S-acylation regulates Kv1.5 channel surface expression

Lian Zhang, Karyn Foster, Qiuju Li, Jeffrey R. Martens

Abstract

The number of ion channels expressed on the cell surface shapes the complex electrical response of excitable cells. An imbalance in the ratio of inward and outward conducting channels is unfavorable and often detrimental. For example, over- or underexpression of voltage-gated K+ (Kv) channels can be cytotoxic and in some cases lead to disease. In this study, we demonstrated a novel role for S-acylation in Kv1.5 cell surface expression. In transfected fibroblasts, biochemical evidence showed that Kv1.5 is posttranslationally modified on both the NH2 and COOH termini via hydroxylamine-sensitive thioester bonds. Pharmacological inhibition of S-acylation, but not myristoylation, significantly decreased Kv1.5 expression and resulted in accumulation of channel protein in intracellular compartments and targeting for degradation. Channel protein degradation was rescued by treatment with proteasome inhibitors. Time course experiments revealed that S-acylation occurred in the biosynthetic pathway of nascent channel protein and showed that newly synthesized Kv1.5 protein, but not protein expressed on the cell surface, is sensitive to inhibitors of thioacylation. Sensitivity to inhibitors of S-acylation was governed by COOH-terminal, but not NH2-terminal, cysteines. Surprisingly, although intracellular cysteines were required for S-acylation, mutation of these residues resulted in an increase in Kv1.5 cell surface channel expression, suggesting that screening of free cysteines by fatty acylation is an important regulatory step in the quality control pathway. Together, these results show that S-acylation can regulate steady-state expression of Kv1.5.

  • quality control
  • potassium
  • channels
  • palmitoylation
  • posttranslational

voltage-dependent K+ (Kv) channels comprise the largest and most diverse class of ion channels. They are critical for numerous cellular and physiological functions including control of resting membrane potential, shaping of action potentials, and the regulation of fluid homeostasis and insulin release. Both the level of cell surface expression and the proper plasma membrane localization are important for Kv channel function. Similar to other ion channels, Kv channels are expressed at a relatively low cell surface density compared with other membrane proteins. Overexpression of cell surface Kv channels can be cytotoxic and in some cases lead to disease. Recent evidence indicates that overexpression of human Kv1.5 leads to enhanced apoptosis in pulmonary vascular smooth muscle cells (4). In addition, overexpression of Kv1.5 in rat cardiomyocytes dramatically shortens action potential duration, producing a phenotype similar to that observed in short QT syndrome (42). In the nervous system, increased expression of Kv3.4 has been linked to altered synaptic activity that may underlie the neurodegeneration observed in Alzheimer's disease (1). Clearly, the number of Kv channels on the cell surface is tightly regulated.

The steady-state cell surface expression of proteins is determined by the balance between the secretory and endocytic pathways. Surprisingly, however, little is known about the mechanisms that regulate Kv channels in these two pathways (20). Progress has been made with regard to a subset of ion channels in identifying mechanisms that regulate posttranslational trafficking. These include endoplasmic reticulum (ER) retention/retrieval motifs, binding to scaffolding proteins, and posttranslational modifications such as phosphorylation and glycosylation (for recent reviews, see Refs. 9 and 20).

Fatty acylation, the covalent attachment of long-chain fatty acids (C14–C20) to side chains of specific amino acid residues, is an essential modification for many proteins. Protein modification with fatty acids can regulate a variety of cellular events including protein trafficking, protein-protein interaction, lipid raft association, and protein signaling. Many types of protein fatty acylation exist. Classification of fatty acyl modifications is dependent on the type of fatty acid, the chemical nature of the linkage, and the time during protein translation when the modification occurs: either cotranslationally or posttranslationally. Two major categories of fatty acylation of eukaryotic proteins, myristoylation and S-acylation, have emerged.

N-myristoylation is a stable, cotranslational linkage that occurs on the NH2-terminal glycine residue via an amide bond (43). The enzymology and consensus recognition sequence (Met-Gly-X-X-X-Ser/Thr) on target proteins of this 14-carbon saturated fatty acid modification are well understood (33).

S-acylation is a reversible posttranslational reaction that involves the addition of fatty acids via an ester linkage to the thiol of a cysteine residue (39). The enzymology of these reactions is not fully understood and may involve the newly recognized aspartate-histidine-histidine-cystein (DHHC) protein S-acyltransferases (26). No clear consensus sequence for these enzymes has emerged. Traditionally, S-acylation is a term often used in the context of palmitoylation; however, it more accurately encompasses the conjugation of other fatty acids (C14 or longer) as well. Recent reports using mass spectrometry have revealed the heterogeneous nature of S-acylation in vivo (18, 19). Interestingly, the nature of the thioesterified fatty acids incorporated in a given protein appears to be cell type specific (32).

The voltage-gated K+ channel Kv1.5 mediates the ultrarapid K+ current (IKur) central to atrial action potential repolarization (27) and also plays a critical role in the regulation of arterial tone (5, 47). Here we demonstrate that the voltage-gated K+ channel Kv1.5 is a target of S-acylation through hydroxylamine-sensitive thioester bonds involving intracellular cysteine residues. Inhibition of S-acylation results in intracellular accumulation and a dramatic decrease in steady-state Kv1.5 channel protein levels. This novel posttranslational modification plays an important role in the regulation of Kv1.5 cell surface expression. Mutational analysis suggests that S-acylation may function by screening free cysteine residues that are unfavorable during Kv1.5 channel biogenesis/transport.

EXPERIMENTAL PROCEDURES

Materials.

