,1 dimethyl urea, the calmodulin antagonist W-7, the ser/thr kinase inhibitor ML-9, and sodium orthovanadate, but not by dithiothreitol, which inhibits formation of zeaxanthin, the blue light photoreceptor of guard cells.
  •  Stomatal opening was accompanied by K+ uptake and starch loss. Similar results were obtained when leaves were exposed to conditions designed to lower intercellular leaf [CO2].
  •  These data suggest that the guard cell chloroplasts transduce reduced [CO2], activating stomatal opening through an ion uptake mechanism that depends on chloroplastic photosynthetic electron transport and that shares downstream components of the blue light signal transduction cascade.

Introduction

Two environmental signals trigger stomatal opening in intact, healthy leaves of C3 and C4 plants: increased light intensity (Sharkey & Ogawa, 1987) and decreased intercellular concentrations of leaf CO2 (ci; Mott, 1988, 1990; Morison, 1998; Assmann, 1999). Experiments with detached leaf epidermis and guard cell protoplasts (GCPs) have shown that the guard cells that flank each stoma have intrinsic mechanisms that equip them to perceive and transduce these signals into stomatal opening. The cellular mechanism(s) by which guard cells translate lowered ci into stomatal opening is/are not well understood. According to Assmann (1999) we do not know with certainty: (1) how many mechanisms guard cells may have for perceiving changes in CO2 concentrations; (2) where in the guard cell such mechanism(s) is/are located; (3) the components of any mechanism for CO2 perception; (4) whether the mechanisms for stomatal opening under reduced CO2 concentrations are the same as those that result in stomatal closure at increased CO2 concentrations; or (5) how CO2-sensing mechanisms may be integrated either with each other or with guard cell signal transduction mechanisms for responding to light, abscisic acid, etc.

Complicating the analysis of CO2 signaling of stomatal opening is the possible interaction of the guard cell CO2 response with light signaling mechanisms for opening. Among the best studied photosensory signal transduction cascades of guard cells is the one activated by low fluences of blue light (photosynthetic photon fluence rate [PPFR] = 0.5–50 µmol m−2 s−1; Zeiger et al., 1987; Zeiger, 2000). Blue light-induced stomatal opening is observed when blue light is applied over a background of red light to whole leaves, isolated guard cells, or GCPs (Zeiger et al., 1987). The blue light photoreceptor is zeaxanthin (Z) (Frechilla et al., 1999), a carotenoid pigment formed when the thylakoid membrane of the chloroplast is energized by light (Zeiger, 2000). Light-induced acidification of the thylakoid lumen activates violaxanthin (V) de-epoxidase, which catalyses the conversion of V to Z (Zeiger & Zhu, 1998; Zeiger, 2000). By some unknown mechanism that may be initiated by blue light-activated cistrans isomerization of Z (Zeiger, 2000), a signal transduction cascade is triggered that involves calmodulin and a ser/thr kinase (Shimazaki et al., 1992; Shimazaki et al., 1993; Assmann & Shimazaki, 1999). The kinase catalyses phosphorylation of the C-terminal end of a P-type H+-translocating ATPase (proton pump) in the guard cell plasma membrane (Kinoshita & Shimazaki, 1999). Phosphorylation promotes binding of a guard cell-specific 14-3-3 protein to the C-terminal end of the pump, stabilizing pump activation (Emi et al., 2001). The resulting proton extrusion causes the guard cell plasma membrane to hyperpolarize to voltages that regulate opening and closing of a variety of ion channels to favor influx of K+ and water (Grabov & Blatt, 1998), guard cell turgor production, and stomatal opening.

Elevated CO2 reduces both guard cell Z content and light-induced stomatal opening (Zhu et al., 1998); thus, it has been suggested that guard cells use Z levels to integrate both blue light and CO2 signals (Zeiger & Zhu, 1998; Zeiger, 2000). In this model, Z levels would be expected to vary inversely with rates of photosynthetic carbon fixation in guard cells. At lower CO2 concentrations, photosynthetic carbon fixation would be decreased, creating: (1) increased demand for Z to ‘de-energize’ the thylakoid membrane; (2) conditions in the thylakoid lumen that would favor Z formation; and (3) increased guard cell sensitivity to blue light (Zeiger, 2000). Elevated CO2 concentrations would favor increased photosynthetic carbon fixation by guard cells, which would reduce acidification of the thylakoid lumen, decrease conversion of V to Z and decrease guard cell sensitivity to blue light (Zeiger, 2000). In this model, CO2 would exert its action through the blue light signal transduction pathway. An attractive feature of this hypothesis is the idea that the chloroplast may be used to integrate light and CO2 responses of guard cells. It has long been thought that guard cell chloroplasts are specialized for transducing the signals that trigger stomatal opening (Outlaw et al., 1981), but other possible mechanisms involving Ca2+ (Webb et al., 1996) or malate (Hedrich & Marten, 1993) have been proposed to explain the effects of CO2 on stomata (for a review, see Assmann, 1999), and even stomatal opening in darkness in response to reduced CO2 concentrations has been documented (Dale, 1961; Mansfield & Heath, 1961).

