Tamoxifen Activates Smooth Muscle BK Channels through the Regulatory β1 Subunit

Estrogen (17β-estradiol; 17βE) and xenoestrogens, estrogenic compounds that are not steroid hormones, have non-genomic actions at plasma membrane receptors unrelated to the nuclear estrogen receptor. The open probability (Po) of large conductance Ca2+/voltage-sensitive k+(BK) channels is increased by 17βE through the regulatory β1 subunit. The pharmacological nature of the putative membrane binding site is unclear. We probed the site by determining whether tamoxifen ((Z)-1-(p-dimethylaminoethoxy-phenyl)-1,2-diphenyl-1-butene; Tx), a chemotherapeutic xenoestrogen, increased Po in clinically relevant concentrations (0.1–10 μm). In whole cell patch clamp recordings on canine colonic myocytes, which express the β1 subunit, Tx activated charybdotoxin-sensitive K+current. In single channel experiments, Tx increased the NPo (Po × number channels; N) and decreased the unitary conductance (γ) of BK channels. Tx increased NPo (EC50 = 0.65 μm) in excised membrane patches independent of Ca2+ changes. The Tx mechanism of action requires the β1 subunit, as Tx increased the NPo of Slo α expressed in human embryonic kidney cells only in the presence of the β1 subunit. Tx decreased γ of the α subunit expressed alone, without effect on NPo. Our data indicate that Tx increases BK channel activity in therapeutic concentrations and reveal novel pharmacological properties attributable to the α and β1 subunits. These data shed light on BK channel structure and function, non-genomic mechanisms of regulation, and physiologically and therapeutically relevant effects of xenoestrogens.

Estrogen (17␤-estradiol; 17␤E) and xenoestrogens, estrogenic compounds that are not steroid hormones, have nongenomic actions at plasma membrane receptors unrelated to the nuclear estrogen receptor. The open probability (P o ) of large conductance Ca 2؉ /voltage-sensitive k ؉ (BK) channels is increased by 17␤E through the regulatory ␤1 subunit. The pharmacological nature of the putative membrane binding site is unclear. We probed the site by determining whether tamoxifen ((Z)-1-(p-dimethylaminoethoxy-phenyl)-1,2-diphenyl-1-butene; Tx), a chemotherapeutic xenoestrogen, increased P o in clinically relevant concentrations (0.1-10 M). In whole cell patch clamp recordings on canine colonic myocytes, which express the ␤1 subunit, Tx activated charybdotoxin-sensitive K ؉ current. In single channel experiments, Tx increased the NP o (P o ؋ number channels; N) and decreased the unitary conductance (␥) of BK channels. Tx increased NP o (EC 50 ‫؍‬ 0.65 M) in excised membrane patches independent of Ca 2؉ changes. The Tx mechanism of action requires the ␤1 subunit, as Tx increased the NP o of Slo ␣ expressed in human embryonic kidney cells only in the presence of the ␤1 subunit. Tx decreased ␥ of the ␣ subunit expressed alone, without effect on NP o . Our data indicate that Tx increases BK channel activity in therapeutic concentrations and reveal novel pharmacological properties attributable to the ␣ and ␤1 subunits. These data shed light on BK channel structure and function, non-genomic mechanisms of regulation, and physiologically and therapeutically relevant effects of xenoestrogens.
Estrogens and xenoestrogens have non-genomic effects mediated by plasma membrane receptors unrelated to the nuclear estrogen receptor (1). One of these is to increase the NP o of BK 1 channels (2), Ca 2ϩ -sensitive members of the voltage-gated K ϩ channel superfamily with important functions in many cells (3). BK channels are composed of ␣ and ␤ subunits. The ␣ subunit forms the K ϩ -selective pore, while ␤ subunits influence the pharmacology, kinetics, and voltage/Ca 2ϩ -sensitivity of BK channels. The ␤1 subunit of smooth muscle BK channels is physiologically important because knockout mice lacking this subunit are hypertensive and demonstrate altered vascular reactivity (4,5).
