Angiotensin II regulates neuronal excitability via phosphatidylinositol 4,5-bisphosphate-dependent modulation of Kv7 (M-type) K⁺ channels

Abstract

Voltage-gated Kv7 (KCNQ) channels underlie important K⁺ currents in many different types of cells, including the neuronal M current, which is thought to be modulated by muscarinic stimulation via depletion of membrane phosphatidylinositol 4,5-bisphosphate (PIP₂). We studied the role of modulation by angiotensin II (angioII) of M current in controlling discharge properties of superior cervical ganglion (SCG) sympathetic neurons and the mechanism of action of angioII on cloned Kv7 channels in a heterologous expression system. In SCG neurons, which endogenously express angioII AT1 receptors, application of angioII for 2 min produced an increase in neuronal excitability and a decrease in spike-frequency adaptation that partially returned to control values after 10 min of angioII exposure. The increase in excitability could be simulated in a computational model by varying only the amount of M current. Using Chinese hamster ovary (CHO) cells expressing cloned Kv7.2 + 7.3 heteromultimers and AT1 receptors studied under perforated patch clamp, angioII induced a strong suppression of the Kv7.2/7.3 current that returned to near baseline within 10 min of stimulation. The suppression was blocked by the phospholipase C inhibitor edelfosine. Under whole-cell clamp, angioII moderately suppressed the Kv7.2/7.3 current whether or not intracellular Ca²⁺ was clamped or Ca²⁺ stores depleted. Co-expression of PI(4)5-kinase in these cells sharply reduced angioII inhibition, but did not augment current amplitudes, whereas co-expression of a PIP₂ 5′-phosphatase sharply reduced current amplitudes, and also blunted the inhibition. The rebound of the current seen in perforated-patch recordings was blocked by the PI4-kinase inhibitor, wortmannin (50 μm), suggesting that PIP₂ re-synthesis is required for current recovery. High-performance liquid chromatographic analysis of anionic phospholipids in CHO cells stably expressing AT1 receptors revealed that PIP₂ and phosphatidylinositol 4-phosphate levels are to be strongly depleted after 2 min of stimulation with angioII, with a partial rebound after 10 min. The results of this study establish how angioII modulates M channels, which in turn affects the integrative properties of SCG neurons.

The M current of neurons is a voltage-gated, non-inactivating K⁺ current that plays a dominant role in regulating neuronal excitability (Marrion, 1997; Gu et al. 2005; Peters et al. 2005). It is named because of its suppression by muscarinic acetylcholine receptor (mAChR) stimulation in sympathetic neurons (Brown & Adams, 1980; Constanti & Brown, 1981). This action is via G_αq/11-coupled M₁ receptors and activation of phospholipase C (PLC) (Haley et al. 1998). In the same neurons, three other types of G_q/11-coupled receptors also suppress the M current, angiotensin II (angioII) AT1 (Constanti & Brown, 1981; Shapiro et al. 1994), bradykinin B2 (Jones et al. 1995) and purinergic P2Y (Akasu et al. 1983; Tokimasa & Akasu, 1990; Filippov et al. 1994) receptors, with several different mechanisms serving to mediate these receptor actions (Delmas & Brown, 2005). Since the identification of the KCNQ(Kv7) gene products as underlying M-type currents, Kv7 channels have been found throughout the peripheral and central nervous system, composed of homo- and heteromeric assembly of the Kv7.2, 7.3 and 7.5 members of this channel family (Wang et al. 1998; Lerche et al. 2000; Cooper et al. 2001; Roche et al. 2002).

In addition to its crucial hormonal action to regulate cardiovascular and renal functions, angioII also acts in the central and peripheral nervous systems as a neuromodulator. Recently, angioII has been shown to augment short-term and long-term synaptic potentiation in sympathetic ganglia cells (Aileru et al. 2004), and a role for angioII in alterations of ganglionic function in hypertensive animal models has been suggested (Yarowsky & Weinreich, 1985; Magee & Schofield, 1994). AngioII was identified early on as a potent modulator of the M current (Constanti & Brown, 1981), and M current suppression in sympathetic neurons has been shown to be mediated by AT1 receptors, G_q/11 and an intracellular second messenger (Shapiro et al. 1994; Haley et al. 1998). The lack of Ca²⁺_i signals provoked by angioII-induced stimulation in those cells has suggested that angioII and muscarinic modulation of M currents are mediated by identical signalling mechanisms (Shapiro et al. 1994).

Growing literature suggests that phosphatidylinositol 4,5-bisphosphate (PIP₂) controls the activity of Kv7 channels, much as for a number of other ion channels and transporters (Suh & Hille, 2005). It is now generally agreed that M channels are sensitive to the abundance of membrane PIP₂, and that activation of PLC, and subsequent PIP₂ hydrolysis, can sufficiently deplete the membrane of PIP₂ to cause the observed depression of M currents by muscarinic stimulation (Suh & Hille, 2002; Ford et al. 2003; Zhang et al. 2003; Suh et al. 2004; Li et al. 2005). However, the primary signal for M current suppression by B2 or P2Y receptor stimulation seems to be not depletion of PIP₂, but rather increases in [Ca²⁺]_i (Cruzblanca et al. 1998; Bofill-Cardona et al. 2000; Delmas et al. 2002), probably in concert with calmodulin (Gamper & Shapiro, 2003; Gamper et al. 2005a).

Given the clear role of M-type channels in the control of neuronal excitability, we probed the effects of modulation of M current by angioII on the firing properties of superior cervical ganglion (SCG) neurons. We found that suppression by angioII of M current produces a robust increase in neuronal excitability through a change in resting potential and spike threshold, and a marked reduction in spike frequency adaptation. To investigate the molecular mechanism of modulation of M current by angioII, Kv7.2/7.3 channels and AT1 receptors were heterologously expressed in Chinese hamster ovary (CHO) cells and in this system we investigated the involvement of PIP₂. Additionally, we used analytical biochemistry to determine whether such stimulation of the cloned receptors can substantially deplete mammalian cells of PIP₂. We conclude that changes in PIP₂ levels are involved in the modulation by angioII of Kv7 channels, and this underlies its regulation of neuronal excitability.

Methods

cDNA constructs

Plasmids encoding human Kv7.2 and rat Kv7.3 (GenBank accession numbers AF110020 and AF091247, respectively) were kindly given to us by David McKinnon (State University of New York, Stony Brook, NY, USA). Kv7.2 and Kv7.3 were subcloned into pcDNA3 (Invitrogen) as previously described (Shapiro et al. 2000). The EGFP-Lyn-PH-PP and EGFP-Akt-PH constructs were kind gifts of Tobias Meyer (Stanford University, Palo Alto, CA, USA). Type 1α PI(4)P5-kinase was kindly given to us by Lutz Pott (Ruhr-University, Bochum, Germany). The CHO cell line stably expressing angiotensin AT1 receptors was kindly given to us by Catherine Monnot (Collège de France, Paris). The human AT1 receptor clone was purchased from the Guthrie Foundation (Sayre, PA, USA).

