Xeno)estrogen Sensitivity of Smooth Muscle BK Channels Conferred by the Regulatory β1 Subunit A STUDY OF β1 KNOCKOUT MICE

Estrogen and xenoestrogens (i.e.agents that are not steroids but possess estrogenic activity) increase the open probability (Po) of large conductance Ca2+-activated K+ (BK) channels in smooth muscle. The mechanism of action may involve the regulatory β1 subunit. We used β1 subunit knockout (β1−/−) mice to test the hypothesis that the regulatory β1 subunit is essential for the activation of BK channels by tamoxifen, 4-OH tamoxifen (a major biologically active metabolite), and 17β-estradiol in native myocytes. Patch clamp recordings demonstrate BK channels from β1−/− mice were similar to wild type with the exception of markedly reduced Ca2+/voltage sensitivity and faster activation kinetics. In wild type myocytes, (xeno)estrogens increased NPo(Po × the number of channels, N), shifted the voltage of half-activation (V 1 2 ) to more negative potentials, and decreased unitary conductance. These effects were non-genomic and direct, because they were rapid, reversible, and observed in cell-free patches. None of the (xeno)estrogens increased the NPo of BK channels from β1−/− mice, but all three agents decreased single channel conductance. Thus, (xeno)estrogens increase BK NPothrough a mechanism involving the β1 subunit. The decrease in conductance did not require the β1 subunit and probably reflects an interaction with the pore-forming α subunit. We demonstrate regulation of smooth muscle BK channels by physiological (steroid hormones) and pharmacological (chemotherapeutic) agents and reveal the critical role of the β1 subunit in these responses in native myocytes.

BK channels are large conductance Ca 2ϩ -sensitive K ϩ channels (1,2). These channels are members of the voltage-gated K ϩ channel superfamily (3) and products of a nearly ubiquitous, alternatively spliced gene (Slo, (4)). Whereas the poreforming ␣ subunit has a wide tissue distribution, BK channel function varies greatly among cell types because of the addition of specific regulatory ␤ subunits (for example, see Ref. 5). In smooth muscle, biochemical purification and reconstitution have shown that ␣ subunits combine with regulatory ␤1 subunits in a 1:1 ratio (6 -8). This combination of ␣ and ␤1 subunits in smooth muscle is manifested functionally. Specifically, the ␤1 subunit imparts greater Ca 2ϩ /voltage sensitivity (9,10), making BK channels important modulators of smooth muscle excitability (11)(12)(13)(14). The ␤1 subunit also functions in responses to various pharmacological agents, including charybdotoxin (15) and dehydrosaponin I (9). Additionally, the ␤1 subunit may bind estrogens and xenoestrogens (i.e. agents that are not steroids but possess estrogenic activity). Valverde et al. (16) have shown that smooth muscle BK channels are activated by 17␤-estradiol and proposed that this action is because of binding on the ␤1 subunit. We have recently shown that tamoxifen (Tx) 1 also activates BK channels (17). We demonstrated in a heterologous expression system (human embryonic kidney 293 cells) that the presence of the ␤1 subunit is required for the effects of tamoxifen and 17␤-estradiol on recombinant BK channels (17). The importance of the ␤1 subunit in the response of native smooth muscle cells to (xeno)estrogens is unknown.
Recently, genetically engineered mice in which the ␤1 subunit has been deleted (␤1Ϫ/Ϫ) have been produced (14), and these animals provide the opportunity to assess the role of the ␤1 subunit in the actions of (xeno)estrogens in native myocytes. ␤1Ϫ/Ϫ mice are hypertensive and demonstrate altered vascular reactivity (13,14), illustrating the physiological importance of the ␤1 subunit. Pharmacological manipulation of the regulatory ␤1 subunit may become a novel approach to treating conditions such as hypertension, because the expression of the ␤1 subunit is thought to be limited to smooth muscle (18,19). In this study, we have sought to determine whether tamoxifen, 4-OH tamoxifen, and 17␤-estradiol activate BK channels in native smooth muscle cells. 4-OH tamoxifen is a very potent anti-estrogen and may be the biologically active product in some tissues (20). However, aside from binding the nuclear estrogen receptor, very little is known about its mechanism(s) of action, particularly non-genomic mechanisms. In this study, we have used ␤1Ϫ/Ϫ mice to test the hypothesis that the regulatory ␤1 subunit is essential for the response of BK channels to (xeno)estrogens, including tamoxifen and its P450 metabolite, 4-OH tamoxifen.

