|Home | About | Journals | Submit | Contact Us | Français|
Transient outward K+ currents are particularly important for the regulation of membrane excitability of neurons and repolarization of action potentials in cardiac myocytes. These currents are modulated by protein kinase C (PKC) activation, and the K+ channel subunit, Kv4.2, is a major contributor to these currents. Furthermore, the current recorded from Kv4.2 channels expressed in oocytes is reduced by PKC activation. The mechanism underlying PKC regulation of Kv4.2 currents is unknown. In this study, we determined that PKC directly phosphorylates the Kv4.2 channel protein. In vitro phosphorylation of the intracellular amino (N)- and carboxyl (C)-termini of Kv4.2 glutathione S-transferase (GST) fusion protein revealed that the Kv4.2 C-terminal was phosphorylated by PKC, while the N-terminal was not. Amino acid mapping and site-directed mutagenesis revealed that the phosphorylated residues on the Kv4.2 C-terminal were Serine (Ser) 447 and Ser537. A phospho-site specific antibody showed that phosphorylation at the Ser537 site increased in the hippocampus in response to PKC activation. Surface biotinylation experiments revealed that alanine mutation to block phosphorylation at both of the PKC sites increased surface expression compared to wildtype Kv4.2. Electrophysiological recordings of the wildtype and both the alanine and aspartate mutant Kv4.2 channels expressed with KChIP3 revealed no significant difference in the half activation or inactivation voltage of the channel. Interestingly, the Ser537 site lies within a possible extracellular regulated kinase (ERK)/mitogen activated protein kinase (MAPK) recognition (docking) domain in the Kv4.2 C-terminal sequence. We found that phosphorylation of Kv4.2 by PKC enhanced ERK phosphorylation of the channel in vitro. These findings suggest the possibility that Kv4.2 is a locus for PKC and ERK cross-talk.
Regulation of transient outward K+ currents in cardiac myocytes and in neurons can have profound effects on membrane excitability. Kv4 family members are the primary subunits underlying the transient outward K+ currents (Ito) in dog and human ventricular myocytes  and Kv4.2 is the main subunit underlying this current in the rat [2, 3] and mouse ventricle myocytes [3, 4]. Moreover in neurons, Kv4.2 is the primary pore-forming subunit of the transient outward K+ currents (IA) in the dendrites of hippocampal pyramidal cells , visual cortical pyramidal neurons  and spinal cord dorsal horn neurons . These currents are regulated by kinase activation. IA in the dendrites of pyramidal and dorsal horn neurons is modulated by PKC activation [8, 9]. In many cell types, including neurons, PKC activates ERK/MAPK , and studies have shown that the effect of PKC on IA in area CA1 dendrites is mediated by ERK activation [9, 11]. This modulation of IA amplitude can regulate the peak of back-propagating action potentials and excitability of dendrites [5, 11, 12]. In addition, activation of a number of cell surface receptors regulate Ito in the heart, possibly through the common effector, PKC [13, 14]. For example, in lower mammals Ito of ventricular and atrial myocytes is regulated by alpha adrenergic receptor activation most likely through PKC activation [15-18].
PKC activation modulates other voltage-gated ion channels, including Na+ channels  and K+ channels [20-23]. These PKC effects on K+ channels include the conversion of a rapidly inactivating human Kv3.4 (h-Kv3.4) current to a non-inactivating current . An additional PKC effect is a reduction in currents mediated by both Kv4.2  and long-Kv4.3 (L-Kv4.3), a splice-variant of Kv4.3 that has a 19 amino acid insert that contains a PKC consensus sequence compared to short-Kv4.3 [13, 14]. In the case of h-Kv3.4 and L-Kv4.3, these effects are due to direct phosphorylation of the primary subunit [13, 22], however, the mechanism of the reduction in Kv4.2 current is unknown.
Activation of the PKC pathway can lead to phosphorylation of Kv4.2 channels and modulation of IA via activation of the ERK/MAPK pathway [9, 11, 24]. We propose that PKC may modulate IA by direct phosphorylation of the Kv4.2 subunit. In this study, we investigate whether the cytoplasmic regions of Kv4.2 N- and C-termini are directly phosphorylated by PKC and the functional effects of the phosphorylation events.
The original rat Kv4.2 and human KChIP3 cDNA were generously provided by Dr. P. J. Pfaffinger. Both constructs were in a cytomegalovirus vector. The C- and N-terminal glutathione S-transferase (GST) fusion proteins were expressed in pGEX vectors.
