Activation of a Metabotropic Glutamate Receptor and Protein Kinase C Reduce the Extent of Inactivation of the K+Channel Kv1.1/Kvβ1.1 via Dephosphorylation of Kv1.1*

Various brain K+ channels, which may normally exist as complexes of α (pore-forming) and β (auxiliary) subunits, were subjected to regulation by metabotropic glutamate receptors. Kv1.1/Kvβ1.1 is a voltage-dependent K+ channel composed of α and β proteins that are widely expressed in the brain. Expression of this channel inXenopus oocytes resulted in a current that had fast inactivating and noninactivating components. Previously we showed that basal and protein kinase A-induced phosphorylation of the α subunit at Ser-446 decreases the fraction of the noninactivating component. In this study we investigated the effect of protein kinase C (PKC) on the channel. We showed that a PKC-activating phorbol ester (phorbol 12-myristate 13-acetate (PMA)) increased the noninactivating fraction via activation of a PKC subtype that was inhibited by staurosporine and bisindolylmaleimide but not by calphostin C. However, it was not a PKC-induced phosphorylation but rather a dephosphorylation that mediated the effect. PMA reduced the basal phosphorylation of Ser-446 significantly in plasma membrane channels and failed to affect the inactivation of channels having an α subunit that was mutated at Ser-446. Also, the activation of coexpressed mGluR1a known to activate phospholipase C mimicked the effect of PMA on the inactivation via induction of dephosphorylation at Ser-446. Thus, this study identified a potential neuronal pathway initiated by activation of metabotropic glutamate receptor 1a coupled to a signaling cascade that possibly utilized PKC to induce dephosphorylation and thereby to decrease the extent of inactivation of a K+ channel.

Various brain K ؉ channels, which may normally exist as complexes of ␣ (pore-forming) and ␤ (auxiliary) subunits, were subjected to regulation by metabotropic glutamate receptors. Kv1.1/Kv␤1.1 is a voltage-dependent K ؉ channel composed of ␣ and ␤ proteins that are widely expressed in the brain. Expression of this channel in Xenopus oocytes resulted in a current that had fast inactivating and noninactivating components. Previously we showed that basal and protein kinase A-induced phosphorylation of the ␣ subunit at Ser-446 decreases the fraction of the noninactivating component. In this study we investigated the effect of protein kinase C (PKC) on the channel. We showed that a PKC-activating phorbol ester (phorbol 12-myristate 13-acetate (PMA)) increased the noninactivating fraction via activation of a PKC subtype that was inhibited by staurosporine and bisindolylmaleimide but not by calphostin C. However, it was not a PKC-induced phosphorylation but rather a dephosphorylation that mediated the effect. PMA reduced the basal phosphorylation of Ser-446 significantly in plasma membrane channels and failed to affect the inactivation of channels having an ␣ subunit that was mutated at Ser-446. Also, the activation of coexpressed mGluR1a known to activate phospholipase C mimicked the effect of PMA on the inactivation via induction of dephosphorylation at Ser-446. Thus, this study identified a potential neuronal pathway initiated by activation of metabotropic glutamate receptor 1a coupled to a signaling cascade that possibly utilized PKC to induce dephosphorylation and thereby to decrease the extent of inactivation of a K ؉ channel.
It is now well accepted that modulation of activity of voltagegated K ϩ channel by protein kinases and phosphatases can regulate neuronal excitability (for recent reviews see Refs. 1 and 2). Previously, we showed that basal activity of an unidentified protein kinase endogenous to Xenopus oocytes as well as activation of protein kinase A (PKA) 1 and PKC modulate the delayed rectifier-type of current through a Kv1.1 (␣ subunit) homomultimeric channel (3)(4)(5). We could correlate the modulation of the channel activity by the unidentified kinase and by PKA with phosphorylation of Ser-446 on the cytoplasmic C terminus of the ␣ subunit. In the case of PKC, however, mutations of the numerous putative phosphorylation sites on the ␣ subunit did not eliminate the modulation by PKC (5).
Mammalian Shaker family homologues such as Kv1.1 may normally exist as heteromultimers of ␣ with ␤ subunits that supply the pore-occluding domain that confers fast inactivation upon coexpression of the two subunits in heterologous systems (6,7). Recently we showed (8) that the extent of fast inactivation of the heteromultimeric Kv1.1/Kv␤1.1 (␣␤) channel expressed in Xenopus oocytes is regulated by the basal and PKAinduced phosphorylations of the ␣ subunits that affect the interaction of the channel with microfilaments. Part of the interaction is probably mediated by a native post-synaptic density-95-like protein of the oocyte that recognizes the C-terminal end of the ␣ subunit (9).
In this work we studied the effect of PKC on the extent of fast inactivation of the Kv1.1/Kv␤1.1 channel. We showed that the PKC effect is opposite that of PKA, and it is mediated by dephosphorylation of Ser-446 that is phosphorylated by PKA.
