Abstract
We have found that phosphorylation of a G-protein-coupled receptor by protein kinase C (PKC) disrupts modulation of ion channels by the receptor. In AtT-20 cells transfected with rat cannabinoid receptor (CB1), the activation of an inwardly rectifying potassium current (Kir current) and depression of P/Q-type calcium channels by cannabinoids were prevented by stimulation of protein kinase C by 100 nm phorbol 12-myristate 13-acetate (PMA). In contrast, activation of Kir current by somatostatin was unaffected, and inhibition of calcium channels was only modestly attenuated. The possibility that PKC acted by phosphorylating CB1 receptors was confirmed by demonstrating that PKC phosphorylated a single serine (S317) of a fusion protein incorporating the third intracellular loop of CB1. Mutating this serine to alanine did not affect the ability of CB1 to modulate currents, but it eliminated disruption by PMA, demonstrating that PKC can disrupt ion channel modulation by receptor phosphorylation.
- cannabinoid
- G-protein-coupled receptor
- calcium channel
- inwardly rectifying potassium channel
- protein kinase C
- phosphorylation
Neurons are continuously exposed to neuromodulators that bind G-protein-coupled receptors (GPCRs) and alter neuronal excitability (Hille, 1994). Many modulators decrease electrical excitability by activating inwardly rectifying potassium currents (Kir currents) and inhibiting voltage-dependent calcium currents (ICa), often via pertussis toxin (PTX)-sensitive G-proteins (North, 1989). Other modulators stimulate phospholipase C, generating diacylglycerol and IP3, which increase intracellular calcium levels and activate protein kinase C (PKC) (Huang and Huang, 1993). Activation of PKC often enhances neurotransmission (Malenka et al., 1986; Capogna et al., 1995). The mechanism and the interaction and balance between neuronal stimulation and inhibition by these two classes of receptors can now be studied directly.
Activation of protein kinase C attenuates the modulation of N- and P/Q-type calcium currents and of Kir currents by G-protein-coupled receptors. For all three channels, the fast modulation is mediated by direct binding of G-protein βγ subunits to the channel itself (Reuveny et al., 1994; Krapivinsky et al., 1995;Herlitze et al., 1996; Ikeda, 1996; De Waard et al., 1997; Zamponi et al., 1997). The attenuation by PKC could therefore involve phosphorylation of the receptor, the G-protein, or the channel. Much work has analyzed attenuation of the modulation of N-type (class B) calcium current by PKC (Golard et al., 1993; Swartz, 1993; Zhu and Ikeda, 1994; Shapiro et al., 1996; Zamponi et al., 1997), which is thought to be attributable to phosphorylation of the linker between domains I and II of the α1 channel subunit (Zamponi et al., 1997). Much less is known about attenuation of modulation of identified P/Q-type (class A) calcium currents. Results from oocyte experiments raise the possibility that effects of PKC on P/Q-type calcium channels are not mediated by phosphorylation of their α1 subunit (Stea et al., 1995). Protein kinase C can also enhance neuronal excitability by inhibiting potassium current Kir currents or their activation (Takano et al., 1995;Velimirovic et al., 1995). One possible explanation for these varied actions of PKC is that it disrupts a signaling step proximal to ion channels.
Cannabinoids produce their behavioral effects as a consequence of binding to a G-protein-coupled receptor, the CB1 cannabinoid receptor (Matsuda et al., 1990; Pertwee, 1993). The abundance of these receptors and the discovery of several endogenous ligands (Devane et al., 1992;Stella et al., 1997) suggests that cannabinoid neuromodulatory systems play important physiological roles (Di Marzo et al., 1994). Indeed, endogenous cannabinoids are released during chronic painful inflammation and also play a role in some types of memory (Terranova et al., 1996; Richardson et al., 1997). Cellular consequences specifically linked to CB1 receptor activation include inhibition of adenylyl cyclase and modulation of ion channels (Pertwee, 1993). It is not known whether PKC activation disrupts cannabinoid modulation of ion currents. In this study we used AtT-20 cells, which express P/Q-type calcium current (Mackie et al., 1995), two species of Kir channel protein, GIRK1 and GIRK2 (J. Redell, B. L. Tempel, and K. Mackie, unpublished results), and modest N-type calcium currents (Mackie et al., 1995). We compared the effects of PKC activation on modulation of P/Q-type calcium channels and of Kir current in cells stably transfected with rat CB1 cannabinoid receptor and endogenously expressing somatostatin receptors. Activation of protein kinase C strongly suppressed modulation of these currents by cannabinoids but only weakly suppressed their modulation by somatostatin. The effect of protein kinase C appears to be a result of phosphorylation of a serine in the third intracellular loop of CB1.