Cell culture and transfections

CHO cells were used for electrophysiological analysis as recently described (Gamper et al. 2005b). Cells were grown in 100-mm tissue culture dishes (Falcon; Becton Dickinson, Mountain View, CA, USA) in Dulbecco's modified Eagle's medium (DMEM) with 10% heat-inactivated fetal bovine serum and 0.1% penicillin-streptomycin in a humidified incubator at 37°C (5% CO₂) and passaged every 3–4 days. Cells were discarded after ∼30 passages. For transfection, cells were plated onto poly-l-lysine-coated coverslip chips and transfected 24 h later with Polyfect reagent (Qiagen, Hilden, Germany) according to the manufacturer's instructions. For electrophysiological experiments, cells were used 48–96 h after transfection. As a marker for successfully transfected cells, cDNA encoding green fluorescent protein (GFP) was cotransfected together with the cDNA of the genes of interest. We found that > 95% of green-fluorescing cells expressed Kv7 currents in control experiments.

Perforated-patch current clamp

SCG neurons were cultured from juvenile rats as previously described (Bernheim et al. 1991; Gamper et al. 2003). Rats were anaesthetized by inhalation of halothane, until completely non-responsive, and decapitated. The experiments were performed in accordance with the guidelines laid down by the University of Texas Health Science Center Institutional Animal Care and Use Committee. Neurons were studied under current clamp with an EPC-9 amplifier and PULSE software using the perforated-patch recording technique (Rae et al. 1991). A stock solution of 50 mg amphotericin-B/1 ml DMSO was prepared and stored in the dark. For recordings, 50–75 μ l of this stock was dissolved in 1 ml of pipette solution using an ultrasonicator. Electrical recordings were made once the access resistance from the pipette to the cell interior fell to < 15 MΩ, usually 1–5 min after achieving a pipette-to-membrane seal resistance of > 2–3 GΩ. Action potentials were evoked and recorded with an EPC-9 amplifier (HEKA Electronic), at a sampling frequency of 2.5 kHz.

Whole-cell and perforated-patch voltage clamp

Whole-cell patch-clamp experiments were performed as previously described (Gamper et al. 2003). To evaluate the amplitude of macroscopic Kv7.2/7.3 currents, CHO cells were held at 0 mV, and 800-ms hyperpolarizing steps to −60 mV were applied, followed by 1-s pulses back to 0 mV. The amplitude of the current was usually defined as the difference between the holding current at 0 mV and the current at the beginning (after any capacity current had subsided) of the 1-s pulse back to 0 mV. In some cells, a more precise measurement was the XE991- or linopirdine-sensitive current at the holding potential of 0 mV. CHO cells have negligible endogenous macroscopic K⁺ currents under our experimental conditions, and 50 μm XE991 or linopirdine completely blocks the K⁺ current in Kv7-transfected CHO cells, but has no effect on currents in non-transfected cells (Gamper et al. 2005b). All results are reported as means ±s.e.m.

Computer modelling

Simulations were performed using NEURON v5.7 (Hines & Carnevale, 2001) on a Macintosh G4 computer. The passive properties of a single-compartment model (50 μm diameter) were chosen to reflect the input resistance and charging time constant of SCG neurons to negative current steps (specific membrane resistivity, 50 000 Ω cm²; specific capacitance, 1 μF cm⁻²). Four voltage-gated and two Ca²⁺-dependent conductances (available for download through ModelDB at http://senselab.med.yale.edu) were inserted into the model as indicated. A fast voltage-gated Na⁺ conductance (nahh.mod) (Traub et al. 1991) was inserted with a density of 4 mS cm⁻². Threshold was adjusted by modifying the model file, shifting the forward and backward activation rate constants by +20 and +10 mV, respectively. Spike repolarization was achieved with a delayed rectifier K⁺ conductance (kdr.mod) (Migliore et al. 1995) at a density of 0.8 mS cm⁻² (E_K, −80 mV). A high-threshold voltage-gated Ca²⁺ conductance (cal.mod) (Migliore et al. 1995) at a density of 1 mS cm⁻² was inserted with the equilibrium potential for Ca²⁺ determined by the Goldman–Hodgkin–Katz equation. Intracellular calcium dynamics were modelled as a single-shell process (cadifus.mod). In our final model, we also employed a BK-like Ca²⁺-dependent conductance (BK; cagk.mod) at a density of 0.1 mS cm⁻². Finally, the M current model (borgkm.mod) (Borg-Graham, 1991; Migliore et al. 1995), with a half-maximal activation of −40 mV, was added with a maximum conductance density of 0.01 mS cm⁻². It was assumed that the M current was active and contributed to the resting potential of the cell. Therefore, the equilibrium potential of the leak conductance was determined with no M current, giving a resting potential (V_rest) of −50 mV. V_rest was −62 mV when g_K,M was maximum.

Phosphoinositide analysis

CHO cells were plated in 10-cm dishes and grown to approximately 90% confluence. Cells were washed with CO₂–bicarbonate-equilibrated F-12K medium (Gibco) and incubated at 37°C for 10 min, and angioII was applied as described in the Results. The incubation was terminated by removing medium and application of 0.75 ml of an ice-cold 1 : 1 mixture of methanol and 1m HCl. Cells were then removed by scraping and mixed vigorously with 0.38 ml ice-cold chloroform in microfuge tubes. Subsequently, phospholipids were deacylated, and the head groups were separated by anion-exchange high-performance liquid chromatography (HPLC) and quantified by suppressed conductivity as previously described (Nasuhoglu et al. 2002).

Solutions and materials

The external solution used to record Kv7 currents in CHO cells contained (mm): NaCl 160, KCl 2.5, CaCl₂ 2, MgCl₂ 1 and Hepes 10; pH adjusted to 7.4 with NaOH. For CHO cells, the regular pipette solution contained (mm): KCl 160, MgCl₂ 5, Hepes 5, 1,2-bis(2-aminophenoxy)ethane N,N,N′,N′-tetraacetic acid (BAPTA) 0.1, K₂ATP 3 and NaGTP 0.1; pH adjusted to 7.4 with KOH. In many experiments, a Ca²⁺-clamping ‘cocktail’ solution (CaBAPPS) was used which contained (mm): BAPTA 20, CaCl₂ 2, KCl 125, MgCl₂ 5, Hepes 5, K₂ATP 3 and NaGTP 0.1, and 100 μg ml⁻¹ pentosan polysulphate; pH adjusted to 7.4 with KOH. For current-clamp perforated-patch experiments on SCG neurons, the pipette solution contained (mm): potassium acetate 90, KCl 20, Hepes 40, MgCl₂ 3, EGTA 3 and CaCl₂ 1; pH adjusted to 7.4 with NaOH. The bath solution contained (mm): NaCl 140, KCl 2.5, CaCl₂ 2, MgCl₂ 1, Hepes 10 and glucose 10; pH adjusted to 7.4 with NaOH. Reagents were obtained as follows: angiotensin II, XE991, ATP GTP, pentosan polysulphate and edelfosine (Sigma); BAPTA (Molecular Probes); DMEM, fetal bovine serum, penicillin-streptomycin (Gibco).