One way to separate other effects of CO2 on stomatal opening from those mediated by its interaction with the blue light signal transduction mechanism would be to examine CO2 responses under red light. In gas exchange studies by Assmann (1988) in which intact leaves of Commelina communis were illuminated with red light prior to administration of low-fluence blue-light pulses, stomatal conductance was c. 150 µmol m−2 s−1 at a red light fluence of only 263 µmol m−2 s−1. This initial stomatal response to red light prior to blue light treatment could be interpreted as a guard cell response to reduced ci due to photosynthesis. Based on these studies, we hypothesized that in epidermis detached from leaves of Vicia faba, stomata illuminated with red light would open more than those of dark controls under reduced concentrations of CO2. Even though under red light GCPs can fix CO2 (Gotow et al., 1988) and isolated guard cells can accumulate sucrose (Poffenroth et al., 1992; Talbott & Zeiger, 1993), we reasoned that any red light-induced opening at low concentrations of CO2 would be unrelated to guard cell photosynthetic carbon fixation and/or Z formation, because: (1) low CO2 concentrations would be expected to reduce photosynthetic rates; and (2) Z is specific for blue light photoreception. Because chlorophyll is the only known red light photoreceptor of guard cells, we hypothesized that any mechanism for opening observed at reduced concentrations of CO2 under red light would be located in the guard cell chloroplast and that it might be sensitive to inhibitors of photosynthetic electron transport such as 3-(3,4-dichlorophenyl)-1, dimethyl urea (DCMU). Finally, we hypothesized that any such response should be insensitive to dithiothreitol (DTT), a known inhibitor of V de-epoxidase and Z formation (Srivastava & Zeiger, 1995). The results of our studies indicate that the guard cell chloroplast harbors at least one thylakoid-dependent, but Z-independent, mechanism for sensing reduced concentrations of CO2.

Materials and Methods

Plants

Vicia faba L. (Loretta) was grown from seed (Johnny’s Seed, Albion, ME, USA) in a greenhouse on the campus of Willamette University in Salem, OR, USA. Seeds (two per 6.3-l pot) were germinated and plants were grown in HP Premier Pro-mix potting soil (Premier Horticulture Ltee, Rivere-du-Loup, Quebec, Canada). Upon germination, plants were watered every other day until the soil was saturated. Plants were grown under natural daylight supplemented with light from sodium vapor lamps (PPFR of lamps = 100 ± 15 µmol m−2 s at the top of the canopy; mean ± SE, n = 30) to give a 20-h daylength. Experiments were conducted between May 15 and September 1 in 1997–2000. Greenhouse temperature, PPFR and relative humidity were measured daily at 12 : 00 h Pacific Standard Time (PST) using a steady-state porometer (Model LI-1600, Li-Cor Inc., Lincoln, NB, USA) held at canopy height. Noon glasshouse temperatures ranged from 19 to 26°C, relative humidity from 26 to 58%, and PPFR from 230 to 1780 µmol m−2 s−1.

Epidermis

At 17 : 00 h on the day prior to each experiment, plants were watered until the soil was saturated and then placed in a darkened cabinet. At 10 : 00 h the next day, leaves from 3- to 5-wk-old plants were removed from insertion levels 3 or 4 above the soil surface and were transported to the laboratory in moist towels in a plastic bag. Epidermis was detached (‘peeled’) from between major veins on abaxial leaf surfaces using forceps. To minimize contamination with mesophyll tissue and prevent dehydration, peeling was performed under water at a 180° angle to the leaf surface (Weyers & Travis, 1981). Peels were trimmed to c. 0.5 × 0.5 cm squares and transferred to 10 ml glass dishes (20–25 per dish) containing 5 ml of 5 mm KCl, 5 mm (2-[N-morpholino] ethanesolfonic acid (Mes), pH 6.5.