MATERIALS AND METHODS
Cell culture. AtT-20 cells stably transfected with rat CB1 (Mackie et al., 1995) or CB1-S317A were plated on poly-l-lysine-coated coverslips and grown in DMEM, 10% heat-inactivated horse serum, 1:200 penicillin/streptomycin, and 400 μg/ml G418 in a humidified environment with 5% CO2 at 35°C. Cells were passaged with 0.05 mg/ml trypsin in PBS and used within 15 passages after the initial clones were isolated.
Electrophysiological recording. Currents were recorded using the whole-cell voltage-clamp technique (Hamill et al., 1980). Pipettes were pulled from microhematocrit glass (VWR Scientific) and were fire-polished. For recording, a coverslip containing cells was transferred to a 200 μl chamber that was constantly perfused (1–2 ml/min) with the appropriate external solution. Solution reservoirs were selected by means of a series of solenoid valves, and solution changes were accomplished in <30 sec. Voltage protocols were generated, and data were digitized, recorded, and analyzed using BASIC-FASTLAB (Indec Systems, Capitola, CA). Liquid junction potentials were uncorrected.
For measuring potassium currents, the pipette solution contained (in mm): 120 KCl, 10 HEPES, 5 EGTA, 3 MgCl2, 3 Na2ATP, 0.3 GTP, and 0.1 leupeptin, pH 7.2, with KOH, whereas the external solution contained (in mm): 40 KCl, 110 N-methylglucamine, 1 CaCl2, 25 HEPES, and 10 glucose, pH 7.35, with NaOH. Fatty acid-free bovine serum albumin (BSA; 3 μm) was added to decrease adsorption of cannabinoids to surfaces. The Kir current was defined as that component of the current sensitive to 1 mmBa2+ elicited during the final 150 msec of a 250 msec hyperpolarizing pulse to −100 mV from a holding potential of −45 mV. Currents were sampled at 1 kHz. Because the magnitude of the Kir current was dependent on cell size, aggregate current data are presented as current densities normalized to cell capacitance.
For measuring barium currents, the pipette solution contained (in mm): 100 CsCl, 40 HEPES, 10 EGTA, 5 MgCl2, 3 Na2ATP, 0.3 GTP, and 0.1 leupeptin, pH 7.2, with CsOH, whereas the external solution contained (in mm): 140 NaCl, 10 BaCl2, 5 CsCl, 1 MgCl2, 10 HEPES, and 10 glucose, pH 7.3, with NaOH. Tetrodotoxin (200 nm) was added to block voltage-dependent sodium currents, 2 μm nifedipine was added to block L-type calcium currents, and BSA was added to decrease adsorption of the cannabinoids. IBa was measured near the end of a 25 msec depolarizing pulse to 0 mV from a holding potential of −90 mV and was defined as the component of current sensitive to 100 μm CdCl2. Currents were sampled at 4 kHz.
To control for possible variations of response with passage number and to avoid one source of systematic bias, experimental and control measurements were alternated whenever possible, and concurrent controls were always performed. Where appropriate, data are expressed as mean ± SE.
Fusion protein generation, purification, and phosphorylation. All fusion proteins were generated in a similar manner using PCR. As an example, the procedure used to generate the IC3 fusion protein follows. An amplicon incorporating the third intracellular loop of the rat CB1 receptor was generated with the following sense (5′-CG GGATCCTGGAAGGCTCACAGCCAC) and antisense (5′-GC GAATTCGGTTTTGGCCAGGCTAAT) primers and digested with EcoRI and BamHI. After ligation into pGEX-3X (Pharmacia, Piscataway, NJ) and transformation, colonies were screened for expression of the appropriate length fusion protein. Mutations were made by direct substitution in the sense primer (S304A) or by the PCR overlap technique (S317A and S323A) (Ho et al., 1989) and were verified by sequencing. Glutathione S-transferase (GST) fusion proteins were purified from bacterial lysates using glutathione-Sepharose (Pharmacia).
