Regulation of Surface Localization of the Small Conductance Ca2+-activated Potassium Channel, Sk2, through Direct Phosphorylation by cAMP-dependent Protein Kinase*

Small conductance, Ca2+-activated voltage-independent potassium channels (SK channels) are widely expressed in diverse tissues; however, little is known about the molecular regulation of SK channel subunits. Direct alteration of ion channel subunits by kinases is a candidate mechanism for functional modulation of these channels. We find that activation of cyclic AMP-dependent protein kinase (PKA) with forskolin (50 μm) causes a dramatic decrease in surface localization of the SK2 channel subunit expressed in COS7 cells due to direct phosphorylation of the SK2 channel subunit. PKA phosphorylation studies using the intracellular domains of the SK2 channel subunit expressed as glutathione S-transferase fusion protein constructs showed that both the amino-terminal and carboxyl-terminal regions are PKA substrates in vitro. Mutational analysis identified a single PKA phosphorylation site within the amino-terminal of the SK2 subunit at serine 136. Mutagenesis and mass spectrometry studies identified four PKA phosphorylation sites: Ser465 (minor site) and three amino acid residues Ser568, Ser569, and Ser570 (major sites) within the carboxyl-terminal region. A mutated SK2 channel subunit, with the three contiguous serines mutated to alanines to block phosphorylation at these sites, shows no decrease in surface expression after PKA stimulation. Thus, our findings suggest that PKA phosphorylation of these three sites is necessary for PKA-mediated reorganization of SK2 surface expression.

Small conductance, Ca 2؉ -activated voltage-independent potassium channels (SK channels) are widely expressed in diverse tissues; however, little is known about the molecular regulation of SK channel subunits. Direct alteration of ion channel subunits by kinases is a candidate mechanism for functional modulation of these channels. We find that activation of cyclic AMP-dependent protein kinase (PKA) with forskolin (50 M) causes a dramatic decrease in surface localization of the SK2 channel subunit expressed in COS7 cells due to direct phosphorylation of the SK2 channel subunit. PKA phosphorylation studies using the intracellular domains of the SK2 channel subunit expressed as glutathione S-transferase fusion protein constructs showed that both the amino-terminal and carboxylterminal regions are PKA substrates in vitro. Mutational analysis identified a single PKA phosphorylation site within the amino-terminal of the SK2 subunit at serine 136. Mutagenesis and mass spectrometry studies identified four PKA phosphorylation sites: Ser 465 (minor site) and three amino acid residues Ser 568 , Ser 569 , and Ser 570 (major sites) within the carboxyl-terminal region. A mutated SK2 channel subunit, with the three contiguous serines mutated to alanines to block phosphorylation at these sites, shows no decrease in surface expression after PKA stimulation. Thus, our findings suggest that PKA phosphorylation of these three sites is necessary for PKA-mediated reorganization of SK2 surface expression.
The small conductance, Ca 2ϩ -activated K ϩ (SK) 2 channels are found in both neuronal and non-neuronal tissue (1). Functionally, the SK channels are best characterized in the central nervous system. Three genes encode the SK channel subunits (SK1, SK2, and SK3) in mamma-lian brain (2). SK channels are blocked by the bee venom toxin, apamin, although SK1 is slightly less sensitive than SK2 and SK3 (2)(3)(4). In neurons throughout the nervous system, the apamin-sensitive SK channels modulate firing frequency by contribution to the afterhyperpolarization (AHP) that follows a single or a train of action potentials (5)(6)(7). SK2 is thought to specifically underlie the medium AHP current (I mAHP ) in hippocampal CA1 pyramidal cells (8). The I mAHP is Ca 2ϩ -dependent with a time constant of 100 -250 ms and sensitivity to apamin (9 -11). The apamin-sensitive I mAHP modulates instantaneous firing rates and sets the interspike duration in action potential trains to produce spike frequency adaptation (12). In addition, SK channels are localized to the dendrites of pyramidal cells in hippocampal area CA1 and pyramidal neurons of the lateral amygdala where they function to shape synaptic potentials and limit Ca 2ϩ influx through NMDA receptors (13,14) as well as plateau potentials evoked by exogenous glutamate application (15). Other studies have linked overexpression of SK2 and enhancement of the I mAHP in hippocampus with neuroprotection (16), attenuation of hippocampal LTP, and memory deficits (16,17). These functions have major implications for a role for SK2 channels in the formation of synaptic plasticity and learning and memory (17,18).
In addition to the important role in pyramidal neurons, SK channels are critical for regulation of firing frequency in neurons in other parts of the nervous system. In magnocellular neurons of the hypothalamus, the characteristic bursting pattern of oxytocinergic cells that precedes milk ejection is modulated by a Ca 2ϩ -dependent, apamin-sensitive I mAHP that appears to be formed by SK3 channels (19,20). Furthermore in immature cerebellum, SK2 channels are highly expressed and determine the spontaneous firing pattern of Purkinje cells (21). In retinal ganglion cells, the apamin-sensitive I AHP that controls firing pattern and excitability is mediated by SK2 channel subunits (22). Hence, understanding the functional regulation of SK2 channel subunits would yield valuable information about the molecular regulation of the I mAHP .
Currently, the molecular mechanism of modulating the I mAHP and neuronal firing frequency is poorly understood. A number of studies have shown that another component of the AHP current, the slow AHP current is regulated by kinase cascades (23,24). One intriguing possible molecular mechanism of regulating the I mAHP is direct kinase phosphorylation of the SK channel subunits conducting the I mAHP . Indeed, direct K ϩ channel phosphorylation is a well known post-translational modification regulating neuronal excitability. Phosphorylation of K ϩ channels is associated with modulation of channel kinetics and cellular localization changes (25). In addition, recent studies have suggested that activity-dependent potassium channel phosphorylation and functional modulation may be controlled by net kinase activity (26).
In the present study, we show that activation of PKA causes a decrease in SK2 surface expression. Furthermore, the PKA phosphorylation sites within the SK2 channel subunit were mapped, and we developed a phospho-selective antibody based on the identified sites. The phospho-selective antibody was used as a tool for further study of PKA modulation of the SK2 subunit. We showed that SK2 was phosphorylated in the COS7 cell expression system and that direct PKA phosphorylation of SK2 subunits in COS7 cells leads to altered cellular localization of the channel subunits, suggesting that the SK2 channel is a downstream target of the PKA cascade. These studies advance our understanding of the molecular regulation of SK channels by posttranslational modification.
