Silencing of NtMPK4 impairs CO2-induced stomatal closure, activation of anion channels and cytosolic Ca2+signals in Nicotiana tabacum guard cells

Authors

  • Holger Marten,

    1. Molecular Plant Physiology and Biophysics, Julius-von-Sachs Institute for Biosciences, Biocenter, Würzburg University, Julius-von-Sachs-Platz 2, D-97082 Würzburg, Germany,
    Search for more papers by this author
  • Taekyung Hyun,

    1. Institute for Pharmaceutical Biology, Biocenter, Würzburg University, Julius-von-Sachs-Platz 4, D-97082 Würzburg, Germany, and
    Search for more papers by this author
  • Kenji Gomi,

    1. Plant–Microbe Interactions Research Unit, National Institute of Agrobiological Sciences, Tsukuba, Ibaraki 305-8602, Japan
    Search for more papers by this author
    • Present address: Center for Ecological Research, Kyoto University, Otsu, Shiga 520-2113, Japan.

  • Shigemi Seo,

    1. Plant–Microbe Interactions Research Unit, National Institute of Agrobiological Sciences, Tsukuba, Ibaraki 305-8602, Japan
    Search for more papers by this author
  • Rainer Hedrich,

    1. Molecular Plant Physiology and Biophysics, Julius-von-Sachs Institute for Biosciences, Biocenter, Würzburg University, Julius-von-Sachs-Platz 2, D-97082 Würzburg, Germany,
    Search for more papers by this author
  • M. Rob G. Roelfsema

    Corresponding author
    1. Molecular Plant Physiology and Biophysics, Julius-von-Sachs Institute for Biosciences, Biocenter, Würzburg University, Julius-von-Sachs-Platz 2, D-97082 Würzburg, Germany,
    Search for more papers by this author

*(fax +49 931 8886158; e-mail roelfsem@botanik.uni-wuerzburg.de).

Summary

Light-induced stomatal opening in C3 and C4 plants is mediated by two signalling pathways. One pathway is specific for blue light and involves phototropins, while the second pathway depends on photosyntheticaly active radiation (PAR). Here, the role of NtMPK4 in light-induced stomatal opening was studied, as silencing of this MAP kinase stimulates stomatal opening. Stomata of NtMPK4-silenced plants do not close in elevated atmospheric CO2, and show a reduced response to PAR. However, stomatal closure can still be induced by abscisic acid. Measurements using multi-barrelled intracellular micro-electrodes showed that CO2 activates plasma membrane anion channels in wild-type Nicotiana tabacum guard cells, but not in NtMPK4-silenced cells. Anion channels were also activated in wild-type guard cells after switching off PAR. In approximately half of these cells, activation of anion channels was accompanied by an increase in the cytosolic free Ca2+ concentration. The activity of anion channels was higher in cells showing a parallel increase in cytosolic Ca2+ than in those with steady Ca2+ levels. Both the darkness-induced anion channel activation and Ca2+ signals were repressed in NtMPK4-silenced guard cells. These data show that CO2 and darkness can activate anion channels in a Ca2+-independent manner, but the anion channel activity is enhanced by parallel increases in the cytosolic Ca2+ concentration. NtMPK4 plays an essential role in CO2- and darkness-induced activation of guard-cell anion channels, through Ca2+-independent as well as Ca2+-dependent signalling pathways.

Introduction

Land plants acquire CO2 for photosynthesis through stomata, which are microscopic pores in the largely gas-impermeable waxy cuticle covering the leaf surface. Because stomata are non-selective for gases, these pores also represent pathways for evaporation of H2O vapour. During drought, excessive loss of H2O can only be prevented through stomatal closure, which results in a lower rate of CO2 uptake. Depending on the environmental conditions, plants thus need to balance the demand for CO2 with the availability of H2O by altering their stomatal conductance. In general, the stomata of C3 and C4 plants open in the light to meet the photosynthetic demand for CO2, and close in darkness, or during drought, to limit loss of H2O. These stomatal responses are executed by two guard cells that surround the pore in the cuticle. Guard cells are able to sense changes in environmental conditions and translate these into stomatal movements (Assmann and Shimazaki, 1999; Hetherington, 2001; Roelfsema and Hedrich, 2005; Schroeder et al., 2001; Shimazaki et al., 2007).

Light is sensed by the guard cells through two signalling pathways that act in parallel with each other (Assmann and Shimazaki, 1999; Roelfsema and Hedrich, 2005; Shimazaki et al., 2007). One pathway is specific for blue light and depends on two phototropins (Kinoshita et al., 2001) that resemble blue light-activated protein kinases associated with the plasma membrane (Briggs and Christie, 2002). In guard cells, blue light-activated phototropins induce a dual effect on plasma membrane ion transport. Blue light stimulates H+-ATPases, through phosphorylation of their C-termini (Kinoshita and Shimazaki, 1999), and inhibits anion channels (Marten et al., 2007a). Changes in the activity of these ion-transport proteins cause the guard-cell plasma membrane to hyperpolarize, which in turn drives K+ uptake through inward-rectifying K+ channels. The accumulation of K+ salts causes the guard cell to swell and bend, thereby pushing the stomatal pore open (Roelfsema and Hedrich, 2005).

