Modulation of Kv3.1b Potassium Channel Phosphorylation in Auditory Neurons by Conventional and Novel Protein Kinase C Isozymes*

In fast-spiking neurons such as those in the medial nucleus of the trapezoid body (MNTB) in the auditory brainstem, Kv3.1 potassium channels are required for high frequency firing. The Kv3.1b splice variant of this channel predominates in the mature nervous system and is a substrate for phosphorylation by protein kinase C (PKC) at Ser-503. In resting neurons, basal phosphorylation at this site decreases Kv3.1 current, reducing neuronal ability to follow high frequency stimulation. We used a phospho-specific antibody to determine which PKC isozymes control serine 503 phosphorylation in Kv3.1b-tranfected cells and in auditory neurons in brainstem slices. By using isozyme-specific inhibitors, we found that the novel PKC-δ isozyme, together with the novel PKC-ϵ and conventional PKCs, contributed to the basal phosphorylation of Kv3.1b in MNTB neurons. In contrast, only PKC-ϵ and conventional PKCs mediate increases in phosphorylation produced by pharmacological activation of PKC in MNTB neurons or by metabotropic glutamate receptor activation in Kv3.1/mGluR1-cotransfected cells. We also measured the time course of dephosphorylation and recovery of basal phosphorylation of Kv3.1b following brief high frequency electrical stimulation of the trapezoid body, and we determined that the recovery process is mediated by both novel PKC-δ and PKC-ϵ isozymes and by conventional PKCs. The association between Kv3.1b and PKC isozymes was confirmed by reciprocal coimmunoprecipitation of Kv3.1b with multiple PKC isozymes. Our results suggest that the Kv3.1b channel is regulated by both conventional and novel PKC isozymes and that novel PKC-δ contributes specifically to the maintenance of basal phosphorylation in auditory neurons.

There is a good correlation between the ability of neurons to discharge at high rates and the expression of Kv3 family potassium channels, particularly the Kv3.1 channel (1)(2)(3)(4)(5). Neurons that express Kv3.1 at high levels include those in the medial nucleus of the trapezoid body (MNTB) 2 within the auditory brainstem (6 -8). The MNTB is key element of neural pathways that detect differences in the level or timing of interaural stimuli to compute sound localization. MNTB neurons are able to fire at frequencies of hundreds of Hz (9,10) and to lock their action potentials precisely to the phase of auditory stimuli at frequencies of up to 2-4 kHz or to rapid fluctuations in the amplitude of high frequency sounds (11). One important physiological specialization enabling MNTB neurons to fire at such high frequencies is the high level of expression of Kv3.1 potassium channels. Genetic knock-out of the Kv3.1 gene, as well as pharmacological and computer modeling studies, confirms that a high threshold component of potassium current in MNTB neurons is carried by Kv3.1 channels and that its elimination impairs the neuronal response to high frequency stimulation (12)(13)(14). Nevertheless, high levels of Kv3.1b current degrade the accuracy of action potential timing at lower frequencies of firing (14 -16).
Two isoforms of the Kv3.1 channel exist, Kv3.1a and Kv3.1b, that are generated by alternative splicing of the Kv3.1 gene. The Kv3.1b channel predominates in the mature nervous system and has a longer carboxyl terminus than that of Kv3.1a (2,17). Activators of PKC significantly reduce the amplitude of Kv3.1b current (14,18,19). Although activation of PKC stimulates phosphate incorporation into several serine residues in Kv3.1b, the specific actions of PKC on Kv3.1b currents have been shown to depend selectively on the phosphorylation of Ser-503 in the carboxyl-terminal region (14).
By using a phospho-specific antibody to serine 503 of Kv3.1b, it has been found that Kv3.1b in MNTB neurons is basally phosphorylated by PKC in a quiet auditory environment, providing maximal timing accuracy at low firing frequencies. In vivo acoustic stimulation of animals, or high frequency stimulation of the afferent input of MNTB neurons in brainstem slices, results in a rapid and reversible decrease in the level of phosphorylation. This dephosphorylation permits neurons to fire at higher rates, albeit with lower temporal accuracy. Thus phosphorylation of Kv3.1b by PKC appears to be a mechanism that rapidly adjusts the intrinsic electrical properties of neurons to the pattern of incoming auditory stimuli (16).

