Reciprocal regulation of Ca²⁺-activated outward K⁺ channels of Pyrus pyrifolia pollen by heme and carbon monoxide

Summary

• •
  The regulation of plant potassium (K⁺) channels has been extensively studied in various systems. However, the mechanism of their regulation in the pollen tube is unclear.
• •
  In this study, the effects of heme and carbon monoxide (CO) on the outward K⁺ (K⁺_out) channel in pear (Pyrus pyrifolia) pollen tube protoplasts were characterized using a patch-clamp technique.
• •
  Heme (1 μM) decreased the probability of K⁺_out channel opening without affecting the unitary conductance, but this inhibition disappeared when heme was co-applied with 10 μM intracellular free Ca²⁺. Conversely, exposure to heme in the presence of NADPH increased channel activity. However, with tin protoporphyrin IX treatment, which inhibits hemeoxygenase activity, the inhibition of the K⁺_out channel by heme occurred even in the presence of NADPH. CO, a product of heme catabolism by hemeoxygenase, activates the K⁺_out channel in pollen tube protoplasts in a dose-dependent manner. The current induced by CO was inhibited by the K⁺ channel inhibitor tetraethylammonium.
• •
  These data indicate a role of heme and CO in reciprocal regulation of the K⁺_out channel in pear pollen tubes.

Introduction

Potassium is required for both pollen germination and pollen tube growth (Brewbaker & Kwack, 1963; Weisenseel & Jaffe, 1976; Feijo et al., 1995). It is known that potassium (K⁺) transport across the pollen plasma membranes is regulated by the K⁺ channels (Feijo et al., 1995; Taylor & Hepler, 1997; Fan et al., 1999; Dutta & Robinson, 2004; Amien et al., 2010). The use of electrophysiology (mainly the patch-clamp technique) to measure ion currents has enabled considerable progress to be made in understanding the nature of plant ion physiology (Ward et al., 2009). Three families of K⁺ channel have been identified in plants: Shaker, tandem-pore K⁺ (TPK) and inwardly rectifying K⁺ (Kir)-like K⁺ channels (Véry & Sentenac, 2003; Lebaudy et al., 2007). The inward Shaker K⁺ channels cloned from Arabidopsis thaliana were found to functionally complement yeast mutant strains defective in K⁺ uptake, and were the first plant K⁺ channels identified at the molecular level (Anderson et al., 1992; Schachtman et al., 1992; Sentenac et al., 1992). TPK K⁺ channels in plants were firstly identified by in silico searches (Czempinski et al., 1997), while the first Kir-like channel in A. thaliana was identified by searching for TPK1-related sequences in A. thaliana genome sequence databases (Czempinski et al., 1999). Plant K⁺ channels have numerous physiological functions, including root K⁺ uptake, K⁺ long-distance transport, control of the stomatal aperture and pollen tube growth (Lebaudy et al., 2007).

The pollen tube is an ideal model system for the study of ion fluxes in plants (Michard et al., 2009). A K⁺ channel of the Shaker family referred to as SPIK (plant Shaker inwardly rectifying K⁺) from A. thaliana has been cloned and characterized in pollen. The SPIK-defective mutant has greatly reduced inwardly rectifying K⁺-channel activity in the pollen plasma membrane, which results in impaired pollen tube growth (Mouline et al., 2002). AtTPK4 plays functional roles in K⁺ homeostasis and membrane voltage control of the growing pollen tube, which are blocked by extracellular Ca²⁺ and cytoplasmic protons (Becker et al., 2004). Pollen tube K⁺ channels may be involved in the control of the osmotic concentration to prevent pollen tube rupture during growth (Fan et al., 1999; Becker et al., 2004). Moreover, a decrease in K⁺ concentration in the cytoplasm during pollen hydration is a prerequisite for protein synthesis and the subsequent onset of pollen germination (Bashe & Mascarenhas, 1984). Several outward K⁺ channels have also been identified in the pollen tube (Fan et al., 2003; Griessner & Obermeyer, 2003); however, their regulatory mechanisms are still unclear.