Monoclonal antibody against the human transferrin receptor was purchased from Zymed (San Francisco, CA). Rabbit antiserum against the 112-amino acid NH2-terminal region of Kv1.5 was provided by the Tamkun laboratory (25). Complete protease inhibitor cocktail tablets were obtained from Roche (Penzenberg, Germany). Proteasome inhibitor Z-Leu-Leu-Leu-al (N-carbobenzoxyl-Leu-Leu-leucinal; MG-132), cathepsin L inhibitor acetyl-l-leucyl-l-leucyl-l-norleucinal (ALLN), serine and cysteine protease inhibitor leupeptin, the monoclonal tubulin antibody, dl-α-hydroxymyristic acid, 2-bromopalmitate, and protein A-Sepharose beads were purchased from Sigma-Aldrich (St. Louis, MO). Cerulenin was purchased from Calbiochem (San Diego, CA). Quick Change Site-Directed Mutagenesis Kit was purchased from Stratagene (La Jolla, CA). N-glycosidase F (PGNase F) with denaturing reaction buffer and NP-40 were purchased from New England BioLabs (Beverly, MA). EZ-Link Biotin-BMCC {(1-biotinamindo-4-[4′-aleimidomethyl)cyclohexanecarboxamido] butane} was from Pierce (Rockford, IL). The biotin-conjugated secondary antibody and streptavidin-conjugated Cy3 used in immunostaining were obtained from Jackson Immunoresearch Laboratories (West Grove, PA).

Cell lines, culture, and transfection.

Ltk cells (L-cells) stably expressing human Kv1.5 (24) under the control of a dexamethasone-inducible promoter were maintained in 60- or 100-mm culture dishes in Dulbecco's modified Eagle's medium supplemented with 10% horse serum, 2% penicillin-streptomycin (Invitrogen, Carlsbad, CA), 0.1% gentamicin (Invitrogen), and 500 μM G418 (Mediatech, Herndon, VA) at 37°C in a humidified atmosphere of 95% air-5% CO2. L-cells in 60-mm culture dishes were transfected at 60–80% confluence with 4 μg of DNA combined with 8 μl of Lipofectamine reagent (Invitrogen) in serum-free Dulbecco's modified Eagle's medium for 3–5 h and then changed to normal media, followed by drug treatments and use for experiments.

Preparation and deglycosylation of membranes.

L-cells stably expressing rat Kv1.5 were washed 2× with potassium-free PBS, scraped in 3 ml of phosphate-sucrose buffer (0.32 M sucrose, 5 mM Na2HPO4, complete protease inhibitors), and dounced 18 times. Homogenates were centrifuged at 1,000 g for 15 min at 4°C. The supernatant was saved, while the pellet was resuspended in 2 ml of phosphate-sucrose buffer and centrifuged again at 1,000 g for 15 min at 4°C. The supernatants were combined and centrifuged at 20,000 g for 60 min at 4°C. Pelleted membranes were resuspended in potassium-free PBS and stored at −80°C. Protein concentration was determined with the Bio-Rad detergent-compatible protein assay per the manufacturer's instructions. Deglycosylation experiments were performed on 20 μg of membranes. In a 10-μl reaction, membranes were incubated in 1× glycoprotein denaturing buffer for 10 min at 100°C. After the incubation, 1 μl of 10× G7 reaction buffer, 1 μl of 10% NP-40, and 1 μl of N-glycosidase F were added and incubated with membranes for 1 h at 37°C. Samples were analyzed by Western blot.

Cell lysis.

Cells were washed two times with potassium-free PBS and scraped in PBS with complete protease inhibitors. Cells were spun at 1,000 g for 4 min to pellet and lysed in lysis loading buffer (50 mM Tris-Cl, pH 6.8, 2.0% SDS, 10% glycerol, with complete protease inhibitors) for 5–10 min. Samples were heated for 5 min at 100°C and run through both a 20- and a 26-gauge needle 10 times each. The sample was centrifuged at 10,000 rpm for 10 min at room temperature. The supernatant was then used for a protein concentration assay, using the Bradford detergent-compatible assay according to the manufacturer's instructions (Bio-Rad, Hercules, CA).

Western blot analysis.

Thirty micrograms of cell lysis samples were run on a 4–12% Bis-Tris acrylamide gel (Invitrogen). After electrophoretic transfer to nitrocellulose, membranes were incubated with the anti-Kv1.5 or anti-tubulin antibodies (1:5,000 or 1:1,000 dilution, respectively). Bound primary polyclonal antibody was detected with a 1:10,000 dilution of horseradish peroxidase-conjugated goat anti-rabbit IgG (Zymed). Bound primary mouse antibodies were detected with a 1:5,000 dilution of horseradish peroxidase-conjugated goat anti-mouse IgG (Zymed). The Renaissance Western blot chemiluminescence reagent was used according to the manufacturer's protocol (Perkin Elmer Life Sciences, Wellesley, MA). Images were captured using the EpiChemi3 Darkroom (UVP, Upland, CA).

Immunoprecipitation and labeling of S-acylation sites.