Recent studies suggest that BK channels are potential targets for 17␤E and xenoestrogens. BK channel NP o is increased by 17␤E, an effect that requires the regulatory ␤1 subunit (2). 17␤E and xenoestrogens reduce coronary vascular tone by inhibiting L-type Ca 2ϩ channels and activating BK channels (6). The pharmacological nature of the putative 17␤E-binding site on the smooth muscle BK ␤1 subunit is unknown. The xenoestrogen Tx, a commonly used chemotherapeutic agent, is an antagonist of the nuclear estrogen receptor (7). It is not known, however, if this clinically important drug increases BK channel NP o . We investigated whether Tx increases BK channel NP o in smooth muscle cells and whether the ␤1 subunit is important for this effect. These findings give insight into BK channel structure and function, non-genomic regulation by xenoestrogens, and Tx-induced side effects.

MATERIALS AND METHODS
Cell Isolation and Preparation-Smooth muscle cells were isolated by enzymatic dispersion described previously (8). Dogs were anesthetized with ketamine, and the colon was removed via a midline incision. Mice were anesthetized with chloroform and killed by cervical dislocation. Human tissue samples were obtained from consenting patients undergoing gastric bypass for the treatment of morbid obesity. Circular muscle of the canine colon and human jejunum was dissected free of mucosa, submucosa, and longitudinal muscle in Ca 2ϩ -free Hanks solution. Strips of muscle were treated with collagenase (345 units/ml; Worthington Biochemical Corp.; Freehold, NJ) in Ca 2ϩ -free Hanks at 37°C to produce suspensions of single cells by gentle stirring. Mouse colon (circular and longitudinal muscle layers) was dissected free of mucosa prior to enzymatic dispersion. Mouse aorta and canine mesenteric vein were enzymatically digested without further dissection.
HEK293 cells (ATCC cell line number CRL-1573; Manassas, VA) were grown in glutamax-supplemented RPMI medium (Life Technologies, Inc., Manassas, VA) with 10% heat-inactivated horse serum (Summit Biotechnology; Fort Collins, CO) in a humidified atmosphere with 5% CO 2 at 37°C. cDNA encoding the ␣ and ␤1 subunits from canine colonic smooth was cloned into the pZEOSV mammalian expression vector (Invitrogen; Carlsbad, CA) as described previously (9). Cells were transiently transfected via electroporation with a total of 40 g of plasmid DNA encoding either the ␣ subunit alone or with a 1:1 mix of plasmids encoding the ␣ and ␤1 subunits. Cells were subcultured on glass coverslips for electrophysiological studies 24 -48 h after transfection.
Patch Clamp Techniques-Currents were recorded from whole cell, inside-out, and outside-out patches. An Axopatch 1D amplifier and CV-4 headstage were used for data acquisition (Axon Instruments; Foster City, CA). Data were acquired in pCLAMP version 5.5.1 (Clampex and Fetchex; Axon). The digitization rate was 4 kHz, and the cut-off frequency for filtration was 1 kHz. Data were analyzed in pCLAMP (version 6; whole cell I-V curves, Clampfit; Axon) and the Analysis of Single Channel Data program (NP o , ␥, and histograms; University of Leuven, Belgium).
Statistical Analysis-Data are presented as the mean Ϯ S.E. from n cells. Statistical analyses were made with paired and unpaired Student's t test, one-and two-way analysis of variance (ANOVA; with repeated measures as necessary). The threshold for statistical significance was set as p Ͻ 0.05 and post-hoc analyses (e.g. Bonferroni test) were applied to indicate differences suggested by ANOVA. Statistical tests were run in Prism (version 3; GraphPad Software, Inc.; San Diego, CA) and Sigma-Stat (version 2; Jandel Scientific Software; San Rafael, CA).

RESULTS
Voltage steps from a holding potential of Ϫ80 mV elicited two types of voltage-dependent current in canine colonic myocytes studied in the whole cell patch clamp configuration with physiological ion gradients (5 mM K ϩ outside; 140 mM K ϩ inside; Fig. 1A). Depolarizations to 0 mV activated delayed rectifier current, while greater depolarizations activated BK current (11). Tx (1 M) produced two effects on voltage-dependent outward currents. First, in the range of potentials from 0 to ϩ40 mV, Tx inhibited the delayed rectifier component of outward current as reported previously (8). Second, Tx increased outward current at potentials positive to ϩ40 mV.