Glutamate is a major excitatory neurotransmitter in the brain. mGluRs participate in synaptic plasticity, both in long term potentiation and long term depression, as well as in neurotoxicity and neuroprotection (for reviews see Refs. [11][12][13][14][15][16]. mGluRs were shown to inhibit several types of K ϩ currents like the M-type current, the Ca 2ϩ -activated current (IK AHP ), a voltage-dependent K ϩ current (I K,slow ), and resting K ϩ currents. mGluRs were shown to activate K ϩ currents in cerebellar granule cells (17). They belong to the superfamily of G protein-coupled seven transmembrane receptors (14,18) and comprise eight members encoded by distinct genes, mGluR1-mGluR8. Group I of mGluRs comprises mGluR1 (the longer splice variant is mGluR1a) and mGluR5 that activate phospholipase C (probably via coupling to G q class of G proteins) and thereby activate a large endogenous Ca 2ϩ -activated chloride current (19,20) and a PKC subtype, possibly PKC- (21), when expressed in oocytes. In this study we identify a potential physiological pathway initiated by activation of mGluR1a coupled to a signaling cascade that possibly utilizes PKC to modulate the extent of inactivation of Kv1.1/Kv␤1.1 channel.
Oocytes, Drug Treatments, and Electrophysiological Recording-Frogs (Xenopus laevis) were maintained and dissected, and their oocytes were prepared as described (22). Previously (8) we showed that coinjection of Kv1.1 (␣) and Kv␤1.1 (␤) mRNAs in a 1 to 30 ratio is enough to saturate ␣ with ␤ and to render over 90% of the channels in the form of ␣␤ in the plasma membrane. Thus, for biochemical studies oocytes were injected (50 nl/oocyte) with 100 -200 ng/l K v 1.1 and 3-7 g/l K v ␤1.1 mRNAs. For electrophysiological studies a ratio of 1 to 100 was used as done previously (8) to ensure saturation of ␣ with ␤ beyond any doubt; thus, 2-5 ng/l K v 1.1 and 200 -500 ng/l K v ␤1.1 mRNAs were injected. 5 ng/l mGluR1a mRNA were injected for both electrophysiological and biochemical studies. The concentrations of the injected S446A mRNA deviated slightly from the above concentrations as it was adjusted to give current amplitudes similar to wild type. Injected oocytes were incubated at 22°C for 1-4 days in ND96 solution (8) and then assayed either electrophysiologically or biochemically. Oocytes injected with mGluR1a were injected with 20 nl of 50 mM K ϩ -EGTA (pH 7.6) 2-6 h before the electrophysiological experiment; this corresponds to ϳ1 mM EGTA in the oocyte. Staurosporine (3 mM  M final) were added to the standard NDE solution in which the oocytes were incubated for at least 2 h before the experiment; all the stock solutions were kept at Ϫ20°C except for that of cyclosporin A which was prepared daily (the solution of calphostin C was exposed to daylight for its activation). Control solutions always included the same concentration of the vehicle as in the corresponding solutions containing the drug of choice. 10 nM ␤-phorbol 12-myristate 13-acetate (␤-PMA) and ␣-PMA were applied by constant bath perfusion after stability of the current had been verified for at least 5 min. 100 M glutamate was bath perfused for 1 min and then washed away.
Two-electrode voltage clamp recordings were performed as described (4). To avoid possible errors introduced by series resistance, only current amplitudes up to 4 A were recorded, and in a given experiment the amplitudes of WT and mutant currents were similar. Also, the currents in mGluR1a-injected oocytes had similar amplitudes (achieved by injection of larger amounts of channel mRNAs). Currents were elicited by stepping the membrane potential from a holding potential of Ϫ80 mV to ϩ50 mV for 160 ms. Current-voltage relationships were obtained by 160-ms depolarizing steps from Ϫ80 mV to the indicated voltages. Net current was obtained by subtraction of the scaled leak current elicited by a voltage step from Ϫ80 to Ϫ90 mV. Oocytes having a leak current of more than 3 nA/1 mV were discarded. Intervals of 20 s between each trace allowed for recovery from inactivation. The time course of the decay phase of the current was fitted with a sum of two exponential components. The slow component had a time constant of several tens of milliseconds, thus representing inactivation distinct from the fast inactivation of the fast component. The time constant of the fast component ( inact ) is referred to below.