Results

Stimulation with angiotensin increases the excitability of sympathetic neurons

As M current in sympathetic and hippocampal neurons plays an important role in controlling excitability (Jones et al. 1995; Gu et al. 2005; Peters et al. 2005; Shen et al. 2005), the first stage of the present work was to systematically study the influence of angioII on the somatic excitability of sympathetic neurons (cultured for 2–5 days) from SCG of neonatal rats. Using perforated-patch recording, we evaluated both the active and passive properties of the neurons in response to current steps before and after application of angioII. Bath exposure to angioII had strong, but transient, effects on SCG neurons. Representative voltage recordings from positive and negative current steps are shown in Fig. 1. In control, a 100-pA current injection elicited a short train of action potentials (APs) that terminated soon after the beginning of the current step (Fig. 1A). The duration of such an AP train is thought to be determined by a balance between the inward persistent Na⁺ current and outward M-type and K_Ca K⁺ currents (Chen et al. 2005; Yue et al. 2005; Yue & Yaari, 2006). Application of angioII for 2 min caused an increase in the number of APs elicited by the 100-pA current step, consistent with suppression of M current. The effect was transient, as after 10 min of angioII exposure, the number of elicited APs returned to near control levels. In control, the number of APs elicited by the 100-pA depolarizing current was 2.6 ± 0.6 (n = 20). Application of angioII for 2 min increased the number of such APs to 4.7 ± 0.8 (P < 0.05, n = 12); and after 10 min of angioII exposure, the number of APs was 2.9 ± 0.4 (n = 12). For comparison, we assayed the firing properties of SCG neurons in the presence of XE991, which completely blocks M-type channels (Zaczek et al. 1998). Application of XE991 robustly increased AP firing, as expected from strong blockade of M current, with a 100-pA depolarizing current step eliciting a maintained train of APs (25.0 ± 1.8; n = 15) that lasted throughout the entire duration of current stimulus (Fig. 1B). As M channels are partially open at V_rest, we expected angioII to make V_rest less negative, and this was indeed the case (Fig. 1A and 2A). Before application of angioII, V_rest was −55.2 ± 0.4 mV (n = 20), and after 2 min of angioII application, V_rest was depolarized to −51.8 ± 0.8 mV (P < 0.01, n = 12). After 10 min of angioII application, V_rest partially recovered to −53.2 ± 0.5 mV (P < 0.05, n = 12). Application of XE991 (50 μm) depolarized V_rest to 50.0 ± 0.6 mV (P < 0.01, n = 15) (Fig. 1B and 2A).

Figure 1. AngioII increases the somatic excitability of SCG neurons.

Cultured SCG neurons were studied under perforated-patch current clamp, and voltage responses were recorded from a set of current pulses (inset). A, voltage sweeps are shown before (top), after 2 min (middle) or after 10 min (bottom) bath application of angioII (500 nm). B, voltage sweeps are shown before (top), during bath application of XE991 (10 μm) (middle) and after washout (bottom).

Figure 2. Analysis of effects of angioII on SCG discharge properties.

Bars show summarized measurements for neurons of the resting potential (A), the threshold current for AP generation (B), or the spike-train duration (D) in control, after 2 min or 10 min bath application of angioII, or after application of XE991. C, shows the number of APs evoked during the current step versus current amplitude for neurons in the above cases. E, shows the instantaneous interspike interval versus spike number relation for control, after 2 min application of angioII and in the presence of XE991. F, the data in E are shown normalized, calculated as the difference between the initial and final interspike interval of the spike train divided by the number of spikes.

We quantified the effects of angioII on excitability by examining the relationship between injected current and AP firing. In this analysis, we continued the use of XE991 as a strong positive control for the action of M channel closure. One such measure of excitability is the amplitude of the current required to reach threshold for action-potential generation (AP_thresh). In control, AP_thresh was 28.8 ± 0.4 pA (n = 20). Application of angioII for 2 min decreased AP_thresh to 25.1 ± 2.8 pA (P < 0.05, n = 12), and after 10 min of angioII exposure, AP_thresh somewhat recovered to 26.1 ± 1.7 pA (Fig. 2B). Exposure to XE991 decreased AP_thresh to 21.1 ± 2.3 pA (P < 0.05, n = 15). AngioII significantly increased the number of APs elicited over a range of injected current amplitudes (Fig. 2C). This measure of the effects of angioII on firing also showed the action to be transient, with the values at all injected current amplitudes returning to near control values after 10 min of angioII exposure. As expected, XE991 robustly increased AP generation at all stimulatory current levels.

We further characterized the regulation by angioII of neuronal discharge by assaying effects on spike-frequency adaptation (SFA) – observed as the slowing and/or cessation of a spike train driven by a long depolarizing stimulus. Under control conditions, the spike train duration in response to a 100-pA current step was 420 ± 85 ms (n = 20), and after application of angioII for 2 min, it was increased to 790 ± 105 ms (P < 0.01, n = 12) (Fig. 2D). The effect was again transient, as after a 10-min exposure to angioII, the spike-train duration recovered to 450 ± 80 ms (n = 12), which was not significantly different from controls. As expected if M current strongly influences SFA, application of XE991 dramatically increased the spike-train duration to 1880 ± 100 ms (n = 15, P < 0.001). Another measure of SFA is the relationship between instantaneous interspike interval and spike number. In control, the instantaneous interspike interval rapidly lengthened during the short spike train, but after exposure of angioII for 2 min, the slope of this relationship was shallower, and with full M channel blockade by XE991 application, the instantaneous interspike interval increased only very slowly over the dramatically lengthened spike-train duration (Fig. 2E). We quantified this parameter by approximating the slope of the relation, calculated as the difference between the initial and final interspike interval divided by the number of spikes in the train. This measure of SFA was depressed by angioII to a value 0.5 ± 0.1 (P < 0.05, n = 12) that of control (Fig. 2F). In the presence of XE991, this measure of SFA was reduced to 0.1 ± 0.1 that of control (P < 0.01, n = 15). AngioII had no significant effect on the spike half-width or amplitude, suggesting that the actions of the hormone were not mediated through modulation of voltage-gated conductances associated with the generation of individual spikes (Hille, 2001). For example, in control the spike half-width was 6.8 ± 0.3 ms, and after 2 min of angioII, it was 6.9 ± 0.3 ms. AngioII also had no significant effect on the responses to negative current steps, indicating that it does not affect the input resistance at voltages negative to M current activation. All of the angioII-induced effects on the discharge properties of these neurons are consistent with an increase in somatic excitability caused by angioII-mediated suppression of M current.