Inhibition of plasma membrane anion channels in guard cells is also exerted by a second signalling pathway that depends on photosynthetically active radiation (PAR). For this response, light probably acts via a decrease in the intercellular CO2 concentration of the leaf, which is mainly due to photosynthetic CO2 fixation in mesophyll cells (Roelfsema et al., 2002). High levels of CO2 stimulate the activity of guard-cell anion channels (Brearley et al., 1997; Roelfsema et al., 2002), and light-dependent reduction of the CO2 concentration thus leads to deactivation of these channels. In the high temperature 1 (ht1) mutant of Arabidopsis thaliana, stomata do not open in CO2-free air (Hashimoto et al., 2006), suggesting that guard-cell anion channels are active independently of the CO2 concentration in this mutant. In line with the proposed PAR signal pathway, the stomata of ht1 mutants are insensitive to PAR, but still display blue light-specific opening.

The CO2-insensitive guard-cell phenotype of the ht1-1 and ht1-2 mutants is due to mutations in a gene with high homology to MAPKK kinases (Ichimura et al., 2002). Both mutations cause reduced HT1 kinase activity (Hashimoto et al., 2006), suggesting that active HT1 kinases are required for deactivation of anion channels in the absence of CO2. The HT1 protein may thus be part of a CO2 signalling pathway in guard cells. Alternatively, this protein kinase may have a role in guard-cell development, as for the YODA MAPKK kinase (Bergmann et al., 2004). YODA was shown to phosphorylate the MAPK kinases MKK4 and MKK5, which in turn activate the MAP kinases MPK3 and MPK6 (Wang et al., 2007). In addition to their role in stomatal development, MPK3 and MPK6 have been implicated in biotic and abiotic stress responses (Asai et al., 2002; Droillard et al., 2002; Kovtun et al., 2000; Menke et al., 2004), similar to the homologous proteins WIPK and SIPK of Nicotiana tabacum (Romeis et al., 1999; Seo et al., 1995, 1999; Zhang and Klessig, 1998).

The NtMPK4-encoded MAP kinase of N. tabacum is involved in systemic acquired resistance and wounding responses (Gomi et al., 2005). Repression of NtMPK4 increased the susceptibility to atmospheric ozone, which to some extent may be due to increased stomatal conductance. Because of this stomatal phenotype of NtMPK4-silenced tobacco plants, we studied the properties of guard cells and their responses to CO2, PAR and abscisic acid (ABA). These studies revealed that repression of NtMPK4 causes loss of stomatal responses to high atmospheric CO2 concentrations and a severely reduced response to PAR. However, stomata of NtMPK4-repressed plants still respond to ABA. An analysis with multi-barrelled intracellular micro-electrodes showed that NtMPK4 is involved in darkness-induced, Ca2+-independent- and Ca2+-dependent activation of guard-cell anion channels.

Results

Repression of NtMPK4 feeds back on guard-cell size

Repression of MPK4 in N. tabacum does not change the stomatal density (Gomi et al., 2005), but stomata in the abaxial epidermis of NtMPK4-silenced plants were found to be increased in size (Figure 1). A fraction of these stomata had stomatal lips that bent towards each other in the centre of the pore, a feature that was not observed for stomata in wild-type plants (Figure 1a,b). Quantification of the stomatal size revealed a broader distribution of guard-cell length for NtMPK4-silenced plants compared to wild-type (Figure 1c,d). The mean length of the stomata increased from 38 ± 0.4 μm (n = 234) in wild-type to 48 ± 0.6 μm in line MPK4IR-11 (n = 228) and 46 ± 0.6 μm (n = 322) in line MPK4IR-6. This suggests that the activity of NtMPK4 is important for the control of elongation growth in guard cells.

Figure 1.

 Silencing of NtMPK4 leads to an increased guard cell size.
(a, b) Images of stomata in epidermal strips from the abaxial epidermis of wild-type (a) and line MPK4IR-11 (c) leaves. Scale bar = 50 μm.
(c, d) Frequency histogram of guard cell length, plotted against the number of cells with the respective length. ‘N’ represents the number of guard cells analysed in the abaxial epidermis of wild-type (c) and line MPK4IR-11 (d). Silencing of NtMPK4 broadens the distribution of guard-cell length and leads to a population of very long guard cells.

Plants transformed with an overactive SIPKK (Salicylic Acid-Induced Kinase Kinase), which is able to phosphorylate NtMPK4 (SIPKKEE-34, Gomi et al., 2005), had stomata with a mean length of 42 ± 0.6 μm (n = 192). Apparently, expression of SIPKKEE does not lead to a phenotype inverse to that of NtMPK4 repression with respect to guard-cell size. The NtMPK4-dependent pathway differs in this respect from that in Arabidopsis involving YODA, MKK4/MKK5 and MPK3/MPK6. High activity of MPK3 and MPK6 results in loss of stomata, whereas the stomatal density increases after loss of activity of both MPK3 and MPK6 (Wang et al., 2007).