The amino- and carboxyl-terminal cytoplasmic domains of Kv4.2 were expressed in E. coli as (GST) fusion protein constructs, purified and phosphorylated in vitro following the methods previously described except for the following modifications . The recombinant proteins were incubated for 20 min at 37°C in reaction mixtures (25 μl) containing 20 ng of the catalytic domain of PKC (Calbiochem), Tris buffer 1 (50 mM Tris-HCl), pH 7.5, 0.5 mM EDTA, 0.5 mM EGTA, 1 mM Na4P2O7, 10 μg/ml aprotinin, 10 μg/ml leupeptin), and ATP mix (100.μM ATP, 100 mM MgCl2, and 10 μCi [γ-32P] ATP). Reactions were stopped by boiling for 5 min with sample buffer. The GST-fusion proteins were separated by SDS-PAGE (10%) and visualized by Coomassie blue staining. Phosphopeptides were identified by autoradiography. Parallel reactions were performed with GST alone. A time course of PKC phosphorylation of the Kv4.2 C-terminal peptide was performed.
In the case of phosphorylation of mutant C-terminal constructs, 32P incorporation into mutant and WT proteins were normalized to total protein for each construct. This was then represented as percentages relative to WT proteins (set at 100%).
Phosphorylation of the Kv4.2 C-terminal wildtype and mutant fusion proteins by ERK (MAP kinase) and PKC were performed as previously described with minor modifications . Briefly, the fusion proteins were incubated for 30 min at 37 °C with 10 μCi [γ-32P] ATP per reaction in the presence of activated ERK only (as control) and activated ERK plus PKC (Stratagene), HEPES buffer (in mM: 25 HEPES, 0.5 EDTA, 0.5 EGTA, 1 Na4P2O7, 10 μg/ml aprotinin, 10 μg/ml leupeptin), 10 mM MgCl2 and 100 μM ATP. The reactions were stopped and samples were loaded in the SDS-polyacrylamide gel and visualized by Coomassie blue staining. Phosphopeptides were identified by autoradiography. To quantify phosphorylation, autoradiographs were analyzed further by densitometry using NIH Image software. 32P incorporation was normalized to total protein for all constructs, which was then demonstrated as a percentage of control (ERK phosphorylation of WT construct) which was set at 100%.
The methodology for the phosphopeptide mapping is as previously described except for the following modifications: the reaction volume was increase by a factor of 10, specific activity was increased (20 μCi [γ-32P]/ 25 μl reaction volume and 50 μM ATP), and the incubation period was increased to 60 min based on the time course results. The phosphorylated C-terminal GST-fusion protein was separated by SDS-PAGE (10%). The Coomassie-stained band corresponding to the Kv4.2 C-terminal phospho-protein was excised and used for phospho-peptide mapping as previously described with minor modifications [25, 27]. An in-gel digest with trypsin or Lys-C was performed and following extraction from the gel, the peptides were separated using reverse phase HPLC (high-pressure liquid chromatography) with absorption monitoring at 214 nm. Counts/min (CPM) in each HPLC fraction was measured as Cherenkov radiation. Phospho-peptides identified as HPLC fractions containing high radioactivity were sequenced using automated Edmann degradation. Radioactivity was determined with each sequencing cycle by scintillation counting. Notable is that mapping was first attempted using tandem mass spectroscopy as previously described . However, using electrospray ionization mass spectroscopy (API 3000 LC/MS/MS System, PE Sciex, Thornhill, ON, Canada) we were unsuccessful at obtaining peptide sequences of the Kv4.2 C-terminal domain that contained the candidate PKC phosphorylation sites. Thus, we employed the methodology as described in the above paragraph.
Peptides were synthesized in the Protein Chemistry Core Laboratory (Baylor College of Medicine) which contained the PKC phosphorylation sites within the Kv4.2 C-terminal cytoplasmic domain. The synthetic peptides ANAYMQSKRNGLLC and SRRHKKSFRIPNAC (referred to hereafter as PKC-A and PKC-B peptides, respectively) had a total of 14 amino acid residues corresponding to Kv4.2 channel residues 441-454 (PKC-A) and residues 531-543 (PKC-B) within the Kv4.2 sequence. The phosphorylation site was located in the middle (amino acid number 7) of the peptide, and a cysteine residue for coupling to carrier proteins was located at the carboxyl-terminus (amino acid number 14) of each peptide.
We performed kinetic characterization of the PKC phosphorylation sites using the synthetic peptides in kinase assays as previously described . The catalytic domain of PKC (20 ng) was used and reactions were incubated at 25°C. The assays used to determine Km and Vmax were linear with respect to time and linear with added kinase (PKC), and less than 10% of the peptide substrate was converted to product. To obtain the concentration curve for each of the peptides, peptide concentrations ranging from 5-400 μM were used. As a control parallel reactions were included using the PKC substrate, a synthetic peptide analogue of a fragment of neurogranin .
