Translocation of an endoproteolytically cleaved maxi-K channel isoform: mechanisms to induce human myometrial cell repolarization

Abstract

Large conductance Ca²⁺- and voltage-activated K+ (maxi-K) channels modulate human myometrial smooth muscle cell (hMSMC) excitability; however, the role of individual alternatively spliced isoforms remains unclear. We have previously shown that the transcript of a human maxi-K channel isoform (mK44) is expressed predominantly in myometrial and aortic smooth muscle and forms a functional channel in heterologous expression systems. The mK44 isoform contains unique consensus motifs for both endoproteolytic cleavage and N-myristoylation, although the function of these post-translational modifications is unknown. The goal of these studies was to determine the role of post-translational modifications in regulating mK44 channel function in hMSMCs. An mK44-specific antibody indicated that this channel is localized intracellularly in hMSMCs and translocates to the cell membrane in response to increases in intracellular Ca²⁺. Immunological analyses using an N-terminally myc-tagged mK44 construct demonstrated endoproteolytical cleavage of mK44 in hMSMCs resulting in membrane localization of the mK44 N-termini and intracellular retention of the pore-forming C-termini. Caffeine-induced Ca²⁺ release from intracellular stores resulted in translocation of the C-termini of mK44 to the cell membrane and co-localization with its N-termini. Translocation of mK44 channels to the cell membrane was concomitant with repolarization of the hMSMCs. Endoproteolytic digest of mK44 did not occur in HEK293 cells or mouse fibroblasts. MK44 truncated at a putative N-myristoylation site did not produce current when expressed alone, but formed a functional channel when co-expressed with the N-terminus. These findings provide novel insight into cell-specific regulation of maxi-K channel function.

Large conductance Ca²⁺- and voltage-dependent potassium channels (maxi-K channels; mK) regulate membrane excitability by generating a potent repolarizing K⁺ current in response to depolarizing stimuli (McCobb et al. 1995; Toro et al. 1998). Maxi-K channel openers have been theorized to be a potential treatment for hypertension, pre-term labour, and ischaemic stroke (Carl et al. 1996; Liu et al. 1998; Brenner et al. 2000; Gribkoff et al. 2001; Sobey, 2001). However, the widespread expression of maxi-K channels has hampered the targeting of this protein for pharmacological intervention. Thus, understanding cell-specific regulation of maxi-K channels will aid in the identification of substances that could regulate maxi-K channels for therapeutic purposes.

One molecular mechanism that may determine the cell-specific function of maxi-K channels is alternative splicing of the transcript encoding the pore-forming α-subunit (Saito et al. 1997). Splicing results in maxi-K channel isoforms that are differentially regulated by Ca²⁺ and voltage, oestrogen, ACTH and post-translational modification by kinases (Dopico et al. 1994; Nara et al. 1998; Xie & McCobb, 1998; Hall & Armstrong, 2000; Tian et al. 2001; Zarei et al. 2001; Holdiman et al. 2002). Our laboratory previously identified and functionally characterized a human maxi-K channel α-subunit isoform (mK44) generated by introduction of 44 additional amino acids into the first intracellular loop of the canonical channel (Korovkina et al. 2001). This isoform is expressed predominantly in myometrial and aortic smooth muscle and demonstrates decreased sensitivity to intracellular Ca²⁺ (Korovkina et al. 2001).

The addition of 44 amino acids into the mK44 isoform introduces several putative consensus motifs for endoproteolytic digest and one putative site for N-myristoylation.

Interestingly, several ion channels and exchangers have been identified as substrates for endoproteolytic cleavage (Gerhardstein et al. 2000; Jovov et al. 2001; Bano et al. 2005; Wachter & Schwappach, 2005). N-myristoylation is a process catalysed by myristoyl CoA–protein N-myristoyl transferase (NMT; EC 2.3.1.97), which covalently attaches a myristic fatty acid residue to N-terminal glycines in proteins. This process is essential for intracellular trafficking of certain proteins to the cell membranes (Sakoda et al. 1995; Tanaka et al. 1995; Vaandrager et al. 1996; Raju et al. 1997) by hiding or unmasking the myristic acid residue in response to a variety of stimuli (switch mechanisms) (Resh, 1999; Senin et al. 2002; Hantschel et al. 2003). Recent studies have shown that myristoylation motifs aid in intracellular targeting of K⁺ channel interacting proteins (KChIP), which results in trafficking of Kv channel proteins (O'Callaghan et al. 2003).

This study aimed to elucidate the functional significance of the endoproteolytic cleavage and N-myristoylation consensus motifs within mK44 in hMSMCs. We conclude that the mK44 isoform of maxi-K channels is inactive in quiescent hMSMCs due to endoproteolytic digest and retention in the endoplasmic reticulum (ER). Translocation of mK44 to the cell membrane in response to increases in intracellular Ca²⁺ (Kupittayanant et al. 2002) coincides with repolarization of hMSMC membranes. N-Myristoylation of the mK44 isoform regulates ER retention and membrane trafficking of mK44. These findings provide further insight into cell-specific regulation of maxi-K channel function in smooth muscle.

