in human myometrial smooth muscle cells (hMSMCs), a splice variant of the large-conductance Ca²⁺- and voltage-activated K⁺ channel from the gene KCNMA (maxi-K channel) containing a 44-amino acid insertion (mK44) employs a unique mechanism to contribute to the regulation of cell membrane potential. In quiescent cells, it is not functional because of retention of the COOH-terminal pore-forming fragment in the endoplasmic reticulum (ER), whereas the NH₂-terminal peptide, which consists of the extracellular NH₂ terminus, the S0 transmembrane domain, and the first 18 amino acids of the first intracellular loop, localizes to the plasma membrane (12). Upon an increase in intracellular Ca²⁺, the COOH terminus translocates to the cell membrane and colocalizes with the NH₂ terminus to reconstitute a functional channel to generate a repolarizing current. Notably, the intracellular linker region between the S0 and S1 transmembrane helices contains an amino acid domain that distinguishes mK44 from other maxi-K channel isoforms (13). Analysis of the primary structure of this domain revealed that the dibasic -RK- motif may render mK44 a substrate for posttranslational modifications by the prohormone convertase family of serine/threonine proteases or insulinase family of metalloproteases, namely, nardilysin convertase isoform 1 [N-arginine dibasic convertase (hNRDc1), EC 3.4.24.61; http://elm.eu.org/].

Endoproteolytic processing is one posttranslational modification that is frequently used to regulate ion channel expression and function. For example, activation/inhibition of Na⁺ or K⁺ channels by serine/threonine or acid proteases may contribute to the pathophysiological mechanisms that underlie hearing loss and Alzheimer's disease, as well as mechanisms that regulate electrolyte metabolism by the kidney (7, 8, 10). Also, extracellular metalloprotease-mediated K⁺ channel modification in blood vessels has been shown to contribute to the regulation of vascular tone and blood flow (11, 17). Recently, it was also suggested that an unidentified protease may modulate K⁺ channels in myometrial smooth muscle (12). There is no evidence that insulinases, in particular, NRDc1, regulate ion channels (5, 8, 16).

Several functions of hNRDc1 in addition to proteolytic activity have been reported. NRDc acts as a receptor for heparin-binding epidermal growth factor-like growth factor, a function that may thereby regulate cell migration and proliferation (9). It also enhances the proteolytic activity of α-secretases in the brain, thereby preventing the accumulation of β-amyloid, a peptide that has been implicated in the pathophysiology of Alzheimer's disease (8). Finally, a stretch of acidic amino acids in the hNRDc1 protein may act as an independent domain to facilitate the formation of complexes with other proteins (14).

In this study, we investigated whether hNRDc1 complexes with mK44 to regulate membrane potential in hMSMCs. We conclude that, in hMSMCs, hNRDc1 contributes to the regulation of cell membrane potential via mK44. In quiescent hMSMCs, hNRDc1 induces retention of the mK44 pore-forming COOH terminus in the ER and attenuates mK44 current. Pharmacological and genetic inhibition of hNRDc1 results in predominantly membrane expression of mK44. Our results suggest that hNRDc1 may also modulate reconstitution of functional mK44 channels in response to an increase in intracellular Ca²⁺. From our data, hNRDc1 is not the protease responsible for the posttranslational endoproteolysis of mK44 (12) but, rather, represents a novel class of proteins that regulate the excitability of hMSMCs via ion channel modulation.

MATERIALS AND METHODS

Tissue collection and cell culture.

Human myometrial tissues from the lower segment of the uterus were collected from women undergoing elective cesarean section under spinal anesthesia during late pregnancy (38–40 wk gestation) in the absence of spontaneous or induced labor contractions. All patients signed written consent forms approved by the University of Iowa Internal Review Board (no. 199809066). Nonpregnant human myometrial tissue was obtained through collaboration with the Cooperative Human Tissue Network (CHTN) of the National Cancer Institute ( http://chtn.nci.nih.gov). hMSMCs were dispersed by incubation with collagenase (Worthington Biochemicals, Lakewood, NJ) and allowed to adhere to tissue culture plates for 24 h. A smooth muscle cell phenotype was subsequently induced by incubation in DMEM-Ham's F-12 (DMEM-F12) medium supplemented with 0.5% FBS. Ltk⁻ mouse fibroblasts were propagated in DMEM-F12 medium supplemented with 10% FBS until they reached 80–90% confluency.

Isolation and cloning of hNRDc1.

Total RNA from human myometrium was reverse transcribed using the First-Strand RT-PCR kit (Stratagene, La Jolla, CA). hNRDc1 (accession no. NM_002525.1) was isolated using the sense primer 5′-ATGCTGAGGAGAGTCACTGTTGCT-3′ and the antisense primer 5′-TTTATTTGACTATTTTATGGTAGGGGT-3′ and subcloned into the pEGFP-N1 expression vector (Clontech). A COOH-terminal fusion construct of hNRDc1 and the red fluorescent protein mCherry was generated by replacement of the green fluorescent protein (GFP)-encoding gene in hNRDc1/pEGFP-N1 with a gene encoding mCherry (18).

Heterologous expression of maxi-K channel and hNRDc1 constructs.

An adenoviral construct of myc/mK44 (accession no. AF_349445.1) was used to transduce hMSMCs at 1 μl/ml of 1–1.2 × 10⁹ plaque-forming units/ml purified construct (12). Cells were harvested 48–72 h after infection. Infection efficiency was monitored using an adenoviral enhanced GFP (eGFP) reporter gene. Constructs in plasmid expression vectors (mK44/eGFP, maxi-K/eGFP, and hNRDc1/mCherry) were transfected into hMSMCs using the GeneJammer transfection reagent (Stratagene) and into Ltk⁻ mouse fibroblasts using Lipofectamine PLUS (Invitrogen, Carlsbad, CA) according to the manufacturers' instructions. Cells transfected in the absence of plasmid DNA were used as negative controls.

Inhibition of hNRDc1 by small interfering RNA.

A small interfering RNA (siRNA) for hNRDc1 and a control (scrambled) siRNA were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). siRNA or scrambled siRNA was transfected into adherent hMSMCs using GeneJammer reagent or into mouse Ltk⁻ fibroblasts using the Lipofectamine PLUS reagent according to the manufacturers' instructions. In overexpression experiments, siRNA was cotransfected with myc/mK44, mK44/eGFP, maxi-K/eGFP, and hNRDc/mCherry. Cells were inspected 48 h after transfection.

Pharmacological inhibition of hNRDc1 activity.

The inhibitor 1,10-phenanthroline (1,10-Phen) was purchased from Sigma-Aldrich. Adherent hMSMCs were incubated with 20 μM 1,10-Phen for 30 min and 24 h, localization of mK44 was examined by immunocytochemistry, and current density was assessed by whole cell patch-clamp analysis.

Electrophysiological experiments.

Mouse fibroblasts were transfected with mK44/eGFP, mK/eGFP, mK44/eGFP + hNRDc/mCherry, or mK/eGFP + hNRDc/mCherry for 24–48 h. Cells were trypsinized and spun at 300 g for 5 min at room temperature (∼22°C) and resuspended in bath solution (see below). All patch-clamping experiments were done at room temperature. One drop of cell suspension was placed in a perfusion chamber filled with a pH 7.4 bath solution containing (in mM) 135 NaCl, 4.7 KCl, 1 MgCl₂, 10 glucose, 2 CaCl₂, and 5 HEPES (for whole cell measurements) or 145 KCl, 2 CaCl₂, 1 MgCl₂, and 5 HEPES (for cell-attached measurements). For whole cell recordings, heat-polished borosilicate glass pipettes (2- to 5-MΩ resistance) were filled with a pH 7.2 solution containing (in mM) 140 KCl, 0.5 MgCl₂, 1 EGTA, 5 ATP, and 5 HEPES. High-resistance (3–30 GΩ) patch seals were achieved for whole cell measurements. Membrane potentials from −80 to +160 mV in 20-mV increments were applied to membrane patches for up to 5 min using an Axopatch 200B voltage-current amplifier, and the elicited currents were recorded using pClamp 9.0 software (Molecular Devices, Sunnyvale, CA). The current levels at the given voltages were measured using the Clampfit program. For cell-attached single-channel measurements, currents were elicited in symmetrical K⁺ at constant holding potentials. Borosilicate glass pipettes (8- to 20-MΩ resistance) were filled with a pH 7.2 solution containing (in mM) 145 KCl, 2 CaCl₂, 1 MgCl₂, and 5 HEPES. Membrane potentials from −60 to +80 mV were applied to membrane patches for up to 2 min. The number of open events and the open-state probability were calculated using the Clampfit program.

Western immunoblotting.

Myometrial tissues were homogenized in lysis buffer [250 mM glucose, 50 mM MOPS, 2 mM EDTA, 2 mM EGTA, 1% Triton X-100, and protease inhibitors (Roche, Basel, Switzerland), pH 7.4] at 4°C. The insoluble fraction was separated by centrifugation (14,000 g at 4°C) for 10 min. Lysates were resuspended in SDS sample loading buffer [40% glycerol, 0.25 M Tris base (pH 7.0), 12% SDS, 20% β-mercaptoethanol, and 0.02% bromphenol blue] and separated on 7.5% acrylamide gels. Proteins were transferred onto nitrocellulose membranes in Towbin buffer (12.5 mM Tris, 96 mM glycine, and 20% methanol, pH 8.5). Nonspecific binding was blocked using 3% skim milk in TBS buffer (in mM: 20 Tris, 138 NaCl, and 4 HCl, pH 7.4), and subsequently membranes were probed with the rabbit polyclonal anti-NRDc antibodies (Santa Cruz Biotechnology) at 1:250 dilution. Incubations were performed at room temperature for 2 h or at 4°C overnight. Signal was detected by incubations in horseradish peroxidase-conjugated antibodies (1:3,000 dilution; Jackson ImmunoResearch, West Grove, PA) at room temperature for 1 h, followed by exposure to enhanced chemiluminescence reagents (Amersham, Uppsala, Sweden) according to the manufacturer's instructions. Experiments were repeated at least three times.

For control of inhibition of protein expression by commercially purchased NRDc siRNA, lysates were prepared from mouse fibroblasts cotransfected with hNRDc1/mCherry + NRDc siRNA or hNRDc1/mCherry + scrambled siRNA. Western immunoblotting was performed as described above. Experiments were repeated four times.

Immunohisto(cyto)chemistry.

Frozen human myometrial tissue in OCT compound (obtained from CHTN, see Tissue collection and cell culture) was cut into 8-μm sections. mK44 was detected as previously described (12). Briefly, mK44 was visualized by consecutive incubations with a sheep anti-mK44 antibody (1:100 dilution) (17) at 37°C for 30 min and a donkey anti-sheep Cy3 antibody (1:1,000 dilution; Jackson ImmunoResearch) at 37°C for 15 min. hNRDc1 was detected by consecutive incubations with goat anti-NRDc (1:50 dilution; Santa Cruz Biotechnology) at 37°C for 30 min and donkey anti-goat-Cy2 (1:1,000 dilution; Jackson ImmunoResearch Laboratories) at 37°C for 15 min. Images were obtained using a confocal laser microscope (model LSM 5, Zeiss, Jena, Germany) and analyzed using LSM 510 image analysis software (Zeiss). Experiments were repeated at least five times.

Adherent nontransfected hMSMCs were fixed in 2% paraformaldehyde at room temperature for 30 min and permeabilized by 0.1% Triton-X 100 at room temperature for 5 min. mK44 and hNRDc1 were detected and visualized as described above (12). Experiments were repeated eight times. Alternatively, hMSMCs were transfected as described above (see Inhibition of hNRDc1 by siRNA and Pharmacological inhibition of hNRDc1 activity). mK44 was detected using a sheep anti-mK44 antibody (1:100 dilution) (17) at 37°C for 30 min and a donkey anti-sheep Cy3 antibody (1:1,000 dilution; Jackson ImmunoResearch) at 37°C for 15 min. Experiments were repeated three times.

Live cell imaging.

Adherent hMSMCs in glass-bottom tissue culture dishes coated with mouse laminin were cotransfected with mK44/eGFP or mK/eGFP and hNRDc/mCherry. Cells placed on the heating stage of a Zeiss 500 laser confocal microscope 48 h after transfection were washed twice in Ca²⁺-free PBS (Invitrogen), and 20 mM caffeine was applied. Cells were observed for 20 min at 37°C with 488- and 543-nm lasers. Membrane-localized intensity of signals generated by eGFP and mCherry was registered every 2 min and quantitated as previously described (12). Experiments were repeated three times.

Statistical analysis.

Live cell imaging and electrophysiological experiments were analyzed by ANOVA and post hoc Bonferroni's test. Differences were considered significant at P < 0.05.

RESULTS

This study explored the potential role of the insulinase family of endoproteases in regulation of myometrial contractility via the mK44 isoform of maxi-K channels. Initial screening of human myometrium and adherent hMSMCs by RT-PCR failed to detect the expression of prohormone convertase family members [i.e., furin, prohormone convertases (PC1, PC2, PC3, PC4, PC5/6, and PC7), or paired basic amino acid-cleaving enzyme isoform 4]. However, in tissue and cells isolated from human myometrium, NRDc1, a multifunctional metalloendopeptidase of the insulinase family with catalytic and noncatalytic activities (5, 8, 16), was detected (data not shown). Thus we explored the potential role of hNRDc1 in regulating mK44 activity and excitability of uterine smooth muscle.

Immunohistochemical analysis of human myometrium isolated from nonpregnant subjects demonstrated that hNRDc1 colocalizes with mK44 in the intracellular compartment of smooth muscle cells (Fig. 1A, left). Line scan analysis (Fig. 1A, left) made it possible to separate hNRDc1- and mK44-generated signals and further confirmed that the proteins colocalize (Fig. 1A, right). Previously published data indicate that mK44 localizes primarily in the ER, thus suggesting that hNRDc1 coexpresses with mK44 in the ER (12). Taken together, these results lend credence to the hypothesis that mK44 may be regulated by hNRDc1 in human myometrial cells. We further determined whether adherent hMSMCs express hNRDc1 and whether the expression patterns of hNRDc1 and mK44 mimic those in the myometrial tissue. We found that this was indeed the case, with hNRDc1 and mK44 colocalizing in the ER (Fig. 1B, left). Line scan analysis (Fig. 1B, left) suggested that, as in the case of the myometrium, hNRDc1 and mK44 colocalize in the ER.

To determine whether hNRDc1 contributes to the retention of mK44 in the ER in quiescent hMSMCs, we used pharmacological and siRNA approaches. Translocation of mK44 to the cell membrane occurred after inhibition of hNRDc1 by 1,10-Phen (Fig. 2A, arrows), but not in the absence of the inhibitor (Fig. 2B). Similarly, downregulation of hNRDc1 expression by siRNA resulted in partial translocation of mK44 to the plasma membranes (Fig. 2C, arrows), whereas scrambled siRNA did not demonstrate this effect (Fig. 2D). Effectiveness of siRNA to inhibit hNRDc1 protein was tested in mouse fibroblasts overexpressing hNRDc1/mCherry [Fig. 2E, (−)]. Expression of hNRDc1 was inhibited by siRNA, but not by scrambled siRNA (Fig. 2E).

The observation of live myometrial smooth muscle cells overexpressing mK44/eGFP (Fig. 3, A and B) and hNRDc1/mCherry (Fig. 3, C and D) suggested that hNRDc1 and mK44 are cotransported to the cell surface in response to caffeine (12). Indeed, we found that the intensity of signals generated by hNRDc1 and mK44 on the cell surface increased simultaneously over the course of 20 min following the addition of 20 mM caffeine (Fig. 3E). Notably, fractions of hNRDc1 and mK44 pools were retained intracellularly, suggesting that the overexpression of hNRDc1 may have an inhibitory effect on membrane translocation of mK44. Our data also suggest that the hNRDc1-dependent ER retention in hMSMCs is specific to mK44, since maxi-K/eGFP channels remained predominantly localized to the cell surface when expressed with hNRDc1/mCherry (see Supplemental Fig. 1 in the online version of this article).

In mouse fibroblasts, hNRDc1 had disparate effects on mK44- and maxi-K-generated K⁺ currents: mK44 current was significantly reduced in the presence of hNRDc (Fig. 4A, left), whereas maxi-K currents were unaffected (Fig. 4A, right). A summary shows reduction of mK44-generated K⁺ currents in the presence of hNRDc1 (Fig. 4B, left). Human NRDc1 did not affect the sensitivity of mK44 to the selective maxi-K channel inhibitor iberiotoxin (IbTX; Fig. 4B, left). In contrast, maxi-K channels generated comparable currents in the absence and presence of hNRDc1 (Fig. 4B, right) but were significantly inhibited by IbTX (Fig. 4B), with magnitude of reduction similar to that of the mK44-generated current brought about by hNRDc coexpression. The higher current density values in Fig. 4B were not consequences of changes in the single-channel conductance (see Supplemental Fig. 2) or the open-state probability; the open-state probability was decreased in mK44 because of its lower sensitivity to Ca²⁺ and voltage compared with maxi-K as reported previously (13).

A chelator of heavy metals, 1,10-Phen inhibits the activity of hNRDc1 (6). In hMSMCs, IbTX-sensitive K⁺ currents were assessed before (Fig. 5, A and B) and after (Fig. 5, C and D) application of 1,10-Phen. An increase in total K⁺ current density in the presence of 1,10-Phen was due to a contribution of an IbTX-sensitive component (Fig. 5E), inasmuch as IbTX inhibited K⁺ currents in the absence and presence of 1,10-Phen to comparable levels (Fig. 5E). The differences in amplitudes of whole cell traces after IbTX (Fig. 5, B and D) did not affect the current-voltage relationship (Fig. 5E) and represent normal variability within hMSMCs.

The 10-kDa NH₂-terminal peptide (aa 1-61) has been shown to complement the mK44 COOH-terminal fragment in generating transmembrane K⁺ current (12). However, immunoprecipitation experiments to isolate the 10-kDa fragment from mouse Ltk⁻ fibroblasts heterologously expressing mK44 channels and hNRDc1 or from hMSMCs that express these proteins endogenously have been inconclusive. We previously showed that hMSMCs overexpressing myc/mK44 generate the peptide of expected size (12). Thus whether hNRDc1 or other proteases modify mK44 by endoproteolysis remains to be further investigated.

Gestational changes in hNRDc1 expression in human myometrium have been studied by Western immunoblotting. hNRDc1 showed the highest expression in laboring myometrium (Fig. 6, top). Signal intensity was reduced in samples from late-pregnant nonlaboring myometrium compared with laboring and nonpregnant myometrium (Fig. 6, top). Equal loading of proteins was confirmed by the detection of GAPDH (Fig. 6, bottom). A 140-kDa fragment corresponding to the expected size of hNRDc1 was not visualized when antibody was preabsorbed with an antigenic peptide (data not shown).

In conclusion, in myometrial smooth muscle as well as in a heterologous expression system, hNRDc1 modifies the expression of, as well as K⁺ currents generated by, mK44 channels. Our data suggest that, in quiescent smooth muscle, hNRDc1 acts to retain mK44 in the ER and may have a role in Ca²⁺-induced translocation of mK44 to the plasma membrane and repolarization of smooth muscle cells. From our data, hNRDc1 is not the protease responsible for endoproteolysis of mK44 but, rather, plays a role in retaining the channel in the ER and modulates mK44 plasma membrane expression of hMSMCs. The interaction of hNRDc1 and mK44 represents a unique mode of regulation of repolarizing K⁺ current and excitability in human myometrial smooth muscle.