The PKC-A and PKC-B synthetic peptides containing the phosphorylated Ser447 and Ser537 PKC sites, respectively were coupled to Keyhole Limpet Hemocyanin and injected into rabbits according to standard protocol at Cocalico Biological Labs (Reamstown, PA). The antisera were screened by western blotting using the phosphorylated and unphosphorylated ovalbumin-coupled synthetic peptides and GST-fusion proteins. The antisera were affinity purified against the phosphorylated PKC-A and PKC-B synthetic peptides using Hi-trap columns (Amersham Pharmacia Biotech) .
The Kv4.2 point mutations were made using the site-directed mutagenesis kit (Stratagene, La Jolla, CA) as previously described [28, 30]. The following primers were used for mutant constructs: S447A: FORWARD 5′-GCAAATGCCTACATGCAGGCCAAGCGGAATGGGTTAC-3′, REVERSE 5′-GTAACCCATTCCGCTTGGCCTGCATGTAGGCATTTGC-3′; S447D: FORWARD 5′-GTGCAAATGCCTACATGCAGGACAAGCGGAATGGGTTACTGAGC-3′, REVERSE 5′-GCTCAGTAACCCATTCCGCTTGTCCTGCATGTAGGCATTTGCAC-3′; S537A: FORWARD 5′-CGGAGACACAAGAAGGCTTTCCGAATC-3′, REVERSE 5′-GATTCGGAAAGCCTTCTTGTGTCTCCG-3′; S537D: FORWARD 5′-GCTCACGGAGACACAAAAAAGATTTCCGAATCCCAAATGCC-3′, REVERSE 5′-GGCATTTGGGATTCGGAAATCTTTTTTGTGTCTCCGTGAGC-3′ (bold underline denotes a difference compared to the original sequence). All mutations were confirmed by DNA sequence analysis through the entire coding region.
Oocytes were harvested as described previously . After ~24 h, oocytes were injected with 3–10 ng of DNA Kv4.2 [wild type (WT) or mutants] + KChIP3 in a 1-to-1 ratio using a Nanoject microinjector (Drummond Scientific), into the nucleus of stage V—VI oocytes. Currents were recorded after 2 days under two-electrode voltage clamp with an Axoclamp 2A amplifier (Axon Instruments) at room temperature. Microelectrodes were pulled from filamented glass (1.5 mm × 0.86 mm; A-M Systems) filled with 3 M KCl. The current electrode had a resistance of 0.30-0.50 MOhms, whereas the voltage electrode ranged from 0.3 to 1.0 MOhms.
Currents were leak subtracted online with P/4 leak subtraction. Data were digitized at 2 kHz and stored on a computer equipped with Digidata 1200 software. Current protocols used to obtain data included: 1) activation: hyperpolarization to −110 mV and then depolarization to +40 mV for 400–800 ms, repeated in −5- or −10-mV step intervals; 2) inactivation: depolarization to 0 mV and then hyperpolarization to −110 mV for 650 ms, changing this step by +5-mV intervals, then depolarization to 0 mV.
The chamber was continuously perfused at a rate of 3–6 ml/min with ND-96 (in mM: 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, and 5 HEPES pH 7.4 with NaOH). Oocytes expressing mutant DNA and WT (control) DNA were always recorded on the same day and were recorded in at least three batches of oocytes. The data from oocytes expressing WT DNA from different days were not different; therefore, all WT data were combined. Data were analyzed with Clampfit, Origin, and Prism programs. Peak currents were obtained and conductance (G) was determined with a reversal potential (Vrev) of −95 mV according to the equation G = Ip/(Vc − Vrev), where Ip is the peak of the current at a given voltage command (Vc). Activation and inactivation curves were fit with a Boltzmann sigmoidal curve with the equation G/Gmax = 1/[1 + exp(X − V1/2)/slope], where X is equal to the test potential (Vm) and Gmax is maximum conductance. The mean ± SE voltage at which half the currents are activated (V1/2) was determined from the Boltzmann fit and compared among the mutants and WT with a One-Way ANOVA and post hoc test.