Protein Expression and Purification-GST fusion protein constructs of the SK2-NT and -CT (GST SK2-NT and GST SK2-CT, respectively) were expressed in Escherichia coli BL21 (DE3) using methods previously described (28). Briefly, single colonies of BL-21 (DE3) cells transformed with plasmids containing the NT, CT, or second intracellular loop domains were selected and grown in LB broth. After growing to an optical density of 0.6 -0.8 (A 600 ), the bacteria were induced by incubation at room temperature with 200 M isopropyl ␤-D-thiogalactopyranoside (Sigma) for 4 h and were harvested by centrifugation at 3,000 ϫ g for 10 min. The cells were resuspended and incubated in STE buffer (in mM): 10 Tris-HCl, pH 8.0, 1 EDTA, 150 NaCl, containing Protease Inhibitor Mixture (Sigma) and 100 g/ml lysozyme (Sigma) for 15 min on ice. 10 mM dithiothreitol and 1.5% N-laurylsarcosine were added, and then the sample was sonicated for a total of 2 min. The lysates were then centrifuged (10,000 ϫ g, 20 min, 4°C) and adjusted to 0.7% N-laurylsarcosine and 2% Triton X-100. The GST fusion proteins were purified using glutathione affinity absorption. Glutathione-agarose beads (Amersham Biosciences) were washed, resuspended in phosphate-buffered saline (PBS, pH 7.4; in mM: 137 NaCl, 2.7 KCl, 4.3 Na 2 HPO 4 , 1.4 KH 2 PO 4 ) and then incubated with the lysates for 1 h at room temperature or overnight at 4°C. The beads were washed three times with PBS buffer by centrifugation (100 ϫ g, 5 min, 4°C). After the final wash, the bead preparation was resuspended in PBS buffer containing Protease Inhibitor Mixture. The recombinant proteins were left on the beads for subsequent experiments and stored at 4°C.
PKA Phosphorylation of GST SK2 Fusion Proteins-GST SK2-NT or -CT fusion proteins were incubated for 30 min at 37°C in reaction mixtures (50 l) containing 70 ng of the catalytic subunit of PKA (Sigma), Tris buffer, and ATP mix buffer(100 M ATP, 100 mM MgCl 2 , and 10 Ci of [␥-32 P]ATP). For time-course studies of phosphorylation of the GST SK2-NT and -CT fusion proteins incubation periods of 5, 10, 30, 60, and 90 min were used. Reactions were stopped by boiling for 5 min at 95°C with sample buffer (30 mM Tris-HCl, pH 6.8, 200 mM dithiothreitol, 40% glycerol, 8% SDS, 0.04 mg/ml bromphenol blue). The GST fusion proteins were separated by 12.5% SDS-PAGE and visualized by Coomassie Blue staining. Phosphopeptides were identified by autoradiography. As a control, parallel reactions were performed for GST alone and PKA alone with ATP mix buffer. The optical density of autoradiography was normalized to the corresponding Coomassie bands (29).
Phosphopeptide Mapping-Identification of the specific amino acids phosphorylated by PKA within the SK2 CT was performed by a modified tandem mass spectroscopy method in the Protein Chemistry Core Facility at Baylor College of Medicine as previously described (30). The product of a 60-min preparative scale reaction (volume of 300 -400 l) of activated PKA, the GST SK2-CT, and ATP was separated by SDS-PAGE (12.5%) stained by Coomassie Blue and digested in-gel by Lys-C. The separated peptide fragments were desalted using a C 18 ZipTip column (Millipore) then analyzed by electrospray ionization mass spectroscopy (API 3000 LC/MS/MS System, PE Sciex, Thornhill, Ontario, Canada). Phosphopeptides were identified by a Ϫ79-dalton precursor ion scan and subsequently sequenced with tandem mass spectrometry.
Phospho-antibody Production-Synthetic peptides corresponding to SK2 residues 558 -580, which included the identified phosphorylation sites (Ser 568 , Ser 569 , and Ser 570 , unphosphorylated and phosphorylated) were produced by the Protein Chemistry Core Facility at Baylor College of Medicine. The peptides were synthesized with a cysteine at the C-terminal end of the peptide to allow conjugation to the carrier protein keyhole limpet hemocyanin via m-maleimidobenzoyl-N-hydroxysuccinimide ester. The conjugates of the phosphopeptides were used for the production of custom antibodies (Cocalico, Inc., CA). After a preimmune blood sample was taken, the conjugates of the phosphopeptide were injected into two albino New Zealand rabbits with the adjuvant Titermax via intradermal injections. The animals were boosted with antigen several times. The phospho-SK2 antiserum was affinity-purified and characterized with Western blotting (28,29).
COS7 Cell Expression Systems and Membrane Preparation-COS7 cells were cultured and transfected with FuGene6 (Roche Diagnostics). In brief, cells were grown in poly-D-lysine-coated 6-well plates and maintained in Dulbecco's modified eagle's medium (Invitrogen) containing fetal bovine serum and penicillin/streptomycin. 1 g of SK2 cDNA was used for transfection in each well with FuGene6 according to manufacturer's instructions. To optimize PKA phosphorylation of fulllength SK2 expressed in the cells, we determined basal PKA phosphorylation of SK2 with medium that contained either 1% or 10% serum. Western blotting using a PKA-substrate phospho-antibody revealed basal phosphorylation in both 1 and 10% serum media. However, 10% serum demonstrated higher basal phosphorylation (data not shown). After PKA stimulation, there was a marked increase in the phosphorylation of PKA substrates with 1% serum relative to basal levels (data not shown). Therefore, to obtain a robust response to PKA activation, 1% serum was used in all subsequent PKA manipulation experiments. Transfection of the SK2 channel was performed on COS7 cells with FuGENE 6 and grown for 36 -48 h. The media was switched from 10% serum to 1% serum ϳ12 h before PKA manipulation. Sham transfections were performed using a blank vector (pDs-Red, Clontech, Palo Alto, CA) or a Kv4.2 construct. COS7 cells were treated with PKA modulators dissolved in Me 2 SO. Me 2 SO alone served as the vehicle control. In all conditions the final concentration of Me 2 SO was 0.2%. The PKA pathway manipulations were are follows: 1) Me 2 SO alone; 2) 15-min incubation with 50 M forskolin (FSK, PKA activator) plus 100 M Ro-201724 (phosphodiesterase inhibitor) (28); 3) 30-min preincubation with 10 M H89 (PKA inhibitor) (31); or 4) preincubated with 10 M H89 for 30 min followed by a 15-min incubation with 50 M FSK/ 100 M Ro-201724. Cell membrane preparation was performed; briefly, plates were put on ice to stop the reaction and washed three times by cold PBS with phenylmethylsulfonyl fluoride (PMSF, 100 M). After harvesting by scraping cells off the plates in BHB buffer (in mM: 20 Tris, pH 7.5, 1 EGTA, 1 EDTA, 1 Na 4 P 2 O 7 , 1 Na 3 VO 4 , 0.1 PMSF), membranes were prepared by centrifuging at 92,500 ϫ g at 4°C for 60 min. The cell pellet was resuspended in 5% SDS with 100 mM dithiothreitol, 10 g/ml pepstatin, 10 g/ml aprotinin, 10 g/ml leupeptin, 100 M PMSF. Sample buffer was then added, and the samples were loaded on a SDS-PAGE gel for Western blotting.