The outward current activated by Tx was carried by K ϩ , as dialysis of cells with NMDG (N-methyl-D-glucamine) or TEA (tetraethylammonium) instead of K ϩ prevented the stimulatory effect of Tx. Qualitatively similar effects of Tx were noted when voltage step protocols were used (holding potential Ϫ80 mV; Fig. 1, A and B) or when membrane voltage was ramped from Ϫ80 to ϩ80 mV (Fig. 1B, inset). Tx activated K ϩ current whether cells were studied with the amphotericin-perforated or conventional dialyzed patch technique, suggesting preservation of the intracellular signaling milieu was not necessary. Tx increased current when cells were dialyzed with a pipette solution containing 10 mM BAPTA and no added Ca 2ϩ . The calculated free Ca 2ϩ was ϳ2 nM (assuming 50 M Ca 2ϩ contamination from water and reagents) and the free BAPTA concentration was greater than 9.5 mM, providing a large buffer against Ca 2ϩ changes. Thus, experiments with BAPTA suggest that the increase in outward current due to Tx was not likely due to changes in Ca 2ϩ .
Tx-activated current was noisy (Fig. 1A), suggesting it was due to openings of channels with a large unitary conductance. Tx-activated current was blocked by TEA (1 mM). For example, outward current activated by steps to ϩ80 mV averaged 1.77 Ϯ 0.31 nA and was increased to 4.14 Ϯ 0.51 nA by 1 M Tx (n ϭ 3; p Ͻ 0.05 versus control by one-way repeated measures ANOVA). Addition of TEA (1 mM) in the continued presence of Tx reduced outward current to 0.95 Ϯ 0.15 nA (p Ͻ 0.05 versus both control and Tx). Charybdotoxin (Ctx; 50 nM), a specific peptide inhibitor of BK channels, also inhibited the current activated by Tx. A summary of five experiments testing the effect of Ctx on Tx-activated current is shown in Fig. 1C. We tested whether the addition of a BK channel blocker abrogated the subsequent effect of 1 M Tx (n ϭ 3; all comparisons by one-way repeated measures ANOVA). The addition of 1 mM TEA reduced current at ϩ80 mV to 22 Ϯ 9% of control (p Ͻ 0.05). Current with the addition of 1 M Tx, in the continued presence of 1 mM TEA, was 43 Ϯ 8% of control (p Ͻ 0.05 versus control, not different from TEA alone). Removal of TEA, FIG. 1. Tx activates charybdotoxin-sensitive K ؉ current in canine colonic myocytes. A, representative current traces from a whole cell patch clamp experiment before and after the addition of 1 M Tx. The bath was nominally Ca 2ϩ -free, and the cell was dialyzed with 10 mM BAPTA. Tx decreased current at 0 mV, but increased current at more positive potentials. B, group data (n ϭ 11) demonstrating the effect of 1 M Tx on K ϩ current. Conditions were the same as in A, and current was measured in the last 50 ms of the voltage steps. Asterisks indicate differences between the groups (two-way repeated measures ANOVA). Inset, representative current traces elicited by a voltage ramp from Ϫ80 to ϩ80 mV. Tx inhibited delayed rectifier current, but activated current at more positive potentials. Scale bars are 1 nA and 1 s. C, group data (n ϭ 5) for current at ϩ80 mV and the effect of 1 M Tx and 50 nM Ctx. Asterisk indicates a significant difference from control, while the cross indicates a difference between Tx and Tx ϩ Ctx (oneway repeated measures ANOVA). D, the time course and reversibility of Tx is shown in the left panel. Current was measured at ϩ80 mV and plotted versus time. Tx increased current in a rapid and reversible manner. The right panel contains averaged current traces (mean of 60 s or 12 sweeps) before, during, and after Tx exposure. in the continued presence of 1 M Tx, increased current to 211 Ϯ 42% of control (p Ͻ 0.05 versus control, TEA alone, and TEA plus Tx). When all drugs were washed out, current at ϩ80 mV returned to within 13 Ϯ 11% of control (not different). The pharmacological profile of the current activated by Tx is consistent with BK channels. Experiments were performed to determine the time course and reversibility of the effects of 1 M Tx on whole-cell current at ϩ80 mV (Fig. 1D). Tx increased current 183 Ϯ 40% (n ϭ 6; p Ͻ 0.05) in 143 Ϯ 35 s. Current returned to within 19 Ϯ 13% of the control value within 3-5 min of Tx removal (not different from control; all comparisons made by one-way repeated measures ANOVA).