COS7 cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 50 units/ml penicillin, 50 μg/ml streptomycin ( all from Invitrogen) in a humidified incubator at 37°C under 5% CO2. Cells were maintained in poly-L-lysine-coated 6-well plates. Cells were transfected with mammalian expression vectors for Kv4.2 with FuGene6 (Roche) transfection reagents using the manufacturers’ protocols. Cells expressing Kv4.2 wildtype and Ser to Alanine (Ala) and Ser to Aspartate (Asp) mutants were stained 48 hours post-transfection using an immunofluorescence protocol . Briefly, after fixing cells with 4% paraformaldehyde for 30 min, and incubation in 0.3% Triton X-100 in PBS for 20 min at room temperature, the cells were blocked by 10% fetal bovine serum in PBS for 60 min at room temperature, and then incubated for 60 min at room temperature or overnight at 4°C with primary antibodies: polyclonal Kv4.2 antibody from Alomone Labs (Jerusalem, Israel) generated against residues 454-469 (SNQLQSSEDEPAFVSK) on the C-terminal or a monoclonal Kv4.2 antibody against an ectodomain . The Alexa 488-conjugated goat anti-rabbit IgG (Molecular Probes) secondary antibody was used for detection of primary antibody binding. Microscopic analysis was performed with a Nikon TE200 inverted fluorescence microscope equipped with a digital camera (Coolmax Fx, RS Photometrics, Tucson, AZ).
Kv4.2 WT+KChIP3 or Kv4.2 AA (Ser447Ala/Ser537Ala) mutant + KChIP3 - transfected COS7 cells were washed with ice-cold PBS containing 0.1 mM CaCl2 and 1 mM MgCl2 (PBS/Ca2+ plus Mg2+, pH 8.0) then treated with sulfo-NHS-LC-biotin (0.5 mg/ml, Pierce) at 4 °C for 40 min as described previously . Cells were lysed in radioimmune precipitation assay buffer (RIPA -150 mM NaCl, 0.5% sodium deoxycholate, 20 mM Na2HPO4, 1% Triton X-100, 0.1% SDS, pH 7.4), supplemented with 0.01 mM PMSF, 0.005 μg/ml leupeptin, and 0.005 μg/ml pepstatin for 1 h at 4 °C. Lysates were centrifuged at 20,000 × g for 30 min at 4 °C, and the protein concentration in the supernatants was determined using the BCA assay kit (Pierce). UltraLink Immobilized NeutrAvidin beads (50 μl, Pierce) were added to each group sample, and the mixture was incubated for 1 h at room temperature or 4 °C overnight. The beads were washed four times with cold RIPA buffer and eluted by boiling with 50 μl of Laemmli loading buffer (Bio-Rad) for 5 min at 95 °C. The eluates (25 μl) were resolved by SDS-PAGE gel and immunoblotted with the Kv4.2 antibody. Immunoreactive bands were visualized using horseradish peroxidase-conjugated secondary antibody (1:10,000). Immunoreactivity values of surface Kv4.2 channels were normalized to levels of actin immunoreactivity in total cell extracts to preclude errors that accompany sample loading and transfer.
The N- and C-terminal phospho-selective antibodies were used to evaluate modulation of PKC phosphorylation of Kv4.2 in rat hippocampal slices. Transverse hippocampal slices were prepared using the methodology of Roberson et al. . Slices were incubated in artificial cerebrospinal fluid (aCSF) containing in mM: NaCl 125; KCl 2.5; NaH2PO4 1.25; NaHCO3 25; CaCl2 2; MgCl2 1; glucose 25) at 32°C for approximately 1 hour. Slices were then incubated in vehicle (DMSO) or 10 μM PDA for 10 minutes. Slices were rapidly frozen on dry ice. The slices were subsequently sonicated, normalized according to protein assay, and then centrifuged (Beckman, 100,000g, 20 min, 4°C) and the pellet was resuspended in the 10% SDS solution with 200mM DTT and protease/phosphatase inhibitors as previously described . After resuspension of the pellet sample buffer was added. The hippocampal membrane proteins were then used for western blotting. 4-5 slices were used for each condition and pooled for sonication.
Immunoreactivity was measured using densitometry (public domain, Scion Image). Densitometry data were analyzed with a Student’s t-test, Mann Whitney test for nonparametric data analysis or with ANOVA and post-hoc test. The GraphPad Prism software package was used for statistical analysis. Error bars represent standard error of the mean.
The catalytic subunit of PKC (rat brain) was obtained from Calbiochem. Glutathione agarose bead and [γ-32P] ATP were obtained from Amersham Pharmacia Biotech. The FuGene 6 Transfection Reagent was obtained from Boehringer Mannheim. PDA (phorbol diacetate) was obtained from Sigma.
The PKC-A and PKC-B peptides were synthesized in the Protein Core Chemistry Laboratory at Baylor College of Medicine, Houston TX.