Metabolic Labeling with [ 35 S]Methionine/Cysteine ([ 35 S]Met/Cys) in
Vivo-This was done essentially as described (4). Following injection of mRNA(s), six to eight oocytes were incubated at 22°C for 2-4 days in NDE containing 0.2 mCi/ml [ 35 S]Met/Cys. Before homogenization oocytes were incubated with either 100 or 10 nM PMA for 25 min in a dark environment or with 100 M glutamate for 1 min, as required by the experiment. Homogenization was done in 150 -300 l of medium consisting of 20 mM Tris (pH 7.4), 5 mM EDTA, 5 mM EGTA, 100 mM NaCl, 50 g/ml phenylmethylsulfonyl fluoride, 1 mM iodoacetamide, 1 M pepstatin, 1 mM 1,10-phenanthroline supplemented with protein phosphatase inhibitors as follows: 50 nM okadaic acid, 0.5 mM vanadate, and 50 mM KF. Yolk was removed by centrifugation at 1000 ϫ g for 10 min at 4°C. After addition of Triton X-100 to a final concentration of 4%, followed by centrifugation for 15 min at 4°C, antiserum was added to the supernatant for 16 h. After 1 h incubation with protein A-Sepharose, immunoprecipitates were washed four times with immunowash buffer (150 mM NaCl, 6 mM EDTA, 50 mM Tris (pH 7.5), 0.1% Triton X-100); the final wash contained no Triton. Samples were boiled in SDS-gel loading buffer and electrophoresed on 8% polyacrylamide-SDS gel (SDS-PAGE).
Plasma membranes were separated mechanically, as described (3). Defolliculated oocytes were placed for 10 min in ice-cold hypotonic solution (5 mM NaCl, 5 mM Hepes, 1 mM phenylmethylsulfonyl fluoride, 1 M pepstatin, and 1 mM 1,10-phenanthroline) supplemented with phosphatase inhibitors, as specified above. Plasma membranes to-gether with the vitelline membranes (extracellular collagen-like matrix) were removed manually with watchmaker's forceps. Ten internal fractions or 20 -30 plasma membranes were processed as described for whole oocytes.
PKC Stimulation in Oocyte Homogenates-This was done essentially as described (23). Yolk-free homogenates of 10 oocytes that were injected with mRNA(s) 2-3 days before homogenization were incubated in 20 mM Tris-HCl (pH 7.4), 5 mM MgCl 2 , 5 mM NaH 2 PO 4 , 1 mM EDTA, 1 mM dithiothreitol supplemented with protease inhibitors (as above) in a final volume of 100 l at 25°C for 40 min in the presence of 100 nM PMA, 1.5 mM CaCl 2 with either 3 mM ATP or 50 M [␥-32 P]ATP (3000 Ci/mmol). Phosphatase inhibitors (as above) were added, and the following steps were as described above for [ 35 S]Met/Cys labeling.
Quantification of Labeling Intensities and Generation of Digitized PhosphorImager Scans-Gels were dried and placed in a PhosphorImager (Molecular Dynamics) cassette for about 1 day. Using the software ImageQuant, a digitized scan was derived, and relative intensities of protein bands were estimated quantitatively by the software Image-Quant as described (4).
Statistical Analysis-Data are presented as means Ϯ S.E.; n denotes the number of oocytes assayed. Student's t test was used to calculate the statistical significance of differences between two populations. (5) we showed that oocytes injected with Kv1.1 (␣) RNA and assayed by the two-electrode voltage clamp technique express a delayed rectifier type of K ϩ current (␣ current) that is modulated by PMA. Incubation with 10 nM PMA causes gradual reduction of the ␣ current amplitude (Ref. 5; e.g. Fig. 1D) that is mediated by activation of the oocyte's PKC.

PMA Decreases the Extent of Inactivation of the Kv1.1/Kv␤1.1 Current-Previously
When Kv1.1 is coexpressed with Kv␤1.1 (␤) a heteromultimeric ␣␤ channel is formed that expresses a rapidly inactivating current that has a fast inactivating component (I i ) and a noninactivating sustained current component (I s ) remaining at the end of a depolarizing pulse to ϩ50 mV (Ref. 8; see illustrations in Fig. 1A). The apparent extent of inactivation (inactivating fraction) is defined as I i /I p (I p ϭ peak current), and the sustained fraction (I s fraction) is defined as I s /I p (8). In this study we investigated the effect of PMA on the ␣␤ current. Fig. 1 (A and C) shows that 10 nM PMA reduced the peak ␣␤ current amplitude; the time course and the extent of the reduction (reaching about 50% of control after 20 min) were similar to those of ␣ current expressed in oocytes taken from the same frogs. PMA did not affect any endogenous currents in oocytes uninjected with RNA ( Fig. 1A, inset). In addition, PMA caused a gradual decrease in the extent of inactivation (Fig. 1B), which is depicted in Fig. 1E as an increase in the I s fraction determined in currents where I s was larger than 400 nA (after leak subtraction, see "Experimental Procedures") to preclude any interference from endogenous currents activated by PMA. The initial I s fraction was 0.41 Ϯ 0.02 before the addition of PMA and increased to more than 150% of its original value after 20 min. The time constant of inactivation changed slightly from 6.8 Ϯ 0.5 ms before PMA to 8.9 Ϯ 0.4 ms 20 min after the start of PMA (seven oocytes assayed; p Ͼ 0.05). No change in I p or the extent of inactivation could be detected in the absence of PMA (Fig. 1, C and E, respectively) or in the presence of 10 nM ␣-PMA, an isomer that does not activate PKC (not shown). As expected, separate analysis of the effect of PMA on each of the two apparent current components (Fig. 1D) shows that the increase in I s fraction is due to the fact that within 20 min the reduction of I i (ϳ40% of its initial value) was larger than the reduction of I s (ϳ65% of its initial value). Actually, the reduction of I s was preceded by an early increase (ϳ130% of its initial value) that peaked at ϳ10 min when I p was still constant and I i had already started to decrease. Taken together, one may deduce that two overlapping effects are initiated by PMA as follows: 1) as already described for the ␣ current (5) which is a PKC-mediated reduction of the total current amplitude consist-ing of I i and I s , and 2) a novel one, which is an increase in I s that brings about smaller net reduction of I s (caused by the first effect) and thus is manifested in a decrease in the extent of inactivation (increase in I s fraction). The voltage sensitivity of the channel was not altered by PMA, as verified by the overlapping activation curves of I p before and during PMA application (inset of Fig. 1C). PMA was assessed in more than 30 frogs altogether (see also below) and was found to always reduce the current amplitude; however, in oocytes of five frogs it failed to affect the extent of inactivation. These experiments were not taken into account.