Excised patch recordings were performed to identify the conductance activated by Tx. In inside-out or outside-out patches from canine colonic myocytes, large conductance (Ͼ200 pS), Ca 2ϩ -and voltage-dependent channel openings were observed (12). The channels were highly selective for K ϩ over Na ϩ . In outside-out patches, submillimolar concentrations of TEA caused a flicker block of these channels. These are properties of BK channels, which have been identified and characterized previously in canine colonic myocytes (12). Tx (1 M) increased NP o of BK channels in excised patches (Fig. 2). Additionally, Tx decreased ␥. For example, in the records of Fig. 2, ␥ was 269 pS and was reduced to 243 pS by 1 M Tx. Similar to the results in whole cell patches with TEA and Ctx (Fig. 1), the single channels activated by Tx were blocked by Ctx (50 nM; Fig. 3A). To determine this, outside-out patches were studied. Tx had the same effect in outside-out patches as it did in inside-out patches (and in cell-attached patches; not shown). As can be seen in the all-points amplitude histogram from 1 min recordings from an outside-out patch, Tx increased NP o and decreased the ␥ of BK channels (Fig. 3B). In the control record, the Gaussian mean for the single channel openings at ϩ40 mV was centered at 10.4 pA. In contrast, the unitary currents in the presence of Tx were centered at 8.9 pA.
Single channel currents were recorded for long durations to accurately assess the effect of Tx on NP o . As the recording in Fig. 4A shows, NP o was stable after patch excision and Tx and 17␤E increased it in a reversible manner. Me 2 SO, the vehicle used for Tx and 17␤E, had no effect on NP o (Fig. 4B). Tx increased BK NP o in a concentration-dependent manner (0.1-10 M) with an EC 50 of 0.65 M (Fig. 5A; n ϭ 15-22). NP o was allowed to reach a steady-state value for ϳ5 min before the addition of Tx. Current was then recorded for 1-10 min under control conditions, and then Tx was added. NP o was measured in the presence of Tx for 1-10 min and the mean change was calculated. These data demonstrate that Tx increases BK chan-  6). B, Tx was more potent than 17␤E in the same patches (n ϭ 10). The effect of 17␤E was measured after 10 min, then 1 M Tx was added. Data were compared by one-way repeated measures ANOVA; asterisks indicate a difference from control, cross indicates a difference between Tx and 17␤E. C, Tx decreases ␥. Currents were measured from inside-out patches with 5 mM K ϩ in the pipette and 140 mM K ϩ in the bath before and after the addition of 1 M Tx. I-V curves were compared by two-way repeated measures ANOVA; asterisks indicate a difference from control (n ϭ 6). D, Tx and 17␤E decrease ␥ at ϩ80 mV. Statistical analyses and symbols are the same as for C. nel NP o ; in vivo, however, most Tx is bound to plasma proteins. Thus, it is uncertain whether the presence of albumin (the major Tx-binding protein), for example, would alter the efficacy and potency of Tx in these in vitro measurements. We performed experiments to determine the Tx concentration-NP o response relationship in the presence of 20 mg/ml albumin. The effects of Tx were still apparent in the presence of protein; however, albumin decreased the efficacy of Tx ϳ60% and caused a slight right shift in the EC 50 (2.0 M; Fig. 5A).