Heme (Fe³⁺protoporphyrin IX chloride) is a complex of iron with protoporphyrin IX and is essential for living organisms, with various biological functions including protein synthesis and cell differentiation (Ponka, 1999). In plants, heme is synthesized from glutamyl-tRNA^glu via 5-aminolevulinic acid, a linear five-carbon molecule. The steps of tetrapyrrole biosynthesis up to protoporphyrinogen IX occur only within the chloroplast, but a fraction of protoporphyrinogen IX leaves the chloroplast by unknown routes and enters the mitochondria, where it serves as a precursor for heme (Papenbrock & Grimm, 2001). In the presence of oxygen and NADPH, hemoxygenases catalyze heme degradation, producing biliverdin, iron and carbon monoxide (CO) (Ponka, 1999). There are at least three isoforms of hemoxygenase: hemoxygenase-1 in rats is classified as a heat-shock protein (HSP32) (Ewing & Maines, 1991); hemoxygenase-2 and hemoxygenase-3 are constitutively expressed in many mammalian cells (Dulak & Józkowicz, 2003). In plants, hemoxygenase-1 has been reported as the sole enzymatic source of CO (Davis et al., 2001; Xuan et al., 2008).

Plasma membrane ion channels are important heme targets in animals, for example, the Ca²⁺-activated K⁺ channel (Tang et al., 2003; Williams et al., 2004). After heme binds to a channel, it decreases the frequency of channel opening, resulting in the K⁺ current being markedly reduced (Tang et al., 2003; Jaggar et al., 2005); but supplying heme and NADPH induces an increase in the frequency of K⁺ channel opening (Williams et al., 2004). In hypoxic conditions, the channel opening frequency is significantly inhibited even in the presence of heme plus NADPH, indicating that activation by heme plus NADPH requires oxygen (Williams et al., 2004). Interestingly, after addition of CO, a product of heme catabolism by hemeoxygenase, the frequency of Ca²⁺-activated K⁺ channel opening can be increased (Riesco-Fagundo et al., 2001; Williams et al., 2004; Jaggar et al., 2005), indicating that heme and CO reciprocally regulate the Ca²⁺-activated K⁺ channel. No study has yet determined whether ion channels are sensitive to heme or CO in plants. However, early reports showed that oxygen pressure approaches zero in the ovary (Linskens & Schrauwen, 1966; Blasiak et al., 2001), and in these hypoxic conditions, heme cannot transfer to CO. If there are any K⁺ channels regulated by the heme/CO system in the pollen tube, they may play a functional role in the pollen tube burst in the ovary, as the K⁺ channel is involved in osmotic control, as mentioned above. In the present study, we identified a K⁺_out channel in pear pollen tube protoplasts. Heme and CO reciprocally regulate K⁺_out channel activity. This regulatory mechanism may play an important role in controlling pollen tube development.

Materials and Methods

Pollen culture conditions

Pollen from pear, Pyrus pyrifolia Nakai cv Hosui, was collected from the Nanjing Agricultural Fruit Experimental Yard and preserved by drying in air at ambient temperature for 12 h and then stored in silica gel at −20°C. Mature pear pollen was incubated in liquid culture for germination and growth. The culture contained the following components: 0.55 mM Ca(NO₃)₂, 1.60 mM H₃BO₃, 1.60 mM MgSO₄, 1.00 mM KNO₃, 440 mM sucrose and 5 mM 2-(N-morpholino)ethanesulfonic acid hydrate (MES) at pH 6.0–6.2 (adjusted with Tris). CO or tetraethylammonium (TEA⁺) was added directly to the medium as required. The pollen was incubated in small Petri dishes at 24 ± 1°C for 6 h. The pollen tube length was measured under a light microscope with the Image-Pro software (Media Cybemetics, Silver Spring, MD, USA).

Isolation of pollen tube protoplasts

Pollen tube protoplasts were isolated as described by Qu et al. (2007), with some modifications. Briefly, 0.05 g of mature pollen grains was cultured in 0.5 ml of liquid germination medium containing 0.55 mM Ca(NO₃)₂, 1.60 mM H₃BO₃, 1.60 mM MgSO₄, 1.00 mM KNO₃, 440 mM sucrose and 5 mM MES–Tris (pH 6.0–6.2, adjusted with Tris) for 3 h, and then washed with standard solution (1 mM KNO₃, 0.2 mM KH₂PO₄, 1 mM MgSO₄, 1 mM KI, 0.1 mM CuSO₄, 5 mM CaCl₂, and 5 mM MES–Tris, pH 5.5) and incubated in the enzyme solution (standard solution plus 1% (w/v) macerozyme R-10 (Onozuka, Tokyo, Japan), 2.0% (w/v) cellulase R-10 (Onozuka) and 1% (w/v) bovine serum albumin) for 50 min at 30ºC to release the pollen tube protoplasts. The pollen tube protoplasts were washed three times with standard bath solution (see the Patch-clamp experiments section, and also the Materials and Methods section), and then re-suspended in the standard bath solution and kept on ice until use in patch-clamp experiments.