EXPERIMENTAL PROCEDURES
Immunocytochemistry-Chinese hamster ovary (CHO) cells stably expressing Kv3.1a, Kv3.1b, or Kv3.1 mutant S503A were maintained in Iscove's modified Dulbecco's medium (Invitrogen) supplemented with 10% fetal bovine serum, 0.1 mM hypoxanthine, and 0.5 mg/ml geneticin (Invitrogen). CHO cells were grown on coverslips for 24 -48 h in a 5% CO 2 incubator at 37°C preceding drug administration. Cells were preincubated for 30 min with PKC inhibitors at the indicated concentrations and then stimulated with the PKC activator PMA for 12 min. For transient transfection of Kv3.1-CHO cells with pcDNA-mGluR1 (a kind gift from Dr. Jarda Wroblewski, Georgetown University), the cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 2 mM L-glutamine, and 4.5% proline. Cells were transfected using Lipofectamine 2000 (Invitrogen) and were treated 10 min with mGluR1 agonist DHPG (300 M) at 24 h after transfection. After drug treatment, CHO cells were quickly washed with ice-cold 0.1 M phosphate-buffered saline and then fixed with 4% paraformaldehyde at 4°C for 10 min. After blocking and permeabilization, cells were incubated with rabbit anti-phosphoserine 503 Kv3.1b polyclonal antibody (1:400) at 4°C for 24 h and subsequently with Alexa Fluor 488 goat anti-rabbit IgG (1:700) at 4°C overnight. The coverslips were then mounted on glass slides with Citifluor Mountant Media (Ted Pella, Redding, CA), and images were taken on an Olympus BX60 microscope by using a SPOT RT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI).
Immunohistochemistry-Coronal sections were cut on a cryostat at a thickness of 35 m. Similarly with immunocytochemistry, sections were incubated with phospho-specific antibody against Kv3.1b at 4°C for 60 h and subsequently with Alexa Fluor 488 goat anti-rabbit IgG at 4°C for 24 h. Sections were scanned by a Bio-Rad model 1024 UV laser confocal microscope system.
Coimmunoprecipitation and Western Blotting-CHO cells or rat brainstems were homogenized in ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% dodecyl ␤-D-maltoside, and a protease/phosphatase inhibitor mixture). The lysate was precleared with 1:1 slurry of protein A-Sepharose and equilibrated in lysis buffer at 4°C for 1 h with shaking. After a short spin, the supernatant was incubated at 4°C overnight with rabbit IgG or antibodies as indicated. Immune complexes were pooled down by protein A-Sepharose at 4°C for 2 h with shaking. Immunoprecipitates were washed and dissolved in SDS sample buffer. The protein was loaded and separated by SDS-PAGE (10%) and then transferred to polyvinylidene difluoride membranes and probed with the indicated antibodies. Goat anti-rabbit horseradish peroxidase-coupled secondary antibodies were used for detection with West Femto chemiluminescence (Pierce).
Image Analysis-Optical density of the immunostaining was measured using ImagePro Plus software (Media Cybernetics, Silver Spring, MD) and was referred to as "level of immunoreactivity" in the figures. OD values were subjected to statistical evaluation using Student's t test or one-way analysis of variance followed by post hoc comparison to confirm significant differences between the groups. Criteria for significance in all analyses were defined as p Ͻ 0.05. Data were presented as mean Ϯ S.E.

RESULTS
Detection of Phosphorylated Kv3.1b Potassium Channels Using a Ser-503 Phospho-specific Antibody-Immunohistochemistry and immunoblotting were performed to evaluate the specificity of the anti-phospho-Ser-503 Kv3.1b antibody that was used throughout this study. CHO cells were stably transfected with the wild-type Kv3.1b gene or with S503A mutant Kv3.1b, in which this PKC phosphorylation site is mutated. As described previously (15), no detectable basal immunostaining was observed in Kv3.1b-CHO cells until treatment with PMA (500 nM), an activator of PKC (Fig. 1A). Moreover, no immunoreactivity was detected after treatment with PMA in cells transfected with the S503A mutant Kv3.1b or with the Kv3.1a splice variant, both of which lack the Ser-503 phosphorylation site. Negative control experiments in which the primary antibody was either omitted or pre-adsorbed with the antigen peptide also revealed no immunostaining in PMA-treated Kv3.1b-CHO cells. In contrast, incubation of the primary antibody with a heterologous peptide did not affect the immunostaining in PMAtreated Kv3.1b-CHO cells (Fig. 1A).