The technique was based on a protocol described by Drisdel and Green (11). Cells were washed two times with potassium-free PBS, scraped, and collected at 1,000 g for 4 min. The pellet was resuspended and solubilized in solubilization buffer (1 ml per 60-mm dish) for 10 min at room temperature [solubilization buffer: 150 mM NaCl, 50 mM Tris, pH 7.4, 5 mM EDTA, 1% Triton X-100, 0.02% NaN3, 250 mM N-ethylmaleimide (NEM), and protease inhibitors]. After solubilization, the lysates were centrifuged (4°C at 13,000 rpm, 20 min). The supernatant was collected and precleared with 75 μl of protein A-Sepharose beads for 1 h with gentle mixing. The beads were then removed by centrifugation at 10,000 g for 1 min at 4°C. The sample was then incubated with anti-Kv1.5 antibodies (1:500 dilution) and 100 μl of protein A-Sepharose at 4°C with gentle mixing overnight. The beads were then collected by centrifugation at 10,000 g for 1 min at 4°C, resuspended in 1 M hydroxylamine in 1 M Tris, pH 7.4, and nutated for 1 h at room temperature. The beads were collected by centrifugation at 10,000 g at 4°C and washed two times for 5 min at room temperature with wash buffer (150 mM NaCl, 50 mM Tris, pH 7.4, 5 mM EDTA, 0.02% NaN3). After the second wash, the beads were collected by centrifugation at 10,000 g at 4°C and resuspended in 1 ml of 2 μM EZ Link Biotin-BMCC and nutated for 2 h at room temperature. After the incubation, the beads were removed by centrifugation at 10,000 g at 4°C, washed one time at 5 min at room temperature with wash buffer, collected once more by centrifugation, and resuspended in 55 μl of 1× SDS buffer. The proteins were eluted off from the beads at 100°C for 5 min. After a brief centrifugation, the supernatant was collected and analyzed by SDS-PAGE and immunoblotting.

Immunostaining.

Immunofluorescent labeling of Kv1.5 was performed essentially as described previously (22). Cells were fixed and permeabilized before incubation with anti-Kv1.5 antibodies (1:1,000). After incubation in biotinylated secondary antibody (1:200) and washing, bound antibodies were visualized with a Cy3-conjugated streptavidin (1:500). Imaging was performed on a Zeiss Axiovert 200M microscope equipped with standard epifluorescence and a Photometrics Cool Snap HQ charge-coupled device camera (Roper Scientific, Trenton, NJ).

Electrophysiology.

L-cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% horse serum and 1% penicillin-streptomycin. The cells were transfected with 0.1 μg of cDNA for wildtype or mutant, following the lipofection method using LipofectAMINE (Invitrogen); 20–24 h after transfection, the cells were trypsinized and used for patch clamp within 8 h. Current recordings were made with a MultiClamp 700A amplifier (Axon Instruments, Foster City, CA) in the whole cell configuration of the patch clamp technique. Experiments were done at room temperature (20–23°C); current recordings were low pass filtered and sampled at 10 kHz with a Digidata 1322A data acquisition system (Axon Instruments). Command voltages and data storage were controlled with pClamp9 software (Axon Instruments). Patch pipettes were pulled from borosilicate glass (Sutter Instrument, Novato, CA) with a two-stage puller (PP-830; Narishige, Tokyo, Japan) and fire polished. The resistance of the patch pipettes was between 2.5 and 3 MΩ in the standard extracellular solution. The cells were bathed in a bath solution containing (in mM) 130 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 10 HEPES, and 10 glucose, adjusted to pH 7.35 with NaOH. The pipettes were filled with intracellular solution containing (in mM) 110 KCl, 5 K4BAPTA, 5 K2ATP, 1 MgCl2, and 10 HEPES, adjusted to pH 7.2 with KOH. After achieving a gigaohm seal, the whole cell configuration was obtained by suction. Eighty percent of series resistance was compensated.

Data analysis.

The holding potential was −80 mV unless otherwise specified. The interpulse interval was 15 s. For activation, the membrane voltage was stepped from −80 to +60 mV at 10-mV increments. For inactivation, the membrane voltage was stepped from −80 to +90 at 10-mV increments followed by a test pulse (+30 mV). The voltage dependence of channel activation was fitted with a Boltzmann equation: y = 1/{1 + exp[−(EV)/k]}, in which k represents the slope factor, E is the applied voltage, and V is the voltage at which 50% of the channels are activated. Results are expressed as means ± SE.

RESULTS

Kv1.5 is posttranslationally modified via a hydroxylamine-sensitive thioester bond.

In transfected fibroblasts, Kv1.5 is expressed as both a nascent protein and a mature glycosylated protein. Western blot analysis of total cell membranes from mouse L-cells stably expressing human Kv1.5 (hKv1.5) revealed the presence of two immunoreactive bands of ∼68 and 75 kDa (Fig. 1A, lane 1). Treatment with N-glycosidase F collapsed the 75-kDa band into a single band of ∼65-68 kDa (Fig. 1A, lane 2). This is consistent with previous reports showing that the Kv1.5 antibody recognizes two bands (72 and 68 kDa) in rat heart membranes (2) and is glycosylated through asparagine (N) linkage in transfected cells (13). For transmembrane proteins, S-acylation typically occurs on intracellular cysteine residues, proximal to transmembrane domains (45). Kv1.5 contains two cysteine residues (C26, C36) in the NH2 terminus and four cysteine residues (C564, C581, C586, C604) in the COOH terminus proximal to either TM1 or TM6 (Fig. 1A).

Fig. 1.