Experimental

In experiments with detached epidermis, peels were incubated under red or far-red (FR) light or in darkness for up to 4 h in solutions containing KCl bubbled with air containing various concentrations of CO2. Potential antagonists or inhibitors of stomatal opening were added to solutions. Peels were removed from solutions periodically for measurement of stomatal apertures and for histological analysis for K+ and starch in guard cells. Dishes containing peels were incubated under inverted plastic cups (4 cm diameter, 4 cm deep) that were covered with foil and clamped down onto light filters or the black surface of the laboratory bench. In light experiments, light was delivered through filters to the bottom of each dish by a 150 W halogen photo optic lamp (model EKE, Osram Corp., Winchester, KY, USA) from a fiber optic illuminator (Model 170 D, Dolan-Jenner Industries, Lawrence, MA, USA). Although chlorophyll is the only known red light photoreceptor of guard cells, because both chlorophylls and phytochromes are red-light photoreceptors, experiments were performed with light of near-red (R) wavelengths or with light of FR wavelengths (Fig. 1). Red light (620–900 nm with c. 50% of energy output between 650 nm and 680 nm; Fig. 1a) was produced with a Kodak 1 A safelight filter. Far-red light (700–900 nm with 98.5% of energy output between these wavelengths; Fig. 1b) was produced using black Plexiglas (Rohm & Haas, Philadelphia, PA, USA). Most experiments with red light were performed at PPFRs of 125–150 µmol m−2 s−1. In all experiments, PPFR was measured with a LiCor quantum/radiometer/photometer (LiCor, Lincoln, NB, USA). To achieve similar photon fluence rates in FR as in R, spectral energy transmittance profiles of R and FR filters were measured with a spectroradiometer (Optronic Laboratories model OL 754, Orlando, FL, USA) calibrated to a standard bulb (Model 752–10, National Institute of Standards and Technology, Optronics Laboratories) using the lamp described at the same PPFR for each measurement. Because the quantum sensor measured wavelengths between 400 and 700 nm only, equivalent fluence rates for R and FR wavelengths were calculated by dividing the integrated area between 653 nm and 680 nm on the FR profile into the integrated area between 712 nm and 740 nm on the same profile. Using the inverse of this ratio (67.5) as a multiplier, a PPFR of 2 µmol m−2 s−1 through the FR filter was calculated to be equivalent to 135 µmol photons m−2 s−1 in the 712–740 nm waveband. Air containing reduced concentrations of CO2 (c. 30–50 µl l−1) was produced by bubbling line air (400–470 µl l−1 CO2) through two flasks connected in series, each containing 400 ml of saturated NaOH. To produce air with 210–225 µl l−1 CO2, line air and air coming from flasks was mixed in a downstream 1-l flask before it was delivered to incubation cups. Air was bubbled through incubation solutions at 6.7 ml min−1 using hypodermic needles (26 gauge) inserted through the tops of the cups and connected to air lines. To prevent positive pressure in cups, two 5-mm holes were drilled at 180° to each other, centered 1 cm beneath the rim at the open end of each cup. The same holes were used to withdraw air during experiments for monitoring CO2 concentrations. Each day, CO2 concentrations were recorded every minute in one of four available chambers using an EGM-1 Environmental Gas Monitor (PP Systems, Hitchin, Herts, UK). Measurements of CO2 concentration were rotated daily among the four chambers. Epidermal peels were maintained on the surface of the liquid by the aeration protocol and thus were not submerged in solutions for any significant length of time (< 1 s).

Figure 1.

Spectral profiles of filters used to illuminate stomata in epidermis detached from leaves of Vicia faba L. and subjected to reduced concentrations of CO2, as described in the Materials and Methods section. (a) Full data for both filters (open circles, red; crosses, far red); (b) reduced scale to shown fine features of the spectral profile of the far-red filter.

Stomatal apertures were measured periodically with a microscope equipped with a digital video camera and monitor (Tallman & Zeiger, 1988). Light from the microscope illuminator was filtered through a green filter to minimize changes in aperture that might be caused by illumination during measurements. Means and standard errors of sample means were calculated using statview 4.01 (Abacus Concepts, Berkeley, CA, USA).

All experimental solutions were made by adding compounds to the incubation solution described above. All chemicals were from Sigma (St Louis, MO, USA). Potassium ion uptake by guard cells was evaluated with sodium cobaltinnitrite staining as described by Green et al. (1990). The starch content in guard cell chloroplasts was evaluated with iodine–potassium iodide (IKI) stain (Tallman & Zeiger, 1988). To determine whether stomatal opening observed at lower CO2 concentrations in epidermis illuminated with red or FR light was dependent on photosynthetic electron transport in guard cell chloroplasts, peels were preincubated in darkness for 1 h in solutions containing or lacking 10 µm DCMU before light treatments were administered. To evaluate whether Z was required for the response to lowered CO2 concentration, peels were incubated in solutions containing or lacking 5 mm DTT. To test whether stomatal opening at lower CO2 concentrations required activation of the guard cell plasma membrane proton pump, peels were illuminated in solutions containing or lacking 10 mm sodium orthovanadate (Schwartz et al., 1991).

Calmodulin and a ser/thr kinase have been implicated as downstream components of the blue light signal transduction pathway (Zeiger, 2000). To assess whether these proteins might be involved in a stomatal response to lower CO2 concentrations under red light, peels were incubated in solutions containing: (1) 1 or 10 µm W-7 (N-(6-aminohexyl)-5-chloro-1-napthalenesulfonamide), an antagonist of calmodulin and calmodulin-dependent kinases; (b) 200 or 500 µm H-7 (1-(5-isoquinolinesulfonyl)-2-methyl-piperazine), an inhibitor of protein kinase C; or (3) 50 or 100 µm ML-9 (1-(5-chloronapthalene-1-sulfonyl)-1H-hexahydro-1,4-diazepine), a specific inhibitor of mammalian myosin light chain kinases and, at higher concentrations, protein kinases A and C.