Stomata of NtMPK4-silenced plants do not close in response to CO2

Leaves of NtMPK4-silenced N. tabacum plants showed a higher stomatal conductance than wild-type (Gomi et al., 2005). However, the origin of this difference in stomatal opening remained unknown, and was therefore studied by gas-exchange measurements. The difference in stomatal conductance between wild-type and NtMPK4-repressed lines was confirmed using ambient concentrations (350 μl l−1) and elevated concentrations (700 μl l−1) of atmospheric CO2. The mean conductance recorded in darkness at 350 μl l−1 was 103 ± 16 mmol H2O m−2 sec−1 (n = 10) for wild-type and 168 ± 16 mmol H2O m−2 sec−1 (n = 4) for line MPK4IR-11. Similar conductance values of 110 ± 15 mmol H2O m−2 sec−1 (= 7) for wild-type and 157 ± 12 mmol H2O m−2 sec−1 (= 6) for line MPK4IR-11 were found in darkness at 700 μl l−1.

Red light at a high photon flux density (400 μmol m−2 sec−1) induced stomatal opening in wild-type, as did lowering the atmospheric CO2 concentration in darkness from 700 to 0 μl l−1 (Figure 2a). These responses were attenuated in lines MPK4IR-11 (Figure 2b) and MPK4IR-6 (data not shown), as red light induced only a small increase in conductance, whereas the stomata of both NtMPK4-silenced lines were completely insensitive to changes in the atmospheric CO2 concentration (Figure 2b). The reduced responses to PAR and CO2 were not due to an inability of MPK4IR plants to change the stomatal conductance, as application of ABA induced stomatal closure in line MPK4IR-11 (Figure 2b) as well as line MPK4IR-6 (data not shown). Silencing of NtMPK4 did not alter the ABA sensitivity of guard cells, based on experiments using epidermal strips (Figure S1).

Figure 2.

 Stomata of NtMPK4-silenced plants remain open in darkness and at high atmospheric CO2 concentrations.
Changes in stomatal conductance (gw, upper traces) and photosynthetic activity (P, lower traces) were measured on the abaxial side of leaves of wild-type (a) and line MPK4IR-11 (b). The black bar indicates the periods during which the leaves were kept in darkness and air with 700 μl l−1 CO2; stomatal opening was first triggered with red light (400 μmol m−2 sec−1) at 700 μl l−1 CO2 and thereafter with CO2-free air in darkness. Finally, stomatal closure was triggered with 10 μm ABA fed via the petiole. The timing of stimulus application is indicated by white bars. Stomatal closure is stimulated by darkness and high atmospheric CO2 concentrations in wild-type, but not in line MPK4IR-11.

Stomata of N. tabacum showed only small conductance changes in response to blue light (25 μmol m−2 sec−1) given in addition to red light (400 μmol m−2 sec−1, Figure S2). This suggests that guard cells of N. tabacum do not display a strong blue light-specific response mediated by phototropins (Kinoshita et al., 2001). Gomi et al. (2005) also reported a lower stomatal conductance of plants expressing overactive SIPKK. However, leaves of greenhouse-grown 4-week-old N. tabacum plants expressing SIPKKEE did not show this difference in conductance compared to wild-type (data not shown). The stomata of SIPKKEE-expressing plants showed normal responses to red light, CO2 and ABA.

CO2 activation of S-type anion channels is lost in line MPK4IR-11

The stomata of the MPK4-silenced lines MPK4IR-6 and MPK4IR-11 do not respond to CO2, and the MPK4IR-6 line shows a more severe reduction in growth than line MPK4IR-11 (Gomi et al., 2005). Larger leaves provide better material for single guard-cell studies, and we therefore used line MPK4IR-11 to record CO2-induced changes in ion-channel activity. These measurements were performed using intracellular double-barrelled micro-electrodes, which were impaled into guard cells of intact N. tabacum plants. In wild-type guard cells, increasing the atmospheric CO2 from 0 to 700 μl l−1 caused a transient increase in inward current (Figure 3a), with an mean peak activity of −109 ± 43 pA (n = 9) at 354 ± 65 sec after stimulus onset. The increase of inward current at the holding potential corresponds to larger instantaneous currents at test potentials ranging from −160 to −20 mV (Figure 3c) and a shift of the steady state current/voltage curve to more negative values (Figure 3e). The slow activation of inward currents at −80 and −60 mV (Figure 3c, middle traces) reveals that the increase in instantaneous current is due to activation of S-type anion channels, as previously shown for Vicia faba (Brearley et al., 1997; Roelfsema et al., 2002). Guard cells of MPK4IR-11 plants showed none or only small CO2-induced changes in current (Figure 3b); the mean peak amplitude in these cells was 8 ± 5 pA (n = 8). In line with the severe reduction of CO2-induced current changes at the holding potential −100 mV, CO2 did not alter currents at test potentials ranging from −180 to 60 mV (Figure 3d,f).

Figure 3.