Kv1.5 is posttranslationally modified via a hydroxylamine-sensitive thioester bond. Kv channel, voltage-gated K+ channel; L-cells, Ltk cells; WT, wildtype. For other abbreviations, refer to text. A: cell membranes from mouse L-cells stably transfected with human Kv1.5 (hKv1.5) were subject to SDS-PAGE before (lane 1) and after (lane 2) incubation with N-glycosidase F for 1 h at 37°C. Western blot analysis demonstrates that the anti-Kv1.5 antibody recognizes 2 bands of ∼68 and 75 kDa. The top glycosylated band (indicated by the arrow at ∼75 kDa) disappeared after N-glycosidase treatment. At right is a cartoon representation showing the topology of a single hKv1.5 channel α-subunit with all 10 cysteine residues labeled. B: immunoblot analysis of biotin-BMCC-labeled Kv1.5 after solubilization in increasing amounts of NEM. The same blot was stripped and probed with anti-Kv1.5 antibody (bottom). The amount of NEM required to block all free sulfhydryl groups was determined to be 250 mM. C: L-cells transiently expressing WT Kv1.5 were lysed in 250 mM NEM. The lysates were incubated in the absence (−) or presence (+) of 1 M hydroxylamine and then labeled by biotin-BMCC. Samples were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with streptavidin. The same blot was stripped and reprobed with anti-Kv1.5 antibody (bottom). A strong hydroxylamine-sensitive signal was detected for WT Kv1.5. D: bar graph showing the average densitometry data from the labeling of S-acylation sites for WT, the C26/36S double mutation (mut-2), the C564/581/586/604S tetra mutation (mut-4), and the mut-6 construct, in which all 6 cytoplasmic cysteine residues were changed to serines. Data are displayed as a percentage of control after subtraction of background [(−) hydroxylamine] and correction for expression levels (intensity of the biotin-BMCC signal is divided by the intensity of the anti-Kv1.5 signal) (n = 3; *P ≤ 0.05). At bottom are representative Western blot data corresponding to Kv1.5 constructs as shown at top. E: bar graph showing the average densitometry data from the labeling of S-acylation sites for WT and constructs with single-point mutations in cysteines in the COOH terminus of Kv1.5. Data are displayed as a percentage of control after subtraction of background [(−) hydroxylamine] and correction for expression levels (intensity of the biotin-BMCC signal is divided by the intensity of the anti-Kv1.5 signal) (n = 3; *P ≤ 0.05). F: bar graph showing the average densitometry data from the labeling of S-acylation sites for WT and constructs with single-point mutations in cysteines in the NH2 terminus of Kv1.5. Data are displayed as described above (n = 3; *P ≤ 0.05).

To determine whether Kv1.5 is a substrate for S-acylation, we used an approach based on hydroxylamine cleavage of the thioester bond between the fatty acid and cysteine residues on the channel protein (11). In this approach, free cysteine residues are first alkylated with NEM. Samples are then treated with hydroxylamine at neutral pH to cleave groups from modified cysteines. The now deprotected free cysteines are then labeled with the biotin-containing cysteine alkylating agent, Biotin-BMCC. This technique is significantly more sensitive than metabolic labeling (11). Free sulfhydryl groups must be blocked by NEM before Biotin-BMCC labeling. To adapt this technique to Kv1.5, we first determined the appropriate concentrations of NEM to block free cysteines before hydroxylamine treatment. As can be seen in Fig. 1B, incubation with 250 mM NEM is sufficient to alkylate free cysteine residues in Kv1.5 and prevent labeling with Biotin-BMCC. To assay for thioester-linked modifications of Kv1.5, immunopurified protein from NEM-treated samples was incubated with or without hydroxylamine, and newly deprotected cysteines were labeled with Biotin-BMCC. A robust hydroxylamine-induced Biotin-BMCC labeling was observed for wildtype Kv1.5 (Fig. 1C). Note that <10% of labeling of wildtype Kv1.5 with Biotin-BMCC was detected for samples with an equivalent amount of protein but not treated with hydroxylamine (Fig. 1C). This demonstrates that this approach can be used successfully to detect thioester modification on Kv1.5.

To prove that the signal is indeed generated from modifications on cysteine residues and to localize regions of the channel that are subject to S-acylation, we created cysteine mutant channel proteins. These included the C26/36S double mutation (mut-2), the C564/581/586/604S tetra mutation (mut-4), and the hexa mutation (mut-6), which includes a mutation of each of the NH2- and COOH-terminal cysteines. The results of our S-acylation assay comparing wildtype with the three different cysteine mutants are summarized in Fig. 1D. Importantly, mutation of all six intracellular cysteine residues to serine (mut-6) resulted in a loss of detectable Biotin-BMCC signal. As indicated, mut-2 and mut-4 eliminated 48 and 58% of the labeling, respectively. In addition, single serine substitutions at each of the COOH-terminal cysteines identified position C604 as the site of S-acylation on the COOH terminus(Fig. 1E), whereas single serine substitutions on the NH2 terminus identified position C26 as the modified site (Fig. 1F). These results show that Kv1.5 channel protein is modified on both the NH2 terminus and the COOH terminus by hydroxylamine-sensitive thioester bonds, consistent with S-acylation. The major source of S-acylation for proteins without enzymatic activity derives from fatty acylation. The S-acylation of Kv1.5 most likely derives from palmitoylation or other thioester-linked fatty acid modifications.

Inhibition of S-acylation decreases steady-state levels of Kv1.5 channel protein, which is dependent on COOH-terminal but not NH2-terminal cysteines.

Pharmacological inhibition of S-acylation is a useful tool to investigate the functional significance of these posttranslational modifications (8). Cerulenin, an inhibitor of fatty acid synthetase, is an effective means to block protein fatty acylation (10, 35). Treatment of L-cells with cerulenin reduced overall Kv1.5 channel protein in a concentration-dependent manner (Fig. 2, A and B). Using fibroblast cells expressing Kv1.5, we isolated cell lysates after 48-h treatment with increasing concentrations of cerulenin. Western blot analysis comparing samples of equal total protein showed a concentration-dependent reduction in the steady-state levels of Kv1.5 channel protein. Tubulin levels were used as a loading control. Results are summarized in Fig. 2B, in which densitometry was used to quantitate the changes in protein levels. Interestingly, cerulenin treatment initially caused a larger decrease in the amount of glycosylated Kv1.5 than nonglycosylated channel protein (Fig. 2B). However, these differences most likely reflect a greater pool of nonglycosylated protein.