Phosphorylation of purified GST proteins was performed as described previously, with 50 ng of purified rat brain PKC (a mixture of α, β, and γ isoforms). Phosphorylated proteins were separated on 10% polyacrylamide gels and identified by autoradiography. Phosphoamino acid analysis was performed using standard techniques (Mackie et al., 1989).
Materials. Tissue culture reagents were from Life Technologies (Gaithersburg, MD); BSA and staurosporine were from Sigma (St. Louis, MO); leupeptin was from Bachem; somatostatin was from Peninsula Labs; and bisindolylmaleimide I (HCl salt), bisindolylmaleimide V, TTX, PMA, 4α-phorbol, and GTP were from Calbiochem (La Jolla, CA). WIN 55,212-2 was a gift from Sterling Research Group.
RESULTS
Stimulation of PKC attenuates cannabinoid activation of Kir current
Our electrophysiological experiments were designed to examine effects of PKC on receptor-mediated modulation of ion channels. To activate PKC, we exposed cells to a bath solution containing 100 nm PMA at 20–22°C for 10 min. Because the actions of PMA generally do not reverse in tens of minutes, it was not necessary to include PMA in any of the subsequent solutions. Control experiments were done with cells similarly exposed to 100 nm4α-phorbol, an inactive analog. In both cases, cells were patch-clamped within 5 min of ending the PMA or 4α-phorbol incubation.
In cells preincubated with 4α-phorbol, 100 nm WIN 55,212-2 activated Kir current in the usual manner in AtT-20 cells stably expressing rat CB1 cannabinoid receptor (Figs.1A,2) (cf. Mackie et al., 1995). However, in cells treated with PMA, WIN 55,212-2 activated Kir current only minimally (Figs. 1B, 2). Incubation of the cells for 5 min with a 1 μm concentration of the protein kinase inhibitor staurosporine before incubation with PMA prevented the PMA effect, providing further evidence that PMA was activating a protein kinase (Fig. 2). A briefer stimulation of protein kinase C by applying phorbol ester for 2.5 min after establishing the whole-cell recording configuration produced similar but weaker effects. In the cells treated with 4α-phorbol, a 100 nm concentration of the cannabinoid agonist WIN 55,212-2 increased the normalized Kir current by 5.2 ± 0.9 pA/pF (n = 23), whereas in cells treated similarly with PMA, 100 nmWIN 55,212-2 increased the Kir current by only 2.2 ± 0.5 pA/pF (n = 16) (data not shown). PMA applied in this manner had no effect on the Kir current before application of WIN 55,212-2 (data not shown).
Somatostatin also activates a Kir current in AtT-20 cells (Pennefather et al., 1988). More current is activated in these cells than with cannabinoid or muscarinic agonists, despite the greater density of cannabinoid compared with somatostatin receptors (Felder et al., 1995). This observation suggests that the coupling of endogenous somatostatin receptors to Kir current in these cells differs from the coupling of transfected CB1 or endogenous m4 receptors to these channels. Thus it was of interest to investigate the effect of protein kinase C stimulation on activation of Kir current by somatostatin. Surprisingly, preincubation with PMA had little effect on activation of Kir current by 10 nm (Figs. 1D, 2) or by 250 nm somatostatin [10.1 ± 1.6 pA/pF; n= 12 (4α-phorbol); vs 9.3 ± 1.4 pA/pF; n = 10 (PMA); data not shown].
