Corresponding author W.-K. Ho: Department of Physiology, Seoul National University College of Medicine, 28 Yonkeun-Dong, Jongno-gu, Seoul 110–799, Korea. Email: firstname.lastname@example.org
We showed in our previous study that in hippocampal CA1 neurons the stimulation of muscarinic receptors inhibited the GIRK current (IGIRK) via a PLC/PKC pathway, whereas group I metabotropic glutamate receptors (mGluR) inhibited IGIRK via a PLA2/arachidonic acid pathway. In this study, we present evidence that receptor-mediated signalling pathways activated by the two Gq-coupled receptors (GqPCRs) converge on the inhibition of GIRK channel–PIP2 interaction. IGIRK was activated in acutely isolated hippocampal CA1 neurons by repetitive application of baclofen, a GABAB receptor agonist, with a 2–3 min interval. When both CCh and DHPG were pretreated before the second IGIRK activation, the magnitude of the second IGIRK was 52.2 ± 2.5% of the first IGIRK, which was not significantly different from the magnitude of inhibition by CCh or DHPG alone. This result shows that the effects of muscarinic receptor and group I mGluR stimulation on IGIRK are not additive but occlusive, suggesting that each pathway may converge to a common mechanism that finally regulates IGIRK. To test the involvement of PIP2 in this mechanism, the effect of CCh and DHPG on IGIRK was tested in cells loaded with exogenous PIP2. The inhibition of IGIRK by CCh or DHPG was almost completely abolished in PIP2-loaded cells. We confirmed that the inhibition of IGIRK by direct application of phorbol ester or arachidonic acid was also completely reversed in PIP2-loaded cells. These results indicate that the decrease in PIP2–channel interactions is the final common mechanism responsible for GqPCR-induced inhibitions of IGIRK mediated by PKC and arachidonic acid.
Phosphatidylinositol-4,5-bisphosphate (PIP2) has recently received much attention as a signalling molecule which regulates the activity of various ion channels and transporters (Hilgemann et al. 2001). It is not yet clear how PIP2 is used by cells to control their activities in normal signalling mechanisms, but two possibilities can be proposed. The first possibility, which has been widely investigated, is that receptor stimulation-induced changes in membrane PIP2 concentration affect the activity of PIP2-sensitive channels. Since PIP2 is a substrate for phospholipase C (PLC), which is activated by Gq-coupled receptor (GqPCR) stimulation, most studies as to how changes in PIP2 might regulate ion channels have focused on investigating whether the effect of GqPCR is mediated via PIP2 depletion. We and others reported that the inhibition of G protein-gated inwardly rectifying K+ (GIRK) channels in cardiac mycocytes by various GqPCRs was well explained by PIP2 depletion occurring during Gq-coupled PLC activation (Cho et al. 2001a, 2005b; Meyer et al. 2001). Muscarinic inhibition of M current (Zhang et al. 2003; Suh et al. 2004; Winks et al. 2005) and Ca2+ current (Gamper et al. 2004) in neurons has also been explained by PIP2 depletion. Although there are many more studies supporting the PIP2 depletion hypothesis, direct evidence demonstrating that PIP2 depletion can indeed occur in response to Gq/PLC activation to the extent that it affects activities of PIP2-sensitive channels is still lacking (Nasuhoglu et al. 2002b; Horowitz et al. 2005). Whether the PIP2 change is global or local is another issue that remains to be solved. These questions involve very complicated issues, including diffusion properties of PIP2 in the membrane (Cho et al. 2005a), localization of lipid kinases and phosphatases and their Ca2+ dependency for activation (Delmas et al. 2002; Gamper et al. 2004).
The other possibility for PIP2 involvement in ion channel regulation is via changes in channel–PIP2 interaction. This is well known as the activation mechanism of GIRK channels by Gi/o-coupled receptor stimulation, in that binding of Gβγ subunits released from Gi/o protein to GIRK channels stabilizes channel–PIP2 interaction leading to channel opening (Huang et al. 1998; Sui et al. 1998). However, whether a decrease in channel–PIP2 interaction is involved in GqPCR-mediated inhibition of PIP2-sensitive channels has not been well studied.