MATERIALS AND METHODS
Murine colonic myocytes were isolated by enzymatic dispersion as described previously (21). The genotype and phenotype of the ␤1Ϫ/Ϫ mice have been characterized previously (14) and corroborated in a similar model by others (13). Controls for this study were wild type C57BL ("Black 6") mice. This is the same strain from which the ␤1Ϫ/Ϫ mice were derived (14). Mice were anesthetized with chloroform and killed by cervical dislocation. The proximal colon was placed in Krebs solution and rinsed free of contents. Krebs contained (in mM) 125 NaCl, 5 After adhering to the glass-bottomed recording chamber, myocytes were suffused with a solution containing (in mM) 140 KCl, 10 HEPES, and 5 Tris, pH 7.1. Ca 2ϩ was added to the bath solution buffered with 1 mM EGTA or N-(2-hydroxyethyl)EDTA to achieve the desired free Ca 2ϩ concentration (MaxChelator, Pacific Grove, CA) (22). Myocytes were approached with heat-polished pipettes having tip resistances of 2-10 megohm when filled with bath solution. BK channel currents were recorded from inside-out patches in symmetrical 140 mM K ϩ or with an asymmetrical gradient (5 mM K ϩ in the pipette with a Na ϩ balance). An Axopatch 1D amplifier with a CV-4 headstage was used for data acquisition (Axon Instruments, Foster City, CA). Data were acquired in pCLAMP 5.5.1 (Clampex and Fetchex, Axon) by an IBM-compatible computer interfaced via a TL-1 analog-digital converter. The digitization rate was 4 times greater than the low pass cut-off frequency for filtration (1 kHz). Data were analyzed using pCLAMP 6 (Clampfit; Axon) and the Analysis of Single Channel Data program (University of Leuven, Leuven, Belgium). NP o and conductance were calculated from all-points amplitude histograms. Data are expressed as the means Ϯ S.E. of n cells. Statistical analyses were made with the Mann-Whitney rank sum test on medians or Student's t test on means as appropriate. The threshold for statistical significance was set as p Ͻ 0.05.

RESULTS
Currents were recorded from inside-out patches of membrane from murine colonic myocytes in symmetrical (140 mM) K ϩ . NP o increased when the patch potential was made more positive or when Ca 2ϩ concentration of the bath was increased (Fig. 1A). In myocytes from wild type mice, the slope conductance of BK channels in symmetrical K ϩ was 258 Ϯ 5 pS (Fig.  1B, n ϭ 6). The I-V relationship in asymmetrical K ϩ (5 mM pipette and 140 mM pipette) was outwardly rectifying, and conductance increased to 265 Ϯ 6 pS at ϩ80 mV (n ϭ 3). The slope conductance of BK channels from ␤1Ϫ/Ϫ mice was 259 Ϯ 6 pS in symmetrical K ϩ (Fig. 1B, n ϭ 9). For ␤1Ϫ/Ϫ mice, BK channel conductance in asymmetrical K ϩ increased to 267 Ϯ 13 pS at ϩ80 mV (n ϭ 3). These characteristics of BK channels in colonic myocytes from ␤1Ϫ/Ϫ mice were not significantly different from those of wild type mice. Furthermore, the characteristics of BK channels from ␤1Ϫ/Ϫ mice are similar to those attributed to the cloned mSlo ␣ subunit (4) and of cerebral arterial myocytes from same animals (14).