Stimulation of PKC attenuates modulation of P/Q-type calcium currents by cannabinoids
AtT-20 cells also express prominent L-, P/Q-, and R-type high-voltage-activated calcium currents (cf. Mackie et al., 1995). Of these, the P/Q-type current is the major current inhibited by cannabinoids and somatostatin (Mackie et al., 1995). Of note, N-type calcium current, defined as inhibited by 1 μmω-conotoxin GVIA (ω-CgTX GVIA), is a minor (<10%) component of the high-voltage-activated calcium current in these cells (Mackie et al., 1995). Using barium as the charge carrier through calcium channels and nifedipine to block current through L-type calcium channels, we determined whether cannabinoid inhibition of P/Q-type calcium channels was blunted by protein kinase C stimulation. Protein kinase C was stimulated using the same preincubation protocols as in the Kir current experiments. In cells pretreated with 4α-phorbol, 100 nm WIN 55,212-2 inhibitedIBa (Figs.3A,4), as expected from previous studies (Mackie et al., 1995). However, pretreatment with PMA markedly attenuated the inhibition by WIN 55,212-2 (Figs. 3B, 4). Further evidence that the attenuation was mediated by protein kinase C included the findings that it was prevented by previous treatment with 1 μm staurosporine or a 100 nm concentration of the more specific protein kinase C inhibitor bisindolylmaleimide I (Fig. 4). The inactive analog bisindolylmaleimide V did not inhibit the PMA effect (Fig. 4).
We repeated the same protocols with somatostatin as the agonist. Protein kinase C activation decreased inhibition of the barium current by somatostatin less effectively than it decreased inhibition by cannabinoid (Fig. 4). Preincubation with PMA reduced barium current inhibition by 10 nm somatostatin from 45 ± 3% (n = 18) to 28 ± 3% (n = 17) (p < 0.0005) (Figs. 3D, 4). Because N-type calcium current comprises only 10% of the high-voltage-activated calcium current, it is likely that the reduction in modulation by PMA reflects decreased inhibition of both P/Q- and N-type calcium channels. Again, further evidence for involvement of PKC comes from the observation that 1 μm staurosporine and 100 nm bisindolylmaleimide I prevented the PMA effect, whereas bisindolylmaleimide V was ineffective (Fig. 4).
Protein kinase C phosphorylates a fusion protein containing the third intracellular loop of the CB1 receptor
Because attenuation by PKC was much greater for cannabinoid-mediated actions than for somatostatin-mediated ones, we reasoned that CB1 receptors might be directly phosphorylated by PKC. We undertook a biochemical approach to test this possibility by constructing glutathione S-transferase fusion proteins for each of the intracellular domains of the rat CB1 receptor. The fusion proteins for intracellular loops I and III (IC3) could be phosphorylated by purified protein kinase C, whereas fusion proteins for the second intracellular loop and C terminus could not (data not shown). Because the third intracellular loop is important for signaling in many G-protein-coupled receptors, further efforts focused on this loop. Phosphoamino acid analysis demonstrated that the IC3 fusion protein was phosphorylated on serine (data not shown). IC3 contains three serines, so three fusion proteins were constructed; in each a different serine was mutated to alanine (S304A, S317A, S323A). Whereas the mutant S317A protein was a poor substrate for PKC, the S304A and S323A proteins were phosphorylated as effectively as the wild-type IC3 fusion protein (Fig. 5). The low amounts of phosphorylation seen in the S317A mutant suggest that PKC can phosphorylate at least one of the two other serines in IC3, but to a lesser degree. These results suggested that S317 is the best PKC site in the third intracellular loop and that its phosphorylation might be responsible for the PKC-mediated disruption of CB1 signaling seen in the electrophysiological experiments. To test this possibility, we mutated S317 to alanine in the full-length rat CB1 and stably expressed it in AtT-20 cells.
Mutation of serine 317 in rat CB1 abolishes PKC-mediated disruption of CB1 signaling
We first tested the coupling of CB1-S317A to Kircurrent. Figure 6, A andC, shows that WIN 55,212-2 could activate Kircurrent in AtT-20 cells expressing the CB1-S317A mutant as effectively as in cells expressing wild-type CB1 (6.3 ± 0.4 pA/pF;n = 6; vs 6.7 ± 0.7 pA/pF; n = 5). Furthermore, whereas preincubation with 100 nm PMA strongly reduced Kir current activation by WIN 55,212-2 in cells expressing the wild-type CB1 receptor, it had no effect on cells expressing CB1-S317A (1.7 ± 0.2 pA/pF; n = 8; vs 6.3 ± 0.4 pA/pF; n = 10) (Fig.6B,C). Thus PKC appears to disrupt activation of Kir current by the CB1 receptor by phosphorylating the receptor on serine 317.