In our previous study, we showed in acutely isolated hippocampal CA1 neurons that stimulation of muscarinic receptors inhibited the GIRK current (IGIRK) via the PLC/protein kinase C (PKC) pathway, whereas group I metabotropic glutamate receptors (mGluR) inhibited IGIRK via the phospholipase A2 (PLA2)/arachidonic acid pathway (Sohn et al. 2007). We found in the present study that muscarinic receptor and group I mGluR-mediated inhibition of IGIRK is not additive but occlusive, suggesting that each pathway may converge on a common mechanism that finally regulates IGIRK. We then found that exogenous PIP2 application effectively abolished the IGIRK inhibitions induced not only by receptor stimulation but also by a PKC activator or arachidonic acid. These results suggest that both the muscarinic receptor–PLC–PKC pathway and the mGluR–PLA2–arachidonic acid pathway inhibit IGIRK by a common mechanism, the decrease in channel–PIP2 interaction.
Preparation of isolated hippocampal neurons
Hippocampal CA1 pyramidal neurons were isolated as previously described (Sohn et al. 2007). Protocols were approved by the Animal Care Committee at Seoul National University. Briefly 9- to 12-day-old Sprague–Dawley rats were decapitated under pentobarbital anaesthesia. The brain was quickly removed and submerged in ice-cold artificial cerebrospinal fluid (ACSF, see below) saturated with 95% O2 and 5% CO2. Transverse hippocampal slices (400 μm thick) were prepared using a vibrating microtome (VT1000S, Leica, Germany). After a 30 min recovery period at 32°C, the slices were treated with protease type XIV (1 mg/5 ml, Sigma, USA) for 30–60 min, and subsequently with protease type X (1 mg/5 ml, Sigma) for 10–15 min at 32°C. The slices were allowed to recover during a 1 h incubation period at room temperature. The CA1 region was identified and punched out under a binocular microscope (SZ40, Olympus, Japan), placed in a recording chamber containing normal Tyrode solution (see below) and mechanically dissociated using a Pasteur pipette to release individual neurons. The dissociated neurons were allowed to adhere to the bottom of the recording chamber for 10–20 min. Cells identified as pyramidal neurons typically had a large pyramidal-shaped cell body with a thick apical dendritic stump.
Solutions and drugs
ACSF contained (mm): NaCl 125, NaHCO3 25, KCl 3, NaH2PO4 1.25, CaCl2 2, MgCl2 1, glucose 10, sucrose 5, vitamin C 0.4, bubbled with mixture of 95% O2 and 5% CO2 to a final pH of 7.4. Normal Tyrode solution contained (mm): NaCl 150, KCl 5, CaCl2 2, MgCl2 1, glucose 10, Hepes 10, adjusted to pH 7.4 with Tris-OH. The high K+ normal Tyrode solution contained (mm): NaCl 95, KCl 60, CaCl2 2, MgCl2 1, glucose 10, Hepes 10, adjusted to pH 7.4 with Tris-OH. Nystatin perforation pipette solution contained (mm): KCl 40, potassium methanesulphonate 120, Hepes 10, adjusted to pH 7.3 with KOH. In supplemental material Fig. 1, the conventional whole cell configuration was used to deliver diC8-PIP2 into the cell with pipette solution containing (mm): potassium gluconate 110, KCl 30, Hepes 20, Mg-ATP 4, Na-vitamin C 4, Na-GTP 0.3, EGTA 0.1, adjusted to pH 7.3 with KOH.
Stock solutions of baclofen (Tocris, UK), 3,5-dihydroxyphenylglycine (DHPG, Tocris), carbachol (CCh, Sigma) and linopirdine (Tocris) were made by dissolving in de-ionized water and were stored stored at −20°C. On the day of the experiment one aliquot was thawed and used. Stock solutions of phorbol-12,13-dibutyrate (PDBu, Biomol), 4α-PDBu (Biomol), bisindolylmaleimide I (GF109203X, Sigma) and arachidonic acid (Calbiochem) were made by dissolving in DMSO (Sigma). The concentration of DMSO in the external solution was maintained below 0.1%.
Delivery of PIP2 to Cells
DiC16-PIP2, diC8-PIP2 or diC16-phosphatidylinositol-4-phosphate (PI-4-P) (Echelon biosciences, USA) was dissolved in de-ionized water to make a stock solution (1 μg μl−1) and stored at −20°C. On the day of experiments one aliquot was thawed and used. Phosphoinositide (PI)–histone carrier complex was prepared by mixing 2 μl of 1 μg μl−1 PI stock solution in de-ionized water with 2 μl of 0.5 mm Shuttle PIP™ carrier-1 solution (Molecular probes, USA). The PI-carrier complex was put into the internal solution to make a final PIP2 (or PI-4-P) concentration of 10–30 μm. Cells were loaded with PIP2 via the patch pipette. Fluorescence (7-nitrobenz-2-oxa-1,3-diazole; NBD)-labelled PIP2 was handled identically. In supplemental material Fig. 1, only diC8-PIP2 (30 μm, without histone carrier) was included in the patch pipette for the conventional whole-cell loading of PIP2.