Fig. 2.

Inhibitors of S-acylation but not myristoylation decrease steady-state levels of channel protein through COOH-terminal cysteines. A: cellular lysates were prepared from L-cells transiently expressing Kv1.5 and treated for 24 h with cerulenin (0, 30, 100, and 300 μM). Samples were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with an antibody raised against Kv1.5. Protein levels were determined by Bradford assay, and an equal amount of protein was loaded for each sample. Tubulin levels were used as an additional loading control. Shown is a representative Western blot that illustrates that treatment with cerulenin results in a dose-dependent reduction of Kv1.5 protein. B: results from densitometry measurements are summarized in the bar graph of data normalized to nontreated control (n = 3; *P ≤ 0.05). C: cellular lysates prepared after 24 and 48 h from bromopalmitate (Bromo; 100 μM)-treated (+) and nontreated (−) L-cells stably expressing Kv1.5 or transiently expressing Kv1.1 were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with an antibody raised against Kv1.5 or V5 tag in the COOH terminus of Kv1.1. Protein levels were determined by Bradford assay, and equal amounts were loaded for each sample. Tubulin levels were used as an additional loading control. D: bar graph showing normalized data from experiments in C. Densitometry measurements were used to compare protein levels for both the glycosylated (Glyco) and nonglycosylated Kv1.5 bands (n = 8) at 24 and 48 h of treatment as well as Kv1.1 levels at 48 h. Signals were normalized to nontreated controls (−). (*P ≤ 0.05) Similar results were recorded from cells transiently expressing Kv1.5 (data not shown). E: bar graph showing the average densitometry data from cells treated for 24 h with 500 μM 2-hydroxymyristate (2-OH-myrist). Data were normalized to nontreated control (n = 3). At bottom are representative Western blot data corresponding to Kv1.5 signal in control and 2-hydroxymyristate-treated cells. Protein levels were determined by Bradford assay, and equal amounts of protein were loaded for each sample. Tubulin levels were used as an additional loading control. Results show that inhibition of myristoylation did not significantly affect steady-state levels of Kv1.5. F: cellular lysates prepared from 100 μM 2-bromopalmitate (48 h at 37°C)-treated (+) and nontreated (−) L-cells transiently expressing Kv1.5 WT, mut-2, mut-4, and mut-6 channel were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with an antibody raised against Kv1.5. Protein levels were determined by Bradford assay, and an equal amount was loaded for each sample. Tubulin levels were used as an additional loading control.

Treatment of cells with a second blocker, 2-bromopalmitate, an inhibitor of fatty acylation with some specificity for palmitoylation (7, 46), also significantly decreased Kv1.5 expression (Fig. 2, C and D). We isolated cell lysates after 24- and 48-h treatment with 100 μM 2-bromopalmitate. Western blot analysis comparing samples of equal total protein showed a dramatic time-dependent reduction in the steady-state levels of Kv1.5 channel protein after both 24- and 48-h treatment. In contrast, levels of Kv1.1 were not affected by 48-h treatment with 2-bromopalmitate (Fig. 2, C and D). Results are summarized in Fig. 2D, in which densitometry was used to quantitate the changes in protein levels. Similar to cerulenin, 2-bromopalmitate treatment initially caused a larger decrease in the amount of glycosylated Kv1.5 than nonglycosylated channel protein (Fig. 2D). As shown, a 24-h treatment of 2-bromopalmitate eliminated ∼84% of glycosylated protein but reduced only ∼65% of nonglycosylated protein. This difference was lost after 48-h treatment, during which levels of both nascent and mature forms were reduced nearly fivefold. Similar results were obtained with COS-1 cells, indicating that this effect was independent of the cell type chosen for heterologous expression (data not shown).

Importantly, saturating concentrations of 2-hydroxymyristate (500 μM), a known inhibitor of myristoylation (29, 30), had no effect on steady-state levels of Kv1.5 channel protein (Fig. 2E). These data are consistent with our finding that Kv1.5 is a fatty acid modified by a hydroxylamine-sensitive thioester bond on cysteine residues (Fig. 1). Next, we tested the hypothesis that the decrease in channel protein induced by the inhibition of S-acylation with 2-bromopalmitate was due to a direct effect on Kv1.5. L-cells transiently expressing either wildtype or cysteine mutants were treated with (+) or without (−) 2-bromopalmitate. Equal amounts of protein were subjected to SDS-PAGE and assayed by Western blot. Mutations that included the four COOH-terminal cysteines, but not the two NH2-terminal cysteines, resulted in a loss of the 2-bromopalmitate effect on steady-state protein levels (Fig. 2F). In addition, truncation of 57 amino acids from the COOH terminus, which removes all four COOH-terminal cysteines, also resulted in a loss of the 2-bromopalmitate effect (data not shown). Taken together, these results show that inhibition of S-acylation, but not myristoylation, dramatically reduces steady-state levels of Kv1.5 channel protein and suggest that S-acylation of COOH-terminal cysteines may be an important regulatory step in the expression of Kv1.5.