Presence of the regulatory ␤1 subunit in smooth muscle BK channels confers greater Ca 2ϩ /voltage-sensitivity and slows gating kinetics, therefore, we tested the Ca 2ϩ dependence of steady-state activation of BK channels from wild type and ␤1Ϫ/Ϫ mice as an index of the presence of the ␤1 subunit (23). Currents were measured from inside-out patches in symmetrical K ϩ with 100 nM free Ca 2ϩ in the bath. Patches were held at 0 mV and depolarized in 10 mV increments until conductance was maximal (Fig. 1C). Currents activated with a different time course in myocytes from control and ␤1 knockout mice. The time constant of activation at ϩ150 mV was 10 Ϯ 2 ms in control myocytes (n ϭ 6) and 4 Ϯ 1 ms in myocytes from ␤1Ϫ/Ϫ mice (n ϭ 10, p Ͻ 0.05). Conductance was normalized and plotted versus voltage (Fig. 1D). The V1 ⁄2 was 117 Ϯ 3 mV in patches from control myocytes. In contrast, V1 ⁄2 was significantly greater in patches from ␤1Ϫ/Ϫ myocytes (151 Ϯ 5 mV, p Ͻ 0.05). Thus, BK channels from the ␤1Ϫ/Ϫ mice demonstrated more rapid activation kinetics and reduced Ca 2ϩ /voltage-sensitivity.
Expression of canine Slo ␣ and ␤1 subunits in human embryonic kidney 293 cells has revealed that tamoxifen activates cloned BK channels only in the presence of the ␤1 subunit (17). We tested whether tamoxifen activated BK channels in colonic FIG. 1. Characteristics of BK channels in colon smooth muscle cells from wild type and ␤1؊/؊ mice. A, representative current traces demonstrate voltage-and Ca 2ϩ -dependent channel openings in a patch of membrane from a wild type colonic myocyte. These records are from an inside-out patch with 100 nM free Ca 2ϩ . The current at Ϫ80 mV was recorded in a bath with 1 mM Ca 2ϩ . B, single channel I-V relationships in symmetrical and asymmetrical K ϩ for wild type and ␤1Ϫ/Ϫ mouse myocytes. Conductance properties were not different between the two groups. C, example of current traces to determine steady-state activation of BK channels in wild type and ␤1Ϫ/Ϫ mouse myocytes (symmetrical K ϩ , 100 nM free Ca 2ϩ ). D, group data for steady-state activation of BK channels in wild type and ␤1Ϫ/Ϫ mouse myocytes. The V1 ⁄2 of activation was significantly more depolarized for BK channels from ␤1Ϫ/Ϫ mice. myocytes from wild type and ␤1Ϫ/Ϫ mice. Single channel records from inside-out patches demonstrated that 1 M tamoxifen increased NP o and decreased the conductance of BK channels from wild type myocytes ( Fig. 2A). NP o and conductance were averaged for at least 3 min (and up to 15 min) of recording before and after the addition of tamoxifen. Longer recordings assure that steady-state NP o is assessed accurately and eliminate contamination of the data from periodic fluctuations in NP o (24). Similar to previous results (16,17), 1 M tamoxifen activated BK channels in the patches from wild type mouse myocytes. The increase in NP o over control was 248 Ϯ 41% (n ϭ 9). The reduction in conductance averaged 12 Ϯ 3%. In contrast to wild type BK channels, 1 M tamoxifen did not activate BK channels in colonic myocytes from ␤1Ϫ/Ϫ mice (Fig. 2B, n ϭ  21). NP o in the presence of tamoxifen was within 2 Ϯ 1% of the control level. The NP o of BK channels of ␤1Ϫ/Ϫ mice was unaffected by higher concentrations of tamoxifen (e.g. Fig. 2B demonstrates the lack of effect of 10 M tamoxifen but was not averaged in the results with 1 M tamoxifen as shown in Fig. 2, C and D). Whereas 1 M tamoxifen did not change NP o significantly, it decreased the conductance of BK channels from ␤1Ϫ/Ϫ mice by 13 Ϯ 1% (not different from the reduction in wild type BK channel). Group data illustrating these effects of 1 M tamoxifen on NP o and conductance are shown in Fig. 2, C and D. The effects of tamoxifen were reversible, and control NP o was restored upon washout of the drug.