Voltage clamp recordings of membrane currents were performed at room temperature using an EPC-10 amplifier (HEKA Elektronik, Germany), while the cells were superfused with the high K+ external solutions (60 mm[K+]o) by gravity flow. To ensure a rapid solution turnover, the rate of superfusion was maintained at ∼5 ml min−1, which corresponded to 50 bath volumes (100 μl) per minute. The holding potential was −80 mV and the recording was performed in a perforated patch configuration by using nystatin (200 μg ml−1, MP biomedicals, USA) at a sampling rate of 10 Hz filtered at 1 kHz. In experiments which tested the effects of CCh, linopirdine (10 μm) was added to the external solution to exclude the influence of muscarinic inhibition of M current. Data were acquired using an IBM-compatible computer running Pulse software v8.67 (HEKA Elektronik). The patch pipettes were pulled from borosilicate capillaries (Hilgenberg-GmbH, Germany) using a Narishige puller (PC-10, Narishige, Japan). The patch pipettes had a resistance of 3–5 MΩ when filled with the above pipette solutions.
Data were analysed using IgorPro (version 4.1, WaveMetrics, USA) and Origin (version 6.0, OriginLab Corp., USA) software. Statistical data are expressed as means ±s.e.m., where n represents the number of cells studied. The peak IGIRK amplitudes were determined as follows. Baseline currents were subtracted from IGIRK traces and normalized to the peak amplitude of the first IGIRK. The normalized IGIRK traces under each experimental condition were averaged to reduce noise and minimize artifacts. This method produced a representative IGIRK trace, as shown in Fig. 1B, from which we could determine the peak IGIRK amplitudes. The significance of differences between the peaks was evaluated using Student's t test with a confidence level of P < 0.05.
CCh- and DHPG-mediated inhibition of IGIRK is not additive but occlusive
In acutely isolated hippocampal CA1 neurons, we recorded membrane currents using a nystatin-perforated whole-cell patch clamp technique at a holding potential of −80 mV, while cells were perfused with a high K+ bath solution (60 mm[K+]o). When baclofen (100 μm), a GABAB receptor agonist, was applied to the bath solution, IGIRK was activated promptly and underwent variable degrees of desensitization (Fig. 1A). We confirmed that the current recovered almost completely when baclofen was applied repetitively at intervals of 2–3 min (Sohn et al. 2007).
The pretreatment of cells with CCh (Fig. 1Aa) or DHPG (Fig. 1Ab) significantly reduced IGIRK. We showed in our previous study that the inhibition of IGIRK by CCh (muscarinic receptor stimulation) and that by DHPG (group I mGluR stimulation) occurred to a similar extent (the data are plotted again in Fig. 1C), but the mechanism was different (Sohn et al. 2007). The effect of CCh was mediated by the PLC/PKC pathway, while the effect of DHPG was mediated by the PLA2/arachidonic acid pathway. To investigate whether the effects of PKC and arachidonic acid on GIRK channels occur independently or via a common mechanism, we tested whether the effect of CCh and that of DHPG are additive or occlusive. The simultaneous application of CCh and DHPG before the second application of baclofen inhibited IGIRK (Fig. 1Ac), but the degree of inhibition appeared similar to that by CCh (Fig. 1Aa) or by DHPG alone (Fig. 1Ab). To demonstrate the average effect, the current trace was normalized to the peak current amplitude of I1 and the data obtained from seven cells were averaged (see methods), and the normalized average current trace of I1 (open circles) during 40 s (dashed rectangles indicated in Fig. 1Ac) was superimposed with that of I2 (filled circles) in the presence of both CCh and DHPG (Fig. 1B). I2,peak was 52.2 ± 2.5% (n= 7) of I1,peak, which was not significantly different from the results with CCh or DHPG alone (Fig. 1C; P > 0.05). These results suggest that the signalling pathways for IGIRK inhibition by CCh and DHPG may converge on a common mechanism downstream of PKC and arachidonic acid. To test whether PIP2 is involved in IGIRK inhibition by CCh and DHPG as a common mechanism, we tested the effect of CCh and DHPG in cells loaded with exogenously applied PIP2.