In contrast to the Kv3.1b-transfected CHO cells, profuse immunostaining was seen on the membrane of MNTB principal neurons in adult rats using the phospho-specific Kv3.1b antibody, even in the absence of pharmacological activation of PKC. This immunolabeling represented specific staining of Kv3.1b phosphorylated at Ser-503 because it was blocked by preincubating the primary antibody with the antigenic peptide but not with the heterologous peptide (Fig. 1B).
Immunoblots of denatured protein extracts were carried out in transfected CHO cells. Consistent with previous studies using antibodies against other regions of the Kv3.1b protein (2, 24), a protein band of ϳ100 kDa was detected in extracts from Kv3.1b-CHO cells using the phospho-specific Kv3.1b antibody pre-absorbed with the heterologous peptide (Fig. 1C). This band was absent when the immunoblots were treated with antibody that was pre-adsorbed with the antigenic peptide or when the cells were treated with Me 2 SO, the vehicle for PMA treatment. Moreover, no band was detected in PMA-treated CHO cells that had been transfected with the S503A mutant Kv3.1b gene. Therefore, this antibody recognized Kv3.1b channels only when the channels were phosphorylated by PKC at Ser-503.
PKC Isozymes Responsible for PMA-induced Phosphorylation of Kv3.1b Channels in CHO Cells-To identify the PKC isozymes involved in the modulation of Kv3.1b channels, a variety of PKC inhibitors was used specifically to inhibit different groups of PKC isozymes (Table 1). Group I inhibitors were those that are known only to inhibit the conventional family of PKCs (cPKC), whereas the group II inhibitors act on both cPKC and nPKC. Gö6976 is a specific inhibitor of cPKCs and has no effect on the activity of nPKC or aPKC isozymes even at micromolar levels (25,26). GF109203X is, at a concentration of 50 nM, a specific inhibitor of cPKCs. At the much higher concentration of 1 M, GF109203X inhibits both cPKCs and nPKCs (27,28). Similarly, Ro 31-7549 is known to antagonize only cPKCs at a concentration of 50 nM but inhibits both cPKCs and nPKCs at the higher concentration of 4 M (29, 30). Finally, rottlerin (10 M) has long been used as a selective inhibitor of nPKC-␦ (31,32). At much higher concentrations, however, rottlerin acts on all classes of PKC.
Quantification of the phospho-Kv3.1b immunoreactivity showed that PMA treatment of Kv3.1b-transfected CHO cells produced a dra-matic increase in the phosphorylation level over control cells treated with Me 2 SO ( p Ͻ 0.0001) (Fig. 2B). The effect of PMA was significantly attenuated by group I inhibitors to values between 59 and 32% of that in PMA-treated cells, i.e.100 nM Gö6976 ( p Ͻ 0.01 versus PMA and p Ͻ 0.05 versus control), 1 M Gö6976 ( p Ͻ 0.0001 versus PMA and p Ͻ 0.05 versus control), 50 nM Ro 31-7549 ( p Ͻ 0.0001 versus PMA and p Ͻ 0.05 versus control), and 50 nM GF109203X ( p Ͻ 0.0001 versus PMA and p Ͻ 0.05 versus control). The group II inhibitors, i.e. 4 M Ro 31-7549 and 1 M GF109203X, completely abolished the effect of PMA, attenuating staining to less than 4.0% of PMA treatment (for both cases, p Ͻ 0.0001 versus PMA and p Ͼ 0.05 versus control). The effects of group II inhibitors were significantly more potent than those of group I inhibitors ( p Ͻ 0.01). The response to PMA was not significantly affected by 10 M rottlerin ( p Ͼ 0.05 versus PMA and p Ͻ 0.001 versus control).