Immunostaining of COS7 Cells-Transfected COS7 cells were treated with the PKA activators or inhibitors and then fixed with 4% paraformaldehyde (Polysciences, Inc., Warrington, PA) for 40 min at room temperature. Cells were rinsed twice with PBS and treated with 0.3% Triton X-100 for 15 min at room temperature. The cells were blocked with 10% fetal bovine serum for 1 h at room temperature followed by SK2 antibody diluted 1:500 in PBS overnight at 4°C. The cells were then washed in 0.1% Triton X-100 or Tween 20 three times for 5 min each. Alexa Fluor 488 goat anti-rabbit IgG (HϩL) secondary antibody (Molecular Probes, Eugene, OR) was diluted to 1:5000 and incubated with cells for 30 -40 min at room temperature. The cells were visualized on the  (H89ϩFSK). A, representative micrographs show the effects of the PKA pathway manipulations on localization of SK2 in COS7 cells by immunostaining with the commercial SK2 antibody (SK2). PKA cascade activation (FSK) decreased SK2 surface staining and increased perinuclear (endoplasmic reticulum and Golgi apparatus) SK2 staining compared with control (Me 2 SO). PKA cascade inhibition (H89) increased surface expression of the SK2 channel subunits compared with control. Transfected COS7 cells treated with H89ϩFSK showed no change in SK2 staining when compared with control (Me 2 SO (DMSO)). No SK2 staining was seen in the sham-transfected COS7 cells (supplemental Fig. S1). Scale bars, 20 m. B, Western blotting with the SK2 antibody was performed on whole cell lysates and on extracted biotinylated surface proteins following the application of PKA pathway modulators. Representative blots of surface and total SK2 subunits are shown. The top panel (Surface) demonstrates that PKA activation (FSK) decreased, while PKA inhibition (H89) increased SK2 surface expression. The bottom panel (Total lysates) shows that there was no difference in total SK2 channel expression. All conditions were normalized to actin (Actin). C, densitometry of the Western blots shows a significant decrease in surface expression of SK2 following PKA activation (FSK, 65 Ϯ 4% of control; *, p Ͻ 0.05) and an increase in SK2 surface expression with PKA inhibition (H89, 168 Ϯ 13% of control; **, p Ͻ 0.01) compared with control (Me 2 SO). There was no significant difference between the control and FSKϩH89 conditions (104 Ϯ 9% of control). Data are expressed as mean Ϯ S.E. One-way analysis of variance with post-hoc analysis was used for comparison, n ϭ 3.
Nikon TE-200 inverted fluorescence microscope. All immunohistochemistry studies were performed in three independent experiments.
Cell-surface Biotinylation-After PKA pathway manipulation, the SK2-transfected COS7 cells were washed with ice-cold PBS containing 0.1 mM CaCl 2 and 1 mM MgCl 2 (PBS/Ca 2ϩ plus Mg 2ϩ , pH 8.0) then treated with sulfo-NHS-LC-biotin (0.5 mg/ml, Pierce) at 4°C for 40 min. Cells were lysed in radioimmune precipitation assay buffer (150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 20 mM Na 2 HPO 4 , 1% Triton X-100, 0.1% SDS, pH 7.4), supplemented with 0.01 mM PMSF, 0.005 g/ml leupeptin, and 0.005 g/ml pepstatin for 1 h at 4°C. Lysates were centrifuged at 20,000 ϫ 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 l, 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 radioimmune precipitation assay buffer and eluted with 50 l of Laemmli loading buffer (Bio-Rad) for 5 min at 95°C. The eluates (25 l) were resolved by SDS-PAGE gel and immunoblotted with the SK2 antibody. Immunoreactive bands were visualized using horseradish peroxidase-conjugated secondary antibody (1:10,000). Immunoreactivity values of surface SK2 channels were normalized to levels of actin immunoreactivity in total cell extracts to preclude errors that accompany sample loading and transfer.
Data Analysis-Immunoreactivity was quantified by densitometry (Scion Corp., version Beta4). The GraphPad Prism software package was used for statistical analysis of the data. Data are expressed as standard error of mean (S.E.). Representative examples are shown under "Results." One-way analysis of variance with post-hoc analysis or Stu-dent's t test was used for comparison. Statistical significance was taken as p Ͻ 0.05.

RESULTS
The PKA Pathway Modulates SK2 Channel Surface Expression in COS7 Cells-We first investigated PKA pathway regulation of SK2 channel trafficking and localization in the COS7 cell expression system. Immunostaining was performed on the SK2-transfected COS7 cells with the commercial SK2 antibody after treating the cells with PKA pathway modulators (Fig. 1A). Immunostaining for SK2 was visualized on the cell-surface membrane and concentrated in presumptive endoplasmic reticulum (ER) and Golgi apparatus in the vehicle control (Me 2 SO). PKA activation using forskolin (FSK (50 M) together with the phosphodiesterase inhibitor Ro-201724 (100 M)) attenuated surface expression and increased perinuclear expression (presumptive ER and Golgi complex) of the SK2 subunits, whereas the PKA inhibitor, H89 (10 M), facilitated SK2 channel trafficking to the cell surface. The decrease in surface expression of SK2 channels by forskolin is attenuated by preincubation with H89. These data suggest that PKA regulates the surface expression of SK2. As controls and in parallel with the studies presented above, we processed untransfected and sham-transfected COS7 cells for SK2 immunohistochemistry. These COS7 cells showed no SK2 staining (sham-transfected COS7 cells shown in supplemental Fig. S1).