Patch-clamp experiments

Whole-cell and single-channel patch-clamp experiments on the pollen tube protoplasts were performed using an Axopatch-200B amplifier (Axon Instruments, Foster City, CA, USA) as described by Qu et al. (2007). Data were analyzed using pCLAMP 9.0 (Axon Instruments). For whole-cell recordings, the standard bath solution contained 1 mM Ca-gluconate, 0.2 mM glucose, 10 mM potassium gluconate, 5 mM MES–Tris (pH 5.5) (osmolality of 800 mOsM adjusted with D-sorbitol). The standard pipette solution contained 10 mM ethyleneglycoltetraacetic acid (EGTA), 100 mM potassium gluconate, 4 mM TrisATP, 10 mM HEPES–Tris (pH 7.1) and 7.5 μM Ca-gluconate to give a free-Ca²⁺concentration of 1 μM (calculated using the chemical speciation program MaxChelator, Stanford University, Stanford, CA, USA, http://maxchelator.stanford.edu/) and an osmolality of 800 mOsM adjusted with D-sorbitol. The whole-cell recordings conducted under these conditions were used as controls. Variations in solution composition for different experiments are indicated in the figure legends. Cell capacitance was measured for each cell with the capacity compensation device of the amplifier. Data were sampled at 2 kHz and filtered at 0.5 kHz. All data were acquired 5 min after a whole-cell configuration was achieved unless stated otherwise. In inside-out single-channel patch-clamp experiments, bath solutions contained 10 mM HEPES–Tris (pH 7.1), 10 mM EGTA, 100 mM potassium gluconate, 4 mM TrisATP and 2.21, 7.5 or 9.87 μM Ca-gluconate to give free-Ca²⁺concentrations of 0.1, 1 and 10 μM, respectively (calculated using MaxChelator) and an osmolality of 800 mOsM adjusted with D-sorbitol. The pipette solution contained 5 mM MES–Tris (pH 5.5), 1 mM Ca-gluconate, 0.2 mM glucose and 10 mM potassium gluconate (osmolality of 800 mOsM adjusted with sorbitol). The data were acquired at 20 kHz and low-pass filtered at 5 kHz. During post-analysis, data were further filtered at 200 Hz. Patch-clamp whole-cell and single-channel recordings were conducted at room temperature (c. 20°C).

Chemicals and stock solutions

All chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA) unless otherwise indicated. Solutions of these chemical reagents were made weekly and kept at −20°C in aliquots. (Ru(CO)₃Cl₂)₂ was freshly dissolved in dimethyl sulfoxide on the day on which the experiments were performed. The medium containing the appropriate concentration of CO donor was made 30 min before the experiments were conducted.

Statistics

Data are displayed as mean ± SEM. Whole-cell currents were assessed using normalizing currents (currents per unit capacitance, pA pF⁻¹) to account for variation in cell surface area. Single-channel events were listed and analyzed by pclampfit 9.0. Valid channel opening was determined using the 50% threshold cross method. The total number of functional channels (n) in a patch was determined when multiple channel events were observed in the patch. In pollen tube growth experiments, the values of pollen tube length were the means of six independent experiments, each of ≥ 50 determinations.