PKC Isozymes Responsible for mGluR1-mediated Phosphorylation of Kv3.1b Channels in CHO Cells-It is known that activation of mGluR1 leads to PKC activation, through increases in membrane-bound diacylglycerol and intracellular inositol 1,4,5-trisphosphate level. In CHO cells cotransfected with mGluR1 and Kv3.1b channels, treatment with the mGluR1 agonist DHPG (300 M) led to an increase in the phosphorylation level of Kv3.1b channels (Fig. 3). This effect was partially inhibited by the cPKC inhibitor Gö6976 (1 M), was fully blocked by the nand c-PKC inhibitor GF109203X (1 M) and the group I/II mGluR   immunoreactivity. Data are presented as means Ϯ S.E. * represents a significant difference with respect to DHPG treatment, and # represents a significant difference between group I and group II inhibitors.
Quantification of the phospho-Kv3.1b immunoreactivity showed that DHPG treatment of mGluR1, Kv3.1b-cotransfected CHO cells produced a very marked increase in the phosphorylation level over control cells that were transfected with Kv3.1b and pcDNA vector alone ( p Ͻ 0.001) (Fig. 3B). The effect of DHPG was significantly attenuated by the group I inhibitor Gö6976 (1 M, p Ͻ 0.001 versus DHPG and p Ͻ 0.05 versus control). The group II inhibitor GF109203X (1 M) completely abolished the effect of DHPG ( p Ͻ 0.001 versus DHPG and p Ͼ 0.05 versus control). The effect of GF109203X was more potent than Gö6976 ( p Ͻ 0.05). The response to DHPG was not significantly affected by the nPKC-␦ inhibitor rottlerin (10 M, p Ͼ 0.05 versus DHPG and p Ͻ 0.001 versus control).
PKC Isozymes That Regulate the Basal Phosphorylation of Kv3.1b in Brainstem Slices-The PKC isozymes that have been identified in the brain and spinal cord are cPKC-␣, -␤I, -␤II, -␥, and nPKC-␦, -⑀, and aPKC- (21,36). As shown in Fig. 4, dense basal phosphorylation of Kv3.1b potassium channels at Ser-503 was observed in MNTB neurons in brainstem slices of 17-19-day-old rats. The basal phosphorylation was diminished by pretreatment with the group I inhibitor, 1 M Gö6976, and was completely abolished by the group II inhibitor, 1 M GF109203X. In contrast to its lack of effect in PMA-treated CHO cells, rottlerin (10 M), an inhibitor of nPKC-␦, also reduced the basal phosphorylation of Kv3.1b channels in MNTB neurons.
Quantification showed that the basal Kv3.1b phosphorylation in MNTB neurons in control slices was reduced to 40% of control levels by 1 M Gö6976 ( p Ͻ 0.001 versus control) and to less than 6% of control by GF109203X (1 M) (p Ͻ 0.0001 versus control and p Ͻ 0.001 versus Gö6976). Rottlerin (10 M) reduced the basal staining to 71% of the control ( p Ͻ 0.05). These data indicate that both cPKCs and nPKCs contribute to the basal phosphorylation of Kv3.1b channels in MNTB neurons and that nPKCs account for at least 34% of the basal staining. Because inhibition of PKC-␦ alone by rottlerin produced a smaller inhibition of the basal staining, these results suggest that both nPKC-␦ and nPKC-⑀ regulate the basal level of Kv3.1b phosphorylation.