 CO2 activates plasma membrane anion channels in wild-type, but not in NtMPK4-silenced guard cells.
(a, b) Current traces for N. tabacum guard cells in intact plants, clamped to a holding potential of −100 mV and stimulated with an increase in the atmospheric CO2 concentration from 0 to 700 μl l−1, as indicated by the bar below the traces. CO2 triggers an increase in inward current in wild-type (a), but not in MPK4IR-11 (b) guard cells.
(c, d) Current traces for the same cells as in (a) and (b) (corresponding symbols), which were clamped stepwise from a holding potential of −100 mV to test potentials ranging from −160 to −20 mV. The traces were obtained before (bsl00043), during (bsl00041) and after (bsl00000) exposure to 700 μl l−1 CO2. CO2 stimulates slow-activating inward-conducting channels, as seen at −80 and −60 mV, in wild-type (c) but not in MPK4IR-11 (d) guard cells. Arrows indicate the 0 pA level.
(e, f) Steady-state current/voltage relationship sampled at the end of the 2 sec test pulses in (c) and (d) (corresponding symbols). In wild-type (e), the introduction of 700 μl l−1 CO2 (bsl00041) caused a shift to more negative current values over the whole voltage range. In guard cells of line MPK4IR-11 (f), no obvious changes in the current/voltage relationship were observed.

Repression of MPK4 impairs dark-induced activation of S-type anion channels

In agreement with the proposed role of CO2 as an intermediate signal in guard-cell responses to PAR, MPK4IR-11 stomata showed a repressed response to high photon flux densities of red light. The influence of PAR on ion-channel activity was studied on single guard cells using a long-distance water immersion objective operated with a drop of solution touching the leaf surface. In the wild-type, switching off red light (500 μmol m−2 sec−1) induced an inward current at the holding potential of −100 mV (Figure 4a), with a mean peak amplitude of −140 ± 42 pA (n = 12) 280 ± 40 sec after PAR withdrawal. Just as with the application of CO2, turning off red light activated channels (Figure 4c,e) that show voltage-dependent activation after stepping from −100 mV to −80 and −60 mV and deactivation at −120 and −140 mV (Figure 4c, middle traces). The dark-induced activation of these channels was reduced in MPK4IR-11 guard cells to a mean value of −29 ± 13 pA (n = 7). In these plants, switching on and off PAR resulted only in minor changes in the instantaneous current/voltage relationship of guard cells (Figure 4f).

Figure 4.

 Switching off red light activates anion channels in wild-type N. tabacum guard cells, but not in line MPK4IR-11.
(a, b) Current traces for N. tabacum guard cells clamped to a holding potential of −100 mV and stimulated by switching off red light (500 μmol m−2 sec−1) provided by the microscope lamp, as indicated by the bar below the traces. Darkness causes an increase in inward current in wild-type (a), but not MPK4IR-11 (b) guard cells.
(c, d) Current traces from the same cells as in (a) and (b) (corresponding symbols ), which were clamped stepwise from a holding potential of −100 mV to test potentials ranging from −160 to −20 mV. The traces were obtained before (bsl00043), during (bsl00041) and after (bsl00000) switching off red light (500 μmol m−2 sec−1). Switching off red light stimulates inward-conducting channels that slowly activate at −80 and −60 mV and slowly deactivate at −120 to −160 mV in wild-type (c), but not in MPK4IR-11 (d) guard cells. Arrows indicate the 0 pA level.
(e) Steady-state current/voltage relationship of wild-type guard cells sampled at the end of the 2 sec test pulses in (c). Data were obtained before (bsl00043), during (bsl00041) and after (bsl00000) switching off red light (500 μmol m−2 sec−1).
(f, g) Instantaneous current/voltage relationship of wild-type (f) and line MPK4IR-11 (g) sampled at the start of the test pulses shown in (c) and (d) (corresponding symbols). Darkness (bsl00041) causes an increase of instantaneous inward currents with linear voltage dependence in wild-type guard cells, but not in line MPK4IR-11.

Dark-induced elevation of cytosolic Ca2+ is associated with anion-channel activation

Guard-cell responses to CO2 have been linked to cytosolic Ca2+ signals, as CO2 triggers Ca2+ increases in Commelina communis (Webb et al., 1996) and slows spontaneous repetitive Ca2+ increases in Arabidopsis thaliana (Young et al., 2006). The Ca2+ response of N. tabacum guard cells was studied using triple-barrelled electrodes that enable simultaneous clamping of the plasma membrane potential and iontophoretic injection of the Ca2+ reporter dye FURA-2 (Marten et al., 2007b). Switching off red light (500 μmol m−2 sec−1) triggered repetitive cytosolic Ca2+ increases that were associated with anion-channel activation in nine of 16 wild-type cells (Figure 5a). In the remaining seven cells, inward current was activated without a change in the Ca2+ level of the guard cells. In ten of 13 MPK4IR-11 guard cells tested, both the increase in cytosolic Ca2+ and changes in plasma membrane current were absent (Figure 6a). Of the remaining cells, two showed only small Ca2+ transients (Δ[Ca2+]cyt < 100 nm), while one showed a similar response to wild-type (Figure 6b).