Tamoxifen shifted the voltage dependence of steady-state activation of BK channels to more negative potentials only in wild type mice (Fig. 3). Inside-out patches were studied in symmetrical 140 mM K ϩ and two Ca 2ϩ concentrations (either 100 nM or 10 M). 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 tamoxifen. Potential was made more positive until conductance was maximal, and then activation curves (i.e. normalized conductance versus voltage) were generated. In patches from control myocytes with 100 nM free Ca 2ϩ , the V1 ⁄2 was 118 Ϯ 3 mV and shifted to 94 Ϯ 2 mV by tamoxifen (n ϭ 9). With 10 M free Ca 2ϩ , the V1 ⁄2 was Ϫ52 Ϯ 3 mV and shifted to Ϫ66 Ϯ 3 mV by tamoxifen in control myocytes (n ϭ 4). Increasing the free Ca 2ϩ concentration, two orders of magnitude made V1 ⁄2 more negative, however, tamoxifen still caused a hyperpolarizing shift in V1 ⁄2 . Tamoxifen decreased conductance by 29 Ϯ 10 and 30 Ϯ 12% in 100 nM and 10 M free Ca 2ϩ , respectively (p Ͻ 0.05). In patches from ␤1Ϫ/Ϫ mice, tamoxifen did not change the voltage dependence of steady-state activation. With 100 nM free Ca 2ϩ , V1 ⁄2 was 152 Ϯ 3 mV under control conditions and 153 Ϯ 4 mV in the presence of tamoxifen (n ϭ 12). With 10 M free Ca 2ϩ in patches from ␤1Ϫ/Ϫ mice, V1 ⁄2 was 34 Ϯ 5 mV under control conditions and 31 Ϯ 4 mV in the presence of tamoxifen (n ϭ 8). However, tamoxifen did significantly decrease the conductance of channels from ␤1Ϫ/Ϫ mouse myocytes by 27 Ϯ 4 and 28 Ϯ 6% in 100 nM and 10 M free Ca 2ϩ , respectively (p Ͻ 0.05, not different from patches from wild type mice).
The effects of tamoxifen were rapid and reversible, suggesting a non-genomic mechanism. Furthermore, the mechanism seems direct, as tamoxifen increased NP o in excised patches, free from cellular signaling mechanisms. The time course of activation by tamoxifen was studied in inside-out patches. The potential of patches from wild type and ␤1Ϫ/Ϫ mouse myocytes was held constant, and the bath-contained 100 nM free Ca 2ϩ . NP o and conductance were measured over 8-s intervals and plotted versus time (Fig. 4). In patches from wild type mice, tamoxifen increased NP o and decreased conductance in a reversible manner (Fig. 4A). NP o reached its maximum (3.5 Ϯ 0.6-fold increase) in 83 Ϯ 9 s (n ϭ 3), and conductance simultaneously reached a minimum (12 Ϯ 3% reduction). In contrast, in patches from ␤1Ϫ/Ϫ mice, tamoxifen decreased conductance (13 Ϯ 3% reduction, n ϭ 3) without affecting NP o (Fig. 4B). The effect on conductance was reversible. Washout commenced within 1 min of tamoxifen removal or approximately the time required to exchange the bath solution. In wild type mice after 5 min of washout, NP o and conductance returned to 94 Ϯ 8% and 100 Ϯ 1% of control, respectively (n ϭ 3, p Ͼ 0.05).