To determine whether a stomatal CO2 response detected in detached epidermis under red light might also occur in intact leaves, individual leaflets were detached from dark-adapted plants, submerged in 60 ml of deionized water and illuminated from the adaxial surface for 1.5 h with 440 µmol m−2 s−1 of red light. This protocol is thought to reduce ci to low levels as leaf photosynthesis is limited by (1) the low concentration of CO2 dissolved in water relative to that required to saturate photosynthesis; and (2) the low diffusibility of CO2 in water vs air. Similar methods have been described elsewhere (Weyers & Meidner, 1990). The time employed and PPFR were sufficient to achieve maximum stomatal opening on the abaxial leaf surface (not shown). After incubation, abaxial epidermis was detached from leaves, apertures were examined and guard cells were stained for K+, as described earlier in this section.

Results

The time course of stomatal opening under R or FR light at three different concentrations of CO2 is summarized in Fig. 2. Over 4 h at CO2 concentrations of 400–450 µl l−1 (Fig. 2a,b), stomata did not open under R or FR light. Stomata apertures reached their peak after 2 h of illumination with R or FR light in air containing CO2 concentrations of 210–225 µl l−1 (Fig. 2c,d); net opening was c. 3 µm (Fig. 2c). Under R light, net opening was c. 5 µm after 4 h at CO2 concentrations of 40–50 µl l−1 (Fig. 2e,f); net opening under FR after 4 h was 6.5 µm (Fig. 2e,f). Small increases in aperture (c. 0.5–1 µm) were detected after 4 h in darkness in air containing 210–225 µl l−1 CO2 (Fig. 2c) and 40–50 µl l−1 CO2 (Fig. 2e). No uptake of K+ by guard cells was detected in epidermis incubated for 4 h under R light in air containing 450 µl l−1 CO2 (Fig. 3A). Substantial K+ uptake was observed after 4 h of illumination with R light in air containing 260 µl l−1 CO2 (Fig. 3b). The greatest amounts of K+ accumulated after 4 h of illumination with R light in air containing 20 µl l−1 CO2 (Fig. 3c). Small amounts of K+ were detected in guard cells in epidermis incubated for 4 h in darkness in air containing 20 µl l−1 CO2 (Fig. 3d). Results under FR were similar (not shown). After 3 h of illumination with R light in air containing 50 µl l−1 CO2, starch was reduced in guard cell chloroplasts (Fig. 4b) over that of guard cell chloroplasts of unincubated guard cells (Fig. 4a) or guard cells treated with air containing 50 µl l−1 CO2 for 3 h in darkness (Fig. 4c).

Figure 2.

Mean stomatal apertures (a,c,e) in epidermis detached from leaves of Vicia faba L. incubated for up to 4 h in darkness (filled squares) or under 125 µmol m−2 s−1 of light enriched in red (open circles) or far-red (closed triangles) wavelengths. Epidermis was incubated in 5 mm KCl, 5 mm Mes, pH 6.5. Each value is the mean of 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day. Values are means ± SE of the sample mean. If no error bar is visible it is contained within the width of the symbol. Below each panel is a corresponding panel (b,d,f) that contains three digital records of CO2 concentrations in air surrounding epidermal peels over the 4-h time-course of each experiment. On each of the three separate days (closed squares, day 1, dark; closed circles, day 2, red; open circles, day 3, far red) that apertures were measured, an infrared gas analyser was used to measure and record CO2 concentrations at 1-min intervals in one of the three experimental chambers in which peels were being incubated.

Figure 3.

Potassium ion uptake by guard cells in epidermis detached from leaves of Vicia faba L. and incubated for 4 h in 5 mm KCl, 5 mm Mes, pH 6.5, under 125 µmol m−2 s−1 of red light in air containing (a) 450 µl l−1 CO2, (b) 260 µl l−1 CO2 or (c) 20 µl l−1 CO2. A dark control treated for 4 h in 20 µl l−1 CO2 is shown in (d). Potassium ion was detected with sodium hexanitrocobaltate III, which closes stomata through its effects as a fixative and an osmoticum. Original magnification, ×200; bar (a), 100 µm.

Figure 4.

Starch in chloroplasts of guard cells in epidermis detached from leaves of Vicia faba L. and not treated (a) or incubated for 3 h in 5 mm KCl, 5 mm Mes, pH 6.5, under 125 µmol m−2 s−1 of red light (b) or in darkness (c) in air containing 50 µl l−1 CO2. Starch was stained with iodine–potassium iodide, which is toxic to guard cells and causes them to slowly lose turgor. Original magnification ×600; bar (a), 50 µm.

After 4 h of illumination under R light in air containing 50 µl l−1 CO2, mean net stomatal opening reached 4.5 µm and 4.7 µm at PPFRs of 150 µmol m−2 s−1 and 200 µmol m−2 s−1, respectively (Fig. 5). In air containing 30–50 µl l−1 CO2, stomatal opening under 150 µmol m−2 s−1 of R or FR light was fully inhibited by 10 µm DCMU (Fig. 6). Under similar conditions, neither R- or FR-induced opening was inhibited by 5 mm DTT (Fig. 7).