Figure 5.

 Darkness induces changes in cytosolic free Ca2+ concentration in N. tabacum guard cells.
(a, c) Traces for the cytosolic Ca2+ concentration (upper traces) and plasma membrane current (lower traces) in guard cells clamped to a holding potential of −100 mV and stimulated by switching off red light (500 μmol m−2 sec−1) as indicated by the bar below the traces. The cytosolic free Ca2+ concentration was calculated from the FURA-2 F345:F390 fluorescent ratio. In 9 of 16 wild-type guard cells (a), the dark-induced activation of anion channels was accompanied by repetitive changes in the cytosolic Ca2+ concentration. In 7 of 16 wild-type guard cells (c), anion channels were activated in darkness without a change of the cytosolic Ca2+ concentration.
(b) False colored images indicating the free Ca2+ concentration, in the same cell as the traces above (corresponding symbols). Switching off red light induces an increase in the cytosolic Ca2+ concentration throughout the cell, which is most obvious in the central area surrounding the nucleus.

Figure 6.

 Darkness-induced changes in the cytosolic Ca2+ concentration are repressed in MPK4IR-11 guard cells.
(a, b) Traces for the cytosolic Ca2+ concentration (upper traces) and plasma membrane current (lower traces) of guard cells clamped to a holding potential of −100 mV and stimulated by switching off red light (500 μmol m−2 sec−1) as indicated by the bar below the traces. The cytosolic free Ca2+ concentration was calculated from the FURA-2 F345:F390 fluorescent ratio. In 10 of 13 MPK4IR-11 guard cells (a), switching off red light did not affect the activity of anion channels nor the cytosolic Ca2+ concentration. In three of the 13 cells (b), switching off red light caused increases in inward current ranging from 30 to 270 pA, accompanied by changes in the cytosolic Ca2+ concentration.

In wild-type guard cells, the dark-induced increases in the cytosolic Ca2+ concentration correlate with increases in anion-channel activity (Figure 5a), but a linear correlation between the cytosolic Ca2+ level and inward current was not observed. Together with the Ca2+-independent activation of anion channels found in other cells (Figure 5b), this suggests that anion channels are primarily activated through a Ca2+-independent signalling pathway. However, the mean darkness-induced increase in anion-channel current was greater in cells showing a Ca2+ signal (Figure 7a) than in those with steady cytosolic Ca2+ levels (Figure 7b). An increase in the cytosolic Ca2+ level may thus further enhance the darkness-induced anion-channel activity.

Figure 7.

 Dark-induced anion channel activation is enhanced in wild-type N. tabacum guard cells showing simultaneous cytosolic Ca2+ increases.
(a, b) Mean darkness-induced changes in the cytosolic Ca2+ concentration (upper traces) and plasma membrane current (lower traces) of guard cells clamped to a holding potential of −100 mV. Changes in the cytosolic Ca2+ concentration and membrane currents were determined relative to the value measured just before switching off red light (= 0) as indicated by the bar below the trace. Errors bars represent SE. The measurements were averaged for those in which darkness induced a change in the cytosolic Ca2+ concentration (a) and those with steady cytosolic Ca2+ levels (b). ‘N’ indicates the number of experiments for which the data were averaged. The mean change in anion channel activity is three times larger for guard cells showing an increase in the cytosolic Ca2+ concentration compared to those with steady Ca2+ levels.

Guard cells depolarize in darkness independently of cytosolic Ca2+ elevation

Ca2+ homeostasis in guard cells depends on the plasma membrane potential, as plasma membrane Ca2+-permeable channels are activated upon hyperpolarization (Grabov and Blatt, 1998; Hamilton et al., 2000; Levchenko et al., 2005; Pei et al., 2000). Because of this relationship, dark-induced Ca2+ signals were studied in current clamp with wild-type N. tabacum guard cells. In hyperpolarized guard cells, switching off red light caused a reversible depolarization, just as previously found for V. faba guard cells (Roelfsema et al., 2002). In four of eight cells, an increase in cytosolic Ca2+ accompanied the change in membrane potential, but the Ca2+ increase occurred much later than the depolarization in two of the four cells (Figure 8a). Apparently, guard cells can depolarize independently of an increase in the cytosolic Ca2+ concentration. Darkness did not elicit membrane potential changes in guard cells of line MPK4IR-11 (n = 7), and no change in the cytosolic Ca2+ concentration was observed in six of seven cells (Figure 8b).

Figure 8.

 Depolarization of the guard-cell plasma membrane is induced by darkness in wild-type, but not in MPK4IR-11 cells.
(a, b) Simultaneous measurement of the cytosolic free Ca2+ concentration (upper traces) and the free-running membrane potential (lower traces) of guard cells stimulated by switching off red light (500 μmol m−2 sec−1), as indicated by the bar below the trace. Switching off red light induced depolarization in wild-type guard cells (a), but not in MPK4IR-11 guard cells (b). The depolarization was accompanied by a change in the cytosolic Ca2+ concentration in four of eight wild-type guard cells, whereas a darkness-induced Ca2+ increase was observed in only one of seven MPK4IR-11 guard cells.