Methods

Generation of constructs

A c-myc tag (myc) was generated by annealing 5′-pGG-CCGCCACCATGGAGCAGAAGCTGATCAGCGAGGA-GGACCTGTGC-3′ sense and 5′-pGGCCGCACAGGTC-CTCCTCGCTGATCAGCTTCTGCTCCATGGTGGCA-3′ antisense primers and fusing a double-stranded tag to the 5′ end of mK44. An adenoviral construct was generated and designated myc/mK44. MK44 was fused with enhanced green fluorescent protein (eGFP) (mK44/eGFP) in the pEGFP-N1 vector (Clontech, Mountain View, CA, USA) for live cell imaging. Other constructs created from the myc/mK44 are listed in Table 1. Briefly, amino acids 1–62 of mK44 were fused with a c-myc epitope (myc/mK1-62). Amino acids 69–1157 of mK44 were fused C-terminally with a V5 epitope (mK69-1157/V5) in pBUDCE4.1 expression vector (Invitrogen Carlsbad, CA, USA) for immunocytochemical studies. Alternatively, the same mutant was fused with eGFP at the C-terminus (mK69-1157/eGFP) in pEGFP-N1 vector (Clontech, Mountain View, CA, USA) for direct visualization in patch-clamp experiments.

Table 1.

Constructs generated from mK44

Name	Description	Use
myc/mK44	mK44 fused with c-myc epitope at its N-terminus	To study expression of wild-type mK44 by immunological methods
mK44/eGFP	mK44 fused with eGFP at its C-terminus	Direct visualization of membrane trafficking of mK44 in live cells
myc/mK1-62	mK1-62 fused with c-myc epitope at its N-terminus	To study expression of the proteolytically cleaved N-terminus of mK44 by immunological methods
mK69-1157/V5	mK69-1157 fused with V5 epitope at its C-terminus	To study expression of the proteolytically cleaved C-terminus mK44 by immunological methods
mK69-1157/eGFP	mK69-1157 fused with eGFP at its C-terminus	Direct visualization of the product of proteolytic digest of mK44 in patch-clamp experiments

Heterologous expression of maxi-K channel constructs

An adenoviral construct of myc/mK44 was used to infect hMSMC cultures at 1 μl ml⁻¹ of 1–1.2 × 10¹² particles ml⁻¹ of purified construct. Cells were harvested 48–72 h post-infection. Infection efficiency was monitored using an adenoviral eGFP reporter gene. Constructs in plasmid expression vectors (mK44/eGFP, mK69-1157/V5, mK69-1157/eGFP and myc/mK1-62) were transfected into hMSMCs and human embryonic kidney cells (HEK293F) cells using GeneJammer transfection reagent (Stratagene, La Jolla, CA, USA) according to the manufacturer's instructions.

Human tissue

Human myometrial tissue from the lower uterine segment was collected from women undergoing elective Caesarean section under spinal anaesthesia at late pregnancy (38–40 weeks gestation) in the absence of spontaneous or induced labour contractions. All patients signed written consent forms approved by Internal Review Board 199809066. Myometrial strips 2 mm × 2 mm × 5 mm long were immediately isolated and incubated in Ca²⁺-free PBS (Invitrogen, Carlsbad, CA, USA) in the presence or absence of 20 mm caffeine (Sigma-Aldrich, St Louis, MO, USA) for 20 min in a tissue culture incubator. Incubation was stopped by placing the strips in 4% paraformaldehyde (PFA) at 4°C overnight. Tissues were cryoprotected by subsequent incubations in 15% and 30% sucrose, embedded, and stored at −80°C. Experiments were repeated 3 times.

Cell cultures

Human myometrial smooth muscle cells (hMSMCs) were isolated from myometrial tissues collected as described above. Cells were dispersed by incubation with collagenase (Worthington Biochemicals, Lakewood, NJ, USA) and placed on laminin-coated coverslips for immunocytochemistry or 35 mm glass-bottom dishes for functional assays. Cells were allowed to adhere to the dish bottom for 48 h, and a smooth muscle cell phenotype was subsequently induced with Dulbecco's modified Eagle's medium (DMEM)–F12 supplemented with 0.5% fetal bovine serum (FBS) for 7 days. HEK293F cells (Invitrogen, Carlsbad, CA, USA) were grown to 60–70% confluency in DMEM–F12 medium supplemented with 10% FBS. Caffeine (20 mm) (Sigma-Aldrich, St Louis, MO, USA) in Ca²⁺-free PBS (Invitrogen, Carlsbad, CA, USA) was added for 20 min. Cells in Ca²⁺-free PBS were used as controls. An inhibitor of NMT, 2-hydroxymyristic acid (MP Biomedicals, Irvine, CA, USA) was added at 100 nm for 48 h. Experiments were repeated at least 4 times.

Membrane potential detection in hMSMCs

Mouse laminin (Sigma-Aldrich, St Louis, MO, USA) was used to coat glass-bottom 35 mm cell culture dishes (MatTek Corporation, Ashland, MA, USA). Differentiated hMSMCs were pre-loaded for 30 min with 2 μm of a membrane potential-sensitive dye DiBAC₄(3) (Invitrogen, Carlsbad, CA, USA) in buffer A containing (mm): 20 Hepes, 140 NaCl, 2 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose. Buffer B, same as buffer A, but devoid of MgCl₂ and CaCl₂, was used in subsequent experimental procedures. Experimental incubations included: (1) buffer B alone, and (2) buffer B supplemented with 20 mm caffeine. DiBAC₄(3) was present in the buffer at all times. Inhibition of maxi-K channels was attained by 40 min incubation in 50 nm iberiotoxin (IbTX) (Sigma-Aldrich, St Louis, MO, USA) in buffer A prior to and during the pre-loading with DiBAC₄(3). IbTX was present during the 20 min experimental incubation in buffer B containing 20 mm caffeine. Cells were observed every 2 min and signal intensity was registered with the 488 nm laser of an LSM 5 Zeiss confocal scanning microscope (Zeiss, Jena, Germany). Changes in signal intensity, reflecting changes in membrane polarization, were averaged at each time point in 3–5 regions of interest using an LSM 5 Image Browser (Zeiss, Jena, Germany). Additionally, the pseudocolored images of the entire field of observation were also quantified. Each experiment was repeated at least 3 times.