High PIP2 mobility enables PIP2 loading via patch pipettes
Cells could be loaded with PIP2 by using two different methods: external application of PIP2 together with carrier and direct application to the cytosol using a patch pipette in the whole cell mode. Since IGIRK could be recorded in a more stable way with the nystatin-perforated patch clamp technique, we preferred using the first method. We found that the best method of PIP2 loading for hippocampal neurons was to apply PIP2 locally using PIP2-containg patch pipettes, as was described in Cho et al. (2005b, 2006). NBD-labelled PIP2 was used to confirm the loading process of PIP2 into the cells, since NBD-labelled probes are fluorescent only when they are incorporated into membranes. A typical example is shown in Fig. 2A. When a patch pipette containing 16 μm of NBD-diC16-PIP2 with histone carrier was attached to the cell membrane (left panel), we could readily observe the bright fluorescence signals from the patch of the membrane and the entire membrane became visible (centre panel) within 1 min. Soon after, the whole cell was labelled (right panel), and fluorescence intensity increased gradually, reaching its maximum within 10 min, suggesting that both lateral diffusion and transmembrane diffusion of PIP2 is rapid in hippocampal neurons. Considering that PIP2 mobility is extremely low in cardiac myocytes (Cho et al. 2005a), this result is particularly interesting since it demonstrates the cell-type specific difference in PIP2 mobility.
Inhibitory effects of CCh or DHPG on IGIRK are reversed by PIP2 loading
During the course of PIP2 loading, we could record current since nystatin was also included in the pipette solutions. Current recordings usually started 10 min after the giga-seal was made. When IGIRK was activated by successive applications of baclofen during PIP2 loading, we could not find any significant difference compared to the results obtained using pipette solutions without PIP2 (Fig. 2B). I2,peak was 89.7 ± 2.0% (n= 8) of I1,peak, not significantly different from that in the control condition (92.3 ± 1.7%; n= 8, Fig. 2D in Sohn et al. 2007). It was also noted that the average amplitudes of I1 were not significantly different between the cells patched with PIP2-free (117.3 ± 22.0 pA, n= 8) and PIP2-containing (113.8 ± 15.8 pA, n= 8) solutions. The fact that additional PIP2 does not potentiate the activation of IGIRK may suggest that PIP2 binding of GIRK channels in the presence of baclofen is already saturated at the normal concentration of PIP2.
Subsequently, we tested if the application of PIP2 could reverse the inhibitory effects of CCh or DHPG on IGIRK. As shown in Fig. 3A, the inhibitory effects of both agonists were almost completely abolished with diC16-PIP2 (16 μm) in the patch pipette. To examine whether this effect is specific to PIP2, or an artifact originating from the loading method using carrier, we tested the effect of loading diC16-PI-4-P with the same carrier, and confirmed that the inhibitory effect of CCh remained unchanged in this condition (I2,CCh,peak= 49.9 ± 1.4% of I1,peak,n= 6). The effect of PIP2 was dependent on the concentration of PIP2 and the recovery was partial at lower concentrations. In the presence of 10 μm diC16-PIP2, the peak amplitude of I2,CCh and that of I2,DHPG were 79.0 ± 1.3% (n= 7) and 74.0 ± 2.1% (n= 7) of I1,peak, respectively (Fig. 3C). The effect of PIP2 was also dependent on the acyl chain length. DiC8-PIP2 was less potent than diC16-PIP2, in that 16 μm diC8-PIP2 did not completely abolish the inhibitory effects of CCh and DHPG on IGIRK (Fig. 3B and C). This result may suggest that PIP2 concentration in the membrane reaches higher levels when pipettes contain PIP2 with a longer acyl chain.