We further investigated this effect with surface biotinylation experiments to quantify the cell-surface expression of SK2 following PKA cascade manipulation in COS7 cells. Western blotting with the commercial SK2 antibody was performed to compare the biotinylated and total pools of SK2 channels in each experimental group (Fig. 1B). Western blotting showed no difference in the total SK2 channel expression in lysates of all treated groups. However, there was a significant reduction in SK2 channel expression on the cell surface following forskolin application relative to control (65 Ϯ 4% of Me 2 SO control, p Ͻ 0.05, n ϭ 3) (Fig. 1C). Western blotting with a second SK2 antibody (generated in the laboratory of J.P.A.) against a different epitope (SK2 residues 538 -547) revealed similar results (data not shown). In contrast, following H89 treatment, there was a significant elevation of SK2 channel expression on the cell surface compared with control (168 Ϯ 13% of Me 2 SO control, p Ͻ 0.01, n ϭ 3). There was no significant difference in surface expression of cells treated with both FSK and H89 (104 Ϯ 9% of Me 2 SO control). Together these results demonstrate that, in SK2-transfected COS7 cells, activation of the PKA cascade attenuates surface expression of SK2 channel subunits, whereas a reduction of PKA activity increases SK2 subunit surface expression. Based on the observation that there are FIGURE 2. Candidate PKA consensus sites within the SK2 channel subunit. The amino acid sequence of the SK2 channel subunit is shown with the intracellular regions underlined. The full-length SK2 subunit consists of 580 amino acid residues with a putative tertiary structure of six transmembrane regions, two intracellular loops and cytoplasmic amino-and carboxyl-terminal domains. We inspected the primary amino acid sequence of the intracellular domains of the SK2 subunit for known PKA consensus sequences (RX 1-2 (S/T)X) as these regions may be accessible to intracellular PKA. Seven amino acid residues (Ser 15 , Ser 136 , Thr 431 , Ser 465 , Thr 510 , Ser 561 , and Ser 568 ) were predicted as candidate PKA phosphorylation sites within SK2 by NetPhos (www.cbs.dtu.dk/services/NetPhos/). The predicted sites are indicated in bold and are numbered.
PKA consensus sites within the SK2 amino acid sequence, we hypothesized that PKA regulates SK2 surface expression by direct phosphorylation of the SK2 subunits.
Candidate PKA Sites within the SK2 Amino Acid Sequence-The SK2 subunit is composed of 580 amino acid residues with six putative transmembrane domains, and cytoplasmic NT and CT regions. We focused our study on SK2 cytoplasmic domains, because these would be accessible to cellular kinases. Analysis of the amino acid sequence of SK2 with NetPhos (www.cbs.dtu.dk/services/NetPhos/) (33) predicted multiple serine and threonine residues (Ͼ95% probability) as candidate PKA phosphorylation sites. Two candidate PKA sites were located within the NT and five within the CT cytoplasmic domains (Fig. 2). No candidate PKA sites were identified within the two cytoplasmic loops.
PKA Phosphorylates the SK2-NT and -CT Domains-To test our hypothesis that the SK2-NT and -CT domains are PKA substrates, recombinant GST fusion proteins of SK2-NT (amino acid residues 1-140, 42 kDa) and -CT (amino acid residues 395-580, 50 kDa) cytoplasmic domains were constructed. In independent reactions, the GST SK2-NT and -CT constructs were incubated with activated PKA and [␥-32 P]ATP in vitro. Reaction products were separated by SDS-PAGE, and bands corresponding to the GST SK2-NT and -CT fusion proteins were identified by Coomassie Blue staining and autoradiography as previously described (28,29). Both the NT and CT fusion proteins demonstrated 32 P incorporation, whereas GST alone and PKA alone controls showed no 32 P incorporation (Fig. 3A). Because the second cytoplasmic loop of SK2 (amino acids 278 -306) contained several serine and threonine residues, this region was evaluated for 32 P incorporation. No 32 P labeling was found in the second intracellular loop of SK2 (data not shown). Based on these findings, we conclude that both the NT and CT domains of SK2 are PKA substrates.
Prior to sequencing the PKA phosphorylation sites within the SK2 domains we determined the time course of phosphorylation of the constructs to ensure that no phosphorylation sites were missed in the mapping studies. The time course of phosphorylation of the SK2 domains by PKA showed that, by 60 min, both the NT and CT fusion proteins had reached saturation levels (Fig. 3B). Therefore, incubations of 60 min are sufficient for complete phosphorylation of the fusion proteins by PKA in vitro.
Identification of the PKA Phosphorylation Sites within the SK2-NT Domain-Because Ser 15 and Ser 136 were the only two predicted PKA phosphorylation sites within the SK2 NT region, we used site-directed mutagenesis to map the sites within the GST SK2-NT fusion protein.
The Ser 15 and Ser 136 residues were mutated to alanine to block phosphorylation at these sites. Three mutant constructs, single mutants S15A and S136A and the double mutant S15A/S136A were made and incubated with active PKA and [␥-32 P]ATP in vitro. There was 32 P incorporation within the S15A mutant, but not within the S136A or S15A/S136A mutants (Fig. 4). These data suggest PKA phosphorylates the SK2-NT domain at Ser 136 .  . Therefore, we used serine to alanine mutants (S15A, S136A, and S15A/S136A) within the GST SK2-NT fusion protein to evaluate PKA phosphorylation of these sites in vitro. The top panel shows a Coomassie Blue-stained gel (Coomassie) with bands at 42 kDa representing the SK2-NT WT and mutant constructs. The lower panel represents an autoradiogram (Autorad) of the Coomassie Blue-stained gel, which shows 32 P incorporation into the WT and S15A mutant but not the S136A and S15A/S136A mutant constructs. These findings suggest that Ser 136 is the only amino acid residue phosphorylated by PKA within the SK2-NT region.