The involvement of PKC in the PMA-mediated decrease of extent of inactivation was tested by using a series of protein kinase inhibitors as follows: a broad specificity protein kinase inhibitor staurosporine acting at the protein kinase catalytic site, a specific PKC inhibitor BIS acting at the PKC catalytic site, and a specific PKC inhibitor calphostin C acting at the diacylglycerol-binding site (24). Oocytes were preincubated for 2-4 h in medium containing 3 M staurosporine or 5 M BIS; in addition, BIS was also injected into the oocytes before PMA application. PMA effects with and without the blockers were compared in oocytes from the same frogs because of variable effects by PMA among oocytes from different frogs (Fig. 1F). Staurosporine and BIS significantly suppressed the decrease in inactivation; however, neither a 4-h incubation with 5 M calphostin C nor injection of this blocker into the oocytes suppressed the PMA effect on the extent of inactivation.
PMA Brings about Dephosphorylation of the ␣ Protein-Previously (3,4), we showed that the ␣ protein is extensively phosphorylated in its basal state in the oocyte and that this phosphorylation is manifested as a shift in its migration in SDS-PAGE; the nonphosphorylated form migrates as a 54-kDa protein, whereas the phosphorylated form migrates as a 57-kDa protein. This phosphorylation was totally abolished when Ser-446 was replaced by alanine (S446A mutant). In this study we looked for PKC-induced phosphorylation of the channel proteins first in homogenates of oocytes and then under in vivo conditions. Fig. 2 shows an experiment in which the oocyte's FIG. 1. Modulation of ␣␤ current by PMA. A, two-electrode voltage clamp recordings in one oocyte injected with ␣ and ␤ RNA that was exposed to continuous bath application of 10 nM PMA. ␣␤ currents were elicited by a 160-ms voltage step from Ϫ80 mV to ϩ50 mV, before (control) and 13 and 25 min after the start of PMA application. Inset shows currents elicited by the same voltage protocol as in A in oocytes uninjected with RNA in the absence and presence of 10 nM PMA (two overlapping current traces). I i , I s , and I p illustrate definitions in the text for the inactivating, noninactivating, and total currents, respectively. B, the same currents superimposed (each normalized to its peak). C, time course of the effect of PMA on the peak amplitude (I p ) of ␣ (filled squares) and ␣␤ (filled circles) currents; shown are average normalized values from 9 and 14 oocytes, respectively, taken from the same 6 frogs. Open squares are of ␣␤ currents in the absence of PMA (4 oocytes taken from the same frogs). Inset shows activation curves of I p of ␣␤ current before (control; circles) and 20 min after (ϩPMA; squares) the start of PMA application (at t ϭ 0). The currents were elicited by 160-ms voltage pulses to the indicated voltages, in 10-mV increments, every 20 s (to allow for recovery from inactivation); the elicited currents were normalized to maximal current at ϩ60 mV. D, time course of the effect of PMA on the sustained (I s ; circles) and inactivating (I i ; squares) current components in the same oocytes of C. E, time course of the changes in I s fraction (I s /I p ) of ␣␤ currents in the presence (circles) and in the absence (squares) of PMA (same oocytes as in C). F, the effect of PMA on I s /I p of ␣␤ currents in the presence of protein kinase inhibitors (see "Experimental Procedures"). The effect of each inhibitor was evaluated by comparing the increase in I s /I p after 20 min application of PMA in the presence (shaded bars) and in the absence (white bars) of the inhibitor (see "Experimental Procedures") in oocytes taken from the same frogs. ss, staurosporine; bis, bisindolylmaleimide; cc, calphostin C. Numbers above bars denote number of assayed oocytes taken from 2 frogs. *, p Ͻ 0.03; **, p Ͻ 0.003: significantly different from values before the application of PMA (C, D and E), or from control (F).