Figure 5.

Mean net stomatal opening (difference between mean stomatal aperture in light and mean stomatal aperture in darkness) in epidermis detached from leaves of Vicia faba L. and incubated for 3 h in 5 mm KCl, 5 mm Mes, pH 6.5 at photosynthetic photon fluence rates (PPFRs) of red light ranging from 10 to 200 µmol m−2 s−1. Means for subtraction were obtained from 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day; y = 2.962 log(x) – 2.652, r = 0.890.

Figure 6.

Effects of 10 µm DCMU on mean stomatal apertures in epidermis detached from leaves of Vicia faba L. Epidermal peels were incubated for 1 h in darkness and then up to 4 h in darkness or under 125 µmol m−2 s−1 of light enriched in red (left panels) or far-red (right panels) wavelengths in air containing 30–50 µl l−1 CO2. Epidermis was incubated in 5 mm KCl, 5 mm Mes, pH 6.5, ± DCMU (crossed circle, light + DCMU; open circle, light – DCMU; crossed square, dark + DCMU; filled square, dark – DCMU). Each value is the mean of 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day. Values are means ± SE of the sample mean. If no error bar is visible it is contained within the width of the symbol. Below each of the top panels is a corresponding panel that contains three digital records of CO2 concentrations in air surrounding epidermal peels over the 4-h time-course of each experiment. Left panel: open circle, day 1; filled square, day 2; open circle, day 3. Right panel: filled circle, day 1; crossed square, day 2; crossed circle, day 3. On each of the three separate days that apertures were measured, an infrared gas analyser was used to measure and record CO2 concentrations at 1-min intervals in one of the four experimental chambers in which peels were being incubated.

Figure 7.

Effects of 5 mm dithiothreitol (DTT) on mean stomatal apertures in epidermis detached from leaves of Vicia faba L. Epidermal peels were incubated for up to 4 h in darkness or under 125 µmol m−2 s−1 of light enriched in red (left panels) or far-red (right panels) wavelengths in air containing 30–50 µl l−1 CO2. Epidermis was incubated in 5 mm KCl, 5 mm Mes, pH 6.5, ± DTT (crossed circle, light + DTT; open circle, light –DTT; crossed square, dark + DTT; filled square, dark – DTT). Each value is the mean of 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day. Values are means ± SE of the sample mean. If no error bar is visible it is contained within the width of the symbol. Below each of the top panels is a corresponding panel that contains three digital records of CO2 concentrations in air surrounding epidermal peels over the 4-h time-course of each experiment. Left panel: open circle, day 1; filled square, day 2; open circle, day 3. Right panel: filled square, day 1; open circle, day 2; filled square, day 3. On each of the three separate days that apertures were measured, an infrared gas analyser was used to measure and record CO2 concentrations at 1-min intervals in one of the four experimental chambers in which peels were being incubated.

When epidermis was illuminated with R light (150 µmol m−2 s−1) in air containing 30–50 µl l−1 CO2, stomatal opening was not inhibited by H-7 in concentrations as high as 500 µm (Fig. 8a). Opening was fully inhibited by concentrations of ML-9 as low as 50 µm (Fig. 8B) and by concentrations of W-7 as low as 10 µm (Fig. 8C). Under similar conditions, opening was also inhibited by 10 mm sodium orthovanadate (Fig. 9).

Figure 8.

Effects of various concentrations of H-7 (a), ML-9 (b), or W-7 (c) on mean stomatal apertures in epidermis detached from leaves of Vicia faba L. Epidermal peels were incubated for up to 4 h in darkness or under 125 µmol m−2 s−1 of light enriched in red wavelengths in air containing 30–50 µl l−1 CO2. Epidermis was incubated in 5 mm KCl, 5 mm Mes, pH 6.5, ± H-7 (a: open circle, no treatment; crossed circle, 200 µm H-7; crossed diamond, 500 µm H-7; filled square, dark), ML-9 (b: open circle, no treatment; crossed circle, 50 µm ML-9; crossed diamond, 100 µm ML-9; filled square, dark) or W-7 (c: open circle, no treatment; crossed circle, 1 µm W-7; crossed diamond, 10 µm W-7; filled square, dark). Each value is the mean of 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day. Values are means ± SE of the sample mean. If no error bar is visible it is contained within the width of the symbol. On each of the three separate days that apertures were measured, an infrared gas analyser was used to measure and record CO2 concentrations at 1-min intervals in one of the four experimental chambers in which peels were being incubated (not shown).

Figure 9.