To examine whether modulation of M current alone could account for the change in excitability, we constructed a computer model of an SCG neuron to analytically evaluate the predicted effects that incremental inhibition of M channels would have on AP firing and SFA. The passive properties used for the model were chosen to approximate the input resistance and membrane time constant observed experimentally. Parameters for ionic conductances were systematically adjusted to approximate the firing behaviour of SCG neurons under control conditions (100% g_K,M), with angioII-mediated 30% suppression of M current (70% g_K,M), and in the presence of XE991 (complete block assumed; i.e. 0% g_K,M). In the initial version of the model, only three voltage-gated conductances were employed: voltage-gated Na⁺ channels, a delayed rectifier K⁺ conductance and M current. However, in this initial simulation, the firing rate when M current was completely blocked was significantly faster than that observed experimentally and the interspike interval was uniform throughout the train, whereas in our SCG neurons, the firing rate still decreased in the presence of XE991. Insertion of a BK-type Ca²⁺-dependent K⁺ conductance (g_K,Ca), along with a high-threshold voltage-gated Ca²⁺ conductance and intracellular calcium dynamics, more closely approximated the firing pattern at all three g_K,M conductance levels. As illustrated in Fig. 3Aa–Ac, the incremental suppression of the M current conductance in the model qualitatively captured the modulation of SCG firing by angioII and XE991 observed experimentally (compare to Fig. 1). With the inclusion of g_K,Ca, the effect of modulating g_K,M on the spike-train duration in this model was linear (Fig. 3Ba), due to the relatively lower contribution of the M current to the total outward K⁺ current included in the model that included the g_K,Ca conductance. Overall, the effects of M channel modulation on SFA in the model were consistent with our experimental observations. Thus, the magnitude of g_K,M in the model strongly affected the relationship between instantaneous interspike interval and spike number, with an increasingly flat slope as g_K,M was decreased (Fig. 3Bb and Bc). In addition, decreasing g_K,M also incrementally depolarized V_rest, in a manner consistent with the experimental observations. In the simulation, a g_K,M of 100%, 70% or 0% produced a V_rest of −53.5, −52.6 and −50.0 mV, respectively. As expected, varying M current had no significant effect on spike amplitude or half-width in the model. In summary, the simulations correlated well with our experimental data suggesting that the modulation of M current alone can account for the modulation of excitability by angioII.

Figure 3. A computational model of an SCG neuron simulates the effect of M channel closure on firing properties.

The data in Figs 1 and 2 were used to constrain the parameters of a computational model of an SCG neuron (see Methods) that simulated the effects on firing of various degrees of M current closure. Aa–Ac, show simulated voltage responses to the set of current injections shown in the inset in the cases of undiminished M current (control), 30% M current blockade (70% g_K,M) or 100% M current blockade (0% g_K,M). Ba, shows the simulated relationship between spike-train duration and percentage g_K,M values. Bb, shows the simulated curves of interspike interval versus spike number at different percentage g_K,M values. Bc, the slopes of the curves in Bb are normalized, as in Fig. 2F.

Inhibition by angiotensin II of cloned Kv7.2/7.3 channels requires PLC

The next stage of our work was to probe the molecular mechanism of action of angioII stimulation on M-type channels. As a model system to elucidate the intracellular mechanism of angioII regulation of M current, we expressed Kv7.2/7.3 heteromultimers in CHO cells by cotransfection of Kv7.2 and Kv7.3, together with AT1 receptors. Kv7.2/7.3 heteromeric channels underlie most M-type currents in sympathetic ganglia cells (Wang et al. 1998; Roche et al. 2002) and their heterologous expression produces voltage-gated K⁺ currents very similar to native M currents in neurons (Wang et al. 1998; Selyanko et al. 2000; Shapiro et al. 2000). Under perforated-patch recording conditions, transfected CHO cells studied using a classic M current–voltage protocol (Fig. 4A) displayed large Kv7.2/7.3 currents that were completely blocked by the specific M channel blocker XE991 (Zaczek et al. 1998). Bath application of angioII (500 nm) caused a strong suppression of the current that rebounded strongly during 10 min of continuous application of the agonist (Fig. 4A). The rebound of the Kv7.2/7.3 currents observed is consistent with the rebound of angioII-mediated effects on firing properties that we observed in SCG neurons, and with known desensitization of AT1 receptors (Thomas & Qian, 2003). As AT1 receptors couple to G_q/11, which commonly activates PLC, we verified its involvement in the angioII action on Kv7.2/7.3 channels using a PLC inhibitor, edelfosine (Powis et al. 1992). Cells were incubated with edelfosine (10 μm) for 30 min before the start of recording. In such a cell, application of angioII did not suppress the Kv7.2/7.3 current at all (Fig. 4B). In control cells, application of angioII suppressed the Kv7.2/7.3 current by 67.5 ± 9.7% (n = 7), but after incubation with edelfosine, the suppression was only 7.7 ± 4.7% (n = 9, P < 0.001; Fig. 4C). Thus, the action of angioII requires PLC activity.

Figure 4. Suppression of Kv7.2/7.3 currents in CHO cells by angioII requires PLC.

CHO cells were transiently transfected with Kv7.2 + Kv7.3 subunits and AT1 receptors and studied under perforated-patch voltage clamp. Cells were incubated for 30 min with vehicle alone (A) or edelfosine (10 μm) (B) before starting the experiment. Plotted are current amplitudes of currents elicited by the pulse protocol shown in the inset. AngioII (500 nm) or XE991 (50 μm) were bath-applied during the period indicated by the bars, and representative current traces are shown on the right. C, bars show summarized current inhibitions for control cells and for those incubated with edelfosine.