Live cell imaging

MK44/eGFP in hMSMCs was visualized using a Zeiss 510 confocal scanning microscope (37°C, 488 nm laser). Cells were washed twice in Ca²⁺-free PBS (Invitrogen, Carlsbad, CA, USA), and 20 mm caffeine was applied. Images were taken every 2 min for 20 min. Four membrane and three intracellular (nucleus excluded) regions of interest were identified. Changes in signal intensity were assessed using a Zeiss LSM 5 Image Browser and mean values calculated for each time point. The experiment was repeated three times.

Antibody production and purification

The 44 amino acid epitope of human mK44 cDNA was obtained by polymerase chain reaction using the original clone as a template (Korovkina et al. 2001). The PCR product was subcloned into the pGEX2T vector (GE Healthcare, Waukesha, WI, USA) and transformed into E. coli strain UT481 for the production of glutathione S-transferase (GST) mK44 fusion proteins. The resulting fusion proteins were injected into sheep (Elmira Biologicals, Iowa City, IA, USA). The ability of anti-mK44 antisera to specifically detect mK44 was confirmed by Western blot analyses of HEK cells transfected with mK44 versus the canonical maxi-K channel lacking the insert.

GST mK44 fusion proteins were purified from E. coli. The antibodies were affinity purified using a commercially available kit (AminoLink Plus Immobilization Kit, Pierce Rockford, IL, USA) and concentrated using commercially available centrifuge concentrator columns (Centriprep-30 Concentrators, Millipore, Billerica, MA, USA). The final concentrates were dialysed against PBS–0.02% NaN₃, and the purity of the antiserum was confirmed by Western blot analysis of transfected cells expressing full-length mK44.

Immunocytochemistry

The sheep polyclonal anti-mK44 antibody raised against the 44 amino acid epitope specific for mK44 was used as the primary antibody. Biotin and fluorophore conjugates of the secondary antibodies and the streptavidin-fluorophore conjugates were purchased from Jackson ImmunoResearch (West Grove, PA, USA). Detection of mK44 was performed as previously described (Korovkina et al. 2001). Briefly, hMSMCs were fixed in 2% paraformaldehyde. The Na⁺–K⁺-ATPase was detected in non-permeabilized hMSMCs with a mouse monoclonal anti-Na⁺–K⁺-ATPase antibody (1 : 50; The University of Iowa Hybridoma Development Bank) at 37°C for 30 min, donkey anti-mouse biotin-conjugated IgG (1 : 1000) at 37°C for 15 min and streptavidin–Cy2 fluorophore (1 μg ml⁻¹) at 37°C for 15 min. Alternatively, the myc epitope was detected in non-permeabilized hMSMCs expressing myc/mK44 using a mouse anti-c-myc antibody (1 : 500; Invitrogen) at 37°C for 30 min and donkey anti-mouse–Cy3 conjugate (1 : 1000) at 37°C for 15 min. Cells were subsequently permeabilized, and mK44 was visualized with a sheep anti-mK44 antibody (1 : 100), donkey anti-mouse–biotin conjugate (1 : 1000) at 37°C for 15 min and streptavidin–Cy5 (1 μg ml⁻¹) at room temperature (RT) for 15 min.

To detect mK69-1157 co-expressed with myc/mK1-62, a rabbit anti-myc antibody (1 : 200; Chemicon, Temecula, CA, USA) and donkey anti-rabbit–Cy3 were used to visualize myc/mK1-62. Cells were permeabilized, and a mouse anti-V5 (1 : 500; Invitrogen) antibody and donkey anti-mouse–Cy5 (1 : 1000) were used for mK69-1157/V5 detection. Endoplasmic reticulum chaperone protein BiP (GRP78) was detected using a goat polyclonal anti-BiP antibody (1 : 500; Santa Cruz Biotechnoligy Inc., Santa Cruz, CA, USA) and donkey anti-goat–Cy5 (1 : 1000).

Immunohistochemistry

Frozen myometrial tissues were sectioned in 4 μm sections. Sections were incubated in 1% PFA and 0.1% Triton X-100 for 5 min at RT and then in 100 mm glycine buffer for 5 min at RT. All aforementioned buffers were prepared in PBS. Detection of mK44, BiP and Na⁺/K⁺-ATPase were performed as described above.