Voltage-gated K+ channel primary subunits are characterized by 6 transmembrane domains with prominent intracellular N- and C-termini . A schematic diagram of the putative Kv4.2 topology (Figure 1A) and the Kv4.2 amino acid sequence (Figure 1B) are shown. Analysis of the amino acid sequence of Kv4.2 with NetPhosK (http://www.cbs.dtu.dk/services/NetPhosK/)  predicted multiple serine and threonine residues (greater than 50% probability) as candidate PKC phosphorylation sites. Candidate sites for PKC phosphorylation are shown in bold and italic. Four consensus sites for PKC phosphorylation exist in the N-terminal and six consensus sites exist in the C-terminal domains of Kv4.2. The putative consensus PKC sequence is Ser/Thr-X-Arg/Lys, where X symbolizes an uncharged residue . The presence of PKC consensus sites within the N- and C-termini suggested that PKC may phosphorylate Kv4.2 at both intracellular termini.
To determine if the Kv4.2 N- and C-termini are directly phosphorylated by PKC, Kv4.2 GST-fusion proteins of the putative intracellular N- and C-termini were expressed in bacteria and purified. The resulting purified proteins were incubated with PKC and 10μCi [γ-32P] ATP in vitro (see methods). Reaction products were separated on an SDS-PAGE gel. Figure 2A shows the bands corresponding to the N- and C-termini GST fusion proteins on the Coomassie Blue stained gel (left panel — Coomassie). The autoradiogram (left panel — Autorad) revealed 32P incorporation into the C-terminal only, suggesting that the C-terminal but not the N-terminal of Kv4.2 was phosphorylated by PKC. In addition, purified GST protein was used as a control and incubated with 32P and PKC. There was no 32P incorporation into the GST protein, indicating that GST was not phosphorylated by PKC (data not shown). In order to determine the optimal time for the saturation of PKC phosphorylation of the Kv4.2 C-terminal substrate, a time course of PKC phosphorylation was done (Figure 2A; right panel). Saturation of phosphorylation is desirable for mapping the phosphorylation sites in order to be certain that sites are not missed during mapping. Phosphorylation of the Kv4.2 C-terminal reached saturation levels in 40 minutes. Therefore, incubations of 60 minutes ensured complete phosphorylation of the protein.
To determine the individual PKC phosphorylation sites in the C-terminal of Kv4.2, phosphopeptide mapping and protein sequencing was performed. For these studies Kv4.2 C-terminal fusion protein was incubated with PKC in vitro as described above except that the reaction volume was increased by a factor of 10. The reaction products were separated on an SDS-PAGE gel. The phosphorylated protein bands were excised, eluted and enzymatically digested. Following digestion, the peptides were separated using reverse phase high pressure liquid chromatography (HPLC) with absorption monitoring at 214, 254, and 280 nm (Figure 2B; left panel — 214 nm shown). Counts per min (cpm) in each HPLC fraction were measured as Cerenkov radiation . The phosphopeptide map for the C-terminal construct demonstrated two peaks in radioactivity, in HPLC fraction 17-19 and 67-75 (Figure 2B; right panel). HPLC fractions 17-19 (black arrow on HPLC trace) and 67-75 (gray arrows on HPLC trace) were sequenced using Edman degradation. Two phosphorylation sites were identified in the C-terminal construct at residues Ser447 and Ser537 (Figure 2C; left and right panels, respectively).
Kinetic characterization of the PKC sites was performed using synthetic peptides (see methods) containing the phosphorylation site and 6 flanking residues on either side. The concentration curves for the PKC Ser537 site revealed that the Kv4.2 Ser537 site was a good substrate for PKC with a Km of 6.7 μM and a Vmax of 771 nmol/min/mg of protein (averages of two experiments performed in duplicate under steady-state conditions, data not shown). The kinetic characterization for the Ser447 site was not obtained since there was no 32P incorporation into the synthetic peptide containing this site. A possible explanation for the lack of 32P incorporation into this synthetic peptide is that there are tertiary structural determinants necessary for phosphorylation at this site that are disrupted in the small (14 amino acids) peptide used for the kinetic experiments. The full C-terminal GST fusion protein construct was not used for the kinetic experiments as it remained on the beads and thus protein concentrations could not accurately be measured, which would be a critical aspect for determining the kinetic values. An additional possible explanation for the lack of phosphorylation in the peptide is that phosphorylation of the Ser537 site may be necessary for phosphorylation at the Ser447 site. Our data from phosphorylation of alanine site mutants suggest that this, however, is not the case. The Ser447 site is still phosphorylated in the Ser537A site mutant protein (Figure 3).