To further confirm the effect of PIP2, we established the conventional whole cell configuration to deliver PIP2 into the cytosol, as used in previous studies (Gamper et al. 2004; Brown et al. 2005; Robbins et al. 2006). Without PIP2 in the patch pipette, I2,CCh and I2,DHPG were 56.4 ± 2.3% (n= 6) and 51.2 ± 1.6% (n= 5) of I1,peak, respectively. In the presence of 10 μm diC8-PIP2 in the pipette solution, CCh-mediated inhibition was not affected (I2,peak= 51.5 ± 1.5% of I1,peak), indicating that the carrier-mediated PIP2 delivery system is more efficient than whole cell PIP2 delivery. Furthermore, diC16-PIP2 is more efficiently loaded to the cell membrane due to its higher partition coefficient (Cho et al. 2006). However, when diC8-PIP2 was increased to 30 μm, the effects of CCh or DHPG were completely blocked. In the presence of 30 μm PIP2, I2,CCh and I2,DHPG were 98.5 ± 2.6% (n= 8) and 98.0 ± 2.5% (n= 7) of I1,peak, respectively (supplemental Fig. 1).
PIP2 reverses IGIRK inhibition by phorbol ester and arachidonic acid
To investigate whether receptor-induced PIP2 depletion is prerequisite for PKC-mediated IGIRK inhibition, we tested if direct activation of PKC by phobol ester inhibits IGIRK. In control, the PKC activator PDBu significantly inhibited IGIRK when applied as a pretreatment (Fig. 4A and B, left), whereas the inactive analogue 4α-PDBu did not inhibit IGIRK (Fig. 4D). The I2,peak in the presence of PDBu was reduced to 54.7 ± 2.0% of I1,peak (n= 7), whereas that in the presence of 4α-PDBu was 90.7 ± 1.7% of I1,peak (n= 4). The effect of PDBu was significantly reduced by GF109203X, a PKC inhibitor (Fig. 4B, right). The result that PDBu-induced inhibition of IGIRK in normal PIP2 concentration occurs to a similar extent with IGIRK inhibition by muscarinic receptor stimulation may suggest that PIP2 depletion is not a prerequisite for M1/M3 receptor/PKC-mediated IGIRK inhibition.
We then tested whether the inhibitory effect of PDBu on IGIRK was reversed by PIP2 loading. As shown in Fig. 4C, the inhibitory effect of PDBu on IGIRK was fully reversed when 30 μm of diC16-PIP2 was included in the pipette solution. The fact that the effect of direct PKC activation, not accompanied by receptor-induced PIP2 depletion, is also reversed by exogenous PIP2 application may suggest that the mechanism of action of PKC is to decrease the PIP2 affinity of the channel.
We repeated similar experiments to test if direct application of arachidonic acid inhibits IGIRK and this effect is reversed by PIP2. When we pretreated the cell with arachidonic acid before the second application of baclofen, IGIRK was reduced dose-dependently. The peak amplitudes of I2 in the presence of 4 μm and 7 μm arachidonic acid (I2,AA) were 78.4 ± 2.1% (n= 4) and 59.4 ± 0.7% (n= 7) of I1,peak, respectively. The inhibitory effect of arachidonic acid was fully reversed in PIP2-loaded cells (Fig. 5A and B), suggesting that arachidonic acid-induced inhibition of IGIRK is also mediated by a decrease in PIP2 affinity of the channel.
We have demonstrated in this study that the inhibition of IGIRK by receptor stimulation and that by direct application of PKC activator or arachidonic acid were reversed completely by exogenous application of PIP2. Together with the results of the previous study showing that wortmannin did not block the recovery of the receptor-mediated channel inhibition (Sohn et al. 2007), we suggest that PIP2 depletion plays a minor role but the decrease in PIP2 affinity serves as the common mechanism involved in IGIRK inhibition by muscarinic receptors and group I mGluR stimulations. This is the first study to show the role of PIP2 as a common mediator in the receptor-specific regulation of GIRK channels by GqPCRs in native neurons. The signalling pathway involved in GIRK channel regulation is illustrated schematically in Fig. 6.
PIP2 depletion versus decrease in PIP2 affinity
The PIP2 depletion hypothesis has been supported by the experiments showing that the receptor-mediated inhibitions of GIRK current are attenuated with a supplement of PIP2 (Cho et al. 2005b). In the present study, we found that CCh- and DHPG-mediated IGIRK inhibitions were reversed by exogenous PIP2 application, but we could not regard this result as evidence that CCh- and DHPG-mediated IGIRK inhibitions were mediated by PIP2 depletion, because the inhibitions were also reversed completely by inhibitors of PKC and PLA2 (Sohn et al. 2007). These results suggest that the involvement of PIP2 occurs downstream of PKC or PLA2, so we hypothesized that PKC or arachidonic acid reduces the affinity for channel–PIP2 interaction. This possibility was supported by the result that IGIRK inhibitions by PDBu and arachidonic acid were also reversed by exogenous PIP2 (Figs 4 and 5). We therefore now think that the effect of exogenous PIP2 on receptor-mediated inhibition of PIP2-sensitive channels should be interpreted carefully: it cannot be regarded as direct evidence that PIP2 depletion is a primary cause of receptor-mediated channel inhibition, and the possibility that the decrease in channel–PIP2 interaction is associated with receptor-mediated signalling mechanisms should also be considered.