PKC was stimulated in homogenates of oocytes expressing ␣ or ␣␤ channels by stimulating with added 100 nM PMA and 1.5 mM CaCl 2 in the presence of added ␥-ATP, as was described by Beguin et al. (23) for the Na,K-ATPase. To our surprise, PKC activation brought about dephosphorylation, rather than phosphorylation, of the ␣ protein both in ␣ and in ␣␤ channels. Dephosphorylation of the basally phosphorylated ␣ protein at Ser-446 was evident from SDS-PAGE analyses of 32 P-labeled or [ 35 S]Met/Cys-labeled channel proteins that were coimmunoprecipitated by a specific anti-␣ antibody. Thus, the 32 P-labeled 57-kDa band that corresponds to the phosphorylated ␣ protein could be precipitated from homogenates incubated with [␥-32 P]ATP (the low intensity band at 57 kDa in control oocytes that do not express the channel may represent an endogenous K ϩ channel; Ref. 3) but could not be detected after the addition of CaCl 2 and PMA (Fig. 2, right panel). Correspondingly, in homogenates of oocytes that were metabolically labeled with [ 35 S]Met/Cys and incubated with cold ATP, the addition of CaCl 2 and PMA induced dephosphorylation of ␣ that was quantified as the reduction in the extent of its phosphorylation. Thus, following addition of CaCl 2 and PMA, the extent of phosphorylation of ␣ (calculated as the ratio of labeling intensities of the 57-kDa over the 54-kDa bands; see Ref. 4) was reduced to 0.33 and 0.4 of that of control (without PMA) in ␣ (compare lanes 5 and 2) and ␣␤ (compare lanes 4 and 1) channels, respectively (Fig. 2, left panel). The ␤ protein was not phosphorylated (Fig. 2, right panel). The total amount of ␣ did not change significantly by the treatment (compare lanes 1 with 4 (27% increase) and lanes 2 with 5 (4% increase)). This experiment shows that phosphorylation of the ␣ protein by an enzyme constitutively active in the oocyte's homogenate was reduced upon stimulation of the oocyte's PKC. Interestingly, a massive dephosphorylation of oocyte's proteins probably occurred upon PKC stimulation, as judged from the extensive reduction of the 32 P labeling of the other proteins nonspecifically pulled into the immune complex (right panel of Fig. 2), whereas their total amount (intensity of their [ 35 S]Met/Cys labeling; left panel of Fig. 2) was not reduced.
We then asked whether PMA induces dephosphorylation in vivo by testing the effect of a 25-min incubation of intact oocytes with PMA. Fig. 3 shows an SDS-PAGE analysis of ␣ and ␤ proteins immunopurified separately from plasma membranes and from internal fractions of oocytes labeled with [ 35 S]Met/Cys (see "Experimental Procedures"). In oocytes treated with PMA (10 or 100 nM; no significant difference between the effects of the two concentrations was detected) the relative intensity of the 57-kDa band decreased, as compared with untreated oocytes, indicating that PMA induced dephosphorylation of the ␣ protein. Dephosphorylation occurred both in the plasma membranes and in the internal fractions. In three similar experiments (including that shown in Fig. 3) the PMA-induced dephosphorylation was quantified as the reduction in the extent of phosphorylation. Thus, in the plasma membranes of oocytes treated with PMA the extent of phosphorylation (ratio of labeling intensities of 57-over 54-kDa bands) was 0.74 Ϯ 0.05 of those of untreated oocytes, and in the internal fractions it was 0.41 Ϯ 0.005. The larger phosphorylation in the plasma membranes versus the internal fractions probably relates to the phenomenon described by us previously that ␣ protein phosphorylated at Ser-446 tends to accumulate into plasma membrane (4). However, in five of nine additional experiments, in which channel proteins were immunopurified from whole oocytes, PMA decreased the extent of phosphorylation to 0.68 Ϯ 0.07 of control. In one experiment we checked if the dephosphorylation could occur in less than 25 min incubation with PMA and found that already within 5 min significant dephosphorylation occurred (not shown). It is noteworthy that PMA did not alter the ␤-binding capacity of ␣, as verified in all the experiments by a quite constant ratio of intensities between the ␣ and the ␤ proteins coprecipitated with the ␣ antibody (Fig. 3).
Dephosphorylation of Ser-446 of the ␣ Protein Underlies the Modulation of Extent of Inactivation by PMA-Next we tested the possibility that dephosphorylation of Ser-446 underlies the decrease in the extent of inactivation by PMA. We performed two experiments in which the modulation by PMA of wild-type (␣ WT ␤) channels was compared with that of the mutant (␣ S446A ␤) channels where Ser-446 in ␣ was replaced by alanine and thereby rendered the channel totally non-phosphorylated in its basal state (4). Although the peak amplitude of ␣ S446A ␤ was decreased by PMA to the same extent as that of ␣ WT ␤ (Fig.  4A), the extent of inactivation of the mutant was unaffected compared with WT where the extent of inactivation decreased (as shown in the increased I s fraction; Fig. 4C). The inactivation of two other phosphorylation-irrelevant serine mutants S489I and S322A (first in the C terminus and second in the loop between S4 and S5 transmembrane segments, respectively; Ref. 5) was not reduced by PMA (Fig. 4D). It is noteworthy that the I s current component of ␣ S446A ␤ did not exhibit biphasic modulation by PMA; rather, PMA caused an apparent decrease of I s that was larger than that of ␣ WT ␤ (Fig. 4B). This substantiates the notion that a PMA-induced increase of I s that overlaps a PMA-induced reduction of total current (including I i and I s ) underlies the decrease in the extent of inactivation of the WT channel caused by PMA.