Mapping of the PKA Phosphorylation Sites within the SK2-CT
Domain-A number of potential PKA phosphorylation sites were predicted within the SK2-CT cytoplasmic domain, hence, we initially used mass spectrometry (MS) to map phosphorylation sites. The PKA-phosphorylated GST SK2-CT fusion protein was digested with Lys-C, desalted on a resin column, and analyzed by matrixassisted laser desorption ionization (MALDI) scanning in tandem with negative ion precursor scanning. MALDI was able to identify pep-tides containing three of the five potential PKA sites: 430 ETWLIYK 436 , 455 KFLQAIHQLRSVK 467 , and 507 RIVTLETK 514 with the predicted PKA sites Thr 431 , Ser 465 , and Thr 510 , respectively (marked in bold). Negative ion mass/charge (m/z) plot of the peptide containing the Ser 465 site indicated that a phosphate group was associated with this peptide (Fig.  5A). m/z plots of the peptides containing predicted sites Thr 431 and Thr 510 showed that they were not associated with phosphate groups (data not shown); therefore, these sites were not phosphorylated by FIGURE 5. PKA phosphorylates the SK2 carboxyl-terminal cytoplasmic domain. A, to map the PKA phosphorylation sites within the SK2-CT domain we performed a large scale phosphorylation reaction with the PKA catalytic subunit and the GST SK2-CT construct. The reaction product was digested in-gel by Lys-C and analyzed with tandem electrospray ionization mass spectroscopy for peptide-associated phosphate groups. A Ϫ79 precursor ion scan and tandem mass spectroscopy revealed that a phosphate group was associated with a peptide corresponding to SK2 residues 455 KFLQAIHQLRSVK 467 with the m/z ratio of 823.0 (arrow). These results indicate that Ser 465 within the GST SK2-CT construct is phosphorylated by PKA. Mass spectroscopy also identified peptides containing SK2 amino acid residues Thr 431 and Thr 510 (predicted PKA sites), but these peptides were not phosphorylated. B, we evaluated phosphorylation of GST SK2-CT wild-type compared with a S465A GST SK2-CT mutant construct. Coomassie Blue (Coomassie) staining shows 49-kDa bands, representing the GST SK2-CT wild-type (WT) and mutant (S465A) constructs. Autoradiography (Autorad) demonstrates robust 32 P incorporation into both of the constructs. These findings suggest that Ser 465 may be a minor site and that there are additional phosphorylation sites within the SK2 CT. C, serine to alanine mutants of the candidate PKA sites, Ser 561 and Ser 568 in combination with Ser 465 within the GST SK2-CT construct were tested for PKA phosphorylation. Coomassie Blue staining and autoradiography show 32 P incorporation into GST SK2-CT mutants: S561A, S568A, S465A/S561A, S465A/S568A, and S465A/S561A/S568A. These data suggest that there are additional PKA phosphorylation sites located within the SK2-CT region.
PKA. Thus, these findings suggest that Ser 465 , but not Thr 431 and Thr 510 , is a PKA phosphorylation site within the SK2-CT. Peptides containing the two other predicted PKA sites, Ser 561 and Ser 568 were not obtained on MALDI scanning, possibly because the Lys-C digest yielded small peptide fragments that were not detected. Digests obtained with other proteases such as pepsin and Glu-C were also unsuccessful at yielding peptides containing these amino acids (data not shown). Hence, we were unable to evaluate PKA phosphorylation at these sites (Ser 561 and Ser 568 ) using this technique.
As an alternative approach, we used site-directed mutagenesis to determine if the remaining predicted sites were phosphorylated by PKA. Site-directed mutation of serine to alanine at position 465, to remove the PKA phosphorylation site did not decrease 32 P incorporation into the GST SK2-CT construct (Fig. 5B). These data suggest that additional PKA phosphorylation sites exist within the SK2-CT region, possibly in the serine-and arginine-rich CT tail. To identify the remaining PKA phosphorylation site(s), GST SK2-CT constructs with mutations of the predicted Ser 561 and Ser 568 sites and the mapped S465 site were generated and incubated with activated PKA and [␥-32 P]ATP in vitro (Fig.  5C). Phosphorylation was still observed in these constructs. These findings suggest that additional PKA phosphorylation sites within the SK2-CT domain are present. In addition, these data also suggest that Ser 465 is a minor phosphorylation site.
We hypothesized that the additional PKA sites reside in the serineand arginine-rich distal SK2-CT tail. Thus, we assessed 32 P incorporation within the GST SK2-CT wild-type compared with three truncated GST SK2-CT fusion proteins, S559stop, S568stop, and T575stop following incubation with activated PKA (Fig. 6A, upper panels). The 32 P incorporation was found within the GST SK2-CT wild-type and T575stop construct but not within the S559stop and S568stop constructs (Fig. 6A, lower panels). These findings suggest that additional PKA phosphorylation site(s) are located distal to residue 567 within the SK2-CT tail. To further define the PKA phosphorylation sites within this region, we mutated residues 568 SSST 571 to 568 AAAA 571 within the GST SK2-CT construct (Fig. 6B, upper panel number 2) and performed PKA phosphorylation reactions with [␥-32 P]ATP. No 32 P incorporation was found within the GST SK2-CT 568 AAAA 571 construct (Fig. 6B, lower panel, lane 2), suggesting that the major phosphorylation sites were restricted to the 568 SSST 571 region. A series of serine to alanine mutations were made within the 568 SSST 571 region to identify the specific residues phosphorylated by PKA (Fig. 6B, upper panel). There was 32 P incorporation within the SK2-CT 568 SAAA 571 , 568 ASAA 571 , and 568 AASA 571 constructs but not in 568 AAAT 571 construct (Fig. 6B, lower   panels, lanes 3-6) . These findings suggest that the residues Ser 568 , Ser 569 , and Ser 570 within SK2 are PKA substrates.
A summary of all of the PKA sites that we mapped within the SK2 cytoplasmic domains are presented in a model diagram of the SK2 channel subunit (Fig. 7). As indicated, only one phospho-site was identified within the SK2-NT domain, while four additional sites were identified within the SK2-CT domain. For the studies testing the functional significance of direct phosphorylation of the SK2 channel by PKA presented below, we focused on the three sites (Ser 568 , Ser 569 , and Ser 570 ) located within the distal SK2-CT domain. This was based on two observations: 1) Ser 136 within the NT domain of SK2 resides at the beginning of the first transmembrane domain, suggesting that this residue may not be accessible to intracellular kinases in the full-length channel, and 2) Ser 465 appears to be a minor phosphorylation site based on our studies using the GST SK2-CT construct (see Fig. 5B).