These mapped sites were confirmed using site-directed mutagenesis of the C-terminal GST fusion protein. Ser537 and Ser447 of the Kv4.2-GST fusion protein were mutated to an alanine to block phosphorylation. PKC incubation in vitro with the mutant C-terminal construct in which both Ser537 and Ser447 were mutated to an alanine (Ser447A, Ser537A, AA) showed no 32P incorporation (Figure 3A), suggesting that these sites were indeed the only sites within the C-terminal that were phosphorylated by PKC (Figure 3B). Furthermore, mutation of the individual sites (Ser447A alone or Ser537A alone) showed that both sites were phosphorylated by PKC and phosphorylation of each site was sufficient to mimic phosphorylation levels of the WT construct. Interestingly, each site was phosphorylated equally well in the absence of phosphorylation at the other site. These data provide additional support that the C-terminal is indeed phosphorylated at both Ser447 and Ser537 in vitro.
We generated site-selective phospho-specific antibodies to determine if PKC phosphorylates the native Kv4.2 channel in the hippocampus at the sites that were identified. Two synthetic phosphorylated peptides were produced, corresponding to amino acids 441ANAYMQSKRNGLLC454 and 531SRRHKKSFRIPNAC543 with phosphate groups attached to the appropriate serine residues (Ser447 and Ser537, respectively). Antisera against the two peptides were generated in separate rabbits. After the terminal bleed, the antisera were affinity-purified using columns with the appropriate peptide. Antibodies were first screened for immuno-reactivity to the synthetic peptides (phosphorylated and unphosphorylated) that were coupled to ovalbumin for the western blotting. Both the PKC-A antibody (Ser447; Figure 4A — left) and the PKC-B antibody (Ser537; Figure 4B — left) were specific for the phosphorylated peptide, as there was no recognition of the unphosphorylated peptide. In addition, each antibody was specific for the peptide for which it was made, and there was no cross-reactivity with the peptide generated for the other site (data not shown). The antibody was tested against the C-terminal fusion protein. While the PKC-A antibody was not phospho-specific for the phosphorylated fusion protein (Figure 4A — right), the PKC-B antibody was indeed specific for the C-terminal fusion protein only after phosphorylation by PKC, indicating that the PKC-B antibody was phospho-specific (Figure 4B — right).
The PKC-B antibody was further tested in the hippocampus to determine if it recognized the full-length phosphorylated Kv4.2 protein and to determine if Kv4.2 is phosphorylated at Ser537 in the hippocampus. Rat hippocampal slices were treated with PDA (10 μM) to stimulate PKC activation or vehicle (control) and then the membranes were extracted for Western blotting. The PKC-B antibody recognized a band at 69 kDa, suggesting basal phosphorylation of Kv4.2 at that site (Figure 4C; Control). In addition, the immunoreactivity of the PKC-B antibody increased in response to PDA application (+96 ± 31% over vehicle DMSO, p < 0.05, n = 4-6, Figure 4C lower), suggesting increased phosphorylation at Ser537 within Kv4.2 sequence in response to PKC activation.
PKC activation causes a reduction in IA in the dendrites of hippocampal pyramidal neurons  and dorsal horn neurons , and Kv4.2 is the major pore-forming subunit of IA in these neurons [5, 7]. Furthermore, activation of PKC with phorbol esters decreases the amplitude of Kv4.2 and Kv4.3 currents in Xenopus oocytes . This could be due to direct PKC phosphorylation of the channel at the sites we have mapped. In order to evaluate the functional effects of the PKC phosphorylation sites within Kv4.2, we generated alanine (A - to block phosphorylation) and aspartate (D - to mimic phosphorylation) mutants of the phosphorylation sites. We made single site mutants (Ser447Ala or Asp or Ser537Ala or Asp) and mutants where both sites were mutated to alanines (AA mutant) or aspartates (DD mutant). These mutant channel constructs and the WT construct were transfected into the COS7 expression system to compare expression and localization of the channel protein. Immunohistochemistry of COS7 cells expressing the various mutant channel constructs revealed that mutation of Ser447 to Ala or Asp or Ser537 to Ala or Asp or both sites to an Ala (Ser447A/Ser537A, AA mutant) or Asp (Ser447D/Ser537D, DD mutant) had no significant effect on channel protein expression (data not shown). However, in order to more specifically study changes in surface localization, we evaluated surface expression of the WT and double AA mutant channels using surface biotinylation (see methods). We observed that there was a significant increase in the surface expression of the AA mutant compared to the WT (+61 ± 25% over WT, p = 0.028, Mann-Whitney test, n=4; Figure 5A). These data suggest that basal phosphorylation of the channel at the Ser447 and Ser537 sites reduces channel surface expression. This reduction in channel surface expression provides a possible mechanism for the reduction in current in response to activation of PKC previously observed .