PKC Isozymes That Mediate PMA-induced Phosphorylation of Kv3.1b in MNTB Neurons-We next tested the actions of PKC inhibitors on PMA-induced phosphorylation of Kv3.1b channels in MNTB neurons. As shown in Fig. 5, treatment of brainstem slices with PMA (700 nM) significantly enhanced Kv3.1b phosphorylation in MNTB neurons. The group I inhibitors (100 nM Gö6976, 1 M Gö6976, 50 nM Ro 31-7549, or 50 nM GF109203X) diminished Ser-503 phosphorylation but did not abolish it. In contrast, group II inhibitors (4 M Ro 31-7549 or 1 M GF109203X) completely abolished K3.1b phosphorylation in the presence of PMA, suggesting that both cPKCs and nPKCs contribute to PMA-induced phosphorylation of Kv3.1b channels. In contrast to the case of basal phosphorylation in MNTB neurons, the PKC-␦ inhibitor rottlerin (10 M) had no effect on PMA-induced phosphorylation. Because PKC-␦ and PKC-⑀ are the only two novel PKC isozymes detected in the nervous system, the present data indicate that PMAproduced phosphorylation of Kv3.1b channels is mediated by both cPKCs and nPKC-⑀ but that PKC-␦ contributes selectively to basal phosphorylation. Quantification (Fig. 5B) showed that Kv3.1b phosphorylation was increased significantly by over 58% when the slices were treated with PMA ( p Ͻ 0.001). Group I inhibitors reduced the PMA-produced phosphorylation to between 42 and 37% ( p Ͻ 0.001 versus PMA treatment for each inhibitor). Kv3.1b phosphorylation in slices treated with PMA combined with the group I inhibitors was significantly higher than that in MNTB neurons exposed to 1 M Gö6976 but not treated with PMA ( p Ͻ 0.05). This suggests that cPKCs alone could not fully account for the PMA-produced increase in Kv3.1b phosphorylation. Meanwhile, group II inhibitors completely abolished PMA-produced phosphorylation, reducing immunoreactivity to less than 5% that in the presence of PMA (for both inhibitors, p Ͻ 0.001 versus PMA treatment and p Ͻ 0.05 versus PMA ϩ group I inhibitors). Moreover, there was no significant difference between the low phosphorylation level of Kv3.1b in these groups and in the neurons that were not exposed to PMA but were treated with the group II inhibitor (1 M GF109203X). No significant difference was found between neurons exposed to PMA alone and those exposed to PMA with 10 M rottlerin ( p Ͼ 0.05).
Time Course of Stimulation-induced Dephosphorylation of Kv3.1b Channels-The trapezoid body contains fibers that provide the major excitatory input to MNTB neurons. Previous studies have shown that conditioning stimulation of the trapezoid body at a high frequency (600 Hz) for 20 s activates a phosphatase that decreases the phosphorylation level of Kv3.1b in MNTB neurons, producing an increase in Kv3.1b current. As a result, the number of overshooting action potentials evoked by intracellular current pulses applied at a high frequency (500 or 600 Hz) was almost doubled. At lower frequencies, however, dephosphorylation of Kv3.1b resulted in a lower timing accuracy of action potentials (15). Therefore, by changing the phosphorylation state of Kv3.1b potassium channels, high frequency stimulation modifies the neuronal response to different stimuli.
By stimulating brainstem slices for different durations at 600 Hz, we have now tested the time course of stimulation-induced dephosphorylation of Kv3.1b channels. As shown in Fig. 6, stimulation for 1 or 2 s failed to alter the phosphorylation level of the Kv3.1b channel, although significant dephosphorylation of this channel was observed when the stimulation was applied for 5-20 s ( p Ͻ 0.05 versus control). Double immunostaining with syntaxin, a presynaptic marker, showed that stimulation-induced dephosphorylation occurred both pre-and post-synaptically (Fig. 6C).
PKC Isozymes That Mediate the Recovery of Phosphorylation following Stimulation-When stimulation was applied for 20 s, partial recovery of phosphorylation was observed within 1 min of the end of the stimulus train, and full basal phosphorylation level was restored within 5 min (Fig. 7). The full recovery process was partially inhibited by the group I inhibitor, 1 M Gö6976, and was completely abolished by the group II inhibitor, 1 M GF109203X (Fig. 8). Rottlerin (10 M), the inhibitor of nPKC-␦, also inhibited the recovery process.