The effect of Tx on increasing BK NP o was similar to the effect of 17␤E described previously (2). In our experiments, however, Tx was approximately 4-fold more potent than the previously reported EC 50 of 2.6 M for 17␤E (2). These pharmacological concentrations of 17␤E are substantially higher than the 100 -600 pM physiological concentration range (13), but the EC 50 for Tx was well within the therapeutic plasma concentration range measured in plasma of human cancer patients (0.1-10 M; Refs. 14 -17). To compare the potency of 17␤E and Tx to increase NP o of BK channels, we tested the two agents in the same excised patches (Fig. 5B). 17␤E (10 M) was applied first and NP o was allowed to reach a steady state. 17␤E was then washed out, and 1 M Tx was added, which increased NP o to a higher level. These studies indicated that Tx is at least one log order more potent than 17␤E, as Tx had a greater potentiating effect than a 10-fold higher concentration of 17␤E.
Tx shifted the voltage dependence of steady-state activation to more negative potentials. Inside-out patches were studied in symmetrical 140 mM K ϩ and varying Ca 2ϩ concentrations. The patch potential was held at 0 mV and stepped to different potentials to determine steady-state activation before and after the addition of 1 M Tx. The patch potential was made more positive until conductance was maximal and then activation curves (i.e. normalized conductance versus voltage) were generated. Tx (1 M) decreased current (Fig. 6A) and shifted the voltage of half-activation (V1 ⁄2 ) to more negative values at the same free Ca 2ϩ concentration (Fig. 6B). Tx had no effect on BK channel activation kinetics. The time constant of activation at ϩ150 mV in 100 nM free Ca 2ϩ was 23 Ϯ 2 and 24 Ϯ 3 ms before and after the addition of 1 M Tx, respectively (n ϭ 18). Experiments were performed in Ca 2ϩ concentrations ranging from pCa 7 to 3. Increasing the free Ca 2ϩ concentration made V1 ⁄2 more negative; however, Tx (1 M) caused the same hyperpolarizing shift in V1 ⁄2 independent of the free Ca 2ϩ concentration (Fig. 6C).
The effect of Tx to increase BK NP o was not limited to canine colonic smooth muscle. We tested the effect of Tx on BK channels in 4 additional smooth muscle cell types from both visceral and vascular tissues. All of the same effects of Tx observed in canine colonic myocytes were noted in smooth muscle cells from the human jejunum, mouse colon, mouse aorta, and canine mesenteric vein. Tx increased BK current in whole cell recordings; however, delayed rectifier current was inhibited, as in canine colon (Fig. 7A). Tx increased the NP o of BK channels in excised patches in a concentration-dependent manner (Fig. 7B) and ␥ was decreased (Fig. 7C). Representative data from these experiments are shown in Fig. 7 and summarized in Table I.
Experiments were performed to determine whether the effects of Tx are due to interactions with the ␣ or ␤1 subunits of BK channels. We expressed the cloned canine Slo ␣ subunit in HEK293 cells, which do not express an endogenous ␤1 subunit (18). Inside-out patches from these cells contained K ϩ channels with a slope conductance of 264 Ϯ 13 pS (n ϭ 7). Tx had no effect on the NP o of single channels in cells expressing Slo ␣, but Tx did reduce ␥ (233 Ϯ 11 pS; p Ͻ 0.05, paired Student's t test; 12 Ϯ 4% reduction; Fig. 8A). Expression of Slo ␣ and ␤1 subunits together produced channels with an average ␥ of 266 Ϯ 8 pS (n ϭ 9). This conductance was not significantly different from the ␥ of channels in cells transfected with ␣ subunits alone (p Ͼ 0.05; unpaired Student's t test). Tx increased NP o of channels in cells expressing both ␣ and ␤1 subunits (Fig. 8B) and decreased ␥ (230 Ϯ 9 pS, p Ͻ 0.05, paired Student's t test; 13 Ϯ 2% reduction). These data suggest that the Tx-induced increase in NP o requires the presence of the regulatory ␤1 subunit, however the attenuation of ␥ does not. DISCUSSION Tx, a commonly used chemotherapeutic agent for the treatment and prevention of estrogen receptor-dependent cancers, increased NP o of BK channels. Activation of BK channels by Tx was observed in tonic and phasic smooth muscle myocytes from mice, dogs, and humans, and it may be a common phenomenon for cells that express the regulatory ␤1 subunit (as do the tissues used here, Ref. 9). The