These experiments indicated that a dephosphorylating activity underlies the modulation by PMA, so we tested the possible involvement of phosphatases. Previously, we showed in in vitro studies that 10 nM okadaic acid (a selective inhibitor of protein phosphatase-1 (PP-1)) inhibits spontaneous dephosphorylation of the ␣ protein in the oocyte homogenate and that protein phosphatase-2B (PP-2B) can dephosphorylate its immunopuri-fied form (4). Here we tested the effect of PMA in the presence of several phosphatase inhibitors according to experimental protocols used successfully by others in different cells (10,38). mGluR1a Mimics the Effect by PMA-Since mGluR1a activates phospholipase C and lately was shown to activate PKC in oocytes (21), we tested whether activation of mGluR1a, coexpressed with the channel in oocytes, will mimic the modulation by PMA. Coinjection of ␣, ␤, and mGluR1a RNAs into oocytes gave rise to an ␣␤ current with an average amplitude of 53 Ϯ 0.04% of the current elicited in oocytes that were injected with the same RNA amounts of ␣ and ␤ without mGluR1a (51 and 61 oocytes tested, respectively; 4 frogs; p Ͻ 0.001). The reduced amplitude could be due to lower expression of the ␣␤ proteins in oocytes injected with mGluR1a; however, no evidence for that was obtained in concomitant biochemical analyses of the level of expression of the proteins in six experiments (see below). The other alternative would be that the reduced amplitudes were intrinsic to the receptor modulation of the current (see below) due to enough receptor molecules being spontaneously active without agonist, thereby continually activating at a significant level a signaling cascade that utilizes PKC (cf. Ref. 30). This notion is supported by the observation that in two experiments the currents elicited in oocytes injected with mGluR1a had significantly larger (ϳ140%) I s fractions as compared with those in oocytes not injected with the receptor.
Exposure of these oocytes to 100 M glutamate for 1 min caused a Ca 2ϩ -dependent Cl Ϫ current as described previously for oocytes injected with mGluR1a alone (20,39). To observe the effect of glutamate on the ␣␤ channel without the interfer- oocytes, respectively, taken from two frogs (as described in legend to Fig. 1,  D-F). Shown are values normalized to that before the start of PMA application (at t ϭ 0 min). D, the maximal effect of PMA on currents through channels mutated at the ␣ subunit, normalized to the effect of PMA on WT currents elicited in oocytes (number denoted above bars) taken from the same frogs (two for S446A, two for S489I; and three for S322A). *, p Ͻ 0.03; **, p Ͻ 0.007, significantly different from values before the application of PMA (A-C) or from WT (D); *, p Ͻ 0.03, a significant difference between S446A and WT.
ence of this current, the oocytes were preinjected with the Ca 2ϩ chelator EGTA. This treatment abolished all the transient and most of the sustained Ca 2ϩ -activated Cl Ϫ current upon application of glutamate and allowed the leak current to return close to its original value after washout of glutamate, as described (21). Fig. 5 shows that the effect of glutamate on the ␣␤ current resembled that induced by PMA regarding the reduction of I p , the biphasic response of I s , and the increase in I s fraction. The extent of reduction in I p was small, as compared with that by PMA, possibly because the initial I p was already reduced by spontaneously active receptors. The onset of the increase in I s fraction was faster (4 min to reach 40% of response as compared with 10 min for that of PMA), and saturation occurred at 10 min. inact was not significantly different between oocytes expressing the channel proteins only and those coexpressing mGluR1a, also the application of glutamate in oocytes injected with EGTA did not change inact in the five experiments described above. In two additional experiments inact increased, especially during the first few minutes following the glutamate application, probably because the EGTA buffering was less effective; these experiments were discarded.
The Reduction in Inactivation by mGluR1a Is Due to Dephosphorylation of Ser-446 -Biochemical analysis of mGluR1a expressing oocytes in three experiments out of six done (see Fig.  6 for a representative example in oocytes of one frog) showed that coexpression of the receptor already caused some reduction in the extent of basal phosphorylation of the ␣ protein; the intensity ratio of the 57-kDa over the 54-kDa bands was reduced to 0.82 Ϯ 0.04 that in oocytes that were not expressing the receptor. This could correspond to the large I s fractions observed in electrophysiologically assayed oocytes injected with the receptor (see above), substantiating the notion that in oocytes of some frogs there are enough receptor molecules that are spontaneously active. Biochemical analysis of all six experiments showed that incubation of the oocytes for 1-5 min in 100 M glutamate before homogenization resulted in extent of phos-phorylation of the ␣ protein that was 0.72 Ϯ 0.05 that in oocytes expressing the receptor but were not exposed to the agonist. Thus, also in this respect mGluR1a mimicked the PMA modulation and induced dephosphorylation of the ␣ protein, suggesting that its effect on the extent of inactivation may result from the dephosphorylation.