Patch-clamp recording of mK44 currents

HEK293F cells expressing mK69-1157/eGFP alone or coexpressed with myc/mK1-62 were trypsinized and spun at 300 g for 5 min at RT. Cells were resuspended in bath solution (listed below). All patch-clamping experiments were done at room temperature (∼22°C). One drop of cell suspension was placed in a perfusion chamber containing a 7.4 pH bath solution containing (mm): 135 NaCl, 4.7 KCl, 1 MgCl₂, 10 glucose, 2 CaCl₂ and 5 Hepes. Heat-polished borosilicate glass pipettes (2–5 MΩ) were filled with a pH 7.2 solution containing (mm): 140 KCl, 0.5 MgCl₂, 1 EGTA, 5 ATP, and 5 Hepes. High resistance patch seals (3–30 GΩ) were achieved for whole-cell measurements. Membrane potentials from −80–120 mV were applied to membrane patches up to 5 min using an Axopatch 200B voltage–current amplifier. The elicited currents were recorded using pCLAMP 6.0 or pCLAMP 9.0 software (Molecular Devices, Sunnyvale, CA, USA). The current levels at the given voltages were measured using the Clampfit program. Experiments were repeated 5 times.

Western immunoblot

Whole cell lysates of hMSMCs were prepared by incubating cell pellets in lysis buffer (mm: 250 glucose, 50 Mops, 2 EDTA, 2 EGTA, 1% Triton X-100 and protease inhibitors (Roche, Bazel, Switzerland), pH 7.4) at 4°C for 30 min. The insoluble fraction was separated by centrifugation 14 000 g at 4°C for 10 min. Supernatants were resuspended in Tricine sample buffer (Bio-Rad, Hercules, CA, USA) and resolved on 10–20% Tris-Tricine gradient gels (Bio-Rad) Proteins were transferred onto PVDF membranes in Dunn buffer (mm: 10 NaHCO₃, 3 Na₂CO₃, 20% methanol, pH 9.9). Non-specific binding was blocked by 3% skimmed milk in TBS buffer (mm: 20 Tris, 138 NaCl, 4 HCl, pH 7.4), and subsequently membranes were probed with the anti-c-myc rabbit polyclonal antibody (1 : 100; Santa Cruz Biotechnoligy Inc.). Incubations were performed either at RT for 2 h or at 4°C overnight. Signal was detected by incubations in HRP-conjugated goat anti-rabbit antibody (1 : 5000; Jackson ImmunoResearch, West Grove, PA, USA) at RT for 1 h and exposure to Enhanced Chemiluminescence kit (Amersham, Uppsala Sweden) according to the manufacturer's instructions. Experiments were repeated 3 times.

Statistical analysis

Results of live cell imaging and membrane potential experiments were subjected to a repeated measures ANOVA. Scheffé's post hoc test was used for individual time point comparisons. The level of significance was set at P < 0.05. Electrophysiological experiments were analysed by Student's t test for unpaired observations (Sigma Plot, 2001). Differences were considered significant at P < 0.05.

Results

Endogenous mK44 channels are localized in the endoplasmic reticulum and translocate to the cell membrane in response to caffeine

In previous studies, our laboratory demonstrated that mK44 forms a functional channel on the cell membrane when expressed heterologously in mouse fibroblasts (Korovkina et al. 2001). We have also shown high expression levels of the mK44 transcript in human myometrium suggesting a functional significance of mK44 in myometrial excitability (Korovkina et al. 2001). Others demonstrated an increase in mK44 transcript in human myometrium at the onset of normal labour suggesting a contribution during labour contractions (Curley et al. 2004). Immunocytochemical analysis using an isoform-specific antibody determined that in human pregnant myometrium, endogenous mK44 channel proteins (red) were intracellular (Fig. 1A). A line scan through the tissue (Fig. 1A, arrow) demonstrated that the channel co-localized with the endoplasmic reticulum (ER) chaperone protein BiP (blue: Volpe et al. 1992) (Fig. 1A and B). The channel did not localize with the membrane-localized Na⁺–K⁺-ATPase (Xie & Cai, 2003) (data not shown). Human MSMCs freshly isolated from pregnant myometrium and propagated for 7 days also demonstrated intracellular localization of mK44 (Fig. 1C and D). This finding indicates that mK44 is intracellular and does not contribute to K⁺ current in quiescent myometrium.

Figure 1. Intracellular localization of mK44 in human myometrium.

A, endogenous mK44 channels (red) in human myometrium localize in the ER with the ER chaperone protein BiP (blue). The white arrow indicates the direction of the line scan analysis of the tissue (see B). B, a line scan of the tissue in A shows that signals generated by BiP and mK44 come from the same intracellular compartment suggesting ER localization of mK44 Scale bar, 10 μm; n = 3. C, endogenous mK44 channels (red) also localize in the ER with the ER chaperone protein BiP (blue) in primary cultured hMSMCs. D, a line scan of the cell in C shows that signals generated by BiP and mK44 come from the same intracellular compartment. The purple colour is due to signal interference from the closely localized red and blue epitopes. Scale bar, 10 μm; n = 5.

Maxi-K channels in smooth muscle are activated by release of Ca²⁺ from sarcoplasmic reticulum (SR) stores (Jaggar et al. 2000). Since mK44 current is sensitive to intracellular Ca²⁺ (Korovkina et al. 2001), we hypothesized that in human myometrium, membrane expression of mK44 could be induced by Ca²⁺ released from caffeine-sensitive stores. To ensure that SR was the only source of Ca²⁺ all the experiments were conducted in Ca²⁺-free buffers. After incubation in 20 mm caffeine for 20 min, mK44 channels in myometrium (red) co-localized with Na⁺–K⁺-ATPase (green) indicating a switch from intracellular to cell surface expression (Fig. 2A and B). Similarly, in response to caffeine mK44 translocated to the cell membrane in adherent hMSMCs (Fig. 2C and D). Control experiments in Ca²⁺-free buffer lacking caffeine showed intracellular localization of mK44 (data not shown). Since the mK44 response to caffeine was similar in myometrial tissue and hMSMCs, hMSMCs were further used for transfection and patch-clamp experiments.