Next, we evaluated the effects of mutation of both of these sites (AA or DD mutant channel) on the kinetics of the Kv4.2 current. Recordings were conducted in oocytes expressing the WT and mutant (AA and DD) channels with KChIP3. There was no significant difference in the kinetics of the WT channel control for AA and DD mutants, therefore, the WT data were combined. Figure 5B shows typical current responses to voltage depolarizations from the WT channel and AA and DD mutant channels. Figure 5C is the summary I-V curve for the WT and mutant channels. Consistent with the increase in surface expression of the AA mutant channel, the average peak current was significantly greater in the AA mutant channel recordings compared to WT and DD mutant channels (p<0.001; Repeated Measures One Way Anova with post-hoc Tukey’s). Furthermore, no significant differences in the half-activation (see table 1 — F(2,39) = 0.31; p = 0.74) or half-inactivation voltage (see table 1 — (F2,27) = 0.28; p = 0.76) were observed between WT and mutant (AA or DD) channels. Together, these data suggest that surface expression of the AA mutant is enhanced, but the kinetics of the channel is not modulated in the mutant channels.
Interestingly, the Ser537 PKC phosphorylation site lies within a putative ERK docking domain in the Kv4.2 C-terminal sequence (Figure 6A), which consists of a series of basic amino acid residues  corresponding to amino acid residues 532-540 in the Kv4.2 sequence. Because the Ser537 site is localized within this domain, we hypothesized that PKC phosphorylation of the Kv4.2 channel could modulate ERK phosphorylation of the channel. To test the possibility we evaluated the interaction of PKC and ERK phosphorylation of the Kv4.2 C-terminal fusion protein construct in vitro. The C-terminal GST fusion protein was incubated with activated PKC for 15 minutes, and subsequently incubated with activated ERK to determine the effect of prior PKC phosphorylation of the C-terminal on ERK phosphorylation. Prior phosphorylation of the C-terminal GST fusion protein by PKC caused a significant increase in phosphorylation of the Kv4.2 C-terminal by ERK (157 ± 47% of PKC/ERK phosphorylation vs 100 ± 45% of ERK phosphorylation alone, p < 0.05, Figure 6A). This effect was blocked in the Ser537A mutant (157 ± 47 % vs. 72 ± 28 %, p < 0.05, Figure 6A) and the AA mutant C-terminal construct (157 ± 47 % vs. 42 ± 12 %, p < 0.01, Figure 6A). These data suggest that PKC phosphorylation of the C-terminus of Kv4.2 facilitates ERK phosphorylation of this protein and that phosphorylation of the Ser537 site is necessary for this effect.
We have previously shown that ERK phosphorylates the Kv4.2 C-terminal at 3 amino acid residues, Thr602, Thr607 and Ser616  and that there are functional effects of phosphorylation of these sites . In addition, phospho-specific antibodies for the individual ERK phosphorylation sites were generated. These antibodies were used to determine whether PKC activation affects ERK phosphorylation at one, two or all three of these sites within the Kv4.2 C-terminal GST fusion protein. We found that there was a significant increase in ERK phosphorylation at the Thr602 and Thr607 sites, but not at Ser616 in response to prior phosphorylation of the GST fusion protein by PKC (Figure 6B). Densitometry showed that phospho-specific immunoreactivity at the Thr602 [+84 ± 46% over WT control (PKC-, ERK+), p < 0.05, n=3] and the Thr607 [+68 ± 24% over WT control (PKC-, ERK+), p < 0.05, n=3] sites significantly increased, while phospho-specific immunoreactivity against the Ser616 site was unaffected (Figure 6B). These data suggest that PKC phosphorylation of the Kv4.2 C-terminus enhances phosphorylation at two of the ERK phosphorylation sites, Thr602 and Thr607 in the Kv4.2 C-terminal.
In this study we showed that the C-terminal of Kv4.2 is phosphorylated by PKC at two sites, Ser447 and Ser537. Using a phospho-selective antibody against the Ser537 PKC site within Kv4.2 we determined an increase in Kv4.2 phosphorylation at this site following PKC pathway activation in hippocampal slices, indicating that PKC couples to Kv4.2 phosphorylation in a system where it is natively expressed. Additional functional data were evident in the COS7 expression system using phospho-site mutants. These studies suggested that PKC phosphorylation of Kv4.2 regulates surface expression of the channel. These studies showed that when the PKC sites within Kv4.2 were mutated to alanines to block phosphorylation there was a significant increase in surface expression of Kv4.2 compared to wildtype channels. In addition, PKC phosphorylation of Kv4.2 enhanced ERK phosphorylation of the channel in vitro, suggesting the possibility that Kv4.2 is a locus for PKC and ERK cross-talk.