Brown et al. (2005) investigated the role of PIP2 depletion and PKC activation in muscarinic inhibition of Kir3.1/3.2 A currents expressed in HEK-293 cells. Because the blocking of PIP2 regeneration with wortmannin inhibited recovery, PIP2 was thought to be hydrolysed by CCh. PIP2 depletion, however, was not considered as a major mechanism of inhibition since the inhibition of PKC almost completely blocked muscarinic inhibition. They interpreted their results as evidence for dual regulation, in that muscarinic inhibition results primarily from PKC-mediated sensitization of the channel to decrease in membrane PIP2 levels, whereas the recovery from inhibition depends primarily on PIP2 regeneration. The result of PKC inhibition is consistent with the result obtained for baclofen-activated GIRK currents in hippocampal neurons (Sohn et al. 2007), suggesting that in both expressed and native GIRK channels, the decrease in membrane PIP2 level is insufficient to cause channel inhibition, but PKC-mediated decrease in channel–PIP2 interaction may play a major role. It may be worthwhile to note that unlike the effect of wortmannin on muscarinic inhibition of GIRK currents expressed in HEK-293 cells, it is not prominent in hippocampal neurons (Sohn et al. 2007). This discrepancy may suggest that regulation mechanisms of PIP2 concentrations involving PIP2 breakdown and PIP2 regeneration differ among different cell types.
In principle, decreases in channel-PIP2 affinity render PIP2-sensitive channels more susceptible to PIP2 depletion, so these two possibilities are not necessarily considered to be mutually exclusive for the mechanism of receptor-mediated channel inhibitions. The present study did not entirely exclude the possibility that PIP2 depletion plays an additional role in receptor-mediated inhibition of GIRK currents in hippocampal neurons. However, it is not easy to evaluate the contribution of each mechanism, since we do not have information about the PIP2 concentration changes caused by CCh and DHPG in hippocampal neurons. Instead, it may be useful to estimate how much concentration drop is needed to cause the half-inhibition of IGIRK by means of PIP2 depletion. It was reported that the EC50 value of PIP2 for GIRK channel activation is 17.7 μm (Jin et al. 2002), and the channel-PIP2 binding is strengthened by more than 10-fold in the presence Gβγ (Huang et al. 1998). Assuming that the EC50 value of PIP2 for baclofen-activated current is 2 μm and that the PIP2 concentration in normal neurons is not different from other cells and is about 10 μm (McLaughlin et al. 2002; Nasuhoglu et al. 2002a), about 80% depletion is required to inhibit baclofen-activated IGIRK by half. If PIP2 concentration is 20 μm under control condition, 90% depletion would be required for the same effect. This situation is very different from that of PIP2-sensitive channels with low PIP2 affinity, such as M channels, in that the EC50 is above the normal PIP2 concentration (87.2 μm in Zhang et al. 2003) and the channel activity can be affected by small changes in PIP2 concentrations. It appears that PIP2 depletion is not an efficient mechanism for controlling PIP2-sensitive channels that have a high affinity for PIP2.
We previously proposed the hypothesis that diffusion properties are important in determining the spatial and temporal profiles of PIP2 depletion in response to Gq/PLC activation (Cho et al. 2005a). Simulation of PIP2 concentration changes in a two-dimensional diffusion model showed that a profound PIP2 depletion restricted to the microdomain adjacent to PLC is expected to occur when PIP2 mobility is low. In contrast, when PIP2 mobility is high, the PIP2 depletion induced by PLC activation is not localized but readily attenuated by diffusion, so that the resulting changes become slower and smaller. We found in the present study that the PIP2 fluorescence signal was detected in the apical dendrite after local application of NBD-labelled PIP2 to the soma using a cell-attached patch pipette (Fig. 2A, arrow), indicating that PIP2 mobility in hippocampal neurons is high. In this condition, PIP2 depletion produced in response to PLC activation is expected to be less profound than that in cardiac myocytes where PIP2 mobility is extremely low (Cho et al. 2005a). Considering that PIP2 depletion plays a major role in receptor-mediated inhibition of GIRK channels in cardiac myocytes, but not in hippocampal neurons, it may be inferred that the difference in PIP2 mobility in different cell types may underlie, at least in part, the differential use of PIP2 when it is involved in receptor-mediated signalling mechanisms.