Results

Characterization of the Pyrus pyrifolia pollen K⁺_out channel

Outward currents were elicited under control conditions (Fig. 1a) when the plasma membrane of pear pollen tube protoplasts was depolarized from −90 to +90 mV within 1 s during whole-cell patch-clamp recording (Fig. 1b). The elicited currents were 94.4 ± 7.8 pA pF⁻¹; n = 12 (all data in this format are mean ± SEM with noted n values) at +90 mV under control conditions (Fig. 1c). We used tail-current analysis to determine the major ionic species contributing to the recorded currents. The measured reversal potential (E_rev) value for the currents was c. −30 mV under control conditions, with 10 and 100 mM K⁺ in the bath and pipette solutions, respectively (Fig. 1d). This value was an approximation to the theoretical equilibrium potential for K⁺ of −59 mV under these conditions, while the theoretical equilibrium potential for Ca²⁺ was c. +89 mV. The results suggested that the outward currents were mainly carried by K⁺. Additionally, some other ions may have made small contributes to the whole-cell currents, as the value of E_rev was more positive than the theoretical equilibrium potential for K⁺. The K⁺_out channel was highly selective for K⁺ compared with Na⁺ (Fig. 1a). When 10 and 100 mM K⁺ in the bath and pipette solutions were substituted with 10 and 100 mM Na⁺, respectively, the elicited currents decreased to 11.7 ± 2.2 pA pF⁻¹ at +90 mV (Fig. 1c; n = 6). Inside-out single channel recording showed that K⁺_out conductance was c. 9 pS at +60 mV (Fig. 2a). Furthermore, the K⁺_out channel open probability (NP_O: N, number of channels; P_O, open probability of one channel) was very sensitive to the intracellular free Ca²⁺ concentration ([Ca²⁺]_cyt). The channel NP_O was significantly higher for 1 than for 0.1 μM [Ca²⁺]_cyt (Fig. 2b; P < 0.01; n = 5), and for 10 than for 1 μM [Ca²⁺]_cyt (Fig. 2b; P < 0.001; n = 5).

Figure 1.  The whole-cell outward K⁺ current in the pollen tube protoplasts of Pyrus pyrifolia cv Hosui. (a) Whole-cell K⁺ and Na⁺ currents evoked by an increasingly depolarized voltage, with pulses from −90 to +90 mV, in pear pollen tube protoplasts. (b) The protocol for the whole-cell patch-clamp experiments. The voltage was held at 0 mV, and then activated from −90 to +90 mV within 1 s. (c) Current–voltage (I–V) relationships for K⁺ (closed circles) and Na⁺ (open circles). Bars show mean ± SEM (n = 12 and 6, respectively). (d) Tail-current analysis revealed the reversal potential of whole-cell conductance. After activation of conductance using a single voltage step from 0 to +60 mV, the voltage was increased in a single step to depolarizing voltages ranging from −60 to +10 mV. The arrow indicates the reversal voltage of c. −30 mV.

Figure 2.  K⁺_out channel activation by cytosolic free Ca²⁺. (a) K⁺ currents in the inside-out single-channel configuration (10 mM K⁺ in the pipette solution and 100 mM K⁺ in the bath solution; Ca²⁺ in the bath solution was kept constant at 0.1, 1 or 10 μM). Step-like current events indicate the opening (1) and closing (0) of a single channel. Different channel opening states are indicated as 1, 2 or 3. The channel opening state was recorded at +60 mV. (b) The K⁺_out channel open probability (NP_O) of the inside-out single channel of K⁺ at 0.1, 1 or 10 μM Ca²⁺ in the bath solution. Bars show mean ± SEM (n = 5). Results differed significantly between 0.1 and 1 μM Ca²⁺, and between 1 and 10 μM Ca²⁺ (P < 0.01 and P < 0.001, respectively; Student’s t-test).

Effect of heme on channel activity

We used an inside-out single-channel configuration to determine whether heme can regulate K⁺_out channel activity. In this configuration, channel-containing membrane patches isolated from the plasma membrane were physically isolated from the rest of the cell and cytoplasmic factors were largely absent. We added heme at a concentration of 1 μM to the intracellular side of the channels (the bath solution of the inside-out recording configuration) and this significantly reduced the NP_O of the K⁺_out channel (Fig. 3a,b; P < 0.001; n = 7). There was no change in unitary current amplitude under the heme treatment (Fig. 3a). Moreover, the heme inhibiting the K⁺_out channel could be washed out (Supporting Information Fig. S1). The inhibition disappeared when 1 μM heme was co-applied with 10 μM intracellular free Ca²⁺(Fig. 3a,b; P < 0.001; n = 5).