Indeed, in concomitant electrophysiological experiments activation of mGluR1a failed to affect the extent of inactivation of the ␣ S446A ␤ channel (Fig. 5C). DISCUSSION This study shows that in Xenopus oocytes activation of PKC by PMA decreases the extent of inactivation (increases the I s fraction) of rat brain Kv1.1/Kv␤1.1 (␣␤) current. It further suggests that the metabotropic glutamate receptor mGluR1a that is coupled to phospholipase C is a plausible candidate to initiate such a cellular process and to modulate the inactivation of the ␣␤ channel in a physiologically relevant environment. Several interesting aspects to this modulation are as follows: (i) it does not involve PKC-mediated phosphorylation but rather a PKC-mediated dephosphorylation of the channel; (ii) it is opposite to the modulation by PKA that increases the extent of inactivation of the channel (8); (iii) PKA and PKC signaling pathways converge onto the same site bringing about phosphorylation and dephosphorylation of Ser-446, respectively.
PKC-induced Dephosphorylation of Ser-446 Decreases the Extent of Inactivation of the ␣␤ Channel-An intriguing mode of modulation of an ion channel is described here that involves induction of a dephosphorylating activity by activation of PKC. Involvement of PKC is evident from the inhibition of the PMA effect on inactivation by two potent blockers, staurosporine (a wide specificity kinase blocker) and BIS (a specific PKC blocker). The fact that calphostin C (a specific PKC blocker) did not inhibit the effect may be interpreted as an indication of the PKC subtype involved, as detailed in the following. PKC isoenzymes have been classified into three groups with different structure and cofactor regulation (31,32). Activation by PMA makes the involvement of "atypical" PKC isoenzymes unlikely. Ca 2ϩ chelation that did not impair the effect by mGluR1a leaves out the possibility of "conventional" PKCs. Of the "new" PKCs that are left as candidates, PKC-is the most plausible isoenzyme to mediate the decrease in inactivation as it is insensitive to calphostin C (33).
Correlation between the decrease in the extent of inactivation and dephosphorylation of Ser-446 of the ␣ subunit was suggested by SDS-PAGE analysis of the ␣␤ proteins after PMA treatment. First, it eliminated the possibility that decreased ␤-binding capacity of ␣ was the cause of the decreased inactivation by PMA. Second, it demonstrated that PMA did not phosphorylate but rather dephosphorylated the channel by decreasing the basal phosphorylation of Ser-446. This analysis was verified specifically for channel proteins residing in the plasma membrane. Final verification of a causal relationship between dephosphorylation and decreased inactivation was provided by biophysical analysis that showed that PMA reduces the extent of inactivation of the nonphosphorylated ␣ S446A ␤ mutant channel but not that of irrelevant mutant channels. The mechanism underlying dephosphorylation is unclear at present. It could be due either to activation of phosphatases or inhibition of kinases but does not seem to involve the phosphatases PP-1, PP-2A, or PP-2B, as the corresponding inhibitors calyculin A, okadaic acid, and cyclosporin A did not inhibit the decrease in inactivation by PMA. Many novel protein serine/threonine phosphatases have been recently discovered, for some the pharmacology has not been determined (34). It should be noted that the reduction of amplitudes by PMA is not dephosphorylation-mediated, as it was not eliminated in the ␣ S446A channel.
This mode of signaling conforms with a concept that kinases and phosphatases do not simply oppose each other; rather, they may work in series to create cascades of signaling processes that eventually will regulate target protein(s). Such modulatory pathways of different levels of complexity and comprising various combinations of enzymes have already been described (e.g. Ref. 40). With respect to PKC, they have been associated with turning on or off the following dephosphorylation activities: PKC phosphorylated and inhibited calcineurin (35), phosphorylated and inactivated the serine/threonine glycogen synthase kinase-3␤ (36), and activated phosphatases PP-1 (37,38) and PP-2 (38). In only a few cases such modulatory pathways have been demonstrated to regulate ion channel activity; in leech neurons PKC activated a cation channel by activating a tyrosine phosphatase activity (41), and in bag cell neurons PKA regulated a voltage-gated cation channel via indirect activation of a protein tyrosine phosphatase (42). A recent study suggests that phosphorylation of purified Na ϩ channels by PKC decreases dephosphorylation of PKA phosphorylation sites by purified calcineurin or PP-2A (43).
mGluR1a-induced Dephosphorylation of Ser-446 Decreases the Extent of Inactivation of the ␣␤ Channel-In this study we show that activation of mGluR1a by glutamate mimics the effect of PMA on inactivation, the onset being faster (4 min to reach 40% of response as compared with 10 min for PMA) and the response saturating. The fast onset of the response could be due to colocalization of the receptor, the channel, and the oocyte's signaling molecules involved in the response in submembranous sites targeted by protein(s) which serve as a scaffold (for review see Ref. 44). Notably, a post-synaptic density-95-like endogenous protein was shown by us to interact with the channel and to affect its extent of inactivation (9). The effect of glutamate on the current amplitudes was small compared with that of PMA, possibly because it was somewhat occluded by the effect exerted by spontaneous coupling of the receptor without glutamate to the signaling machinery, since coexpression of the receptor with the channel reduced the current amplitudes significantly even in the absence of agonist.