Although muscarinic (Beech et al. 1991) and angiotensin-induced (Shapiro et al. 1994) stimulations do not cause rises of Ca²⁺_i in SCG neurons, stimulation of heterologously expressed G_q/11-coupled M₁ mAChRs induces robust Ca²⁺_i transients in mammalian cells (Shapiro et al. 2000). Likewise, stimulation of AT1 receptors expressed in CHO cells induced large Ca²⁺_i signals, as observed using fura-2 for Ca²⁺ imaging (data not shown), consistent with robust activation of PLC. As the aim of our work in CHO cells was to elucidate the mechanisms of action of angioII in SCG cells, which does not involve Ca²⁺_i signals (Shapiro et al. 1994), we performed two types of experiments to estimate the extent of the suppression of Kv7.2/7.3 current when such rises in [Ca²⁺]_i are prevented. The first type utilized whole-cell recording conditions with a pipette solution containing either a low concentration of BAPTA, or one containing the [Ca²⁺]_i-clamping CaBAPPS solution (Shapiro et al. 2000), containing a high concentration of BAPTA and pentosan polysulphate (PPS), a blocker of inositol 1,4,5-trisphosphate (IP₃) receptors (Tones et al. 1989) that prevents release of Ca²⁺ from intracellular stores. Using the pipette solution containing 0.1 mm BAPTA, angioII (500 nm) caused a slow, but strong suppression of the Kv7.2/7.3 current (Fig. 5A). The suppression was somewhat less than that seen in perforated-patch mode, and the current did not show an obvious rebound during angioII exposure. Using CaBAPPS as the pipette solution, CHO cells transfected with Kv7.2 + 7.3 channels and AT1 receptors responded to angioII with a substantial suppression of the Kv7.2/7.3 current that was slightly, but not significantly, lower than that with 0.1 mm BAPTA in the pipette (Fig. 5A and B). As [Ca²⁺]_i is clamped at ∼50 nm, and cannot change, the latter inhibition must be due to a mechanism that does not involve Ca²⁺_i signals. The second experiment used thapsigargin, a blocker of sarco(endo)plasmic reticulum Ca²⁺-ATPase (Thastrup et al. 1990), in order to deplete Ca²⁺_i stores and thus prevent Ca²⁺_i signals (Tse et al. 1994; Cruzblanca et al. 1998; Shapiro et al. 2000). These experiments were performed on CHO cells stably expressing AT1 receptors and transiently transfected with Kv7.2 + 7.3 channels, studied under perforated-patch recording. After pre-application of thapsigargin (2 μm) for 3 min, exposure of the cells to angioII (in the continued presence of thapsigargin) still significantly suppressed the Kv7.2/7.3 current (Fig. 5C).

Figure 5. Stimulation of angioII AT1 receptors suppresses Kv7.2/7.3 channels independently of [Ca²⁺]_i.

CHO cells were transiently transfected with Kv7.2 + 7.3 subunits and AT1 receptors (A and B), or CHO cells stably expressing AT1 receptors were transiently transfected with Kv7.2 + 7.3 subunits (C). The whole-cell configuration was used with pipettes containing either the normal (low BAPTA) (A) or ‘cocktail’ (see Methods) (B) pipette solution, or the perforated-patch mode was used (C). AngioII (500 nm), XE991 (50 μm) and thapsigargin (2 μm) were bath applied during the periods shown by the bars. Current traces in control, and after application or angioII or XE991 are shown on the right, using the indicated voltage-pulse protocol. D, shows the summarized current amplitudes (left) or percentage inhibitions (right) for these three groups of cells.

The results from all of these experiments are summarized in Fig. 5D. We evaluated both the tonic Kv7.2/7.3 current amplitudes and the current suppression by angioII. Tonic current amplitudes were not significantly different for the cells studied (1) under whole-cell clamp with 0.1 mm BAPTA in the pipette, or (2) with the CaBAPPS solution in the pipette, or (3) those treated with thapsigargin under perforated patch clamp. For these three groups of cells, the suppression of the Kv7.2/7.3 current was similar. For cells studied under whole-cell clamp with the standard, or the CaBAPPS, pipette solutions, or under perforated-patch using thapsigargin, the Kv7.2/7.3 currents were inhibited by 47.4 ± 7.9% (n = 7), 32.0 ± 6.6% (n = 12) and 30.6 ± 8.1% (n = 5), respectively (Fig. 5D). We conclude that with Ca²⁺_i signals blocked by either manoeuvre, AT1 receptor stimulation suppresses the Kv7.2/7.3 current by 30–40%. This is somewhat less than that produced by mAChR stimulation under similar conditions (Shapiro et al. 2000), but still substantial. In the Discussion, we suggest several reasons why the inhibition seen here is different from that seen in Fig. 4.

Over-expression of PI(4)5-kinase strongly reduces modulation by angioII

A number of studies have suggested that G_q/11-mediated muscarinic modulation of M-type channels and voltage-gated Ca²⁺ channels in sympathetic neurons is via depletion of membrane PIP₂ (Suh & Hille, 2005). Given that muscarinic and angiotensin-induced stimulation in such neurons share the lack of induced Ca²⁺_i transients, we performed several tests to determine whether the modulation by angioII of cloned Kv7.2/7.3 channels observed in the absence of Ca²⁺_i signals likewise involves depletion of PIP₂. In all of these experiments, cells were studied under whole-cell clamp using the CaBAPPS pipette solution to prevent Ca²⁺_i signals. PIP₂ is produced in the membrane by the sequential phosphorylation of phosphatidylinositol (PI) by PI4-kinase and PI(4)5-kinase. Over-expression of PI(4)5-kinase in sympathetic neurons or cardiomyocytes has been shown to strongly blunt G_q/11-mediated modulation of endogenous M current or Kir3 currents, respectively, almost certainly by increasing tonic membrane [PIP₂] (Bender et al. 2002; Winks et al. 2005). We found that over-expression of PI(4)5-kinase strongly diminished modulation by angioII of Kv7.2/7.3 currents in CHO cells (Fig. 6A). In all of these types of experiments, we compared the modulation to that shown in Fig. 5 using the Ca²⁺-clamping CaBAPPS pipette solution (i.e. 32.0 ± 6.6%, n = 12; replotted in Fig. 6D). In cells over-expressing PI(4)5-kinase, the modulation was reduced to only 11.2 ± 3.6% (P < 0.001, n = 10). We interpret this effect as resulting from a large increase in tonic membrane abundance of PIP₂ (Bender et al. 2002), such that hydrolysis of PIP₂ by PLC activation does not reduce [PIP₂] sufficiently in PI(4)5-kinase over-expressing cells to cause significant channel inhibition (Li et al. 2005; Winks et al. 2005). As the single-channel open probability (P_o) of Kv7.2/7.3 channels at saturating voltages is in the range 0.13–0.31 (Selyanko et al. 2001; Tatulian & Brown, 2003; Li et al. 2005), an increase in tonic [PIP₂] might result in greater Kv7.2/7.3 current amplitudes if maximal P_o is governed by [PIP₂]. However, this was not the case. For control cells, the current amplitude at 0 mV was 919 ± 189 pA (n = 12), and it was not significantly different in cells over-expressing PI(4)5-kinase (1085 ± 299 pA; n = 10). Both of these results with PI(4)5-kinase over-expression are very similar to those seen for muscarinic modulation of endogenous M current (Winks et al. 2005).

Figure 6. Over-expression of a PIP₂ kinase or phosphatase affects the current amplitudes and modulation by angioII of Kv7.2/7.3 channels.