Inhibition of S-acylation causes Kv1.5 to accumulate intracellularly before degradation, while mutation of COOH-terminal but not NH2-terminal cysteines results in an increase in cell surface expression.

To investigate the effect of 2-bromopalmitate on the subcellular localization of Kv1.5, we immunostained L-cells stably transfected with Kv1.5. Cells were cultured for 24 h with no treatment or with 100 μM 2-bromopalmitate and imaged by fluorescence microscopy. Figure 3A, top, shows the punctate cell surface distribution of Kv1.5. In contrast, cells treated for 24 h with 2-bromopalmitate (Fig. 3A, bottom) showed a distinct perinuclear expression pattern with nearly no detectable plasma membrane staining. Treatment of cells for 48 h resulted in little detectable protein, consistent with Western blot results (data not shown). Similarly, measurement of Kv1.5 current density by patch clamp analysis demonstrates that whole cell steady-state current levels are dramatically reduced in cells treated with 2-bromopalmitate (Fig. 3B). These results show that inhibition of S-acylation results in the redistribution of Kv1.5 to intracellular compartments, suggestive of retention before degradation. The lower levels and mislocalization suggest that 2-bromopalmitate may facilitate Kv1.5 degradation.

Fig. 3.

2-Bromopalmitate causes intracellular accumulation of channel, while protease inhibitors block the degradation of Kv1.5. Mutation of COOH-terminal, but not NH2-terminal, cysteines results in a loss of the effect of 2-bromopalmitate and an increase in cell surface channel protein. A: immunofluorescence localization of Kv1.5 in L-cells stably expressing channel protein shows a punctate cell surface distribution (top) that is altered with 100 μM 2-bromopalmitate for 24 h at 37°C (bottom). Cells were immunostained with anti-Kv1.5 antibody (1:1,000) and detected using a Cy3-conjugated fluorochrome. Shown are representative images captured on a Zeiss Axiovert 200M epifluorescence microscope. B: whole cell Kv1.5 current recordings (−80-mV holding potential; 10-mV steps to +60 mV) were made from control cells stably expressing hKv1.5 and cells treated with 100 μM 2-bromopalmitate. Currents were normalized for differences in cell size and plotted as means ± SE (n = 4). C: cellular lysates prepared from L-cells stably expressing Kv1.5 were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with an antibody raised against Kv1.5. Protein levels were determined by Bradford assay, and equal amounts were loaded for each sample. Tubulin levels were used as an additional loading control. Cell treatments, from left to right, are as follows: 1) control 48 h, 2) 100 μM 2-bromopalmitate, 3) 100 μM 2-bromopalmitate + 30 μM N-carbobenzoxyl-Leu-Leu-leucinal (MG-132; MG), 4) 100 μM 2-bromopalmitate + 50 μM N-acetyl-l-leucyl-l-leucyl-l-norleucinal (ALLN), 5) 100 μM 2-bromopalmitate + 100 μM leupeptin (Leu), 6) 30 μM MG-132, 7) 50 μM ALLN, and 8) 100 μM leupeptin. D: L-cells transiently expressing Kv1.5 WT (Wt), mut-2, mut-4, and mut-6 channel were treated with 100 μM 2-bromopalmitate for 48 h at 37°C and then immunostained with anti-Kv1.5 antibody (1:1,000) and detected using a Cy3-conjugated fluorochrome. Shown are representative of images captured on a Zeiss Axiovert 200M epifluorescence microscope. E: representative whole cell current recordings (−80-mV holding potential; 10-mV steps to +60 mV) from control cells transiently expressing hKv1.5 and cells transiently expressing mut-6 with and without 2-bromopalmitate. Currents were normalized for differences in cell size and plotted as means ± SE for WT (n = 20), mut-6 (n = 14), and mut-6 + treatment with 100 μM 2-bromopalmitate (n = 3). Mean current density of mut-6 is 2-fold larger compared with Kv1.5 WT (*P < 0.05), but 2-bromopalmitate did not affect the current density of the mut-6.

Degradation of intracellularly retained protein can involve a number of proteolytic systems as part of the quality control mechanisms in the secretory pathway (44). We tested the involvement of the ubiquitin-proteasome and lysosomal pathways in the degradation of Kv1.5 following 2-bromopalmitate treatment. MG-132, a potent inhibitor of 26S proteasome activity, prevented degradation and led to the accumulation of nonglycosylated channel protein (Fig. 3C). The accumulation of immature protein suggests that S-acylation plays a role in the biosynthetic processing of Kv1.5. Similarly, ALLN, a modest proteasome inhibitor that blocks cathepsins, also prevented the degradation. In contrast, the lysosome inhibitor leupeptin failed to rescue the degradation induced by 2-bromopalmitate. Inhibitors alone had no significant effects on steady-state levels of channel protein (Fig. 3C). Collectively, these results suggest that S-acylation prevents Kv1.5 from entering a degradation pathway involving the proteasome.

We also tested the hypothesis that the intracellular accumulation induced by the inhibition of S-acylation with 2-bromopalmitate was due to a direct effect on Kv1.5. L-cells transiently expressing either wildtype or cysteine mutants were treated with or without 2-bromopalmitate. Mutations that included the four COOH-terminal cysteines, but not the two NH2-terminal cysteines, resulted in a loss of the 2-bromopalmitate effect on intracellular accumulation (Fig. 3D).