Several types of K ϩ channels have been shown to be regulated variably by activation of group I of mGluRs in neurons of different brain areas (reviewed in Ref. 15). However, the exact mechanisms involved in the regulations have not been clearly established. A PKC-mediated regulation was inferred only in the inhibition by mGluR1 of IK AHP in dentate granule neurons (45) but was shown not to be involved in the same effect in CA3 pyramidal neurons (46). In every expression system examined group I receptors stimulate phospholipase C as revealed by an increase in phosphoinositide turnover and Ca 2ϩ release from internal stores (15). In Xenopus oocytes PKC (probably PKC-) was explicitly shown to mediate inhibition by mGluR1a of an inwardly rectifying K ϩ channel (21). In this study the signaling pathway underlying the effect of mGluR1a on the extent of inactivation seems to be mediated by PKC (possibly PKC-) and involves dephosphorylation of Ser-446. It differs from the signaling pathway underlying the effect on amplitudes which is not eliminated in the ␣ S446A ␤ channel and thus does not involve dephosphorylation.
Modulation of the Extent of Inactivation by Dephosphorylation, Possible Mechanisms-It is evident from this study that stimulation of PKC causes a decrease in the extent of inactivation of the ␣␤ current by dephosphorylating Ser-446. This effect is opposite that shown by us for a constitutively active kinase (yet unidentified) or for a stimulated PKA that causes an increase in the extent of inactivation by phosphorylating Ser-446 that impairs interaction between the channel and the microfilaments (8). Part of the interaction with the microfilaments was shown to be mediated via a post-synaptic density-95-like protein that interacts with the C-terminal end of Kv1.1 (9). A similar phenomenon was described in mammalian cells  2) or from uninjected oocytes (lane 4). Oocytes were either incubated for 2 min (ϩ) or not incubated (Ϫ) with glutamate (glut) (as indicated above the lanes) just before homogenization, in the presence of phosphatase and protease inhibitors. Thin arrows point to migration of molecular mass markers in kDa. Thick arrows point to ␣ and ␤ proteins. Bottom row denotes for each of the lanes the ratio of labeling intensities of the 57-kDa over the 54-kDa bands that represent the ratio of the phosphorylated over the nonphosphorylated forms of the ␣ protein.
for the K ϩ channel Kir 2.3 that dissociates from post-synaptic density-95 upon PKA phosphorylation of a serine residue at its C terminus (47). We proposed and have now confirmed 2 a kinetic model that assumes two modes of gating of the ␣␤ channel, inactivating and a noninactivating. In the noninactivating mode the channel's interaction with microfilaments results in impaired inactivation and gives rise to the sustained current component (I s ), whereas in the inactivating mode the channel does not interact with the microfilaments and gives rise to the inactivating current component (I i ); the equilibrium between the modes is influenced by the extent of phosphorylation of Ser-446. In this context it is expected that the PKCinduced dephosphorylation shifts the equilibrium toward the noninactivating mode which is manifested in the increase in the sustained fraction of the current.
Another mechanism that can possibly underlie the increase in the sustained fraction by PKC-induced dephosphorylation is suggested by the biphasic response to PMA that exhibited an initial increase followed by a decrease in the sustained current amplitude. It was converted in the mutant channel that is not phosphorylated (␣ S446A ␤) to a monophasic response exhibiting a larger reduction of current amplitude. This suggests that the dephosphorylation induced an increase in the sustained current amplitude that overlapped a dephosphorylation-independent reduction. In the above context it means that dephosphorylation affected ␣␤ channels residing in the noninactivating mode. Strikingly, the biphasic response to PMA of the sustained current closely resembled that of the delayed rectifier current of homomultimeric channels consisting of ␣ subunits with deletion of most of the N terminus, whereas the response of the current of intact homomultimeric ␣ channels is monophasic exhibiting only reduction of amplitude (I ␣ in Fig.  1D; Ref. 5). Thus, the deletion of the N terminus of ␣ subunits in homomultimeric channels may have similar consequences on the channel structure, and hence on its modulation, to those caused by binding of the N terminus of ␣ to ␤ subunit(s) in ␣␤ heteromultimeric channels.
In conclusion, activation of PKC induces dephosphorylation of Ser-446 that results in decreased extent of inactivation of Kv1.1/Kv␤1.1. Such a mechanism may turn out to be physiologically relevant in hippocampal pyramidal cells where both Kv1.1 (48) and mGluRs were shown to play a role in long term changes of synaptic function (16).