CHO cells were transiently transfected with Kv7.2 + 7.3 channels and AT1 receptors, together with the PI(4)5-kinase (A), Lyn-PH-PP (B) or Akt-PH (C) constructs. AngioII (500 nm) or XE991 (50 μm) were bath applied during the periods shown by the bars. Current traces in control, and after application or angioII or XE991 are shown on the right, using the indicated voltage-pulse protocol. D, shows the summarized current amplitudes (left) or percentage inhibitions (right) for these three groups of cells. **P < 0.01.

Over-expression of a PIP₂ phosphatase strongly reduces Kv7.2/7.3 current amplitudes and blunts modulation by angioII

We then probed the role of PIP₂ in regulation of Kv7.2/7.3 channels by inversely manipulating [PIP₂]; that is, by over-expressing a construct (Lyn-PH-PP) containing a PIP₂-specific phospholipid 5′-phosphatase (Stolz et al. 1998) that selectively reduces plasma membrane PIP₂ concentrations, but not other intracellular pools of PIP₂ or other plasma membrane phosphatidylinositol phosphates. It is expressed as a fusion protein to a myristoylation/palmitoylation sequence of the tyrosine kinase Lyn that causes plasma membrane localization, and to enhanced GFP (EGFP) to visualize expression (Raucher et al. 2000). If Kv7.2/7.3 channels are regulated by membrane PIP₂ abundance, then one might expect tonic current amplitudes to be reduced in cells over-expressing a PIP₂ phosphatase, and this was indeed the case (Fig. 6B). For cells cotransfected with Kv7.2 + 7.3 subunits and Lyn-PH-PP, whole-cell current amplitudes were reduced to 304 ± 28 pA (P < 0.01, n = 7). Expression of Lyn-PH-PP also reduced the extent of angioII modulation of those reduced currents, for which cells application of angioII (500 nm) inhibited the currents by only 5.3 ± 1.7% (P < 0.01 n = 7; Fig. 6D). This result is similar to that obtained using this same construct as a test for PIP₂ depletion as the mechanism for suppression of G protein-coupled inwardly-rectifying (GIRK) K⁺ and N-type Ca²⁺ currents by G_q/11-coupled thyrotropin-releasing hormone (TRH) and M₁ muscarinic receptors, respectively (Lei et al. 2001; Gamper et al. 2004). The reduction of modulation in the cells over-expressing a PIP₂ phosphatase in those studies, and reported here, suggests a tonic PIP₂ abundance that is greatly reduced in cells over-expressing a PIP₂ phosphatase and that subsequent activation of PLC cannot reduce the PIP₂ abundance much further. The latter could well be due to concurrent receptor stimulation of phosphoinositide kinases or to feedback signals that enhance PIP₂ synthesis when [PIP₂] is sensed to be low (Xu et al. 2003; Gamper et al. 2004; Winks et al. 2005).

As a control, we also used another construct (Akt-PH) that contains the PH domain of the serine/threonine kinase Akt, also fused to EGFP. Akt binds to, and is activated by, PI-3 kinase-generated phosphatidylinositol 3,4-bisphosphate (PI(3,4)P₂) and phosphatidylinositol 3,4,5-trisphosphate (PI(3,4,5)P₃), but not PIP₂ (also known as PI(4,5)P₂) (Franke et al. 1997). Thus, Akt-PH is a useful control for these experiments because it interacts not with PIP₂, but with other doubly or triply phosphorylated phosphoinositides. Fluorescence from the Akt-PH construct was localized diffusely in the cytoplasm, as expected in cells not subjected to strong PI3-kinase stimulation (Venkateswarlu et al. 1998). Akt-PH expression had no effect on either modulation by angioII or on current amplitudes (Fig. 6C). For those cells, inhibition by angioII was 34.4 ± 10.8%, and current amplitudes were 894 ± 234 pA (n = 8; Fig. 6D). These data are consistent with modulation by angioII occurring via depletion of membrane PIP₂, and reinforce the conclusion that Kv7.2/7.3 channels are sensitive to PIP₂ abundance.

Rebound of angioII suppression of Kv7.2/7.3 currents requires PIP₂ resynthesis

The effects of stimulation by angioII on the firing properties of SCG neurons studied under perforated patch clamp showed considerable rebound after 10 min of angioII exposure (Figs 1 and 2), as did the suppression of the Kv7.2/7.3 current in CHO cells studied under perforated patch clamp as shown in Fig. 4A. Our working hypothesis is that both arise from desensitization of the AT1 receptors, and subsequent turn-off of G_q/11 and PLC. If PIP₂ depletion is involved in modulation by angioII of M-type channels, as suggested by the results shown in Figs. 5 and 6, then resynthesis of PIP₂ should be required for rebound of the current seen during long exposures to angioII. To test this hypothesis, we compared such current rebound in the presence and absence of wortmannin, which at micromolar concentrations blocks PI4-kinase (Willars et al. 1998; Xie et al. 1999). As in the experiments shown in Fig. 4, CHO cells were cotransfected with Kv7.2 + 7.3 channels and AT1 receptors and studied under perforated-patch. As before, application of angioII induced a strong suppression of the Kv7.2/7.3 current that rebounded nearly back to baseline during 10 min of AT1 receptor stimulation (Fig. 7A). However, when cells were acutely treated with wortmannin (30 μm) just before and during application of angioII, little rebound of the Kv7.2/7.3 current was observed (Fig. 7B). There was little difference, however, in the extent of initial current suppression. Thus, angioII suppressed the Kv7.2/7.3 currents in control cells by 69.5 ± 7.7% (n = 19) and in cells treated with wortmannin by 62.7 ± 8.6% (n = 9; Fig. 7C; not significantly different at the P = 0.05 level). However, blockade of PI4-kinase strongly prevented rebound of the current during long applications of angioII. Thus, in control cells, current rebound after 10 min of angioII exposure was 73.6 ± 7.4% (n = 19), but in cells treated with wortmannin, the rebound was only 9.2 ± 1.9% (P < 0.001, n = 7; Fig. 7C). The blockade of current rebound by wortmannin provides further evidence that Kv7.2/7.3 channels are sensitive to PIP₂ abundance, and that modulation by angioII involves PIP₂ depletion.

Figure 7. Current rebound from AT1 receptor desensitization requires PIP₂ re-synthesis.

A and B, CHO cells stably expressing AT1 receptors were transiently transfected with Kv7.2 + 7.3 subunits and studied under perforated patch. AngioII (500 nm), XE991 (50 μm) and wortmannin (30 μm) were bath applied during the periods shown by the bars. Current traces at the indicated times are shown on the right, using the voltage-pulse protocol shown above. C, shows the percentage inhibitions (left) or the percentage current rebound after the presumed desensitization of the AT1 receptors (right) for these two groups of cells. ***P < 0.001.