Effects of 10 mm sodium orthovanadate on mean stomatal apertures in epidermis detached from leaves of Vicia faba L. Epidermal peels were incubated for up to 4 h in darkness or under 125 µmol m−2 s−1 of light enriched in red (left panels) or far-red (right panels) wavelengths in air containing 30–50 µl l−1 CO2. Epidermis was incubated in 5 mm KCl, 5 mm Mes, pH 6.5, ± orthovanadate (crossed circle, light + vanadate; open circle, light – vanadate; crossed square, dark + vanadate; filled square, dark – vanadate). Each value is the mean of 90 measurements, 30 on each of three separate days from 10 measurements on each of three epidermal peels each day. Values are means ± SE of the sample mean. If no error bar is visible it is contained within the width of the symbol. Below each of the top panels is a corresponding panel that contains three digital records of CO2 concentrations in air surrounding epidermal peels over the 4-h time-course of each experiment. Left panel: open circle, day 1; filled square, day 2; open circle, day 3. Right panel: open circle, day 1; filled square, day 2; filled circle, day 3. On each of the three separate days that apertures were measured, an infrared gas analyser was used to measure and record CO2 concentrations at 1-min intervals in one of the four experimental chambers in which peels were being incubated.

When intact leaves were submerged in water and illuminated from the adaxial surface for 1.5 h with R light (PPFR = 440 µmol m−2 s−1), guard cells on the abaxial leaf surface accumulated K+ (Fig. 10a) and stomata opened (Fig. 10b).

Figure 10.

Accumulation of potassium ions in guard cells in intact leaves of Vicia faba L. submerged in water and illuminated from the adaxial surface with 440 µmol m−2 s−1 of red light for 1.5 h. At the end of the 1.5-h illumination, epidermis was peeled from the abaxial leaf surface and stained with sodium hexanitrocobaltate III (a) or examined with a microscope for stomatal opening (b). The hexanitrocobaltate stain closes stomata through its effects as a fixative and an osmoticum. Magnification, ×300; bars, 50 µm. Results are representative of those from three separate experiments.

Discussion

The data reported here indicate that in addition to activating the Calvin cycle (Gotow et al., 1988), sucrose accumulation (Poffenroth et al., 1992; Talbott & Zeiger, 1993) and Z formation in guard cells (Zeiger, 2000), red light can also activate a K+ uptake mechanism for stomatal opening under reduced concentrations of CO2 (Figs 2 and 3). Like the blue light response of guard cells, the red light-activated CO2 response also triggers starch degradation in guard cell chloroplasts (Tallman & Zeiger, 1988) (Fig. 4). Stomatal opening and K+ uptake occur in inverse proportion to CO2 concentration (Figs 2 and 3). The response saturates at a PPFR of 150–200 µmol m−2 s−1 (Fig. 5), a light level similar to that on the abaxial surface of a leaf in full sun (Yera et al., 1986). The observation that the response is similar in R and FR light (Fig. 2) and that it is inhibited equally in R and FR light by pretreatment of guard cells with DCMU (Fig. 6) indicates that: (1) the sensory transduction mechanism for the response is localized in the guard cell chloroplast; (2) the response is dependent on chloroplast electron transport; and (3) the photoreceptor is chlorophyll. The ratio of FR to R through the R filter was 1.1 : 1 (Fig. 1a) and the ratio of FR to R ratio through the FR filter was 67.5 : 1 (Fig. 1b). Despite the large differences in FR to R ratios between the two filters, opening responses in R and FR were similar (Figs 2, 6, 7 and 9), suggesting that phytochrome does not act as a photoreceptor in the response. The PPFR of wavelengths between 680 nm and 700 nm was only 2–3 µmol m−2 s−1 through the FR filter at a FR PPFR of 125–150 µmol m−2 s−1 (Fig. 1b). The observation that opening responses were similar under FR and R treatments (Figs 2, 6, 7 and 9) may suggest (1) that a primary signaling event in photosystem II (PSII) is highly amplified and/or (2) that photosystem I (PSI) and its light-harvesting complex play a significant role in the signal transduction cascade. If PSI plays such a role, it is probably not through cyclic electron flow, since the response under FR is inhibited by DCMU (Fig. 6).