PKA Phosphorylates Full-length SK2 Channel Subunits in the COS7
Expression System-Because our results showed that PKA phosphorylates SK2 in vitro, we hypothesized that SK2 is phosphorylated by PKA in a cell system. We generated a phospho-selective antibody against a synthetic peptide containing phosphorylated residues Ser 568 , Ser 569 , and Ser 570 (Fig. 8A). The synthetic phosphorylated peptide was injected into rabbits, and antisera were collected and affinity purified. The purified antibody was screened against the unphosphorylated and phosphorylated ovalbumin-coupled synthetic peptides using Western blotting as previously described (28,29). The antibody showed specific immunoreactivity to the phospho-but not the unphospho-peptides, and the immunoreactivity was blocked by preincubation with the antigenic peptide in a concentration-dependent manner (Fig. 8B). Preincubation with 0.26 nM of the antigenic peptide (phospho-peptide) blocked antibody recognition. The antibody also showed specific immunoreactivity to the phosphorylated but not the unphosphorylated GST SK2-CT fusion protein that was blocked by preincubation with the antigenic peptide (0.26 nM, Fig. 8C). These initial screening studies suggest that the antibody is phospho-selective.
The antibody was used to assess PKA phosphorylation of the full-length SK2 channel subunit in the COS7 expression system. COS7 cells transiently expressing wild-type SK2 were incubated with forskolin (50 M, FSK, together with 100 M Ro-201724) or H89 (10 M) for PKA activation or inhibition, respectively, of the endogenous PKA cascade. Membranes were prepared from SK2-transfected COS7 cells after PKA pathway manipulations. Western blots were probed with the phospho-SK2 antibody and a commercial SK2 antibody that is not sensitive to the phosphorylation state of SK2 (same antibody used in Fig. 1 studies). Actin immunoreactivity was used for normalization. Densitometry revealed a significant increase in immunoreactivity with the phospho-SK2 antibody following forskolin stimulation compared with vehicle (Me 2 SO)-treated control COS7 cells (259 Ϯ 21% of control, p Ͻ 0.001, n ϭ 4). This effect was blocked by preincubation with H89 (H89 plus FSK, 149 Ϯ 16% of control, not significant compared with control, compared with FSK: p Ͻ 0.001, n ϭ 4) (Fig. 9A, left  panels). There was no significant change in immunoreactivity using the commercial SK2 antibody following PKA pathway manipulation (Fig. 9A, right panels), indicating that this manipulation does not result in a change in SK2 protein expression. No immunoreactivity was seen in lanes containing membranes from sham-transfected COS7 cells treated with vehicle (Sham) or forskolin (Sham plus FSK) that were probed with the phospho-SK2 or commercial SK2 antibodies (Fig. 9A). Based on this series of experiments, we conclude that PKA activation couples to direct phosphorylation of residues Ser 568 , Ser 569 , and Ser 570 within the SK2 channel subunit.
The Me 2 SO-treated SK2-transfected COS7 cultures demonstrated basal levels of immunoreactivity with the phospho-SK2 antibody, suggesting that there are basal levels of SK2 phosphorylation, potentially explaining the increase in surface expression produced by H89 treatment (Fig. 1). However, another possibility is that the antibody has some recognition of the unphosphorylated SK2 channel. We considered the latter possibility unlikely, because the antibody was affinity-purified against the phospho-SK2 peptide used to make the antibody and the initial screening of the phospho-SK2 antibody did not reveal detection of the unphosphorylated synthetic peptide or fusion protein (Fig. 8, B  and C). However, to confirm the selectivity of the phospho-SK2 antibody, alkaline phosphatase (AP, 5 units/l) to remove phosphates was incubated with SK2-transfected COS7 lysates after PKA pathway manipulations. An aliquot of the lysate from each condition was incubated in parallel to the AP-treated samples for comparison. Western blotting of the lysates was performed using the phospho-SK2 antibody. In addition, we used an antibody that recognizes a wide variety of PKAphosphorylated substrates. As expected, PKA stimulation with forskolin and inhibition with H89 lead to modulation of the immunoreactivity of a 49-kDa band corresponding to SK2 in Western blotting with the phospho-SK2 antibody (Fig. 9B, top panel). The PKA substrate antibody also detected modulation of the immunoreactivity of several bands of varying molecular weight (Fig. 9B, middle panel). Alkaline phosphatase treatment abolished immunoreactivity for both the phospho-SK2 and PKA substrate antibodies. There was no change in total actin levels. These findings indicate that the antibody is phospho-selective and suggest that there are basal levels of SK2 phosphorylation in COS7 cells transiently expressing SK2. As an additional control for the SK2 anti-  To measure PKA phosphorylation of the SK2 subunits, phospho-selective antisera were generated and the affinity-purified antibody was screened by Western blotting. A, a phosphorylated SK2 CT peptide fragment representing amino acids 558 -580 of the SK2 CT cytoplasmic domain was synthesized for antibody generation (phosphorylated residues are marked in bold and italics). This peptide and the unphosphorylated form of the synthetic peptide were used for antibody screening. B, the phosphorylated and unphosphorylated synthetic peptide fragments of the SK2-CT were coupled to ovalbumin for use in Western blotting. In each gel shown, lanes 1 and 3 contain the unphosphoovalpeptide (full-strength and 1:5 dilution, respectively), whereas lanes 2 and 4 contain the phospho-ovalpeptide (full-strength and 1:5 dilution, respectively). The purified phospho-SK2 antibody showed recognition of the phosphorylated but not the unphosphorylated ovalbumin-coupled peptide (right panel labeled Ab only). The immunoreactivity with the purified phospho-SK2 antibody was blocked by preincubation of the antigenic peptide (Pep) in a concentration-dependent manner (panels labeled Pep (2.6 nM), Pep (0.26 nM), and Pep (0.1 nM)). C, phospho-selectivity of the affinity-purified antibody was screened by Western blotting using the unphospho-and phospho-SK2-CT GST constructs. The left panel (Ab) shows antibody recognition of the phospho-SK2-CT GST construct (lane 1) with no recognition of the unphospho-SK2-CT GST fusion protein (lane 2). In parallel studies immunoreactivity of the phospho-antibody to the phospho-SK2-CT GST construct was blocked by preincubation of the antibody with the phospho-peptide antigen (right panel, AbϩPep (0.26 nM)).
body, COS7 cells expressing the SK2 subunit with the PKA sites within the distal CT cytoplasmic domain of SK2 (Ser 568 , Ser 569 , and Ser 570 ) mutated to alanines (SK2-AAA) were treated with Me 2 SO (vehicle) or forskolin. Western blots with the phospho-SK2 antibody demonstrated no immunoreactivity against the SK2-AAA channel subunit (Fig. 9C,   top panel). Parallel lysates probed with the commercial SK2 antibody demonstrated that the SK2-AAA subunit was expressed (Fig. 9C,  bottom panel).