effect of Tx to increase BK NP o required the presence of the regulatory ␤1 subunit, but the exact site of action is unknown. In the absence of the regulatory ␤1 subunit, Tx decreased ␥ without affecting NP o . The activating effect of Tx is similar to that reported for 17␤E on BK channels containing the ␤1 subunit (2) and novel ␤ subunits FIG. 6. Tx shifts the voltage dependence of steady-state activation to more negative potentials. A, representative current traces (the average current from three trials on the same patch) from an inside-out patch in symmetrical 140 mM K ϩ and 100 nM free Ca 2ϩ . The patch potential was held at 0 mV and stepped from ϩ30 to ϩ180 mV in 10 mV steps before and after the addition of 1 M Tx, which decreased current. B, activation curves (normalized conductance versus voltage) before and after the addition of 1 M Tx, which shifted V1 ⁄2 to a more negative value. The curves are different between ϩ50 and ϩ170 mV (n ϭ 18; two-way repeated measures ANOVA). C, Tx shifts V1 ⁄2 independent of the free Ca 2ϩ concentration (pCa is the minus log of the molar concentration). Control and Tx groups are different at each Ca 2ϩ concentration (paired Student's t test; n ϭ 5-18). (19). However, Tx was at least 10 times more potent than 17␤E. Further, the EC 50 for Tx (0.65 M) falls within the plasma concentration range measured during chemotherapy (0.1-10 M; Ref. 14). This contrasts with the peak plasma concentrations of 17␤E (Ͻ1 nM; Ref. 13)) and the reported EC 50 (2.6 M) for 17␤E effects on BK channels (2). Both Tx and 17␤E decreased the ␥ of BK channels, a novel effect seen in addition to changes in NP o . Effects of Tx on BK channel NP o were observable at concentrations as low as 0.1 M, apparent in the presence of 20 mg/ml albumin (a Tx-binding protein), and the effects were rapid and reversible.
The speed and reversibility of the Tx effects on BK channels suggest they are due to a non-genomic mechanism at the plasma membrane (20) (e.g. it has been proposed that there is an extracellular binding site on the regulatory ␤1 subunit; Ref. 2). It is unknown whether Tx or 17␤E binds directly to ␤1 subunits or to an accessory protein, the influence of which is conveyed to BK channels via the ␤1 subunit. However, the similarities between the effect of Tx and 17␤E on BK NP o suggest both agonists may affect the channels via the same binding site, the location of which has not been determined conclusively. The data of Valverde et al. (2) indicate that 17␤E activates BK channels in inside-out patches independent of cellular signaling mechanisms. Our data confirm and further show that Tx has a similar ability to activate BK channels in inside-out patches. Tx and 17␤E are very lipophilic and cross membranes readily. The intended therapeutic action of Tx is to cross the cell membrane and antagonize the nuclear estrogen receptor. This explains why the same effects on BK NP o and conductance were observed in whole cell, cell-attached, outsideout, and inside-out patches. However, Tx (and 17␤E) also acts non-genomically in patches of excised membrane. The actual site of action could be at the extracellular surface, at the cytoplasmic face, or within the plasma membrane itself. The exact site is presently unknown and might be resolved in the future by using site-directed mutagenesis. What can be determined from the work of Valverde et al. (2) is that a membrane-impermeant form of 17␤E conjugated to albumin elicits effects only when applied to the extracellular side of the membrane. The lack of effect from the cytoplasmic face for albumin-conjugated 17␤E may be due to some steric hindrance and not indicative of a known position for the putative binding site. Regardless, the ␤1 subunit is an integral, requisite part of the signaling mechanism. Further, the effects of Tx and 17␤E are likely direct and do not involve other signaling mechanisms such as nitric oxide production or phosphorylation (well recognized mechanisms that affect BK NP o ), as the effects are observed in cell-free patches in the absence of appropriate substrates for those signal transduction cascades. Additional strength for a direct effect of Tx and 17␤E is that they have no effect on the NP o of the ␣ subunit alone, which possesses a stimulatory protein kinase G phosphorylation site (18).