Figure 3.  Heme directly inhibited the K⁺_out channel open probability. (a) K⁺ currents in the inside-out single-channel configuration (10 mM K⁺ in the pipette solution; 1 μM Ca²⁺ and 100 mM K⁺ in the bath solution) treated with 1 μM heme (bath solution) and 1 μM heme plus 10 μM Ca²⁺. Step-like current events indicate the opening (1) and closing (0) of a single channel. Different channel opening states are indicated as 1 and 2. The channel opening state was recorded at a voltage of +60 mV. (b) The K⁺_out channel open probability (NP_O) of an inside-out single K⁺ channel under control conditions and for treatments with 1 μM heme (n = 7), and 1 μM heme plus 10 μM Ca²⁺ (n = 5) in the bath solution; bars show mean ± SEM. Comparisons between the control and the 1 μM heme treatment; the control and the 1 μM heme plus 10 μM Ca²⁺ treatment; and the 1 μM heme and 1 μM heme plus 10 μM Ca²⁺ treatments showed significant differences (P < 0.001, P < 0.01 and P < 0.001, respectively; Student’s t-test). (c) Concentration-dependent K⁺_out channel inhibition by heme at +60 mV. Each point is the average of five to eight cells. (d) Effect of 1 μM heme on K⁺_out channel activity in the pear pollen tube when present in the bath solution with the outside-out configuration. The difference between this treatment and the control was not significant (P = 0.893; Student’s t-test). Bars show mean ± SEM (n = 5).

Inhibition of the K⁺_out channel by heme did not require any exogenous enzyme or cofactor. The half-maximal inhibitory concentration (IC₅₀) range was 100–200 nM (Fig. 3c), equal to the equivalent inhibitory concentrations for the Ca²⁺-activated K⁺ channel (Tang et al., 2003). The inhibitory effect on the K⁺_out channel was specific to intracellular free heme. Application of heme at the same concentration to the extracellular face of the membrane (the bath solution of the outside-out recording configuration) affected neither the single channel conductance nor the NP_O of the K⁺_out channel (Fig. 3d; not significant; n = 5).

Effect of co-applied NADPH and heme on channel activity

NADPH is the reduced form of nicotinamide adenine dinucleotide phosphate (NADP⁺). Application of 1 μM NADPH alone had no effect on the K⁺_out channel activity (Fig. 4; not significant; n = 6). However, when 1 μM heme was co-applied with 1 μM NADPH, the inhibitory effect of heme was reversed; NP_O increased to 212 ± 12% of the value under control conditions (Fig. 4; P < 0.001; n = 8) and unitary conductance remained unchanged. Oxygen, NADPH and hemeoxygenase convert heme to biliverdin, iron and CO, as mentioned above. We further examined the effects of co-application of 1 μM heme and 1 μM NADPH on the K⁺_out channel activity in patches maintained under 100 nM tin protoporphyrin IX (SnPP) treatment, which will inhibit the action of hemeoxygenase. In the SnPP treatment, heme plus NADPH decreased the NP_O of the K⁺_out channel by > 70% compared with controls (Fig. 4; P < 0.001; n = 7), similar to the effect for application of heme only (Fig. 3b). SnPP alone caused only c. 10.8% inhibition of activity, which was not significantly different compared with the control (Fig. 4; ns; n = 4).

Figure 4.  Effect of co-applied chemicals on the K⁺_out channel. Recordings were made of K⁺_out channel activity with the inside-out single-channel configuration (10 mM K⁺ in the pipette solution; 1 μM Ca²⁺ and 100 mM K⁺ in the bath solution). Treatment with 1 μM NADPH alone did not affect K⁺_out channel activity (P = 0.730; Student’s t-test). Bars show mean ± SEM (n = 6). Treatment with 100 nM tin protoporphyrin IX (SnPP) alone had little effect on K⁺_out channel activity (which decreased by c. 10.8%); the activity was not significantly different from that of controls (P = 0.102; Student’s t-test). Bars show mean ± SEM (n = 4). In the treatment with 1 μM NADPH plus 1 μM heme, the channel open probability (NP_O) of the K⁺_out channel increased to 212 ± 12% compared with control conditions (P < 0.001; Student’s t-test). Bars show mean ± SEM (n = 8). Treatment with 1 μM NADPH, 1 μM heme and 100 nM SnPP together induced a decrease in the NP_O of the K⁺_out channel of > 70%, resembling that seen on application of heme alone (P < 0.001, compared with the control; Student’s t-test). Bars show mean ± SEM (n = 7).