We next tested the coupling of CB1-S317A to calcium channels. Figure7, A and C, shows that cannabinoid agonist inhibited calcium channels in AtT-20 cells expressing the CB1-S317A mutant to the same extent as in cells expressing wild-type CB1 (30.8 ± 1.5%; n = 5; vs 31.1 ± 1.1%; n = 5). However, whereas preincubation with 100 nm PMA strongly reduced inhibition of the calcium channels by WIN 55,212-2 in cells expressing wild-type CB1, the same treatment had a much smaller effect on cells expressing CB1-S317A (7.2 ± 3.6%; n = 7; vs 24.6 ± 2.3%; n = 11) (Fig. 7B,C). Evidently PKC strongly disrupts the inhibition of calcium channels by the CB1 receptor when it phosphorylates the CB1 receptor on serine 317.
Although most of the attenuation of CB1 inhibition of total barium current was prevented by the S317A mutation, a small effect of PMA remained (Fig. 7C). We hypothesized that this might reflect CB1 modulation of the small number of N-type calcium channels present in these cells (Mackie et al., 1995). PKC attenuation of this kind of modulation is likely to be a consequence of phosphorylation of the α1B channel subunit and thus would not be rescued by mutation of the CB1 receptor. To test this possibility we blocked the N-type current by 1 μm ω-CgTX GVIA. Figure7C shows that in the ω-CgTX GVIA-treated CB1-S317A cells, WIN inhibited the barium current to a similar extent in the 4α-phorbol- and PMA-treated cells. These results suggest that the modest reduction by PMA of barium current inhibition by cannabinoids in the S317A mutant is attributable to action on N-type calcium channels, consistent with the hypothesis that by phosphorylating the linker between domains I and II of α1B (Stea et al., 1995;Zamponi et al., 1997), PKC weakens G-protein βγ subunit binding to the channel.
DISCUSSION
We have discovered a potent physiological consequence of phosphorylating cannabinoid receptors by protein kinase C. Modulation of Kir current and P/Q-type calcium channels by cannabinoids is exquisitely sensitive to inhibition by protein kinase C, and the inhibition is largely attributable to the phosphorylation of the CB1 receptor on a single serine in the third intracellular loop.
This sensitivity provides a mechanism for neuromodulators that activate protein kinase C (such as substance P or glutamate acting at metabotropic glutamate receptors) to attenuate the effects of cannabinoids on neuronal excitability. The CB1 receptor mediates inhibition of adenylyl cyclase (Howlett, 1985) and calcium currents (Mackie and Hille, 1992; Mackie et al., 1995) and activation of potassium currents (Henry and Chavkin, 1995; Mackie et al., 1995). Because most of these actions tend to decrease the excitability of a neuron and synaptic transmission, stimulation of PKC provides a mechanism whereby neurotransmitters coupled with protein kinase C can restore neuronal excitability and synaptic strength when endogenous cannabinoid levels are high. Whereas the CB1 receptor was the subject of these studies, our results probably apply to other G-protein-coupled receptors. It will be interesting to learn how widespread this kind of crosstalk is.
It is well established that modulation of N- and P/Q-type calcium channels by G-protein-coupled receptors can be disrupted by activation of protein kinase C (Golard et al., 1993; Swartz, 1993; Zhu and Ikeda, 1994; Stea et al., 1995; Shapiro et al., 1996). For α1B(N-type) calcium channels, the suppression has been attributed to direct phosphorylation of the channel. PKC phosphorylates the domain I–II linker of the α1B subunit at a site (or sites) possibly involved in binding of G-protein βγ subunits (Stea et al., 1995; Zamponi et al., 1997). This covalent modification of the target channel would be expected to suppress actions of free G-βγ subunits regardless of which agonist and receptor had generated them. However, with P/Q-type channels, we have found that the effects of cannabinoids are very sensitive to PKC activation, whereas those of somatostatin are affected slightly. [Because N-type current is only a minor (<10%) component of the calcium current in these cells, and somatostatin inhibits >40% of the calcium current, most of the current that somatostatin inhibits must be P/Q type.] The receptor-specific nature of actions on P/Q current modulation suggests that activation of PKC does not materially reduce the sensitivity of P/Q channels to G-βγ subunits. This conclusion is in accord with earlier findings that PKC activation does not change the amplitude of currents from expressed α1A subunits, whereas it does change the amplitudes of those from α1B subunits (Stea et al., 1995; Zamponi et al., 1997). When interpreting our results, it must be kept in mind that there are alternatively spliced forms of α1A (e.g.,Sakurai et al., 1995) and that these may be affected differently from the form(s) expressed in AtT-20 cells. The insensitivity of P/Q channels in these cells to a direct action of PKC makes them an effector of choice in future studies to characterize receptor-specific actions of PKC.