Tx also affects other ion channels. Tx inhibits volume-sensitive Cl Ϫ and delayed rectifier K ϩ currents in canine colonic smooth muscle (8). Tx also inhibits voltage-gated Na ϩ (21), FIG. 7. Tx activates BK channels in various smooth muscle cell types. A, representative whole cell currents in a human jejunal myocyte were recorded before and after the addition of 1 M Tx. Delayed rectifier current was inhibited BK current was activated. B, single channel records in an inside-out patch taken from a mouse colonic myocyte. The holding potential was ϩ60 mV and the bath contained 100 nM Ca 2ϩ . Tx increased NP o in a concentration-dependent manner. C, single channel records from an inside-out patch of canine mesenteric vein smooth muscle. The holding potential was ϩ80 mV, and the bath contained 100 nM Ca 2ϩ . Tx increased NP o and decreased the ␥ (asymmetrical K ϩ ; 5 mM pipette, 140 mM bath). FIG. 8. Tx activates BK channels only in the presence of the regulatory ␤1 subunit. A, representative current traces of inside-out patches from HEK 293 expressing either the ␣ subunit alone or the ␣ and ␤1 subunit together. Currents were recorded at ϩ80 mV before and after the addition of 1 M Tx. Note free Ca 2ϩ concentration between the two groups is different; due to the reduced Ca 2ϩ -sensitivity of the ␣ subunit expressed alone, a higher Ca 2ϩ concentration was required to elicit basal activity. B, group data showing the effect of Tx (1 M) on channels composed of ␣ subunits alone (n ϭ 7) or ␣ and ␤1 subunits together (n ϭ 9). Tx activates BK channels only in the presence of the regulatory ␤1 subunit. The asterisk indicates a significant difference between control and Tx for the ␣ ϩ ␤1 group (paired Student's t test), while the cross indicates a difference in normalized NP o between ␣ and ␣ ϩ ␤1 groups in the presence of Tx (unpaired Student's t test). L-type Ca 2ϩ (8,22), and nonselective cation channels (23,24). Tx inhibits protein kinase C (25) and calmodulin (26) and antagonizes histamine receptors (27). Thus, the effects of Tx in vitro or in vivo are likely to be multifaceted and are difficult to identify specifically. However, none of these other actions explain the effects of Tx to increase BK channel NP o . Tx has a number of effects on smooth muscle tissues. Tx relaxes myometrial arteries and increases uterine blood flow (28). Tx inhibits spontaneous and agonist-induced myometrial contractions (29) and hyperpolarizes and relaxes cerebral arteries (30). These effects are consistent with the action of Tx on BK channels. High extracellular K ϩ , which limits the degree to which openings of BK channels can hyperpolarize smooth muscle, abrogates the relaxing effect of Tx on myometrium (29). BK channels comprise one element of a negative feedback loop controlling the tone of cerebral arteries. Specifically, release of Ca 2ϩ from the sarcoplasmic reticulum and subsequent activation of BK channels relaxes cerebral arteries (31). It is likely that the hyperpolarization of cerebral arteries (30) and relaxation of myometrium and myometrial arteries (28,29) by Tx is due, in part, to the activation of BK channels. BK channel activation would drive membrane potential toward E K , reducing the P o of L-type Ca 2ϩ channels independent of (and in addition to) possible direct inhibitory effects on voltage-gated Ca 2ϩ entry. It is possible that Ctx might antagonize Tx-induced relaxation and hyperpolarization reported previously (28 -30).
The effects of Tx on ion channels may be responsible for some therapeutic side effects. While Tx chemotherapy is generally well tolerated, side effects include vasomotor (facial flushing; Ref. 32), cardiac (Q-T interval prolongation; Ref. 17) and neurological (tremor and seizure; Ref. 17) symptoms. It is likely that the effect of Tx on ion channels (e.g. BK) plays at least some role in producing these effects. However, our findings also suggest that novel pharmacological targeting of the ␤1 subunit with xenoestrogens may be useful for treatment of diseases that have a smooth muscle component, such as hypertension, myocardial ischemia, asthma, impotence, and constipation. These findings give insight into BK channel structure and function, signaling roles, and pharmacological properties of the ␤1 subunit, the non-genomic regulation of BK channels by estrogen and xenoestrogens, as well as Tx-induced side effects.