Interestingly, both the mut-4 and mut-6 channel constructs expressed significantly more channel protein (Fig. 2F) and appeared to have increased cell surface labeling (Fig. 3D). In addition, cysteine substitution at position 604 resulted in a large increase in channel protein (Supplemental Fig. S1A; supplemental data are available at the online version of this article), while single substitutions at NH2-terminal cysteines had no effect (Supplemental Fig. S1B). To determine whether this increase in protein represented functional Kv1.5 channel, we measured the current levels in transiently transfected L-cells. Shown in Fig. 3E are representative whole cell current recordings from cells expressing either wildtype Kv1.5 or the six-cysteine mutant. The current amplitude of mut-6 was significantly greater compared with control. A plot of the average current density showed a statistically significant increase in the current density at potentials positive to −20 mV when all six cytoplasmic cysteine residues were changed to serine (Fig. 3E). For example, the average current density at +60 mV measured 166 ± 37 pA/pF and 349 ± 61 pA/pF for wildtype and mut-6, respectively. In contrast, there was no change in the steady-state activation (V1/2 control = −16.0 ± 2.0 mV, k = 8.7 ± 1.0; V1/2 mut-6 = −15.8 ± 1.6 mV, k = 6.6 ± 0.3) or inactivation (V1/2 control = −20.1 ± 2.7 mV, k = −8.5 ± 0.9; V1/2 mut-6 = −23.6 ± 2.2 mV, k = −6.9 ± 0.7) properties of the channel. These results are consistent with our biochemical data showing large increases in channel protein levels. Importantly, unlike the wildtype (Fig. 3B), 2-bromopalmitate had no effect on current density of the mut-6 construct (Fig. 3E). Collectively, these data indicate that COOH-terminal, but not NH2-terminal, cysteines are important for S-acylation-dependent regulation of cell surface expression. Although we cannot exclude the possibility that there are antagonistic effects of S-acylation of different cysteine residues, our data also suggest that S-acylation is not required, but rather may regulate cell surface expression by providing a quality control checkpoint that is dependent on the thioacylation of intracellular cysteines.

S-acylation occurs in the biosynthetic pathway of nascent channel protein.

S-acylation has been shown to occur on a wide variety of cellular proteins, and the intracellular sites of thioacylation reaction can be quite diverse (3). Determining the cellular location where specific proteins are fatty acylated is difficult and therefore has only been determined for a small subset of proteins (3). In our system, blocking S-acylation results in a decrease of steady-state levels of channel protein. The mechanism of this effect may occur early by inhibiting a step along the biosynthetic/exocytic pathway. Alternatively, blocking S-acylation may exert its effects on the channel by influencing the endocytic pathway, thereby decreasing the lifespan of the protein. We took advantage of our dexamethasone-inducible stable cell model to gain information about the timing of channel S-acylation during the maturation of Kv1.5 protein. As shown in Fig. 4A, we examined the biosynthetic processing of Kv1.5 after a 16-h pulse of dexamethasone. The early nonglycosylated channel form is dominant within the first 16 h of induction. However, by 64 h, the majority of the channel has been converted to the mature glycosylated form. The labeling of S-acylation sites via a hydroxylamine-sensitive thioester bond can be detected as early as 16 h in channel protein biosynthesis/transport (Fig. 4B). As the protein matures at 40 and 64 h, the increase in detectable biotin-BMCC signal reflects the transition from nascent protein to the glycosylated form as seen with the anti-Kv1.5 antibody signal.

Fig. 4.

S-acylation of Kv1.5 occurs early in channel biosynthesis/transport during the time in which 2-bromopalmitate reduces steady-state protein levels. A: time course experiment to follow the maturation of the glycosylated channel protein. Kv1.5 expression was induced for 16 h, and cells were harvested at the indicated times. B: immunoblot analysis of biotin-BMCC-labeled Kv1.5 after incubation in the absence (−) or presence (+) of 1 M hydroxylamine. The same blot was stripped and probed with anti-Kv1.5 antibody (bottom). The analysis was performed at 3 different time points, as indicated at top. C: Kv1.5 channel expression was induced for 48 h, and 2-bromopalmitate was added at the indicated times. Results show that inhibition of S-acylation during the first 24 h, but not the second 24-h period, dramatically reduces channel expression.

Next, we examined the timing of the bromopalmitate effect. L-cells were treated with 2-bromopalmitate at increasing times after channel induction. Control cells were induced for 16 h and allowed to express channel for a total of 48 h. Consistent with our previous results, a majority of channel protein existed as the mature glycosylated form. Functional cell surface channel expression, as determined by the current density, reached a maximum during this time interval (see Supplemental Fig. S2A). Addition of 2-bromopalmitate to the cell culture during the first 24 h resulted in a 52% decrease in channel expression. In contrast, addition of 2-bromopalmitate during the time period of 24–48 h did not significantly alter channel expression levels (Fig. 4C). Treatment of cells with 2-bromopalmitate for the entire 48-h period resulted in a large reduction in channel expression. Live cell labeling with antibodies directed against an extracellular epitope of Kv1.5 were used to quantitate surface expression at various time points after transient expression. The data summarized in Supplemental Fig. S2B indicate that the wildtype and mut-6 channels reach the cell surface in the same time frame. Together, these results indicate that inhibition of S-acylation affects only newly synthesized channel protein, perhaps during biogenesis or early ER/Golgi transport.