Discussion

NtMPK4 is important for guard cell responses to CO2 and PAR

The major function of stomata is to enable uptake of CO2 into photosynthetic active tissues. In line with this role, guard cells sense changes in the intercellular CO2 concentration. High CO2 levels lower the stomatal conductance in the short term by stomatal closure and in the long term by reducing the stomatal density. In Arabidopsis, the long-term response involves the high carbon dioxide (HIC)-encoded fatty acid elongase (Gray et al., 2000), but apart from this gene product little is known about the signalling pathway by which CO2 controls stomatal development (Bergmann and Sack, 2007; Hetherington and Woodward, 2003). In comparison to sparse information on CO2-dependent regulation of stomatal development, more is known about the mechanism by which CO2 affects stomatal movement (Roelfsema and Hedrich, 2005; Vavasseur and Raghavendra, 2005).

MAP kinase signalling pathways negatively regulate the development of stomata, and thus affect stomatal patterning (Bergmann and Sack, 2007). In addition to patterning, however, MAP kinases pathways are also important for stomatal responses to CO2 and PAR (Hashimoto et al., 2006) and may be involved in ABA signalling (Gudesblat et al., 2007; Leung et al., 2006). A role for AtMPK4 in guard-cell responses has already been suggested by Petersen et al. (2000), who observed a high activity of the MPK4 promotor in guard cells of Arabidopsis. In tobacco, the NtMPK4-encoded MAP kinase appears to be essential for CO2-induced stomatal closure. Stomata of NtMPK4-silenced plants remain open at high atmospheric CO2 levels and show a repressed closure response after switching off PAR. The lack of sensitivity to PAR is in line with the proposed role of CO2 in a feedback loop, linking the stomatal aperture to photosynthetic activity (Roelfsema and Hedrich, 2005). Switching off PAR stops photosynthesis in the leaf and causes an increase in the intercellular CO2 concentration (Hanstein et al., 2001). The dark-induced increase in CO2 subsequently triggers stomatal closure in the wild-type (Heath, 1959; Roelfsema et al., 2002; Stahl, 1920), but not in the stomata of NtMPK4-repressed plants. Even though the PAR response of NtMPK4 plants was severely repressed, a residual change in stomatal conductance remained after switching off the illumination. This suggests that N. tabacum stomata receive other PAR-generated signals in addition to changes in the intercellular CO2 level. Light also affects the apoplastic concentration of several ions such as K+, Cl and Ca2+ (Felle et al., 2000; Roelfsema and Hedrich, 2002), which could serve as intermediate signals in addition to CO2-based signals.

In Arabidopsis, loss of stomatal CO2 responses was found for the ht1 mutants, which show a reduced activity of the HT1-encoded protein kinase (Hashimoto et al., 2006). The HT1 gene has been classified as a MAPKK kinase (Ichimura et al., 2002), and functions in a MAP kinase pathway that is essential for CO2 responses in guard cells. In accordance with the phenotype of NtMPK4-silenced plants, the stomata of the ht1 mutant do not respond to PAR. Despite these similarities, the two lines display opposite phenotypes with respect to the degree of stomatal opening. In the ht1 mutant, the stomata remain closed at low atmospheric CO2 concentrations, whereas those of NtMPK4-silenced plants stay open at high CO2 levels. This suggests that the signalling pathway involving HT1 acts as a negative regulator of guard-cell CO2 responses, while that of NtMPK4 represents a positive regulator.

The overactive SIPK kinase SIPKKEE phosphorylates NtMPK4 and therefore can act upstream of NtMPK4 (Gomi et al., 2005). In line with this function, N. tabacum plants expressing SIPKKEE show a reduced response to ozone, whereas NtMPK4-silenced plants are over-sensitive to this stimulus. In contrast to the ozone sensitivity, we observed no differences in the CO2 responses of wild-type and SIPKKEE-expressing plants. This suggests that the activity of NtMPK4 is essential, but not sufficient for CO2-induced stomatal closure. Possibly other MAP kinases are activated in parallel with NtMPK4, and their activity is also required for guard-cell responses to CO2.

Anion channels are activated through Ca2+-dependent and independent mechanisms

Plasma membrane anion channels have been identified as important targets downstream in the CO2 signalling pathway in guard cells of V. faba (Brearley et al., 1997; Raschke et al., 2003; Roelfsema et al., 2002). The signalling chain seems to be evolutionarily conserved, and S-type anion channels are also activated by CO2 in the guard cells of N. tabacum (Figure 3). Anion-channel activation in the presence of CO2 was suggested to involve changes in cytosolic Ca2+ (Webb et al., 1996), as CO2 induces increases in the Ca2+ level and plasma membrane anion channels are activated by cytosolic Ca2+ (Hedrich et al., 1990; Schroeder and Hagiwara, 1989). Simultaneous recordings of the plasma membrane conductance and Ca2+ imaging indicate that elevation of the cytosolic Ca2+ level does indeed lead to enhanced anion-channel activity (Figure 7). The Ca2+ response induced by switching off PAR was observed in only one of eleven MPK4IR-11 guard cells, suggesting that this response is predominantly mediated through changes in the intercellular CO2 concentration.