Effect of CO on channel activity

To determine whether the effects of heme and NADPH might be attributable to the heme catabolism product CO, we investigated the effects of (Ru(CO)₃Cl₂)₂ (a CO donor) on K⁺_out channel activity. On application of this CO donor (100 nM), NP_O was significantly increased, and multiple channel openings were often elicited (Fig. 5a,d; P < 0.001; n = 7), but the single channel conductance of the K⁺_out channel was not modified. The whole-cell recording showed that CO induced channel activity in a dose-dependent manner in the range from 10⁻⁵ to 10 μM (Fig. 5b). The K⁺ currents induced by 100 nM CO were strongly inhibited by 10 mM TEA⁺ (Fig. 5c), which is widely used for reversible blocking of the K⁺ channel. Unlike CO, the gas nitric oxide (NO) did not affect the NP_O of the K⁺_out channel (Fig. 5d; ns; n = 7). Application of 5 μM of the control compound RuCl₂(DMSO)₄, which does not release CO, failed to significantly increase NP_O (Fig. 5d; ns; n = 6). The other heme catabolic product, biliverdin (100 nM), had no effect on the K⁺_out channel activity (Fig. 5d; ns; n = 5). Such results showed that modulation of K⁺_out channel activity by CO resembled that seen on co-application of heme and NADPH, suggesting that the stimulatory effect of these agents might reflect the generation of CO by hemeoxygenase. Furthermore, these findings suggest that heme (a substrate of hemeoxygenase) and CO (a product of hemeoxygenase) act competitively in their effects on the K⁺_out channel.

Figure 5.  Effect of carbon monoxide (CO) on K⁺_out channel activity. (a) Recordings of K⁺ currents with the inside-out single-channel configuration (10 mM K⁺ in the pipette solution; 1 μM Ca²⁺ and 100 mM K⁺ in the bath solution) treated with 100 nM CO in the bath solution. Step-like current events indicate the opening (1) and closing (0) of a single channel. Different channel opening states are indicated as 1, 2 and 3. The channel opening state was recorded at a voltage of +60 mV. (b) Whole-cell K⁺ current evoked by CO (10⁻⁵ to 10 μM) at +90 mV. Each point is the average of five to eight cells. (c) Current–voltage (I–V) relationships for K⁺ induced by 100 nM CO (closed circles), or 100 nM CO plus 10 mM tetraethylammonium (TEA⁺) (squares) (control, open circles). Bars show mean ± SEM (n = 6). The induced K⁺ currents were strongly inhibited by 10 mM TEA⁺. (d) Treatment with 100 nM CO alone significantly increased the K⁺ channel open probability (NP_O) (P < 0.001; Student’s t-test). Bars show mean ± SEM (n = 7). Treatment with 100 nM biliverdin did not affect the K⁺ channel NP_O (P = 0.893; Student’s t-test). Bars show mean ± SEM (n = 5). Treatment with NO gas did not affect the K⁺ channel NP_O (P = 0.338; Student’s t-test). Bars show mean ± SEM (n = 7). Treatment with 5 μM RuCl₂(DMSO)₄ did not affect the K⁺ channel NP_O (P = 0.963; Student’s t-test). Bars show mean ± SEM (n = 6).

We further tested the specificity of the effect of CO on the K⁺_out channel. Current traces from one pear pollen tube protoplast recorded before and after exposure for 1 min to 100 nM CO donor are shown in Fig. 6(a). CO significantly increased the K⁺_out current; however, the inward K⁺ current was not affected by the CO donor at this concentration even after 20 min of exposure (Fig. 6b; n = 7). These results indicate that CO selectively activates the K⁺_out channel in the pear pollen tube.

Figure 6.  Carbon monoxide (CO) selectively activates the K⁺_out channel. (a) Whole-cell patch-clamp recording from a pear (Pyrus pyrifolia cv Hosui) pollen tube protoplast using a protocol of steps between −180 and +60 mV from a holding voltage of −58 mV. Current traces for before (left) and 1 min after (right) CO treatment are shown. (b) Current–voltage (I–V) relationships for the K⁺ currents induced by 100 nM CO (closed circles; control, open circles). Bars show mean ± SEM (n = 7). The K⁺_out current was significantly activated, while the inward K⁺ current was not changed by CO treatment.

Effect of CO on pear pollen germination and pollen tube growth

To elucidate the potential mechanism of the effect of CO on pollen viability, we examined the effect of exogenous CO on in vitro pollen germination and pollen tube growth by adding it to the culture medium. CO had a dose-dependent effect on pear pollen germination inhibition (Fig. 7a). The addition of 100 nM CO to the culture medium reduced tube length by 68% ± 11% by 6 h post-treatment (Fig. 7b). A low concentration of 100 μM TEA⁺ decreased the CO-evoked current (Fig. S2a,b). In pollen tube growth testing, 100 μM TEA⁺ partially reversed the pollen tube growth inhibition by CO (Fig. 7b); but when the concentration of TEA⁺ was increased to 10 mM in the 100 nM CO treatment, there was greater inhibition of pollen tube growth (Fig. S2c).