Mechanisms for modulating ion channel–PIP2 interaction
It is well known that channel–PIP2 interaction is crucial for the activities of inwardly rectifying K+ (Kir) channels, but their affinities for PIP2 vary widely among different type of Kir channels (Huang et al. 1998; Zhang et al. 1999). More recently, it was suggested that differences in their affinity are responsible for differences among Kir channels in their specific regulation by a given modulator (Du et al. 2004). For example, Kir 3.4 or Kir 2.3 channels which have low affinity for PIP2 are inhibited by ACh or by phorbol 12-myristate 13-acetate (PMA), a PKC activator, whereas Kir 2.1 or Kir 3.4 channels in the presence of Gβγ, which have high affinity for PIP2, are little affected. Given that the activity of Kir 3.4 channels in the presence of Gβγ corresponds to ACh-activated IGIRK in native myocytes, the results of Du et al. (2004) are compatible with our previous report showing that ACh-activated IGIRK in native myocytes is not affected by PDBu (Cho et al. 2001b). In contrast, the result in this study showed that baclofen-activated IGIRK was inhibited by PDBu (Fig. 4). This discrepancy may reflect that GIRK channel–PIP2 interactions in the presence of Gi/o-coupled receptor agonists in hippocampal neurons are not as strong as those in cardiac myocytes.
Several studies have suggested the role of PKC in the inhibition of GIRK current (Sharon et al. 1997; Stevens et al. 1999; Hill & Peralta, 2001; Leaney et al. 2001; Brown et al. 2005), but the molecular mechanisms are not yet clear. Mao et al. (2004) were successful in showing that in GIRK channels expressed in Xenopus oocytes, phosphorylation of the serine residues (S185 in Kir3.1 and S191 in Kir3.4) is responsible for channel inhibition by PMA or substance P. Because these serine residues are adjacent to the amino acid residues at the proximal C-terminal that were shown to be responsible for the Kir channel–PIP2 interaction (Huang et al. 1998; Zhang et al. 1999), it could be speculated that the phosphorylation of these sites by PKC decreases the PIP2 affinity. This idea is supported by a recent work by Keselman & Logothetis (2007), and our study suggests that neuronal GIRK channels in native cells may be regulated by the same mechanism.
On the other hand, the molecular mechanism of arachidonic acid action on channel–PIP2 interaction was understood in relation to Nai-dependent activation of GIRK channel. It was shown that endothelin receptor-mediated inhibition of Kir3 channels involved the PLA2/arachidonic acid pathway (Rogalski et al. 1999). Subsequently, a critical aspartate residue (D223 in Kir3.4) which is known to be neutralized by Na+ ion to facilitate PIP2 binding (Ho & Murrell-Lagnado, 1999) was shown to be required for arachidonic acid sensitivity (Rogalski & Chavkin, 2001). From these results together with our results, it could be suggested that arachidonic acid may interfere with the binding of PIP2 to Nai-neutralized aspartate residues and eventually decrease the PIP2 affinity of the GIRK channel.
We showed in the present study that the effect of CCh on baclofen-activated IGIRK in hippocampal neurons and that of DHPG are not additive, suggesting the possibility that the effects of PKC-dependent channel phosphorylation and the action of arachidonic acid on Nai-dependent channel activation occlude each other. Nevertheless, the molecular mechanism of how this kind of occlusion occurs is not yet known. Interestingly, a 3D structural model of Kir channels reveals that the amino acid residues involved in PIP2 binding and those involved in the channel modulations by a diverse group of regulatory molecules are localized with a striking proximity (for reviews see: Logothetis et al. 2007). It can be speculated that the physical proximity between these residues may lead to the occlusion, so that decreased channel–PIP2 interaction by PKC phosphorylation may preclude arachidonic acid action and vice versa. The possibility that occlusion occurs at the level of Gq protein needs also to be considered.
This work was supported by the National Research Laboratory Program (R0A-2004-000-10295-0) funded by the Korean Ministry of Science and Technology. A. Lim was supported by the BK21 Program from the Korean Ministry of Education and Human Resources Development.