Despite the observed relationship between cytosolic Ca2+ changes and anion-channel activity, CO2 also activates anion channels in the absence of Ca2+ signals. This suggests that, in addition to a Ca2+-dependent pathway, anion channels are also stimulated through a Ca2+-independent mechanism. The existence of this Ca2+-independent pathway is supported by the observation that darkness induces a depolarization of guard cells, independently of an increase in the cytosolic Ca2+ concentration (Figure 8). Apparently, CO2 acts primarily on anion channels via a Ca2+-independent pathway, but the activity of anion channels is enhanced when the Ca2+ concentration increases.

In guard cells of N. tabacum, the darkness-induced activation of anion channels as well as the induction of Ca2+ signals depend on NtMPK4. Future studies may identify components acting downstream of NtMPK4, which in turn could interact with anion or Ca2+ channels in guard cells. The search for proteins acting upstream of NtMPK4 may identify the MAP kinase chain that forwards the CO2 signal, and eventually could uncover the receptor that senses CO2 in guard cells.

Experimental procedures

Plant material and experiments with epidermal strips

Nicotiana tabacum var. Samsun plants were grown in the greenhouse, and additional light was provided by HQL pressure lamps (Philips, http://www.lighting.phillips.com, Powerstar HQI-E, 400 W), with a photon flux density of 160 μmol m−2 sec−1 and a day/night cycle of 12/12 h. All measurements were performed using the 2nd to 5th true leaves of 4- to 6-week-old-plants. The isolation of NtMPK4-silenced lines IR-6 and IR-11 and SIPKEE-expressing lines 33 and 34 has been described previously (Gomi et al., 2005).

For measurements of guard-cell size and stomatal apertures, epidermal strips were peeled from the abaxial side of the leaf. Directly after peeling, the epidermis was attached to microscope slides covered with medical adhesive (Medical Adhesive B, Aromando, http://www.amt-med.de). For determination of guard-cell size, microscope slides were transferred to Petri dishes filled with 5 mm KCl, 5 mm potassium citrate pH 6, 0.1 mm CaCl2 and 1 mm MgCl2. Stomata were photographed using a cooled CCD camera (CoolSNAP HQ, Roper Scientific, http://www.roperscientific.com) and guard-cell size was measured using a microscopy standard.

For stomatal opening assays, cover slips with epidermal strips were transferred to Petri dishes containing 50 mm KCl, 0.1 mm CaCl2 and 5 mm MES/BTP (2-(N-Morpholino) ethanesulfonic acid/Bis-Tris Propane) pH 6.0 and a range of ABA concentrations. Stomatal opening was induced with white light provided by a halogen lamp (HLX Xenophot 15 V 150 W−1, Osram, http://www.osram.com) at a photon flux density of 250 μmol m−2 sec−1. The epidermal strips were photographed after 3 h using a CCD camera, and the stomatal apertures were determined using Metamorph software (Molecular Devices, http://www.moleculardevices.com).

Gas-exchange measurements

Sections of pre-darkened N. tabacum leaves were enclosed in a sandwich-type cuvette (diameter 2.1 cm) with windows on the upper and lower side. The abaxial side of the leaves was directed to the upper side of the cuvette and exposed to a gas stream of 0.5 l min−1. Measurements were performed at a temperature of 24°C and relative humidity of 47% in darkness; when red light was used, the temperature and relative humidity altered to 25°C and 44%, respectively. Red light (400 μmol m−2 sec−1) was provided by a halogen lamp, passed through a long-pass glass filter (edge wavelength 610 nm) and projected onto the adaxial side of the leaf using fibreoptics. The radiation was measured using a LI-COR 250 light meter (quantum sensor LI-190) (LI-COR, http://www.licor.com). Atmospheric CO2 concentrations ranging from 0 to 700 μl l−1 were obtained by passing air through a soda lime column and injection of CO2 directly into the air stream. Transpiration and CO2 uptake were measured using an infra-red gas analysis technique (Binos 1, Heraeus, http://www.heraeus.com). Experiments were started at 08:00 h.

Experimental set-up for impalement

Plants were placed next to the table of an upright microscope (Akioskop 2FS, Zeiss, http://www.zeiss.com/). The adaxial side of a leaf was mounted on a Plexiglas® holder in the focal plane of the microscope using double-sided adhesive tape. In this experimental configuration, the abaxial epidermis was directed upwards and guard cells within this cell layer were accessible for impalement with double- or triple-barrelled microelectrodes (Levchenko et al., 2005; Roelfsema et al., 2001). Guard cells were impaled with micro-electrodes using a piezo-driven micro-manipulator (MM3A, Kleindiek Nanotechnik, http://www.nanotechnik.com). The impalement was monitored using either a water-immersion objective (Achroplan 40×/0.80 W, Zeiss) or a dry objective (Epiplan LD 50 ×/0.50 Zeiss). The water-immersion objective was operated with a drop of solution in contact with the leaf surface, containing 5 mm KCl, 5 mm potassium citrate pH 6, 0.1 mm CaCl2, 1 mm MgCl2. In experiments with the dry objective, the leaf surface below the objective was exposed to a flow of air (0.4 l min−1) passed through a suspensor (14 × 2 mm). Guard cells were stimulated by switching from CO2-free air to air containing 700 μl l−1 CO2 (Roelfsema et al., 2002). Leaves were illuminated with red light provided by the microscope lamp (12 V 60 W−1, Zeiss), passed through a long-pass glass filter (λ½ 610 nm, RD610, Schott, http://www.schott.com) and focused on the adaxial side on an area of 2 cm in diameter with a photon flux density of 500 μmol m−2 sec−1.