Quantification showed that 1 M Gö6976 reduced the recovery level of Kv3.1b phosphorylation to 38% ( p Ͻ 0.001 versus control recovery with no inhibitors). The recovery level was reduced to less than 5% by 1 M GF109203X ( p Ͻ 0.001 versus control recovery or versus the recovery in the presence of 1 M Gö6976) and to 76% by rottlerin (10 M)  N ϭ 4, n ϭ 2). N indicates the number of analyzed sections, and n is the number of rats used in each group. Scale bar, 100 m. Histogram shows the quantification result. Data are presented as means Ϯ S.E. * denotes a significant difference compared with the PMA treatment, and # denotes a significant difference between group I and group II inhibitors. ( p Ͻ 0.01). These data indicate that cPKCs, nPKC-␦, and nPKC-⑀ all contribute to the recovery from stimulation-induced dephosphorylation of Kv3.1b channels in MNTB neurons.
Extracts were also were prepared from control rat brainstem and from brainstem treated with PMA (1 M). In these cases, as in CHO cells, anti-PKC-␣ and anti-PKC-⑀ antibodies coimmunoprecipitated phospho-Kv3.1b from PMA-treated tissue. In contrast to the case of CHO cells, however, phospho-Kv3.1b was also immunoprecipitated from untreated tissues, which is in keeping with the finding that Kv3.1b channels are basally phosphorylated in MNTB neurons but not in transfected CHO cells (Fig. 9, left panels). Consistent with results showing a role for PKC-␦ in basal phosphorylation in MNTB neurons, antibodies against PKC-␦ precipitated phospho-Kv3.1b from both PMA-treated and -untreated tissues. Consistent results were also obtained in the reciprocal coimmunoprecipitation experiments using tissue extracts (Fig. 9, right panels). PKC-␣ and PKC-⑀ were immunoprecipitated from both PMA-treated and -untreated brainstems by anti-phospho-Kv3.1b antibodies, with greater amounts detected in the PMA-treated brainstems. PKC-␦ was also coimmunoprecipitated with phospho-Kv3.1b in both PMA-treated and -untreated brainstems.
In summary, the coassembly patterns of Kv3.1b with specific PKC isozymes were consistent with results obtained above using specific inhibitors. Kv3.1b is regulated by both conventional and novel PKC isozymes. Of the novel PKCs, only nPKC-⑀ contributes to PMA-induced increases in the phosphorylation over basal levels.

DISCUSSION
By using both immunocytochemistry and coimmunoprecipitation approaches, we have shown that the Kv3.1b potassium channel is basally phosphorylated at serine 503 in resting MNTB neurons but not in Kv3.1b-transfected CHO cells. High frequency stimulation of the major  synaptic input pathway to MNTB neurons for 5 s or more results in the dephosphorylation of Kv3.1b channels, and the recovery of basal phosphorylation occurs over 1-5 min. Our results using selective inhibitors indicate that at least three classes of PKC isozymes participate in the integrated regulation of Kv3.1b channels in MNTB neurons. cPKC and nPKC-⑀ mediate the PMA-induced increase in phosphorylation of Kv3.1b channels over basal levels. These same enzymes also contribute to PMA-or metabotropic receptor-induced increases in phosphorylation in CHO cells. In contrast, nPKC-␦ contributes to maintaining the basal phosphorylation of Kv3.1b in MNTB neurons but does not participate in pharmacological stimulation of phosphorylation over basal levels in either MNTB neurons or CHO cells.
It has also been well documented that many biological processes are mediated by more than one PKC isozyme. For example, cPKCs and nPKC-␦ all regulate dopamine transporter-associated currents (37). PKC-␣ and -⑀ are both required for Grb2-associated Binder-1 tyrosine phosphorylation in response to platelet-derived growth factor (38). Both cPKC and nPKC are involved in thymocyte-negative selection (39), and both PKC-␣ and -⑀ contribute to Raf/ERK (extracellular signal-regulated kinase) activation (40).
Differential roles for the novel PKC isozymes PKC-⑀ and PKC-␦ have also been reported in the regulation of dopamine transporters (37). Moreover, nPKC-⑀ but not nPKC-␦ is activated during phorbol esterinduced changes in synaptic strength in MNTB neurons (41,42). Because the regulation of nPKC-⑀ and nPKC-␦ isozymes by cofactors is very similar, it is likely that factors such as targeting proteins are responsible for such different activities of the two isozymes. For example, the subcellular localization of different PKC isozymes may differ so that PKC-␦ has more restricted access to phorbol ester. Alternatively, the lack of participation of PKC-␦ in PMA-produced Kv3.1b phosphorylation may reflect the saturation of PKC-␦ activity under basal conditions.