Unlike the blue light response of guard cells, the red light-activated CO2 response is not dependent on Z formation in guard cell chloroplasts, as shown by failure of 5 mM DTT to inhibit opening (Fig. 7). The observation that DTT inhibits blue light-induced opening but not the red light-induced CO2 response, even though a number of components of the signal transduction pathway are shared with those of the blue light response pathway (see below), supports the arguments that: (1) DTT is a very specific inhibitor of Z formation in guard cells (Srivastava & Zeiger, 1995); and (2) that Z is the blue light photoreceptor of guard cells responsible for stomatal opening (Frechilla et al., 1999). As with the blue light response, the red light-activated CO2 response depends on activation of the guard cell plasma membrane proton pump (Fig. 9). In patch clamp studies using GCPs of Vicia faba, Serrano et al. (1988) reported that red light activated the plasma membrane proton pump. It seems plausible that in that study, experimental conditions may have elicited a response similar to the one observed in our experiments. A recent study with red light suggests that guard cell chloroplasts can supply ATP to the plasma membrane H+ pump in guard cells in which the pump has been activated by fusicoccin (Tominaga et al., 2001). However, when patch-clamped GCPs of Vicia faba were treated with DCMU to inhibit red light-induced proton pumping, pumping was not restored by addition of ATP to the patch pipette (Serrano et al., 1988). In our experiments with FR filters, the pump was activated at PPFR of red light too low to drive large amounts of ATP synthesis (2–3 µmol m−2 s−1), suggesting that amplification of primary light/CO2 signals through a signal transduction cascade is required for the response. Indeed, the red light-activated CO2 response appears to share some downstream components of the blue light signal transduction pathway that lie between the chloroplast and the plasma membrane. Similar to the blue light response, the red light-activated CO2 response was not inhibited by H-7 (Fig. 8a), an inhibitor of protein kinase C (Assman & Shimazaki, 1999); but was inhibited by ML-9 (Fig. 8b), which in the relatively high concentrations used, inhibits ser/thr kinases (Assmann & Shimazaki, 1999), and by W-7 (Fig. 8c), a calmodulin antagonist and inhibitor of calmodulin-dependent protein kinases (Shimazaki et al., 1992; Shimazaki et al., 1993; Assmann & Shimazaki, 1999). Results of experiments with whole leaflets submerged in water and illuminated with red light to reduce the CO2 concentrations around guard cells (Fig. 10) suggest that guard cells possess this mechanism in intact leaves and that it is not an artifact created by detachment of epidermis. Such a mechanism could explain why in intact leaves of Commelina communis, Assmann (1988) observed stomatal conductances of up to 150 mmol m−2 s−1 at relatively low red light fluences.

It should be noted that a small amount of stomatal opening was observed in darkness at lower CO2 concentrations (Fig. 2c,e) and K+ uptake was detected in guard cells incubated in darkness at very low CO2 concentrations (Fig. 3d). Whether this represents (1) a light-independent CO2 response; or (2) an effect of lower CO2 concentrations operating unamplified through the components of the red light-activated signal transduction pathway described earlier in this discussion awaits investigation.

To our knowledge, this is the first direct evidence that red light can trigger K+ uptake by guard cells. The response requires reduced CO2 concentrations and localizes to the guard cell chloroplast. It is striking that all mechanisms described to date for light- or CO2-induced stomatal opening are harbored in the guard cell chloroplast.

Results of experiments with GCPs from Vicia faba illuminated with red light demonstrate that guard cells can fix CO2 into intermediates of the Calvin cycle, including phosphoglyceric acid, sugar monophosphates, sugar diphosphates and triose phosphates (Gotow et al., 1988). Neither the amount of Rubisco activity in guard cell extracts of Pisum (Reckmann et al., 1990) nor in intact leaves of Vicia faba grown at lower relative humidity (60%; Lu et al., 1997) were sufficient to produce photosynthetic intermediates at rates likely to support the magnitude of stomatal opening reported for those species. Nevertheless, an immunohistochemical assay of guard cells of 41 species showed that a majority contained Rubisco (Madhavan & Smith, 1982). Based on recent results, Outlaw and De Vlieghere-He (2001) proposed that accumulation in the guard cell apoplast of sucrose produced by the mesophyll under conditions of high vapor pressure differences between leaves and the air (VPD) (and hence, high rates of transpiration) may downregulate transcription of guard cell genes for Rubisco and other enzymes involved in carbon fixation and metabolism. Were this hypothesis to be supported experimentally, it could explain: (1) the wide variability in the amount of guard cell Rubisco observed among species in immunohistochemical studies (Madhavan & Smith, 1982); and (2) why some investigators detected Calvin cycle activity in guard cells of Vicia faba using plants grown under conditions that would be expected to produce lower VPD (Gotow et al., 1988), while others using plants grown under conditions expected to produce higher VPD did not (Lu et al., 1997). This hypothesis is also consistent with the suggestion of Zeiger that Calvin cycle activity in guard cells may modulate the blue light response (Zeiger & Zhu, 1998; Zeiger, 2000). When greenhouse-grown plants (higher VPD) were moved to an environmental chamber (lower VPD) the sensitivity of their stomata to decreased concentrations of CO2 increased dramatically (Talbott et al., 1996). If low VPD = high guard cell Rubisco activity (Outlaw & De Vlieghere-He, 2001), as proposed by Zeiger and colleagues (Zeiger & Zhu, 1998; Zeiger, 2000) and observed by Talbott et al. (1996), illuminated guard cells of chamber-grown plants would be expected to accumulate more Z and become more sensitive to blue light than those of greenhouse-grown plants at the same PPFR. The sucrose regulation hypothesis could also explain why, in the experiments reported here, stomata illuminated with red light at higher CO2 concentrations did not open (Fig. 2a) by producing sugars photosynthetically, since temperature and humidity in the summer months in the greenhouse would have been expected to produce higher VPD.