Serine to Alanine Mutation of the PKA Sites within the Distal Tail of SK2 Blocks PKA Pathway Regulation of SK2 Surface Expression-To determine whether phosphorylation of SK2 at the sites that we mapped underlies the PKA effects on channel trafficking (Fig. 1), we used the SK2-AAA (S568A, S569A, and S570A; see above) mutant channel construct in surface biotinylation experiments. In these studies, changes in surface biotinylation in response to PKA activation were tested with the SK2-AAA channel mutant expressed in COS7 cells to determine if the mutation of these sites blocked the effect of PKA pathway activation on SK2 localization. Western blotting of surface-biotinylated proteins with the commercial SK2 antibody showed that there was no significant change in surface expression of the SK2-AAA channel subunit following forskolin stimulation (Me 2 SO: 100 Ϯ 18% and FSKtreated: 92 Ϯ 9%, n ϭ 3) (Fig. 10). As a control, parallel studies with vehicle and forskolin stimulation studies were performed in COS7 cells transfected with the wild-type SK2 channel subunit. As expected, the surface expression of the wild-type SK2 subunit significantly decreased following forskolin application (not shown, see Fig. 1). These findings support our hypothesis that the PKA pathway regulates SK2 surface expression through direct phosphorylation of the primary subunit.

DISCUSSION
The studies presented here demonstrate that the PKA pathway regulates surface expression of SK2 homomeric channels in the COS7 expression system. PKA activation decreased SK2 surface localization, whereas PKA inhibition had the opposite effect. To determine the mechanisms of this effect, we further investigated whether SK2 was directly phosphorylated by PKA and whether this phosphorylation was responsible for the effect. Through a series of mapping and surface biotinylation experiments, residues Ser 568 , Ser 569 , and Ser 570 within the distal CT of the SK2 cytoplasmic domain were identified as the PKA phosphorylation sites responsible for the trafficking effect of PKA activation on SK2 surface expression. We also showed that PKA activation increases phosphorylation of expressed SK2 channel subunits at these sites in COS7 cells using a phospho-selective antibody. Furthermore, biotinylation and immunostaining studies showed a decrease in surface expression of SK2 following activation of the endogenous PKA cascade. Modulation of SK2 channel surface expression by PKA was eliminated in the SK2 construct with serine to alanine mutation of the three PKA sites within the distal CT tail. These results support an important role for these sites in PKA-mediated alterations in localization and trafficking of the SK2 channel subunit. The functional significance of PKA phospho-regulation at Ser 136 and Ser 465 sites in the full-length SK2 subunit in a cell system remains to be shown.
Ion channel trafficking involves the secretory and endocytic pathways to and from the surface membrane and the regulatory elements in this process are diverse. Recently it has been shown that the activity of protein kinase pathways is an important determinant of potassium channel trafficking (34). Our findings demonstrate that PKA activation leads to a decrease in SK2 channel subunit expression at the COS7 cell surface through direct phosphorylation of the SK2 subunits. Kinase regulation of SK2 surface expression has not previously been shown; however, SK2 association with calmodulin is obligatory for SK2 surface expression in COS7 cells (35). It is possible that phosphorylation of the sites that we mapped leads to conformation changes that disrupt calmodulin binding. It is unlikely that the PKA effect is related to phosphorylation of FIGURE 9. PKA pathway modulation couples to altered levels of SK2 channel subunit phosphorylation in COS7 cells. PKA phosphorylation studies were performed on COS7 cells transfected with the full-length SK2 wild-type (SK2 WT) or full-length SK2 channel subunits with Ser 568 , Ser 569 , and Ser 570 residues mutated to alanine (SK2-AAA). A, following PKA manipulations, PKA phosphorylation of SK2 WT in transfected COS7 cells was measured by Western blotting membrane fractions using the phospho-SK2 antibody (P-SK2, left panel). PKA activation with forskolin (FSK; 50 M together with 100 M Ro-201724, a phosphodiesterase inhibitor) significantly increased SK2 phosphorylation relative to vehicle (0.2% Me 2 SO) control (259 Ϯ 21% of control; ***, p Ͻ 0.001), which was significantly attenuated by PKA inhibition with H89 preincubation (10 M, H89ϩFSK)(149 Ϯ 16% of control, compared with FSK; ***, p Ͻ 0.001). Phospho-SK2 immunoreactivity following preincubation with H89 alone (H89) was unchanged compared with control. In parallel studies, blots probed with the commercial SK2 antibody (not sensitive to phosphorylation sites) showed no differences in SK2 expression (SK2, right panels). Untreated (Sham) and FSK-stimulated (Sham ϩ FSK) untransfected COS7 cells showed no SK2 immunoreactivity. B, since there were basal levels of immunoreactivity in the control and H89 conditions, we evaluated whether the affinity-purified phospho-SK2 antibody recognized unphosphorylated channel proteins. Aliquots of SK2 WTtransfected COS7 cells treated with PKA pathway modulators were incubated with or without alkaline phosphatase (AP, 5 unit/l). Representative Western blots of total lysates probed with the phospho-SK2 (P-SK2), a commercial phospho-PKA substrate (PKA substrate), or actin (Actin) antibody are shown. Treatment of the samples with AP following PKA manipulation (AP treated) eliminated immunoreactivity with the phospho-SK2 antibody. Probing the blots with the actin antibody showed equal protein loading across lanes. Blots probed with the PKA substrate antibody showed that there was an increase in phosphorylated proteins following PKA activation (FSK) compared with basal levels in the control (Me 2 SO (DMSO)). H89ϩFSK reduced the increase in PKA substrate phosphorylation compared with FSK only. In parallel experiments, application of AP to the samples eliminated the immunoreactivity seen with the PKA substrate antibody. C, as an additional confirmation of the phospho-selectivity of the P-SK2 antibody, we performed PKA activation experiments in COS7 cells expressing the full-length SK2-AAA mutant channel subunit. Western blotting with the P-SK2 antibody following PKA cascade activation (FSK) compared with control (Me 2 SO (DMSO)) demonstrated no immunoreactivity with whole cell lysates (top panel). Parallel lysates probed with the SK2 antibody (not phospho-selective) demonstrated that the SK2-AAA subunit was expressed (bottom panel). Data are expressed as mean Ϯ S.E. One-way analysis of variance with post-hoc analysis was used for comparison, n ϭ 4. calmodulin, because there are no PKA consensus sequences within the calmodulin amino acid sequence.