Our results also provide insight into the role PKC has in inhibiting activation of Kir currents. They are the first demonstration that phosphorylation of a G-protein-coupled receptor can disrupt its activation of a Kir current. In some studies, activation of Kir currents by G-protein-coupled receptors is found to be disrupted by PKC stimulation, whereas in others it is not. In a nice series of studies in cultured nucleus basalis and locus coeruleus neurons, Yamaguchi et al. (1990), Farkas et al. (1994), andVelimirovic et al. (1995) found that Kir currents are inhibited by substance P and neurotensin. In locus coeruleus neurons, in which a Kir current is activated by somatostatin and Met-enkephalin, activation is prevented by substance P. Whereas the molecular identity of these currents has not been identified (Takano et al., 1996), the actions of substance P (and probably neurotensin) in nucleus basalis neurons are mediated by protein kinase C (Takano et al., 1995). In contrast, activation of Kir current in acutely isolated dorsal raphe neurons by the 5-HT1A receptor is not sensitive to stimulation of PKC (Chen and Penington, 1996), demonstrating heterogeneity in the interaction of PKC with Kir current activation. In AtT-20 cells we have also found receptor-specific effects; CB1-mediated activation of Kir current was sensitive to PKC stimulation, whereas somatostatin activation was not. AtT-20 cells express mRNA for at least five subtypes of somatostatin receptor (Kaupman et al., 1993). Presumably the G-protein-coupled signaling of at least one of the somatostatin receptors expressed in AtT-20 cells is not interrupted by PKC. By RT-PCR, AtT-20 cells express GIRK1 and GIRK2 but not GIRK3 or GIRK4 (CIR) (J. Redell, B. L. Tempel, and K. Mackie, unpublished results). GIRK1 and GIRK2 have multiple consensus sequences for PKC phosphorylation. Nevertheless, activation of Kir currents by cannabinoids in cells expressing the S317A-CB1 receptor and activation by somatostatin were unaffected by stimulation of PKC. Thus any phosphorylation by PKC of these GIRK proteins does not perturb their activation by the G-βγ subunits liberated after CB1 receptor stimulation.
In summary, cannabinoid modulation of P/Q- and N-type calcium and Kir channels is exquisitely sensitive to disruption by protein kinase C stimulation. Based on our results, one might even imagine that activation of phospholipase C-linked receptors could be viewed as a potential “antidote” to psychoactive effects of marijuana. Disruption of CB1-mediated P/Q-type calcium channel inhibition and Kir current activation by PKC is chiefly a consequence of phosphorylation of the third intracellular loop of the CB1 receptor. This is the first demonstration that the phosphorylation of a G-protein-coupled receptor by protein kinase C can disrupt the modulation of ion channels by the receptor. The high sensitivity of the response provides a point of potential crosstalk and integration of diverse signals, such as that between endogenous cannabinoids and neuromodulators that activate phospholipase C-linked receptors. It seems quite likely that modulation of channels by certain other G-protein-coupled receptors will be attenuated by PKC via a similar mechanism.
Footnotes
This work was supported by the W. M. Keck Foundation, an Alexander von Humboldt Stiftung Fellowship, Universidad Nacional Autonoma de Mexico DGAPA and Consejo Nacional de Ciencia y Tecnologia, and National Institutes of Health Grants NS01588, DA08934, DA00286, and NS08174. We thank A. Perdichezzi and B. Murphy for help with phosphorylation assays, D. Anderson and L. Miller for technical help, W. Catterall, H. Cruzblanca, J. Isaacson, E. Kaftan, K.T. Kim, D.-S. Koh, J. Roche, and M. Shapiro for reading this manuscript, Y. Lai and A. Nairn for providing protein kinase C, and Sterling Winthrop for providing WIN 55,212-2.
Correspondence should be addressed to Dr. Ken Mackie, University of Washington, Department of Anesthesiology, Box 356540, Seattle, WA 98195-6540.