Mimicking or blocking phosphorylation of Kv4.2 at the PKC sites (Ser447 and Ser537) did not modulate the kinetics of the Kv4.2-mediated current, as demonstrated by electrophysiological recordings of the DD and AA Kv4.2 channel mutants. Specifically, there were no differences in the half-activation and -inactivation voltages between the WT and double mutant channels (AA and DD). However, as noted above we demonstrated that surface expression of the AA mutant increased as indicated by an enhanced peak current in oocytes expressing the AA mutant channel. This suggests that basal phosphorylation of the channel by PKC (or ERK/MAPK) reduces trafficking or maintenance of the channel in the membrane.
We found that prior phosphorylation of the Kv4.2 C-terminal by PKC caused an increase in ERK phosphorylation at two of the three ERK/MAPK sites, Thr602, Thr607, but not Ser616 . This is interesting in the context of our previously reported results which showed that phosphorylation at 602 and 607 sites had an overall inhibitory effect on the current, which mimicked the effect of ERK/MAPK activation seen in the dendrites of CA1 pyramidal neurons , whereas phosphorylation at Ser616 caused an opposite effect . The Thr607Asp (T607D) mutation mimicked the effect of ERK seen in neurons, which includes a right shift in the activation curve and an overall decrease in current. This suggests that at least the Thr607 site is the relevant site phosphorylated in response to PKC activation in neurons . Moreover, the Thr607 site effect is dominant when all 3 sites are phosphorylated. Phosphorylation of the C-terminal by PKC prior to ERK may augment phosphorylation at the Thr607 site, increasing the balance of phosphorylation at that site over the Ser616 site to ensure a decrease in current. This does not preclude phosphorylation at Ser616 in response to different physiological stimulation.
Previous data suggest that filamin is a scaffold protein that links Kv4.2 and the actin cytoskeleton. Furthermore, mutation of amino acids 601-604 region in the Kv4.2 C-terminal to alanines completely abolished the interaction of Kv4.2 and filamin in a yeast two-hybrid assay . This suggests that this region is important for the interaction of Kv4.2 with filamin and possibly for Kv4.2 localization and stabilization in the membrane. Phosphorylation of the ERK amino acid residues in this region on the Kv4.2 C-terminal (Thr602 and Thr607) may alter the interaction of filamin and Kv4.2 such that Kv4.2 becomes unstable at the surface membrane. Enhanced phosphorylation of these sites mediated by phosphorylation by PKC may play a role in ‘tagging’ Kv4.2 subunits for internalization. Further studies are necessary to determine the exact role of the interaction of PKC and ERK phosphorylation sites and the interaction with filamin.
Kv4.2 is the primary subunit that contributes to IA in the dendrites of CA1 pyramidal cells, and the threshold for induction of synaptic plasticity in the CA1 pyramidal cell dendrites is lower in the Kv4.2 KO animals as compared to control . This suggests that a decrease in IA can augment induction of synaptic plasticity. IA is regulated by activation of PKC, and the effect is blocked by the ERK/MAPK inhibitor U0126 . This suggests the possibility that direct phosphorylation of the Kv4.2 subunit by both PKC and ERK/MAPK may modulate the current. Indeed, our phospho-site specific antibody demonstrates that Kv4.2 is phosphorylated at Ser537 in the hippocampus in response to PKC activation.
We have previously shown that the Kv4.2 current is modulated by direct phosphorylation of Kv4.2 by ERK/MAPK . These data provide further support for interplay in the regulation of the A-type currents by various kinases and provide another possibility of Kv4.2 as a site of signal integration in neurons and cardiac myocytes. The modulation of ERK phosphorylation of the channel by PKC phosphorylation may be a mechanism of fine-tuning the current in response to a physiological stimulus. Moreover, a greater decrease in current mediated by both direct phosphorylation of the channel by PKC and ERK/MAPK may be a mechanism to dynamically regulate current amplitude and induction of synaptic plasticity.
This project described was supported by grants from the NIH/NIMH MH064620 to LAS; and NIH/NINDS NS039943 to AEA, NS37444 to JDS, and the Child Neurology Foundation (AEA and LAS). The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institute Neurologic Disease and Stroke or Mental Health or the National Institutes of Health.