Figure 2. MK44 translocates to the cell membrane in response to a release of intracellular Ca²⁺.

A, endogenous mK44 channels (red) in human myometrium localize on the cell membrane with the Na⁺–K⁺-ATPase (green) following 20 min incubation with 20 mm caffeine in Ca²⁺-free buffer. A white arrow indicates the direction of the line scan analysis of the tissue (see B). B, a line scan of the tissue in A shows that signals generated by Na⁺–K⁺-ATPase and mK44 come from the same compartment suggesting membrane localization of mK44. Scale bar, 10 μm; n = 3. C, in hMSMCs mK44 translocates to the cell membrane in response to caffeine. MK44 (red) co-localizes with membrane-associated Na⁺–K⁺-ATPase (green) after 20 min incubation with 20 mm caffeine. The yellow colour is due to signal interference from the closely localized red and green epitopes. The white arrow indicates the direction of the line scan analysis of the cell (see D). D, a line scan of the cell in C shows co-localization of mK44 with Na⁺–K⁺-ATPase of the cell membrane. Scale bar, 10 μm; n = 4.

Trafficking of mK44 to the plasma membrane induces hMSMC repolarization

Trafficking of mK44 to the cell membrane in live hMSMCs was visualized using an mK44/eGFP fusion protein (Table 1; Fig. 3A). In the presence of caffeine, membrane-associated regions of the cells showed an increase in eGFP-generated signal intensity, which was significantly different from the intracellular region from 6 min, and plateaued in 10 min (Fig. 3A, membrane). GFP signal intensity decreased in intracellular and perinuclear regions reflecting a movement of mK44/eGFP from the ER to the cell membrane (Fig. 3A, intracellular). These observations suggest that trafficking of mK44 to the membrane may be a mechanism for activating and inducing repolarization of hMSMCs.

Figure 3. Repolarization of hMSMC membranes occurs concomitantly with accumulation of mK44 on the cell membrane.

A, adherent hMSMCs heterologously expressing mK44/eGFP were incubated in 20 mm caffeine in Ca²⁺-free buffer for 20 min. Live-cell images were captured every 2 min during the incubation. The changes in signal intensity in four membrane-associated (membrane) and three intracellular (intracellular) regions of interest were analysed by repeated measures ANOVA and Scheffé's post hoc test. # Statistically significant differences (P < 0.05) between membrane and intracellular regions. Membrane-associated signal intensity plateaued at 10 min (n = 4). B, hMSMCs pre-loaded with DiBAC₄(3) dye were incubated in Ca²⁺-free buffer supplemented with 20 mm caffeine (caffeine) for 20 min. In separate experiments, hMSMCs were incubated with 50 nm IbTX for 40 min to block maxi-K channels followed by incubations with IbTX and caffeine (caffeine + IbTX). Incubations with 50 nm IbTX (IbTX) were used as controls. The changes in signal intensity were assessed every 2 min. Signals taken from 5 regions of interest were analysed by repeated measures ANOVA and Scheffé's post hoc test. *Statistically significant differences (P < 0.05) among (caffeine) (caffeine + IbTX) and (IbTX) experiments. An increase in signal intensity indicated depolarization, a decrease indicated repolarization (n = 3).

Live hMSMCs were pre-loaded with a membrane potential-sensitive dye DiBAC₄(3) and incubated for 20 min in buffer B (see Methods) or buffer B supplemented with 20 mm caffeine. Cells were observed every 2 min during the incubation period. Changes in signal intensity reflect changes in membrane potential with an increase indicating depolarization and a decrease indicating repolarization. In the presence of caffeine, hMSMCs depolarized for 6 min. Depolarization plateaued from 6 min to 8 min, followed by membrane repolarization (Fig. 3B, caffeine). After 18 min, membrane potential was significantly lower than peak depolarization, although it did not return to baseline levels (Fig. 3B, caffeine). The repolarization phase depended on the activation of maxi-K channels and was abolished by IbTX (Fig. 3B, caffeine + IbTX). Similarly, depolarization was sustained for 20 min when hMSMCs were incubated in Ca²⁺-free buffer (data not shown). IbTX in the absence of caffeine did not induce comparable membrane depolarization reflecting the fact that in the absence of external Ca²⁺ maxi-K channels do not contribute significantly to the resting membrane potential in hMSMCs (Fig. 3B, IbTX). Interestingly, the onset of repolarization occurred at 6 min, concurrent with the beginning of accumulation of mK44 on the cell membrane (Fig. 3A). This suggests that movement of mK44 channels to the cell membrane results in functional current on hMSMCs and may be one mechanism for inducing repolarization.