DISCUSSION

This study represents one of the first reports showing fatty acid modification of Kv channel α-subunits and demonstrates a unique role for S-acylation in the regulation of Kv channel expression. Fatty acid modification has been demonstrated for a small subset of voltage-gated ion channel proteins and auxiliary subunits such as sodium channel α-subunits (36), β-subunits of voltage-gated calcium channels (6), Kv channel-interacting protein (KChIP) (41), and recently Kv1.1 (15). Each of these proteins was found to be palmitoylated using metabolic labeling. However, the possibility for the attachment of other fatty acid species under basal conditions, during which cells are not exposed to a large excess of lipid, was not determined. Notably, the report of palmitoylation of Kv1.1 (15) required expression of large amounts of purified channel protein from Spodopter frugiperda (Sf9) insect cells, which are apparently incapable of acylating proteins with other fatty acids such as stearate (34). Interestingly, our results show that pharmacological inhibition of fatty acylation does not affect Kv1.1 expression levels. Initially, our efforts to assay palmitoylation of Kv1.5 in mammalian cells by metabolically labeling with radiolabeled palmitate were unconvincing. This is not surprising, given the difficulty of detecting palmitoylation for low-abundance integral membrane proteins, such as ion channels (12). The method for detecting thioacylated proteins used in this study is highly sensitive and avoids the potential complication of incubation with excess lipid (11). This technique, however, does not allow us to determine the molecular identity of the thioester modification, and therefore we have been careful to describe the Kv1.5 modification as S-acylation. Recent reports indicate that thioacylation can include a variety of attached fatty acid moieties including palmitate, stearate, and oleate (18). Future experiments using mass spectrometry approaches may be necessary to identify the nature of the fatty acid linked to Kv1.5.

Potential mechanisms of S-acylation-dependent regulation of Kv1.5 protein levels.

S-acylation of voltage-gated channel auxiliary subunits appears to share a common function in controlling the plasma membrane targeting of their associated channels. The functions of fatty acid addition to polytopic channel α-subunits are less clear. Recent reports suggest that palmitoylation of Kv1.1 at position C243 in the S2-S3 linker region of the channel may modulate voltage sensing (15). Our results show that thioacylation of Kv1.5 is not required for steady-state function of Kv1.5 but rather represents a mechanism of regulation for expression of fully glycosylated mature channel protein.

S-acylation might control channel expression by regulating the lifespan of the protein. This may occur as reported for the cation-dependent mannose-6-phosphate receptor, in which palmitoylation prevents the receptor from entering the lysosomal pathway during protein recycling (40), or as reported for the CCR5 receptor (a chemokine receptor), in which palmitoylation prolongs the half-life and slows rapid turnover of the receptor (31). However, our data show that inhibiting S-acylation after the channel has matured and presumably reached the cell surface does not affect steady-state levels of channel protein. In addition, mutation of intracellular cysteine residues does not reduce channel expression but rather increases channel current density and steady-state protein levels. These results together with the finding that we can detect S-acylation of nascent channel protein during initial stages in biogenesis (Fig. 4B) suggest that fatty acid addition has an important role in the formation or trafficking of new channel proteins. This result is similar to that reported for the α-subunit of the sodium channel, in which palmitoylation occurs in the early stages of biosynthesis (36). Importantly, our results suggest that this early stage modification of Kv1.5 prevents the channel from accumulating intracellularly and entering the degradation pathway. Therefore, fatty acid modification of the full-length channel via a thioester bond may function by screening free cysteine residues that are unfavorable during Kv1.5 channel biogenesis/transport. It has been shown that free cysteines on the intracellular termini of Shaker potassium channels are in close enough proximity to permit the formation of intersubunit disulfide bonds (37). This mechanism might be similar to the masking of ER retention/retrieval signals by the binding of chaperone or scaffold proteins or through intersubunit interactions (20). Interestingly, posttranslational modification by phosphorylation at sites flanking ER retention motifs has been shown to suppress ER retention and promote export (38).

Our results indicate that S-acylation of a single COOH-terminal cysteine, but not NH2-terminal cysteines, is important for the regulation of steady-state channel protein levels. It is notable that the COOH terminus has been implicated in the surface expression of Kv channels (17). The identified COOH-terminal expression motif is located in close proximity to the cysteine at position 604. It remains to be determined whether S-acylation at this position affects the function of the DLRRSL motif or vice versa. Aberrant trafficking of ion channels is responsible for an increasing number of channelopathies that have been implicated in the pathophysiology of a growing list of human disease (14). Previous work demonstrating the targeting of Kv1.5 to caveolar lipid rafts suggested that protein-lipid interactions should be considered as a mechanism of channel localization (24). Initially, experiments in this study were performed to address the potential role for fatty acid modification in targeting Kv1.5 to lipid raft microdomains. We hypothesized that fatty acid lipid anchors may increase the hydrophobicity of channel proteins and thereby stabilize their associations with the tightly packed lipid environment of a raft. However, although inhibition of S-acylation by 2-bromopalmitate treatment caused a significant reduction in steady-state levels of channel protein, we still detected Kv1.5 association with detergent-resistant membrane fractions (data not shown). Therefore, we do not believe that S-acylation controls association with raft microdomains. Nevertheless, our results demonstrate a clear role for S-acylation in the regulation of steady-state levels of Kv1.5 channel protein. We have shown that interfering with this normal lipid modification promotes proteolytic degradation and decreases cell surface expression.

As evidenced in the literature, there is increasing interest in the potential role for cellular lipids in the regulation of Kv channel localization and function (16, 21, 23, 28). This report is one of the first to show fatty acid modification of Kv channel α-subunits and demonstrates a unique role for S-acylation in the regulation of Kv channel expression. Importantly, it suggests that alteration of cellular lipids, either by disease, diet, metabolic state, or the use of drugs, can potentially affect the expression of Kv channels.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-070973 (J. R. Martens).

Acknowledgments

We thank Dr. Jorge Iniguez-Lluhi for careful review of this manuscript.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract