Estradiol Binding to Maxi-K Channels Induces Their Down-regulation via Proteasomal Degradation*

Estrogens exert their biological action via both genomic and non-genomic mechanisms. Proteins different from classical estradiol receptors are believed to mediate the latter effects. Here we demonstrate that the maxi-K channel functions as an estrogen-binding protein in transfected HEK293 cells. Whole-cell maxi-K channel currents and protein expression were attenuated by exposure to either 17 (cid:1) - or 17 (cid:2) -estradiol. This effect was dose-dependent for 17 (cid:2) -estradiol at concentrations ranging from 10 n M to 1 (cid:3) M , while 17 (cid:1) -estradiol inhibited channel expression only at 1 (cid:3) M . These effects were mediated by direct low affinity binding of estradiol to the maxi-K channel but not to its accessory (cid:2) 1-sub-unit, as revealed by cell membrane estradiol binding assays. However, specific binding of estradiol to the channel was facilitated by the presence of the (cid:2) 1 subunit. Addition of MG-132, a blocker of proteasomal degradation, stabilized channel expression. These data suggest that channel down-regulation is mediated by estrogen-induced proteasomal degradation, similar to the pathway used for estrogen receptor degradation. Membrane expression of endogenous maxi-K channels in cultured vascular smooth muscle cells was also attenuated by prolonged exposure to 17 (cid:1) - and 17 (cid:2) -estradiol. Thus our studies

Estrogens exert their biological action via both genomic and non-genomic mechanisms. Proteins different from classical estradiol receptors are believed to mediate the latter effects. Here we demonstrate that the maxi-K channel functions as an estrogen-binding protein in transfected HEK293 cells. Whole-cell maxi-K channel currents and protein expression were attenuated by exposure to either 17␣-or 17␤-estradiol. This effect was dose-dependent for 17␤-estradiol at concentrations ranging from 10 nM to 1 M, while 17␣-estradiol inhibited channel expression only at 1 M. These effects were mediated by direct low affinity binding of estradiol to the maxi-K channel but not to its accessory ␤1-subunit, as revealed by cell membrane estradiol binding assays. However, specific binding of estradiol to the channel was facilitated by the presence of the ␤1 subunit. Addition of MG-132, a blocker of proteasomal degradation, stabilized channel expression. These data suggest that channel down-regulation is mediated by estrogen-induced proteasomal degradation, similar to the pathway used for estrogen receptor degradation. Membrane expression of endogenous maxi-K channels in cultured vascular smooth muscle cells was also attenuated by prolonged exposure to 17␣-and 17␤-estradiol. Thus our studies demonstrate that estrogen binds to maxi-K channels and may directly regulate channel expression and function. These results will have important implications in understanding estradiol-induced effects in multiple tissues including vascular smooth muscle.
Existing epidemiological data suggest a beneficial role for estrogens in maintaining cardiovascular protection in pre-menopausal women. However, other data show that estrogen-containing preparations can negatively affect the health of premenopausal as well as post-menopausal women (1)(2)(3)(4)(5)(6). These results indicate the existence of complex estradiol-signaling mechanisms, possibly mediated by proteins other than classical intracellular estradiol receptors (7).
One protein that has been hypothesized to be a potential effector of the action of estrogen is the large conductance Ca 2ϩand voltage-activated K ϩ channel (maxi-K channel). This chan-nel provides a buffering repolarizing current in response to various physiological stimuli (8 -11). The mechanism by which estrogen interacts with this channel is unknown. However, both 17␤-estradiol and tamoxifen, a partial agonist of nuclear estradiol receptors, have been shown to activate maxi-K channels in both heterologous expression systems and smooth muscle cells, resulting in greater repolarizing currents (12,13). Because acute incubation with either 17␤-estradiol or membrane-impermeable estrogen derivatives activate maxi-K channels, signal transduction pathways other than those mediated by classical nuclear estradiol receptors ␣ (ER␣) 1 or ␤ (ER␤) are likely to be involved. Valverde et al. (14) showed that activation of maxi-K channels by 17␤-estradiol occurs when the channel is associated with its accessory ␤1-subunit. This effect was not observed in the presence of 17␣-estradiol or when the maxi-K channel was expressed alone. This observation led to the hypothesis that the ␤1-subunit directly binds estradiol and enhances channel activity by an unidentified mechanism (14). However, other studies have demonstrated that ␤1-subunitindependent decreases in maxi-K channel unitary conductance occur in response to (xeno)estrogens, implicating that either subunit can be a target for estrogen (15). Both studies emphasized the cell membrane-limited action of estrogen on the maxi-K channel, thus excluding a role for classical nuclear estradiol receptors. Other data indicate that ER␣ or ER␤ may participate in membrane-associated signaling pathways when they are routed to the cell membrane by differential phosphorylation or binding to caveolins. Novel membrane-associated estradiol receptors are also a possible explanation for short term estrogen action (16 -18).
Short term exposure of maxi-K channels to 17␤-estradiol or to partial agonists of estradiol receptors is thought to activate the repolarizing current. Here, we report that exposure to either 17␣or 17␤-estradiol induces both a decrease in maxi-K channel current and a down-regulation of channel protein expression in heterologous systems. Although 17␣and 17␤-estradiol produce similar results, higher concentrations and longer exposure times are required for 17␣-estradiol to mediate these effects. The presence of the ␤1-subunit did not alter channel expression, and its expression was inhibited by both estradiols. Further investigation into the mechanism underlying this response implicates estradiol-induced proteasomal degradation of the channel, which occurs after direct binding of estrogen to the maxi-K channel. Studies conducted in coronary artery smooth muscle cells suggest that 48 h of exposure to estrogens reduces maxi-K channel protein expression. The estrogen-induced loss of repolarizing current may have poten-FIG. 1. Dose-dependent inhibition of the maxi-K channel current by estrogens. Whole-cell recordings were elicited from Ϫ80 mV holding potential in 20 mV steps from Ϫ100 mV to ϩ160 mV in HEK293 cells stably (n ϭ 14) and transiently (n ϭ 3) expressing maxi-K channels. Whole-cell currents were measured in the absence (A and B, control) and presence of 1 M 17␣-(A, 17␣) and 17␤-estradiol (B, 17␤). Currents were significantly attenuated in the presence of 1 M 17␣-estradiol (C; n ϭ 6) and 1 M 17␤-estradiol (D; n ϭ 8) compared with control recordings. Asterisks indicate significant difference (p Ͻ 0.05) at a given membrane potential. Similar effects were seen in transiently expressed maxi-K channels current (data not shown, n ϭ 3). Whole-cell current expression measured in the presence of 10 nM 17␣-estradiol (E, n ϭ 4) or 10 nM 17␤-estradiol (E, n ϭ 3) did not differ from control recordings. tially negative functional consequences in tissues expressing endogenous maxi-K channels and may be one explanation for the complications associated with therapeutic use of estrogencontaining preparations.

EXPERIMENTAL PROCEDURES
Materials and Reagents-HEK293F cells, fetal bovine serum (FBS), gentamicin, and fungizone were from Invitrogen. Human coronary artery smooth muscle cells were purchased from Cell Applications (San Diego, CA). Diethylstilbestrol, 17␣-estradiol, and 17␤-estradiol, anti-␣smooth muscle actin-fluorescein isothiocyanate antibody, iberiotoxin, and polyethyleneinime were purchased from Sigma. Mouse hybridoma supernatant against a conserved region of the maxi-K channel was generated in our laboratory using a synthetic peptide for amino acids 913-926 of the channel. A sheep polyclonal antibody was generated by using a glutathione S-transferase fusion protein encoding the first intracellular loop of the channel (GenBank TM accession number AF349445). Mouse monoclonal anti-V5 antibody was obtained from Invitrogen, and rabbit polyclonal anti-␤1-subunit antibody was from Affinity Bioreagents (Golden, CO). Biotin-conjugated IgG and streptavidin-conjugated fluorophores were purchased from Jackson Immu-noResearch (West Grove, PA). Human recombinant fibroblast growth factor basic and human epidermal growth factor were obtained from Clonetics (San Diego, CA). The QuikChange site-directed mutagenesis kit and GeneJammer transfection reagent were purchased from Stratagene (La Jolla, CA). The nuclear extract kit was obtained from Active Motif (Carlsbad, CA). Proteasome inhibitor MG-132 and Anopore tissue culture inserts were purchased from Fisher Scientific. [6, H]17␤-Estradiol was obtained from Amersham Biosciences.
Heterologous Expression of the Maxi-K Channel-An adenoviral construct of the maxi-K channel (GenBank TM accession number U11058), encoding the epitope (GenBank TM accession number AF349445), was generated. HEK293F cells, grown to 60 -70% confluence on 0.02-m Anopore tissue culture inserts in phenol red-free DMEM:F-12 medium supplemented with 10% hormone-free FBS, were infected with 10 l of purified construct (1-1.2 ϫ 10 12 particles/ml). Heterologous expression of the maxi-K channel with its accessory ␤1-subunit was performed using 10 l of an adenoviral construct of the human maxi-K channel ␤1-subunit (GenBank TM accession number NM 004137) at a concentration 10 7 particles/ml. Infection efficiency was monitored by eGFP reporter gene expression driven by a separate promoter. Cells cultured in estradiol-free medium as well as cells infected with the adenoviral vector alone were used as controls. One M or 10 nM 17␣-or 17␤estradiol was added to cultures 24 h post-infection and incubation continued for 30 min or 24 h. Proteasome-dependent proteolytic degradation was inhibited by incubating cell cultures with 1 M MG-132 and either 1 M 17␣-estradiol for 24 h or 17␤-estradiol for 30 min. Incubation with MG-132 alone was used as a control.
Generation of the Maxi-K Channel Mutant-The maxi-K channel putative PEST motif mutant was generated by substitution of the -SPKK-amino acid sequence for -AAAA-(SPKK642-645AAAA) using the QuikChange site-directed mutagenesis kit. The mutant was fused at the C terminus with a V5 tag and expressed in HEK293F cells using GeneJammer transfection reagent. Expression of the mutant channel was analyzed by immunocytochemistry following a 30-min incubation with 1 M 17␤-estradiol as described above.
Live Cell Imaging-The maxi-K channels were fused at the C terminus with eGFP tags and expressed in HEK293F cells. Channel expression and intracellular trafficking in response to 17␤-estradiol were examined after 48 h. Cell culture dishes were mounted on a Zeiss 510 confocal scanning microscope (Zeiss, Germany) at 37°C. Cell cultures were subsequently incubated for 45 min in medium containing 1 M 17␤-estradiol and 15 mM HEPES. Images were taken in 1-min intervals.
Electrophysiology-All electrophysiological experiments were performed at room temperature. HEK293 cells expressing the maxi-K channel stably (under G418 selection) or transiently were grown to confluence in 35-mm dishes in phenol red-free DMEM supplemented with 10% hormone-free FBS. Once confluent, cells were trypsinized and resuspended in bath solution. Whole-cell recording was performed and analyzed as described previously (19,20) using a holding potential of Ϫ80 mV, prepulsed to Ϫ100 mV, and currents elicited at step potentials from Ϫ80 to ϩ160 in 20-mV intervals. Iberiotoxin (200 nM) was used to confirm the presence of maxi-K channel current. 17␣-and 17␤-estradiol were added directly to the bath at a final concentration of either 10 nM or 1 M.
Estrogen Binding Assay-HEK293F cells were transfected with either the maxi-K channels or the maxi-K channel accessory ␤1-subunits or both constructs. Cells expressing the adenoviral vector alone and MCF7 cells expressing endogenous nuclear ER␣ were used as controls. All cultured cells were grown in serum-free, phenol red-free medium 24 h prior to harvesting. Preparation of cellular membranes was described previously (21). Nuclear extracts of MCF7 cells were isolated using nuclear extract kit. Triplicate aliquots of 10 g of the cell membranes or nuclear extracts were used in competed and non-competed binding assays. Nonspecific binding (competed) was assayed using 10 nM [6,  Statistical Analysis-Data were expressed as mean Ϯ S.E. Statistical significance of difference between control and experimental groups was calculated by paired t test. p value Ͻ0.05 indicated significant difference between groups.

RESULTS
HEK293 cells, which do not express endogenous estrogen receptors, were utilized to study the effects of 17␣and 17␤estradiol on maxi-K channels. Whole-cell recordings were performed in both HEK293 cell lines expressing the maxi-K channels either transiently or stably. In stable cell lines, both 1 M 17␣-and 17␤-estradiol attenuated whole-cell maxi-K channel currents (Fig. 1, A and B). Maxi-K current was decreased ϳ45% by 17␣-estradiol at ϩ160 mV, although current was significantly inhibited at lower depolarizing voltages (Fig. 1C). A similar effect was observed with 17␤-estradiol with an attenuation of maxi-K channel current by ϳ38% at ϩ160 mV (Fig.  1D). Similar levels of current inhibition were seen in transiently transfected cells. Current inhibition was not detected within 10 min after addition of 10 nM 17␣-and 17␤-estradiol to the bath solution (Fig. 1, E and F). Currents were recorded for up to 15 min until seal integrity was compromised. Due to reports of endogenous K ϩ current in HEK293 cells (23), iberiotoxin, a maxi-K channel inhibitor, was added to the bath solution to determine the contribution of the maxi-K channel to the estrogen-sensitive whole-cell K ϩ current. The maxi-K channel comprised ϳ77% of the total K ϩ current in HEK293F cells and ϳ100% of the estrogen-sensitive component (data not shown).
The effects of estrogens on protein expression of transiently expressed maxi-K channels was studied immunocytochemically. A thirty minute incubation with 1 M 17␤-estradiol induced a profound (ϳ80%) attenuation of channel expression as compared with estradiol-free controls (Fig. 2, A, B, and F). At a concentration of 1 M, 17␣-estradiol demonstrated a similar level of inhibition, but only after 24 h exposure (Fig. 2, C and  F). Although down-regulation of maxi-K channels has also been observed following a 30 min incubation with 10 nM 17␤-estradiol (ϳ45%, Fig. 2, D and F), 17␣-estradiol at the same concentration did not demonstrate an inhibition of channel expression after either a 30 min or 24 h incubation (Fig. 2, E and F). At a concentration of 100 nM, 17␤-estradiol inhibited channel expression by ϳ60%, while 17␣-estradiol was again ineffective at this concentration (data not shown). Preincubation of cells with 17␤-estradiol (10 nM or 1 M) for 24 h prior to infection inhibited maxi-K channels by ϳ80%. 17␣-Estradiol produced a similar effect but only at 1 M (data not shown). Neither estrogen affected infection efficiency as assessed by GFP reporter gene expression. Additionally, an inhibitory effect of 1 M 17␤-estradiol was observed in live HEK293F cells expressing maxi-K channels fused C-terminally with an eGFP tag. The response was detected as early as 10 min after addition of 17␤-estradiol, although some cells demonstrated a delayed reaction detected after 40 min (Fig. 3, A and B). The difference in timing of the response may be related to the presence of the large C-terminal tag affecting the estrogen sensitivity of the channel. This response was clearly estrogen-dependent, since cells incubated in estradiol-free medium did not show a similar decrease in signal intensity (Fig. 3, C and D).
The activation of the maxi-K channels by estradiol has been reported to require the presence of ancillary ␤1-subunits (14,15). We speculated that the accessory ␤1-subunit may be a stabilizing factor for the maxi-K channel in an estradiol-enriched environment. Endogenous ␤1-subunit transcript was not detected in HEK293F cells by reverse transcriptase-PCR (data not shown). Estradiol-induced down-regulation of maxi-K channels was observed in cells co-expressing the maxi-K channel with its accessory ␤1-subunit (Fig. 4, A and B) after a 30-min incubation with 17␤-estradiol (Fig. 4C) or a 24-h incubation with 17␣ estradiol (Fig. 4E). In addition, ␤1-subunit expression was attenuated in a similar manner (Fig. 4, D and  F). Thus the presence of the ␤1-subunit yielded the same degree of channel suppression observed when maxi-K channels were expressed alone (Fig. 4G). Estradiol-mediated down-regulation of maxi-K channel membrane expression may indicate that proteasome-dependent proteolytic degradation may be responsible, since this mechanism is involved in down-regulation of the ligand-activated ER␣ (24). According to this hypothesis, inhibition of the proteasome complex would stabilize the maxi-K channel protein on the cell membrane. We examined the effects of the proteasome inhibitor, MG-132, on either a 17␤-estradiol- (Fig.  5, A and B) or a 17␣-estradiol-induced (Fig. 5, C and D) downregulation of the maxi-K channel. The data demonstrate that incubation in the presence of MG-132 was sufficient to prevent down-regulation of maxi-K channels by estradiols (Fig. 5E). A similar effect was detected using 10 M lactacystin, another proteasome inhibitor (data not shown). Analysis of the secondary structure of maxi-K channel using several Web-based databases (ExPASy, PESTFind) revealed the presence of a putative PEST motif -RLEDEQPSTLSPKK-in its intracellular C terminus, which may be mediating recognition and degradation of the maxi-K channel by the proteasome (25,26). The functional significance of this motif was determined by substituting the -SPKK-amino acids with -AAAA-(SPKK642-645AAAA) and assessing the response of the mutant to 17␤estradiol (Fig. 6, A and B). Resistance to estrogen-induced proteolytic degradation and a tendency toward increased channel expression was observed following a 30-min incubation with 1 M 17␤-estradiol (Fig. 6, A-C). These data indicate that the proteasome mediates channel degradation in response to estrogens.
Ligand binding is required for the estradiol receptor to be targeted to, and degraded by, proteasomes (24,27). It has also been reported that the estrogen sensitivity of the maxi-K channel is conferred by its accessory ␤1-subunit (14,15). Therefore, we attempted to determine whether estrogen binding to the maxi-K channels, independent of the ␤1-subunit, induces proteolytic degradation of the channel protein. Specific binding was compared between sham-transfected controls (Fig. 7, sham) and each experimental group. Maxi-K channels (Fig. 7, maxi-K), but not ␤1-subunits (Fig. 7, ␤1), demonstrate specific binding when incubated with 10 nM [6,7-3 H]17␤-estradiol. However, accessory ␤1-subunits co-expressed with maxi-K channels (Fig. 7, maxi-Kϩ␤1) significantly increased specific binding of estradiol to the channel. MCF7 cells expressing endogenous estradiol receptors (Fig. 7, MCF7) were used as a positive control and also demonstrated specific binding. The presence of estradiol receptors in these cells was confirmed by immunoblotting (data not shown).
The physiological relevance of the estrogen effect on maxi-K channels was confirmed in cultured human coronary artery smooth muscle cells. Similar to what was detected in the heterologous system, prolonged exposure to 1 M 17␤-estradiol (Fig. 8, A and B) and 17␣-estradiol (Fig. 8C) resulted in significant down-regulation of endogenous maxi-K channels. Downregulation of the maxi-K channels occurred despite the presence of endogenous estrogen receptors in smooth muscle cells.
The data presented in this paper suggest that estrogens have an inhibitory effect on the expression of the maxi-K channel protein and on the current generated by this channel. Downregulation of the channel appears to be induced by the binding of 17␤or 17␣-estradiol to the channel, followed by subsequent proteasome-mediated proteolytic degradation. The ancillary ␤1-subunit did not prevent estrogen from down-regulating the maxi-K channel. Although the ␤1-subunit itself does not function as a binding site for 17␤-estradiol, it appears to greatly facilitate estrogen binding to the maxi-K channels.

DISCUSSION
Controversy stirred by the findings of recent clinical trials on hormone replacement therapy has renewed interest in the physiological mechanisms underlying estrogenic protective effects in peripheral tissues that might explain cardiovascular and cognitive health benefits of estrogens in premenopausal women (1,28). It is becoming apparent that estrogen per se is not sufficient for such protection, and the ultimate effects of this steroid may differ depending on unidentified signaling pathways. In the present study, we present one potential pathophysiological foundation for the increase in risk of various complications associated with continuous exposure to estradiol (as in prolonged use of estrogen-containing preparations, environmental pollution with estrogenic compounds, etc.). We provide evidence that exposure to estrogens may cause a significant down-regulation of the membrane expression of maxi-K channels. These channels generate significant repolarizing K ϩ current to buffer cell excitability and mediate a variety of functions in normal and diseased tissues including brain, smooth, and skeletal muscle. Consequences of such down-regulation would be tissue-specific and may be potentially negative for the function of normal tissue. In vascular smooth muscle, for example, these channels maintain the normal myogenic tone of the vascular wall. Loss of this channel may lead to an impaired vascular relaxation and response to contractile stimuli.
Although estrogen-induced degradation of the maxi-K channel protein has not been described, previous studies have reported variable effects of estrogens on the maxi-K channel current. Increases in repolarizing current and open-state probability, as well as decreases in unitary conductance, have been demonstrated (12,15,29). In our experiments the effects of 17␤-estradiol on maxi-K channel expression were consistently inhibitory over a wide range of concentrations, whereas 17␣estradiol inhibited expression only at a high concentration. If both estrogens use the same binding site(s) to inhibit maxi-K channels, 17␣-estradiol may bind with lower affinity (24). Although the exact binding affinity of 17␣and 17␤-estradiol for the maxi-K channel has not been evaluated, the dose response data for 17␤-estradiol are consistent with a binding affinity lower than that reported for intracellular estrogen receptors. It is important to mention that maxi-K channels lack a region homologous to the ligand-binding domain of the estradiol receptor. Thus, other motifs may be recognized by estrogen as potential binding sites. Elucidating this site may provide new insights into estrogenic effects not mediated by nuclear estradiol receptors.
The data presented here do confirm a role for the maxi-K channel accessory ␤1-subunit in regulation of the maxi-K channels sensitivity to estrogen, although our results do not suggest direct binding of estradiol to this subunit. Analysis of our binding assays revealed that total estradiol binding in cells expressing the ␤1-subunits alone was significantly higher when compared with other experimental groups and controls, as reported previously (14). However, nonspecific binding was not statistically different from total binding. It is possible that the ␤1-subunits expressed alone may change the biophysical properties of the cell membrane to induce trapping rather than specific binding of estradiol. The enhanced level of specific binding observed in maxi-K channels co-expressed with their ␤1-subunits implies a role for these subunits in inducing conformational changes in the maxi-K channel protein favorable for specific estradiol binding. Two mechanisms potentially explain the observed increase: 1) conformational change increases the affinity of the binding site or 2) a conformational change reveals more binding sites on the maxi-K channels available to estradiol. Further research is necessary to determine the mechanism responsible for this increase. It is also possible to speculate, although it was not specifically addressed here, that an increase in specific binding would allow for easier recognition of estrogen-bound maxi-K channels by the proteasome and, thus, a faster turnover rate.
It has been confirmed by numerous studies that endogenous estrogens exert effects beyond their reproductive actions. Our study, as well as previously published work (30), suggest that the multiplicity of peripheral actions of estrogen may be explained by its recognition by proteins, in addition to the classical estrogen receptors, that are capable of low affinity binding. Estrogen-bound proteins are then directed along estradiol signaling pathways and degraded by the proteasome similar to classical nuclear estradiol receptors (27). Consequences of such down-regulation for cell or tissue function would depend on the particular protein involved and the degree of inhibition. The presence of nuclear estradiol receptors does not prevent low affinity binding to other proteins, as shown by the experiments on the effects of estrogens on the expression of endogenous maxi-K channels in smooth muscle cells. Although conditions necessary for estradiol binding to proteins besides the estradiol receptors have not been investigated, it is reasonable to speculate that saturation of "traditional" estradiol binding sites (i.e. estradiol receptors) by high concentrations of estrogen and continuous supply of free unbound estradiol would favor low affinity binding. Use of therapeutic or environmental estrogenic compounds may produce such conditions due to relatively high doses and a lack of diurnal variations characteristic of ovarian production of estrogen (31,32).
In summary, estrogens down-regulate maxi-K channel cell surface expression and repolarizing current. Down-regulation involves specific binding of estrogen to the maxi-K channel and induction of a proteasome-dependent proteolytic degradation. This action can be elicited independently of classical nuclear estradiol receptors or the accessory ␤1-subunit, but in the presence of ␤1-subunits specific binding of estradiol to maxi-K channels is increased significantly. A similar down-regulation has been reported in response to chronic exposure to glucocorticoids, although this study did not address the possibility that this response could be mediated by binding of this class of steroids to the channel (33). The results from this study demonstrate the importance of elucidating the mechanisms of nonreproductive action of estradiol given the widespread therapeutic and environmental use of agents with estrogenic activity.