DISCUSSION

In this study, we present evidence in support of the hypothesis that, in the human myometrium, hNRDc1 modifies the cellular localization of the mK44 channel and its response to intracellular Ca²⁺ and, thus, contributes to excitation-contraction coupling. Our data lead us to conclude that the regulatory properties of hNRDc depend primarily on its noncatalytic function and that this entails an interaction between hNRDc1 and mK44 proteins. Whether NRDc1 directly modulates endoproteolytic posttranslational modification of mK44 in smooth muscle is not supported by this study; however, in vitro studies indicate that a number of other ion channels may be regulated by proteases. The consequences of such processing vary according to the protease and may not directly affect the channel protein but, rather, involve changes to the composition of cell membrane microdomains, changes to intracellular transport, or the assembly of channels into multimers (1, 7, 10, 15). Studies show that a deficiency of the serine protease hepsin (TMPRSS1) reduces membrane expression of maxi-K channels in auditory hair cells and that this mechanism may contribute to the profound hearing loss in TMPRSS1^−/− mice (7). Furin and prostasin, both of which are trypsin-like serine proteases, activate amiloride-sensitive epithelial Na⁺ channels (ENaCs) after the cleavage of inhibitory peptides from the extracellular domains of the α- and γ-subunits. However, although these proteolytic fragments of ENaCs are found in mouse renal tissues, the physiological significance of this proteolytic processing, which is known not to alter membrane expression of the channel (4), remains unclear. Another example is the voltage-sensitive Ca²⁺ (Cav2) channel: its generation of current in vitro is enhanced as a result of increased membrane expression of cleaved Cav2 channels (2). However, in this case, the protease involved in cleavage of Cav2 channels is unknown. Identification of these proteases that regulate ion channel activity and expression will elucidate novel mechanisms of modulation of ion channel function.

Our data indicate that, in uterine smooth muscle, the nardilysin convertase hNRDc1 functions in excitation-contraction coupling and that it does so by 1) retaining mK44 in the ER of quiescent cells and 2) modulating the cell membrane expression of mK44 in response to a release of Ca²⁺ from intracellular stores. Our data suggest that, in quiescent hMSMCs, hNRDc1 is necessary for retention of mK44 in the ER (Figs. 2 and 4). Observations described in Fig. 3 may indicate that, in addition to ER retention, hNRDc1 functions to chaperone mK44 to the plasma membrane. The physiological significance and molecular mechanism(s) underlying ER retention and membrane cotransport of hNRDc1 and mK44 in response to an increase in intracellular Ca²⁺ require further investigation. However, the preliminary evidence suggests that this mechanism(s) may be important in regulation of myometrial excitability in gestation. Our data show a decrease in hNRDc1 expression in late gestation and an increase at labor (Fig. 6). An inhibition of hNRDc1 expression promotes assembly of a functional mK44 channel on the cell membrane and, thus, may be beneficial in maintaining quiescence of gravid uterus. In contrast, in labor, high expression of hNRDc1 may lead to the retention of mK44 in the ER and, thus, facilitate uterine contractions.

Both hNRDc1 and mK44 may contribute to Ca²⁺ sensing. A stretch of acidic amino acids exists in hNRDc1 and ion channels that are Ca²⁺ sensitive, the latter of which has been shown to regulate membrane excitability (3). Similar to other maxi-K isoforms, mK44 encodes domains that confer Ca²⁺ sensitivity to the channel (19). However, intracellular localization of mK44 in quiescent smooth muscle suggests that sources of Ca²⁺ other than those released into a subplasmalemmal space induce membrane translocation of mK44 (12, 20). Research from other laboratories indicates that, in uterine smooth muscle, different pools of intracellular Ca²⁺ may make a different contribution to excitation-contraction coupling (20). Thus ER retention and intracellular transport of mK44 mediated by hNRDc1 may contribute to a novel mechanism that regulates contractility specific to myometrial smooth muscle.

Absence of the NH₂-terminal peptide of mK44 in mouse fibroblasts that coexpress the channel and hNRDc1 cells suggests that, in smooth muscle, yet unidentified protein(s), along with hNRDc1, may contribute to retention and/or membrane transport of mK44. The role of N-myristoyl transferase, as suggested by our data (12), and/or a putative role of other accessory proteins, such as 14-3-3, requires further elucidation (21). Additional studies investigating the extent of protease action on maxi-K channel isoforms will continue to reveal novel mechanisms at work in regulation of the activity of this channel. Our data suggest that although hNRDc1 modulates the intracellular transport of the channel, the NH₂-terminal mK44 peptide is necessary for the proper tetramerization of mK44 on the cell membrane and that this function may be independent of hNRDc1 (12). The evidence presented here suggests that, in the myometrium, hNRDc1 contributes to contraction-relaxation by regulating cell membrane expression of a smooth muscle-specific isoform of the maxi-K channel.

GRANTS

This work was supported by National Institute of Child Health and Human Development Grant HD-037831 (to S. K. England) and National Center for Research Resources, Clinical and Translational Science Award M01-RR-00059 (for human tissue attainment).

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.