Figure 7.  Effect of carbon monoxide (CO) on pollen germination and pollen tube growth in pear (Pyrus pyrifolia cv Hosui). (a) Effect of CO on pear pollen germination. The values are the means of six independent experiments, each consisting of ≥ 50 determinations. The error bar represents the SEM. (b) Pollen tube length at 6 h post-culture. A CO concentration of 100 nM inhibited pollen tube growth in vitro. 100 μM tetraethylammonium (TEA⁺) partially reversed the pollen tube growth inhibition by CO (P < 0.01, Student’s t-test). The values are the means of six independent experiments, each consisting of ≥ 50 determinations. The error bar represents the SEM.

Discussion

In the present study, we identified a Ca²⁺-activated outward K⁺ channel in the pear pollen tube and showed that heme directly decreased the probability of K⁺_out channel opening, and that the reaction could be reversed by an increased intracellular Ca²⁺ concentration. However, heme plus NADPH treatment increased the K⁺_out channel activity. With SnPP treatment, the inhibition of the K⁺_out channel by heme occurred even in the presence of NADPH. CO, a product of heme catabolism by hemeoxygenase, directly activated the K⁺_out channel in a dose-dependent manner. Although both Ca²⁺ and CO were able to activate the pear pollen tube K⁺_out channel, whether CO and Ca²⁺ act at the same site in the channel remains to be determined.

Outward K⁺ currents have been identified in many plant tissues (Iijima & Hagiwara, 1987; Schroeder et al., 1987; Moran et al., 1988; Spalding et al., 1992; Van Duijn et al., 1993; Wegner & Raschke, 1994; Demidchik et al., 2010), including pollen and pollen tubes (Fan et al., 2003; Griessner & Obermeyer, 2003). Two plasma membrane outward rectifier K⁺ channels, stelar K⁺ outward rectifier (SKOR) and guard cell outward rectifier K⁺ (GORK), have been studied at the molecular level in plants. SKOR is localized primarily to xylem parenchyma cells and is thought to enable the efflux of K⁺ from these cells into the xylem for transport to the shoot (Gaymard et al., 1998). GORK is strongly expressed in guard cells and mediates K⁺ loss to drive stomatal closure (Ache et al., 2000; Hosy et al., 2003). It is also expressed in other tissues, including the root epidermis (Ivashikina et al., 2001), activated by hydroxyl radicals in stress-induced programmed cell death (Demidchik et al., 2010). Similarly to SKOR and GORK, the opening of the pear K⁺_out channel requires membrane depolarization. In addition, the pear pollen tube plasma membrane K⁺_out channel was activated by Ca²⁺ (Fig. 2), similar to the A. thaliana outward K⁺ channel KCO1 (Czempinski et al., 1997). However, whether the pear pollen tube K⁺_out channel is structurally related to these outward K⁺ channels was not determined in the present study.

An intriguing feature of the pear K⁺_out channel is that it is reciprocally regulated by heme and its product, CO, which has the same effect as heme/CO on the Ca²⁺-activated K⁺ channel in animals (Williams et al., 2004). The K⁺_out channel was inhibited by application of micromolar concentrations of heme to the cytoplasmic surface of the membrane (Fig. 3). Levels of free heme in the cell are controlled by a fine balance of biosynthesis and catabolism and are likely to be below micromolar concentrations (Kim & Doré, 2005). Previous studies on the usage of heme in animal systems employed concentrations of 1–200 nM heme in the cytoplasmic medium to investigate the interaction of heme with ion channels (Tang et al., 2003; Williams et al., 2004; Jaggar et al., 2005). The IC₅₀ of 100–200 nM heme reported here is consistent with these values and comparable with the dissociation constant values of other hemoproteins and heme-binding proteins (Taketani et al., 1998; Hirotsu et al., 1999; Ponka, 1999; Tang et al., 2003). Thus, inhibition of K⁺_out activity by heme will occur under conditions similar to those for other heme–protein interactions. In addition, other proteins (such as hemeoxygenase), which may be spatially close to the K⁺_out channel, can provide a platform for heme binding and modification of K⁺_out channel gating.