Direct phosphorylation of ion channels or auxiliary subunits contributes to protein trafficking to the surface membrane (36,37). However, a number of studies have shown that ER retention signals also play an important role in the trafficking of membrane proteins to the cell surface (38 -40). The ER retention signals, particularly the RXR sequence in ion channels, are found mainly at the extreme CT of proteins. These amino acid sequences function to retain newly formed proteins in the ER and allow for retrieval from the Golgi apparatus. The retention signals become masked to ensure that only completely assembled channels are delivered to the surface membrane (38). In addition, Zhou et al. (41) reported that both phosphorylation and RXR type ER retention signals are required for PKA-mediated potentiation of sodium currents in heart. It also has been reported that trafficking of the Kir1.1 channel to the surface membrane by direct phosphorylation overrides the independent RXR type ER retention signal (42). There are three RXR motifs near the PKA sites within the distal CT tail of the SK2 subunit, corresponding to 558 RSR 560 , 564 RRR 566 , and 565 RRR 567 . Interestingly, immunostaining of COS7 cells expressing an SK2 S559stop truncation dem-onstrated relatively greater localization at the cell membrane surface and less channel protein retained in the ER-Golgi complex compared with that of full-length wild-type SK2 (data not shown), demonstrating that important regulatory elements for SK2 trafficking reside within the SK2 distal CT tail. Based on our preliminary results and other studies, one possible mechanism is that phosphorylation at the PKA sites leads to conformational changes that unmask the ER retention signal motifs within distal tail of the SK2 subunit.
The effects of PKA phosphorylation of SK2 in excitable tissues where the channels are natively expressed have not been elucidated; however, our findings suggest a molecular locus for previous reports of the regulation of the I mAHP , which is thought to be formed at least in part, by SK2 channels (18). The enhancement of theta-burst LTP induction in hippocampal slices by brain-derived neurotrophic factor (BDNF) is associated with an increase in direct serine phosphorylation within SK2 and a reduction in the I mAHP (43). The specific kinase that is activated and serines that are phosphorylated within SK2 in BDNF-enhanced theta-burst LTP remain to be elucidated. However, BDNF transiently activates PKA in the hippocampus (44). Thus, SK2 may be a direct target of the PKA cascade in BDNF-modulated neuronal plasticity. Moreover, functional studies in rat jaw-closing spinal motor neurons have coupled activation of the metabotrophic 5-HT 1A receptor to down-regulation of the apamin-sensitive I mAHP (45). The exact pathway mediating this effect in spinal motor neurons is unknown, but activation of the 5-HT 1A receptor induces downstream cAMP accumulation and PKA activation (46). Therefore, activation of 5-HT 1A receptors in spinal motor neurons may result in direct phosphorylation of the SK2 subunit and modulation of the I mAHP .
The phospho-selective antibody that we have generated may provide a tool to assess PKA phosphorylation of SK2 in cell types that natively express SK2 channels. In preliminary studies we have begun to evaluate PKA pathway coupling to SK2 phosphorylation in hippocampal tissue. 3 In these studies, the phospho-SK2 antibody recognizes a band at roughly 49 kDa that increases with PKA pathway activation by forskolin. The commercial SK2 antibody recognizes a band at the same molecular weight that does not modulate with forskolin stimulation. These findings suggest that PKA phosphorylation of SK2 occurs in a physiologically relevant system and that the phospho-antibody is selective in a tissue type where SK2 channels are natively expressed.
In conclusion, we have demonstrated that direct PKA phosphorylation of SK2 regulates the localization of SK2 channel proteins in the COS7 expression system. The altered cellular distribution of SK2 channels may play a role in the modulation of the I mAHP and regulate excitability in tissues where the channels are expressed, such as the nervous system. Given the recently defined role of SK2 in regulation of NMDA receptor-mediated synaptic potentials and Ca 2ϩ influx (13,14), modulation of SK2 surface expression by PKA phosphorylation may have dramatic consequences on cellular plasticity and learning and memory. FIGURE 10. Serine to alanine mutation of the SK2 distal CT phosphorylation sites blocks the PKA-mediated decrease in SK2 surface expression. To evaluate the effects of PKA phosphorylation of the three contiguous serines Ser 568 , Ser 569 , and Ser 570 on SK2 trafficking, phosphorylation reactions were performed on COS7 cells transfected with the full-length SK2 triple-site mutant (SK2-AAA), which eliminated the PKA phosphorylation sites within the CT. As a control, parallel experiments were performed with the full-length SK2 wild-type channel. COS7 surface expression of the SK2-AAA subunit following vehicle or PKA activation with forskolin (50 M) was quantified by Western blotting of biotin-captured surface proteins with the commercial SK2 antibody (SK2) that is not sensitive to the phosphorylation state of SK2. Representative blots of surface (Surface) proteins and total (Total) lysates from COS7 cells expressing the SK-AAA mutant channel subunit treated with vehicle (0.2% Me 2 SO (DMSO)) or forskolin (FSK) are shown. Normalized data of the changes in the surface expression of the SK2-AAA channel subunit following PKA stimulation are shown in the bar graph. PKA activation had no significant effect on surface localization of the SK2-AAA mutant subunit (Me 2 SO: 100 Ϯ 18% and FSK: 92 Ϯ 9% of Me 2 SO, n ϭ 3) in COS7 cells. Forskolin stimulation significantly decreased surface expression of the wild-type SK2 channel subunit compared with vehicle control (not shown; same results as shown in Fig. 1). There was no difference in total SK2-AAA channel protein (blot marked as Total) or actin (Actin). All conditions were normalized to actin. These findings suggest that the contiguous phosphorylation sites ( 568 SSS 570 ) are critical for PKA-induced down-regulation of SK2 surface expression. Data are expressed as mean Ϯ S.E. Student's paired t test was used for analysis.