Stimulation of angiotensin AT1 receptors in CHO cells depletes PIP₂

In the case of PIP₂-depletion, a significant reduction in PIP₂ abundance would be required to cause the suppression of the currents that is seen. To ascertain whether such angioII-mediated reduction in PIP₂ abundance actually occurs, we used our CHO cell line stably expressing AT1 receptors to analytically measure changes in phosphoinositide abundance induced by stimulation of these cells with angiotensin. Using HPLC analysis, we assayed the abundance of several different phosphoinositides, including PIP₂, phosphatidylinositol 4-phosphate (PIP), PI and phosphatidic acid (PA) before and after exposure of the cells to angioII (2 μm) for 2 or 10 min. We found that stimulation by angiotensin for 2 min indeed greatly reduced the abundance of PIP₂, as well as PIP (Fig. 8). In accord with the rebound of SCG discharge properties and with the rebound of Kv7.2/7.3 currents in CHO cells studied under perforated patch, there was substantial rebound in the abundance PIP₂ and PIP after 10 min of exposure to angioII. Results are given as percentages of total anionic phospholipids, a pool that includes phosphatidylserine and cardiolipin, which are not directly involved in PLC activity. The percentage of PIP₂ was reduced from 0.72 ± 0.07% in control to 0.15 ± 0.04% after 2 min of angioII exposure (n = 4). Receptor stimulation also caused a modest decline in the percentage of PIP, from 1.00 ± 0.23% before to 0.42 ± 0.04% after 2 min of angioII exposure (n = 4). Unexpectedly, the percentage of PI also displayed a significant decline, falling from 53.3 ± 0.5% before to 49.7 ± 0.5% after 2 min of angioII exposure. As diacylglycerol (DAG), the membrane-bound by-product of PIP₂ hydrolysis, is rapidly converted to PA by DAG kinase, the percentage of PA increased upon receptor stimulation, as expected. The percentage of PA rose from 7.7 ± 0.5% in control to 13.7 ± 0.4% (n = 5) after 2 min of angioII exposure. We found the abundance of PIP₂, and to a lesser extent PIP, slowly rebounded during maintained stimulation with angioII. Thus, the percentage of PIP₂ and the percentage of PIP recovered to 0.40 ± 0.04% (n = 4) and 0.49 ± 0.03% (n = 5) after 10 min of angioII exposure, respectively. The percentage of PI, however, continued to fall during maintained receptor stimulation, falling to 40.3 ± 0.2% (n = 4) after 10 min of angioII exposure. The percentage of PA also continued to rise during the maintained stimulation, rising further to 23.2 ± 0.5% (n = 4) after 10 min of angioII exposure (Fig. 8). On the one hand, the partial recovery of the percentages of PIP₂ and PIP during the 10-min angioII treatment is consistent with the known desensitization of AT1 receptors and with the rebound of SCG discharge properties, and of suppression of Kv7.2/7.3 currents in CHO cells studied under perforated patch, during similar 10-min angioII applications. However, we cannot exclude the possibility that this partial recovery is due to the agonist-mediated stimulation of PIP and PIP₂ synthesis (Xu et al. 2003; Gamper et al. 2004). Indeed, the continued fall in the percentage of PI and the continued rise in the percentage of PA could suggest that PIP₂ hydrolysis continued unabated during the longer exposure to angioII, arguing against receptor desensitization as the reason for the partial recovery of the percentages of PIP₂ and PIP during the 10-min treatment. Another explanation may be PIP₂ synthesis in a membrane domain that the channels do not have access to, such as organelle membrane. Further, more spatially compartmentalized experiments may be required to fully correlate the time course of the abundance of lipid molecules during receptor stimulation with the varying degrees of desensitization seen in neuronal discharge properties, and in M-type channel modulation.

Figure 8. Stimulation of cloned AT1 receptors depletes PIP₂.

CHO cells stably expressing AT1 receptors were stimulated with angioII (2 μm) for 2 or 10 min. Bars show the fractional abundance of anionic phospholipids at the indicated times, expressed as a percentage of the total pool. The left ordinate is for the PIP₂ and PIP data, and the right ordinate is for the PI and PA data.

Discussion

In this work, we investigated the physiological and molecular mechanism of angioII-dependent modulation of excitability by M-type channels. Consistent with an important role of M currents in regulating neuronal excitability, stimulation by angioII greatly increased excitability of SCG neurons and reduced SFA. These effects were simulated in a computational model where only the M current conductance was modulated. We also investigated whether depletion of PIP₂ is the likely mechanism of modulation of Kv7 channels by stimulation of angioII AT1 receptors. Indeed, manipulation of tonic PIP₂ levels by manipulation of PIP₂ kinase or phosphatases in CHO cells heterologously expressing cloned Kv7.2/7.3 channels and AT1 receptors had effects consistent with PIP₂ involvement in the action of angioII. Finally, HPLC analysis confirmed that stimulation of cloned AT1 receptors strongly reduces the fractional PIP₂ (and PIP) abundance, similar to that caused by muscarinic stimulation in the same cells (Horowitz et al. 2005; Li et al. 2005). Taken together, these results were consistent with PIP₂-regulated M channels and with AT1-receptor stimulation causing depletion of membrane PIP₂ levels.

Of the four well-described G_q/11-coupled receptors in SCG neurons that have been shown to modulate M current, suppression by the muscarinic M₁ receptor has received by far the most attention. Such muscarinic suppression has now been established as being primarily due to depletion of membrane PIP₂ levels (reviewed by Delmas et al. (2005)). In SCG neurons, the mechanisms of M₁ receptor- and AT1 receptor-mediated suppression of both M-type K⁺ and N-type Ca²⁺ channels seem to be very similar. The similarities include a similar slow, G_αq/11-mediated inhibition of both channels and the lack of rise in [Ca²⁺]_i induced by stimulation of either receptor (Shapiro et al. 1994; Haley et al. 1998). In contrast, bradykinin B2 and purinergic P₂Y receptors, which cause increases in [Ca²⁺]_i and suppress M current only when IP₃-mediated Ca²⁺_i signals are allowed to occur (Cruzblanca et al. 1998; Bofill-Cardona et al. 2000), probably do not deplete PIP₂, as a result of Ca²⁺_i-mediated stimulation of PIP₂ synthesis (Gamper et al. 2004; Winks et al. 2005). Thus, in our thinking, G_q/11-coupled receptors are divided into two classes which act on channels via distinct mechanisms. Both classes activate PLC, and hydrolyse PIP₂, but our working hypothesis is that the receptors that cause [Ca²⁺]_i rises in neurons act primarily via those Ca²⁺_i signals, not via depletions of PIP₂, whereas the class that does not release Ca²⁺ primarily acts via PIP₂ depletion, a difference that is thought to be due to compartmentalization of signalling proteins into distinct microdomains (Delmas et al. 2005).