When heme was co-applied with NADPH, the effect on NP_O depended on hemeoxygenase activity. K⁺_out channel activity was stimulated by hemeoxygenase under physiological conditions, but was hampered when hemeoxygenase activity was inhibited by SnPP (Fig. 4). In the presence of NADPH and oxygen, heme could be catalyzed by hemeoxygenase to produce biliverdin, iron and CO. Application of CO directly stimulated K⁺_out channel activity, mimicking the effects of heme when co-applied with NADPH under physiological conditions (Figs 4, 5). It has been proposed that the heme/CO system regulates the activity of the Ca²⁺-activated K⁺ channel through hemeoxygenase-2 in animals (Williams et al., 2004). We conclude that the heme/CO system also plays an important role in regulating K⁺ flux in pear pollen.

Another intriguing feature is that the heme-sensitive Ca²⁺-activated K⁺ channel in animals possesses a conserved heme-binding sequence motif CXXCH (where X is any amino acid) (Tang et al., 2003). We do not know whether this motif is present in any subunit of K⁺_out in pear pollen; but in A. thaliana, the GORK channels have a similar amino acid sequence of CELCH (amino acids 472–476). We still do not know whether the GORK channel has a similar regulation mechanism to the K⁺_out channel in the pear pollen tube. In future work, we intend to clone the pear pollen K⁺_out channel gene and investigate the regulatory effects of the heme/CO system on the A. thaliana GORK channel.

CO decreased the pollen germination rate and pollen tube growth by promoting pollen tube K⁺ outflow, and low concentrations of the inhibitor TEA⁺ partially reversed the inhibition of pollen tube growth by CO (Fig. 7). However, when the concentration of TEA⁺ increased to 10 mM, there was greater inhibition of pollen tube growth (Fig. S2c), because TEA⁺ is a K⁺ channel inhibitor, hampering both K⁺ inward and outward fluxes, and inhibition of K⁺ influx also restrains pollen tube growth (Fan et al., 2001).

In addition to control of pollen tube growth, the regulation of the K⁺_out channel by heme/CO may have a potential functional role in later stages of plant sexual reproduction. The last phase of pollen tube growth in the ovary is tube rupture and release of the tube contents. Previous studies have shown that the plasma membrane ion channel/pump is involved in pollen tube burst in the ovary. Pollen tubes of the plasma membrane Ca²⁺ pump ACA9 (autoinhibited Ca²⁺ ATPases 9) mutant aca9 fail to burst and discharge their contents into the ovary (Schiott et al., 2004) and inward K⁺ channel Zea mays 1 (KZM1) activation has been found to depolarize the maize pollen tube plasma membrane, inducing pollen tube burst (Amien et al., 2010). Pollen tubes consume tremendous amounts of energy, requiring rapid oxygen uptake during elongation (Tadege & Kuhlemeier, 1997). In the ripe style of Hippeastrum hybridum, oxygen pressure is high in the stigma and the upper part of the style, but suddenly decreases at the base of the style c. 5 mm from the ovary, approaching zero in the ovary (Linskens & Schrauwen, 1966; Blasiak et al., 2001). In the absence of oxygen, the catabolism of heme to CO will be hampered, and then the opening of the pear pollen tube K⁺_out channel will be inhibited, inducing pollen tube depolarization and an increase in turgor, and ultimately bursting the pollen tube. Therefore, decreased oxygen in the ovary may be a reason for pollen tube tip burst. Our results can also provide an alternative explanation of the failure of the aca9 mutant to achieve fertilization in the ovary (Schiott et al., 2004). The A. thaliana genome encodes 14 Ca²⁺ pumps; ACA9 is expressed primarily in pollen and localized to the plasma membrane, and mutant pollen fails to discharge sperm, despite having reached the embryo sac (Schiott et al., 2004). We speculate that in the aca9 mutant pollen, intracellular Ca²⁺ is not pumped to the outside. Thus, high concentrations of cytosolic free Ca²⁺ increased Ca²⁺-sensitive K⁺_out channel activity, even in the presence of heme in hypoxic conditions, inducing the failure of pollen tube rupture.

In summary, we have identified a Ca²⁺-activated outward K⁺ channel in pear pollen tube protoplasts and demonstrated that the K⁺_out channel was reciprocally regulated by heme and its catabolic product CO. These results contribute to the understanding of the potential role of the heme/CO system in controlling pollen tube development.

Acknowledgements

We thank International Science Editing for helping us to improve the manuscript. This work was supported by the National Science and Technology Ministry (2008BAD92B08-3-2), the National Department Public Benefit Research Foundation (2009030044) and the National Natural Science Foundation of China (31071759).