Fast-spiking neurons are found in many brain regions, including the hippocampus, basal ganglia, neocortex, reticular thalamus, medial vestibular nucleus, and auditory nuclei. The ability of such neurons to generate high frequency discharges requires very rapidly activating and deactivating delayed rectifier potassium channels (43). The narrow action potentials that are shaped by such potassium channels minimize Na ϩ channel inactivation, maintaining the excitability of the soma and initial axonal membrane segments. At the somata of MNTB neurons, Kv3.1b channels provide this function and allow rapid repolarization of action potentials with very little relative refractory period (13,14).
The amplitude of Kv3.1 potassium currents plays an important role in determining the firing pattern of MNTB neurons. Although low levels of Kv3.1 current help achieve accurate timing of action potentials at low frequencies, high levels of Kv3.1 current enable neurons to follow high frequency inputs (14,15). The Kv3.1b potassium channel can be regulated by pharmacological activation of PKC, which decreases Kv3.1b current (14,15). Under physiological conditions, this channel is also modulated by neuronal activity. In response to high frequency auditory or electrical stimulation, basal phosphorylation of the Kv3.1b channel in MNTB neurons is rapidly decreased, resulting in an enhanced ability of neurons to fire at high frequencies, albeit with lower temporal precision (15). Thus changes in the phosphorylation level of this ion channel allow neurons to rapidly adapt their intrinsic membrane properties to a continuously changing environment. It has been shown that phosphatases PP1/PP2A contribute to stimulation-induced dephosphorylation of the Kv3.1b channel, and the present study has identified the PKC isozymes mediating channel phosphorylation. PKC activity modulates a variety of voltage-dependent channels in different types of neurons (44 -46) and regulates other aspects of synaptic functions such as desensitization of receptors (47) and synaptic strength (41,42). The roles of specific PKC isozymes in such cellular processes are still under investigation. The conventional PKC-␣ isozyme has been shown to modulate several ion channels including volume-sensitive chloride channels (48), store-operated Ca 2ϩ channels (49), K Ca channels (50 -52), the TRPC1 channel, and store-operated Ca 2ϩ entry (53).
Activation of PKC is associated with intracellular relocalization of the enzyme from the cytosolic fraction to the plasma membrane (54). Both cPKCs and nPKCs are activated by phorbol esters, and the second C1 membrane-targeting motif within PKC is responsible for phorbol ester-dependent translocation. It is known that subcellular localization and functions of conventional PKCs are under the modulation of "priming" phosphorylation (55), although whether this applies to novel PKC isozymes is not yet clear. One attractive hypothesis is that Kv3.1b channels in different locations, such as presynaptic terminals and perisynaptic specializations (2,56), are associated with different PKC isozymes.
Association of Kv3.1b with specific PKC isozymes could also reflect differences in the subunit composition of the substrate channel proteins. Although the high threshold current at the soma of MNTB neurons is likely to represent a homotetramer of Kv3.1 subunits (14), many neurons, including MNTB neurons and the cochlear bushy cells that provide synaptic input to the MNTB, coexpress Kv3.1 with other Kv3 family members, i.e. Kv3.2, -3.3, and -3.4 (41,57), some of which are also regulated by PKC (58). The regulation of such Kv3.1b-containing heteromers, like that of other channel heteromers (59,60), is likely to differ from that of Kv3.1b homomers and may therefore also be associated with different PKC isozymes.
Although it is known that many neuronal ion channels are subject to modulation by protein kinases (61)(62)(63), the biological significance of such modulation is often obscure. The effect of changes in phosphorylation of Kv3.1b on neuronal firing patterns has a clear physiological interpretation, and the present data provide information on the PKC isozymes that control its phosphorylation state in different conditions. Further studies are required to determine the mechanisms that link the channels to specific enzymes, and the way that the phosphorylation state of different channels is integrated within a neuron to regulate its intrinsic excitability.