Another chloroplast mechanism for stomatal opening is accumulation of sucrose in guard cells of V. faba under red light by a DCMU-sensitive mechanism (Poffenroth et al., 1992; Talbott & Zeiger, 1993). Calvin cycle activity is one potential source of that sucrose. Sucrose could also accumulate from red light-induced starch breakdown in guard cell chloroplasts, but no starch breakdown (judged by IKI staining) was evident during red light-induced stomatal opening in air containing ambient concentrations of CO2, while starch breakdown was detected in concurrent experiments with blue light (Tallman & Zeiger, 1988). It is, of course, possible that a concurrent Calvin cycle activity and starch degradation would be futile insofar as it would lead to no net change in the amount of starch in guard cell chloroplasts illuminated with red light. Another possible source of sucrose in intact leaves is the guard cell apoplast (Lu et al., 1997). The GCPs of P. sativum do have sufficient sucrose transport capacity to support stomatal opening at rates described in the literature (Ritte et al., 1999), but no direct measurements of transport of photosynthetically produced sucrose from guard cell apoplast to symplast have been made in any species.

Finally, the guard cell chloroplast is the site of production of Z, the only pigment identified as a blue light photoreceptor for triggering the blue light photosensory signal transduction cascade of guard cells. In early experiments with low fluences of blue light (Zeiger et al., 1987; Assmann, 1988), red light was used to saturate any photosynthetic responses of guard cells so that the blue light response could be evaluated independently of its potential effects on guard cell photosynthesis (Zeiger et al., 1987). It was discovered that pre-illumination with red light was required before a blue light response could be measured. It has since been determined that the photosynthetic electron transport that results from red light illumination drives formation of Z as light-driven acidification of the thylakoid lumen activates the de-epoxidase enzyme that converts V to Z (Zeiger & Zhu, 1998).

In most species, the presence of chloroplasts in guard cells is the most visible feature that differentiates them from neighboring epidermal cells. A wealth of published data now point to the central role of the guard cell chloroplast and to the importance of photosynthetic electron transport in regulating stomatal responses to light and CO2. Their importance to guard cell signal transduction may explain why guard cell chloroplasts fail to senesce even when mesophyll cells in the same leaf have yellowed completely (Zeiger & Schwartz, 1982; Ozuna et al., 1985). In addition to its role as an environmental sensor (Outlaw et al., 1981), emerging data point to the guard cell chloroplast as an evolutionary repository specialized for integrated signal transduction to regulate stomatal aperture in a variety of developmental, environmental and physiological circumstances (Zhu et al., 1995; Quinones et al., 1996).

While it is only possible to speculate as to why evolution has preserved a multiplicity of mechanisms for triggering and/or regulating stomatal opening, it is obvious that guard cells are subject both to the developmental programs of the plants in which they exist and to an environment that fluctuates dramatically and that has the potential to regulate stomatal development (Lake et al., 2001). As plants germinate, grow to maturity, and then flower, fruit and set seed, guard cells move through their own developmental programs, first as components of young leaves, next as elements of expanding leaves and then as constituents of senescing leaves. They may be located near or far from a vein, on the top or the bottom surface of a leaf, near the top or the bottom of the canopy, near to or far from other various organs on the plant. Their proximity to light, temperature, nutritional, hormonal and water potential gradients change with plant growth, leaf age and position in the canopy and with fluctuations in the environment that include length of the photoperiod, humidity, soil moisture, temperature, etc. Whether, at any point in its developmental history, each and every guard cell or population of guard cells possesses each and every mechanism for opening that has been discovered to date is unknown, but published data suggest otherwise. Diurnal changes in guard cell physiology (Dodge et al., 1992; Talbott & Zeiger, 1998), variations in responses related to growth conditions (Talbott et al., 1996; Outlaw & De Vlieghere-He, 2001), species differences in guard cell biochemistry (Madhavan & Smith, 1982), and seasonal variations in guard cell responses (Lohse & Hedrich, 1992) are now well documented. Indeed, the experiments described here were performed over 4 years from mid-May to 1 September of each year because in other months of the year, using greenhouse-grown plants, large dark-opening of stomata was observed at lower concentrations of CO2 (not shown). Staining revealed that K+ uptake was not the basis for dark movements and guard cells illuminated with red light at lower concentrations of CO2 still accumulated K+ in large amounts. These types of variations reinforce the notion that guard cells are dynamic by nature and that they can change character dramatically over even relatively short periods of time. If so, then any individual experiment may simply provide a picture of how a guard cell conditioned by its environmental history responds to experimental protocols that are imposed on that history. In future, many careful studies under rigorously controlled conditions will probably be needed if we are to define how the developmental history of the guard cell shapes it response capacity for various signals, and how it is shaped for the future by the conditions to which it is currently exposed.

Acknowledgements

We thank E. Zeiger and L. Talbott for helpful comments on the manuscript and S. D. Davis and Jeffrey Warren for assistance with spectroradiometry. Supported in part by NSF MCB-9900525 to G.T.

Ancillary