Micro-electrodes and electrical configuration

The micro-electrodes were produced from borosilicate glass capillaries (inner diameter 0.58 mm, outer diameter 1.0 mm, Hilgenberg, http://www.hilgenberg-gmbh.com) as described previously (Roelfsema et al., 2001). The capillaries were aligned, heated, twisted 360°, and pre-pulled on a customized vertical electrode puller (L/M-3P-A, List Medical Electronic). The final pull was executed on a horizontal laser puller (P2000, Sutter Instruments Co., http://www.sutter.com). The electrodes were filled with 300 mm KCl for voltage measurement and current injection, and the 3rd barrel of triple-barrelled electrodes was filled with 2 mm FURA-2 for calcium fluorescence imaging. The tip resistance ranged from 90 to 160 or from 90 to 200 MΩ for double- and triple-barrelled electrodes, respectively. The reference electrode, a 50 mm KCl and 1.5% agarose salt bridge, connected to an AgCl/Cl half cell, was placed in the perfusion solution between the objective and cuticle for stimulation with red light, or in a drop of 50 μl perfusion solution on leaf surface during CO2 exposure experiments. All barrels were connected with Ag/AgCl half cells to head-stages (HS 180, BioLogic, http://www.bio-logic.info), with an input resistance of approximately 1011Ω. Head-stages were coupled to a micro-electrode amplifier (VF-102, BioLogic) and the membrane potential was clamped using a differential amplifier (CA-100, BioLogic). Voltage steps were controlled by Pulse software (Heka, http://www.heka.com) using an ITC-16 interface (Heka). Data were filtered at 250 Hz and sampled at 1 kHz during short pulses, or filtered at 10 Hz and sampled at 33 Hz for long-term registration.

Ratiometric fluorescence microscopy

The ratiometric fluorescent calcium indicator dye FURA-2 (Fluka, Sigma-Aldrich, http://www.sigmaaldrich.com/) was loaded by iontophoretic micro-injection into guard cells, applying an injection current of approximately −350 pA to the 3rd micro-electrode barrel. During FURA-2 injection, the cell was constantly clamped to −100 mV and thus the injection current from the 3rd barrel was automatically compensated for by a current from the 2nd barrel. In approximately six of seven cells, the dye was loaded into the cytoplasm, but it localizes in the vacuole in a minority of injected cells (Marten et al., 2007b). Ratiometric fluorescence spectroscopy measurements were performed using metafluor software (Universal Imaging). FURA-2 was excited by 100 msec flashes of UV light at 345 and 390 nm with a time interval of 0.8 sec (VisiChrome high-speed polychromator system, Visitron Systems, http://www.visitron.de). The emission signal was passed through a beam splitter (FT 395, Zeiss), filtered using a 510 nm band-pass filter (D510/40 m, AHF Analysentechnik, http://www.ahf.de) and captured using a cooled CCD camera (CoolSNAP HQ). Background fluorescence levels at both wavelengths were taken from a reference region within the unloaded neighbouring guard cell or neighbouring epidermal cells. The intracellular free Ca2+ concentration was calculated according to the method described by Grynkiewicz et al. (1985) using the equation:

image( [1])

where Kd represents the binding constant of FURA-2 for Ca2+, R represents the 345/390 nm excitation ratio, and Rmin and Rmax correspond to the ratio of Ca2+-free and Ca2+-saturated FURA-2, respectively. Fmin and Fmax are the fluorescence intensity measured at 390 nm with Ca2+-free and Ca2+-saturated FURA-2, respectively. We used here a Kd of 270 nm, as determined in vitro by Levchenko et al., (2005). Rmin and Fmin were defined as the values obtained after simultaneously injecting FURA-2 and 1,2-Bis (2-aminophenoxy) ethane-N,N,N,N′-tetraacetic acid (BAPTA) at 0 mV into guard cells of intact plants. The values for Rmax and Fmax were obtained by clamping the plasma membrane from 0 to −250 mV, inducing a massive and saturating influx of Ca2+ through stimulation of hyperpolarization-activated Ca2+ channels (Levchenko et al., 2005).

Acknowledgements

We thank K. Neuwinger and E. Reisberger (University of Würzburg) for technical assistance with the gas-exchange recordings. This work was supported by grants from the Deutsche Forschungsgemeinschaft to R.H. and M.R.G.R. (SPP 1108 HE 1640/19).

Ancillary