Heterologous myc/mK44 is expressed as two separate components

To further investigate molecular mechanisms regulating the trafficking of mK44, a c-myc epitope was fused to the N-terminus of mK44 (myc/mK44). Myc/mK44 was heterologously expressed in hMSMCs. Immunocytochemical analysis, using myc and mK44 antibodies, revealed disparate localization of the N- and C-termini. While the C-terminal portion of myc/mK44 accumulated in the ER as was observed for endogenous mK44 (Fig. 1A and B), the N-terminus localized to the cell membrane (Fig. 4A and B). Membrane expression of the N-termini was additionally confirmed by co-localization with Na⁺–K⁺-ATPase and detection of the extracellular c-myc epitope in live non-permeabilized cells (data not shown). In the presence of 20 mm caffeine, the C-termini co-localized with their N-termini on the cell membrane (Fig. 4C and D) suggesting that mK44 may undergo endoproteolytic cleavage in hMSMCs.

Figure 4. The N- and C-termini of myc/mK44 are differentially expressed in hMSMCs but co-localize on the cell membrane in response to caffeine.

A, myc/mK44 were heterologously expressed in hMSMCs and their cellular localization determined by immunocytochemistry. The N-terminal myc epitope (red) localized to the periphery of the cell in hMSMCs, whereas the mK44 epitope (green) localized intracellularly. A white arrow indicates the direction of the line scan analysis of the cell (see B). B, line scan of a cell in A demonstrates intracellular localization of the mK44 epitope and peripheral localization of N-terminal myc epitope. Scale bar, 10 μm; n = 6. C, cultured hMSMCs expressing myc/mK44 were incubated in 20 mm caffeine in Ca²⁺-free buffer for 20 min. The N-terminal myc epitope (red), and the mK44 epitope (green) co-localized on the cell membrane. The yellow colour is due to signal interference from the closely localized red and green epitopes. A white arrow indicates the direction of the line scan analysis of the cell (see D). D, line scan of the cell in C demonstrating that both mK44 and myc epitopes localize to the periphery of the cell. Scale bar, 5 μm; n = 5.

We hypothesized that endoproteolytic cleavage is a post-translational modification specific for mK44 but not other known isoforms of maxi-K channels. Analysis of an mK44-specific amino acid sequence in the first intracellular loop (Fig. 5A, dashed line) using Web-based proteomics databases (http://elm.eu.org/; http://au.expasy.org/tools/peptidecutter/; http://au.expasy.org/tools/scanprosite/; http://au.expasy.org/tools/myristoylator/) revealed several putative post-translational modification motifs for endoproteolytic cleavage and one putative site for N-myristoylation (Fig. 5A, underline). Western immunoblot of HEK293F cells and hMSMCs infected with a construct of myc/mK44 was performed to detect the presence of the myc-tagged proteins. A fragment of ∼10 kDa, predictive of the N-terminal peptide of myc/mK44, was detected in lysates from hMSMCs (Fig. 5B). No fragments were detected in sham-infected HEK293 cells. In contrast, the full-length myc/mK44 (∼140 kDa) was identified in HEK293 cells. This indicates that endoproteolytic digest may be specific for hMSMCs since HEK293 cells (see Supplemental material, Fig. 1) and mouse fibroblasts (Korovkina et al. 2001) express mK44 on the cell membrane exclusively.

Figure 5. The mK44 undergoes endoproteolytic cleavage in the first intracellular loop.

A, schematic representation of the first intracellular loop of mK44 shows the putative proteolytic digest sites indicated by the dashed line. The consensus site for N-myristoylation is underlined. MK44 contains S0–S10 hydrophobic transmembrane domains (accession no. U11058 and AF349445). B, whole-cell lysates from hMSMCs and HEK293 cells heterologously expressing myc/mK44 were analysed by Western immunoblot for the presence of the N-terminal myc epitope. The myc-tagged peptide was detected in lysates from hMSMCs (∼10 kDa). Full-length myc/mK44 was present in lysates from HEK293 cells (∼150 kDa). Sham-transfected HEK293 cells were used as controls (n = 3).

The hypothetical products of endoproteolytic digest of mK44 generate K+ current

The caffeine-induced repolarization of hMSMCs (see Fig. 3B, caffeine) suggests that endoproteolytic fragments of mK44 reconstitute on the cell membrane to form a functional channel. To test this hypothesis, we generated constructs of the hypothesized products of endoproteolytic digest of mK44: (1) the N-terminal peptide myc/mK1-62, which would generate a peptide of ∼10 kDa, and (2) the pore-forming C-terminal mK69-1157/eGFP, which contains the N-terminal myristoylation motif. A methionine was engineered into the first position in this construct (Fig. 6A). MK69-1157/eGFP did not form a functional channel in HEK293F cells despite membrane localization and the presence of an intact pore region (Fig. 6B, left panel; and Fig. 6C, mK69-1157/eGFP). However, upon coexpression of N-terminal myc/mK1-62 with mK69-1157, a K⁺ current was measured (Fig. 6B, middle panel; mK69-1157/eGFP + myc/mK1-62). This current had a phenotype characteristic of maxi-K channels and was sensitive to IbTX (Fig. 6B, right panel; myc/mK1-62 + mK44/eGFP + IbTX). As expected, myc/mK1-62 did not generate K⁺ current probably due to the absence of a pore-forming domain (data not shown). These data indicate that endoproteolytic fragments of mK44 reconstitute on the cell membrane to generate a repolarizing current. This also suggests that mK69-1157/eGFP targets to the plasmalemmal membrane in the correct orientation to form functional channels in the presence of the N-termini.

Figure 6. The hypothetical products of endoproteolytic cleavage of mK44 reconstitute to produce functional maxi-K channel current.