The angioII-induced increase in excitability of SCG cells, as characterized by depolarization of V_rest, reduced threshold, increased AP number, and reduced SFA, is in accord with a prominent role by M channels in the control of neuronal excitability (Jones et al. 1995; Cooper et al. 2001; Shen et al. 2005). Indeed, in our computational model, a moderate reduction of g_K.M is predicted to have a profound effect on the firing pattern of SCG neurons. Changes to the electrophysiological properties of SCG resulting from stimulation by angioII all indicate a lowered threshold for AP generation and a reduction in the braking effect of M channels on neuronal discharge. This is consistent with long-standing evidence suggesting a strong excitatory effect of angioII on a variety of neurons, including the classic indicator of SCG outputs to the nictitating membrane of the cat (Buckley, 1972; Reit, 1972). The latter study, observing the contraction of the ipsilateral membrane as an indicator of ganglionic excitation, found angioII-induced stimulation to be strongly excitatory, as was bradykinin-induced stimulation. Much new work also indicates strong control of M-type channels specifically in SFA (Yue & Yaari, 2004; Gu et al. 2005; Peters et al. 2005), which we observed as a dramatic increase in spike-train duration, and a decrease in inter-spike intervals caused by M channel closure.

The functional view of sympathetic neurons has generally been one of a relay circuit, conveying information through a cholinergic synapse towards a peripheral target. Less attention has been paid to their integrative capabilities, in spite of the fact that they have long been one of the major model systems for studying the modulation of membrane excitability through control of M channels. Indeed, the expression of a long-lasting form of synaptic enhancement at cholinergic synapses further argues for an integrative role (reviewed by Alkadhi et al. (2005)). Why would a synapse change strength if its postsynaptic potential always triggers a spike? There is evidence to suggest that synaptic gain (i.e. the increased output firing rate relative to the frequency of synaptic input), can be produced in autonomic ganglia by the presence of subthreshold excitatory input (McLachlan & Meckler, 1989; Karila & Horn, 2000; McLachlan, 2003). Through a combination of experimental work and computational modelling, it has recently been shown that enhancing membrane excitability, such as through the modulation of M channels, augments such synaptic gain (Wheeler et al. 2004).

In our experiments on cloned Kv7.2/7.3 channels expressed in CHO cells, we saw a marked difference in the extent of current rebound during long exposures to angioII. Thus, of 44 CHO cells studied under whole-cell voltage clamp, only 9% displayed any obvious rebound of the suppressed current, whereas of 40 cells studied under perforated patch clamp, 70% displayed this aspect of the Kv7.2/7.3 channel modulation. The rebound of Kv7.2/7.3 current suppression seen in perforated patch clamp experiments is in accord with the similar rebound of altered SCG discharge properties observed similarly in perforated patch mode, and with studies showing robust desensitization of angioII AT1 receptors. Might the whole-cell recording with the CaBAPPS pipette solution have masked a desensitization of AT1 receptor-mediated modulation of M-type channels present in intact cells? In fact, such AT1-receptor desensitization has been described and is mediated via G-protein-receptor kinases and β-arrestins (Olivares-Reyes et al. 2001; Thomas & Qian, 2003), molecules that may dialyse from the cytoplasm during whole-cell recording, explaining the differential current rebound between whole-cell and perforated-patch recording modes. Also implicated in AT1 receptor desensitization are intracellular Ca²⁺ signals, in concert with calmodulin (Sallese et al. 2000). Thus, in the perforated-patch current-clamp recordings of SCG cells, as for the perforated-patch experiments on CHO cells not treated with thapsigargin (in which the intracellular milieu was undisturbed), desensitization of AT1-receptor actions was more pronounced.

We also observed a greater inhibition of the cloned Kv7.2/7.3 channels expressed in CHO cells in which [Ca²⁺]_i was not controlled, probably due to Ca²⁺-dependent stimulation of PLC when [Ca²⁺]_i rises are permitted. Indeed, Horowitz et al. (2005), assaying the Ca²⁺-dependence of PIP₂ hydrolysis by monitoring translocation of the widely used PLCδ-PH PIP₂/IP₃ probe (Nahorski et al. 2003), found the translocation, and thus PIP₂ hydrolysis, to be significantly reduced when [Ca²⁺]_i was clamped with BAPTA, or stores were depleted with thapsigargin, as in the experiments reported herein. Moreover, recent work implicates PIP₂ depletion as the mechanism of desensitization of TrpM4 channels induced by rises in [Ca²⁺]_i that activate PLC, in this case without stimulation of any receptor (Zhang et al. 2005; Nilius et al. 2006). Thus, although we have shown Kv7.2/7.3 channels to be sensitive to [Ca²⁺]_i when partnered with calmodulin (Gamper & Shapiro, 2003), we suggest the increase in inhibition seen in Figs 4 and 7 to be mostly from the acceleration of PLC activity from the likely robust Ca²⁺_i signals in those intact cells.

Overall, the HPLC results are very similar to those recently obtained from stimulation of muscarinic M₁ receptors using the same cells (Horowitz et al. 2005; Li et al. 2005). The decrease in the percentage of PIP₂ caused by angioII-induced stimulation of AT1 receptor-expressing CHO cells in this study was similar in magnitude to that caused by muscarinic stimulation of M₁R-expressing CHO cells in those studies. In addition, the rebounds in the phosphoinositide abundance caused by muscarinic and angiotensin-induced stimulation were similar. Although M₁ receptor-mediated suppression of M current is not known to desensitize, desensitization of muscarinic receptors in general, and M₁ receptors in particular, is well documented (van Koppen & Kaiser, 2003). However, we cannot be sure that there is not some bias towards rebound in our HPLC analysis. What we can conclude from these measurements is that stimulation of both types of receptors can substantially deplete CHO cells of PIP₂, consistent with PIP₂ depletion causing inhibition of these PIP₂-sensitive channels. Furthermore, with Ca²⁺_i signals prevented, muscarinic stimulation caused 70–80% inhibition of expressed Kv7.2/7.3 channels (Shapiro et al. 2000; Suh & Hille, 2002; Gamper et al. 2003), but stimulation by angioII in the present study only caused about half of that. If suppression of the Kv7.2/7.3 current is by PIP₂ depletion, and such depletions are similar, then one would expect a more parallel degree of inhibition. One explanation for the difference could be the clustering of certain receptors into microdomains containing local ‘hotspots’ of PIP₂, as has been suggested for PIP₂ regulation of Kir K⁺ channels (Cho et al. 2005). In that scenario, the local concentration of PIP₂ near channels could change less upon application of angioII, than upon muscarinic stimulation, even though global depletions measured by HPLC would be the same. Further measurements of the mobility of PIP₂ molecules in the membrane, of the homogeneity of lipid molecules, and of the biochemical affinities of phosphoinositides for the channel proteins, will be required to fully address these issues.