A, the hypothetical products of endoproteolytic digest of mK44 and constructs used in this study: a wild-type maxi-K channel containing the 44 amino acid sequence specific for this isoform (dashed, accession no. AF349445); mK1-62, the N-terminal peptide of mK44 with a predicted molecular weight of ∼10 kDa (see A), and mK69-1157 featuring a putative N-myristoylation site (dashed). ‘M’ represents the N-terminal methionine. B, potassium currents generated by 20 mV depolarizing steps −80 to +120 mV in HEK293 cells expressing mK69-1157 alone (mK69-1157/eGFP), or co-expressed with the N-terminal peptide myc/mK1-62 (mK69-1157/eGFP + myc/mK1-62). K⁺ current generated by mK69-1157/eGFP + myc/mK1-62 is inhibited by IbTX (mK69-1157/eGFP + myc/mK1-62 + IbTX (n = 5). C, mean current levels at a given voltage are plotted for five cells. *Statistically significant differences (P < 0.05) between currents generated in the absence (mK69-1157/eGFP) or presence of myc/mK1-62 (mK69-1157/eGFP + myc/mK1-62) (n = 5).

N-Myristoylation retains the C-terminal end of mK44 in the endoplasmic reticulum

To determine whether N-myristoylation regulates trafficking of the C-terminal end of mK44 to the plasma membrane, we generated the mK69-1157/V5 construct. This construct was truncated at the putative N-myristoylation site. A methionine was engineered into the first position of mK69-1157/V5 (see Fig. 6A) thus masking the N-terminal glycine necessary for N-myristoylation. This construct was cotransfected with myc/mK1-62 into hMSMCs and localization assessed by detection of the V5 and myc tags. MK69-1157/V5 was not retained in the ER and co-localized with its N-termini on the plasmalemmal membrane (Fig. 7A and B). A wild-type myc/mK44 translocated to the cell membrane in hMSMCs in the presence of 100 nm 2-hydroxymyristic acid, to inhibit NMT (Fig. 7C and D). These data suggest that N-myristoylation is one mechanism for retaining the C-termini of mK44 in ER.

Figure 7. N-terminal myristoylation is a mechanism to retain mK44 in the ER in hMSMCs.

A, myc/mK1-62 (red) and mK69-1157/V5 (green) heterologously expressed in hMSMCs co-localize on the cell surface. B, line scan of the cells in A demonstrating localization of both epitopes to the same region on the cell membrane. Scale bar, 10 μm; n = 4. C, the N- (red) and C-termini (green) of heterologous myc/mK44 co-localize on the cell membrane in hMSMCs in response to inhibition of N-myristoyl transferase by 100 nm 2-hydroxymyristic acid for 48 h. A white arrow indicates the direction of the line scan analysis of the cell (see D). D, line scan of the cell in C demonstrating both mK44 and myc epitopes expressed on a periphery of the cell. The yellow colour in panels A and C is due to signal interference from the closely localized red and green epitopes. Scale bar, 10 μm, n = 7.

Discussion

We previously identified a novel maxi-K channel isoform that is prevalent in human myometrial smooth muscle. Our studies in heterologous expression systems demonstrated that mK44 channels are located on the cell surface and have an attenuated sensitivity to Ca²⁺ (Korovkina et al. 2001). However, in the present experiments using hMSMCs, we observed that mK44 undergoes endoproteolytic digest resulting in disparate localization of the N- and C-termini. Several protease/peptidase families were suggested by Web-based proteomics databases as candidates to digest the mK44 isoform, including arginine-C proteinase (tissue kallikrein), proline endopeptidase, subtilisin/kexin family serine proteases and nardilysin convertase. Some of these endoproteases are known to act in concert with the aminopeptidases (Cheong et al. 2002; Fontes et al. 2005). A complex of proteases and peptidases may release the N-terminal peptide and unmask the N-myristoylation motif in mK44 resulting in intracellular retention of the pore-forming fragment of mK44.

Interestingly, reconstitution of functional current by the pore-forming fragment of mK44 (mK69-1157) required the presence of the N-terminal peptide, but physical continuity of the channel protein was not necessary. The hypothetical products of endoproteolytic digest, mK1-62 and mK69-1157, lack six amino acids linking these two fragments in wild-type channels. Nevertheless they were able to generate a repolarizing K⁺ current. HEK293F cells do not express endogenous maxi-K channels thus the current phenotype and sensitivity to IbTX indicate that this current was due to heterologous maxi-K channels on the cell membrane. In these experiments, mK69-1157 demonstrated membrane expression, yet was not functional despite having an intact pore and tetramerization domain (Quirk & Reinhart, 2001). These data indicate that in addition to targeting a particular membrane microdomain to express mK44, the N-terminal peptide is able to elicit tetramerization of this isoform. Similar observations have been made for Gef1p chloride channels in S. cerevisiae (Wachter & Schwappach, 2005). These authors concluded that endoproteolytic digest of Gef1p was an evolutionary errant since the non-cleavable mutant was also functional. Due to the multiplicity of putative cleavage sites, it is difficult to assess the exact location of cleavage within the mK44 protein. However, the observation that endoproteolytic cleavage occurred in hMSMCs and not heterologous systems suggests the importance of this post-translational modification for smooth muscle function.

In the present studies, our data indicate that caffeine induces translocation of the C-terminal end of mK44 to the cell membrane. Caffeine activates membrane-localized maxi-K channels in smooth muscle by releasing sarcoplasmic Ca²⁺ through caffeine-sensitive Ca²⁺ channels and generation of Ca²⁺ sparks that are in spatial proximity to maxi-K channels (Jaggar et al. 2000). Other studies have shown that in smooth muscle, Ca²⁺ sparks are generated within milliseconds (Pabelick et al. 1999). In our experiments, membrane translocation of mK44 peaked at 10 min. Thus, mK44 trafficking could not occur in response to a Ca²⁺ rise in subplasmalemmal space. However, Young and Zhang demonstrated a deep cytosolic Ca²⁺ pool in hMSMCs, and that maxi-K channel activity is regulated by Ca²⁺ derived from this pool (Young & Zhang, 2004). It is possible that ER-localized mK44 channels are sensitive to rises in cytosolic Ca²⁺ and respond by the trafficking and reconstitution of functional channels. Both the low Ca²⁺ sensitivity of mK44 (Korovkina et al. 2001) and the relatively low levels of free Ca²⁺ in cytosol may contribute to the slow trafficking response of mK44 to caffeine. The low sensitivity to Ca²⁺ and evidence of caffeine-induced trafficking indicate that mK44 responds to increases in intracellular Ca²⁺ at supraphysiological concentrations, at levels which may otherwise induce sustained uterine contraction (Kristian & Siesjo, 1998).

Addition of a myristic acid residue regulates membrane localization of proteins (Vaandrager et al. 1996; Raju et al. 1997). N-Myristoylation may explain ER localization of mK44 since inhibition of NMT induced its translocation to the cell membrane. In addition, mK69-1157, containing the N-terminal methionine which prevents myristoylation of mK44, was also membrane-localized. The putative GGVIGC sequence may have a role in trafficking of mK44 in response to caffeine by hiding or unmasking the myristic acid residue in response to a variety of stimuli (switch mechanisms) (Resh, 1999; Senin et al. 2002; Hantschel et al. 2003). A myristoyl–Ca²⁺ switch could be one explanation for mK44 function. Another possible mechanism for regulating trafficking of mK44 to the cell membrane is the reversible (de)myristoylation, shown previously to determine the cellular localization of myristoylated alanine-rich C-kinase substrate (MARCKS) (Manenti et al. 1994).

Caffeine induces hMSMC contraction as the average cell length after caffeine incubation was approximately 40–50 μm versus 120–130 μm in cultures without caffeine. Caffeine also induces relaxation of both human and rat myometrium pre-contracted with oxytocin or K⁺ (Apavdin et al. 1998; Fu et al. 1998); however, the mechanism by which this occurs is unknown. We present a novel mechanism for regulating maxi-K channels that is myometrial smooth muscle- and isoform-specific and can induce alterations in myometrial smooth muscle excitability. In hMSMCs, de novo synthesized mK44 is not a functional channel, but rather is retained in the ER in a quiescent state. A maxi-K channel isoform identified in rat, which is generated by insertion of 33 amino acids into a similar site in the first intracellular loop, also exhibits ER localization (Zarei et al. 2001). One possibility is that the relaxant effects of caffeine on myometrium are due to unique post-translational modifications and membrane translocation of isoforms with similar splicing locations within the maxi-K channel.

Accumulation of mK44 on the cell membrane occurred simultaneously with hMSMC repolarization. The repolarization was not due to the depletion of SR Ca²⁺ stores since in the presence of IbTX, a sustained depolarization was observed for 20 min. Similarly, a sustained depolarization was measured in Ca²⁺-free buffer. Studies have shown that IbTX increases intracellular Ca²⁺ in hMSMCs (Anwer et al. 1993). Thus, the combined action of IbTX and caffeine to mobilize Ca²⁺ may be one mechanism to sustain depolarization in hMSMCs. IbTX alone did not induce depolarization comparable to that observed in the presence of caffeine. One possible explanation is that maxi-K channels do not contribute significantly to the resting membrane potential in hMSMCs in the absence of external Ca²⁺, but are activated by the intracellular Ca²⁺ release by caffeine. There is controversy about a role for maxi-K channels in maintaining the resting membrane potential in smooth muscle. Studies in urinary smooth muscle have demonstrated that maxi-K channels regulate membrane potential at rest (Heppner et al. 1997). Opposite results were obtained in some vascular smooth muscle studies (Prior et al. 1998).

Thus, we have shown a novel mechanism for regulating hMSMC membrane potential. The pattern of mK44 channel expression was induced by endoproteolytic digest within the first intracellular loop. The N-terminal peptide produced by the digest localized to the cell membrane. Reconstitution of the N- and C-termini of mK44 formed a functional maxi-K channel on the cell membrane in response to caffeine. Membrane expression of the N-termini is necessary to reconstitute a functional mK44 channel. We conclude that endoproteolytic digest and translocation of mK44 to the cell membrane in hMSMCs is one mechanism for rapidly increasing repolarizing K⁺ current to buffer smooth muscle cell excitability.

Supplemental material

The online version of this paper can be accessed at:

DOI: 10.1113/jphysiol.2006.106922

//jp.physoc.org/cgi/content/full/jphysiol.2006.106922/DC1 and contains supplemental material consisting of a figure entitled

Supplemental Data, Figure 1

‘Myc/mK44 expressed on the cell membrane in HEK293F cells.’

This material can also be found as part of the full-text HTML version available from http://www.blackwell-synergy.com