In vivo metabolism of tamoxifen produces 4-OH tamoxifen, a very potent anti-estrogen, that may be the biologically active product in some tissues (20). However, aside from binding the nuclear estrogen receptor, very little is known about its mechanism(s) of action. To determine whether this hydroxylated P450 metabolite of tamoxifen interacts with the ␤1 subunit to increase BK channel NP o , currents were recorded before and after the addition of 1 M 4-OH tamoxifen. In myocytes from wild type mice, 4-OH tamoxifen (1 M) increased BK channel NP o (Fig. 5A) and decreased the single channel conductance 12 Ϯ 3% (n ϭ 9). In contrast, 4-OH tamoxifen had no significant effect on the NP o of BK channels from ␤1Ϫ/Ϫ mice (Fig. 5A). It did, however, decrease unitary conductance 13 Ϯ 4% (n ϭ 4). In vivo, most 4-OH tamoxifen is bound to proteins such as albumin, therefore, the amount of free 4-OH tamoxifen is reduced. To mimic this situation and determine whether binding to proteins might negate the effect of 4-OH tamoxifen, we added 20 mg/ml albumin to a solution containing 1 M 4-OH tamox- ifen and tested the effect on BK channel NP o . Albumin decreased the effectiveness of 4-OH tamoxifen, but NP o was still increased significantly (2.2 Ϯ 0.6-fold increase, n ϭ 4). 4-OH tamoxifen also shifted the activation curve of wild type BK channels to more negative potentials (Fig. 5B). In patches from wild type mice with 100 nM free Ca 2ϩ , the V1 ⁄2 was 109 Ϯ 3 mV and shifted to 93 Ϯ 5 mV by 1 M 4-OH tamoxifen (n ϭ 9). 4-OH tamoxifen decreased conductance by 30 Ϯ 5% (p Ͻ 0.05). As a comparison to our previous results with tamoxifen (17), a parallel set of experiments was performed on myocytes from the canine colon. We constructed a complete concentration-re- Reconstitution of human and canine Slo ␣ and ␤1 subunits in heterologous expression systems has indicated that 17␤-estradiol activates BK channels only in the presence of the ␤1 subunit (16,17). The regulatory ␤1 subunit may confer sensitivity to 17␤-estradiol upon recombinant smooth muscle BK channels, however, this has not been determined in native myocytes. Thus, it is unknown what molecular components of the BK channel are necessary for activation by 17␤-estradiol in native smooth muscle cells. We tested the effects of 17␤-estradiol on native smooth muscle cells from wild type and ␤1Ϫ/Ϫ mice. 17␤-estradiol (10 M) activated BK channels from wild type mouse colonic myocytes (Fig. 6A). BK conductance was slightly but significantly decreased by 10 M 17␤-estradiol (4.02 Ϯ 0.69%, p Ͻ 0.05, n ϭ 7, Fig. 5B). In contrast to the effects observed in wild type myocytes, 17␤-estradiol did not activate BK channels in colonic myocytes from ␤1Ϫ/Ϫ mice (Fig. 5A). Fig. 6B shows that although NP o was not changed by 17␤-estradiol, BK conductance was decreased 3.34 Ϯ 0.20% by 17␤-estradiol (p Ͻ 0.05, n ϭ 9). The activation of BK channels by 17␤-estradiol, but not the reduction in conductance, depends on the presence of the regulatory ␤1 subunit. DISCUSSION We have shown that (xeno)estrogens activate smooth muscle BK channels, important regulators of smooth muscle tone FIG. 3. Tx shifts the voltage dependence of steady-state activation in the presence of the ␤1 subunit. A, current traces from a representative inside-out patch taken from a wild type mouse (symmetrical 140 mM K ϩ , 10 M free Ca 2ϩ ). The patch was held at 0 mV and stepped from Ϫ120 to ϩ30 mV under control conditions. The addition of 1 M Tx decreased the current and shifted the V1 ⁄2 to more negative potentials. To measure steadystate activation in the presence of Tx, the patch was stepped from Ϫ140 to ϩ10 mV. B, representative currents from an insideout patch taken from a ␤1Ϫ/Ϫ mouse (conditions are the same as in A). Ca 2ϩ / voltage sensitivity was reduced, therefore, the patch was stepped from Ϫ40 mV to ϩ110 mV to measure activation. Tx (1 M) decreased current but did not shift the V1 ⁄2 . C, group data showing activation curves for wild type and ␤1Ϫ/Ϫ BK channels at 100 nM and 10 M free Ca 2ϩ . Tx shifted the activation curves of wild type but not ␤1Ϫ/Ϫ BK channels. Conditions were similar to those in A with the exception that the patch potential was ϩ80 mV. Application of 1 M Tx had no effect on NP o but decreased ␥. This effect was reversible upon washout.
FIG. 5. Activation of BK channels by 4-OH tamoxifen requires the regulatory ␤1 subunit. A, group data showing that 4-OH tamoxifen (1 M) increases the NP o of BK channels in myocytes from wild type (n ϭ 9) but not ␤1Ϫ/Ϫ (n ϭ 4) mice. Currents were recorded from inside-out patches of in symmetrical 140 mM K ϩ and 100 nM free Ca 2ϩ . B, the voltage dependence of steady-state activation is shifted to more negative potentials by 4-OH tamoxifen.
(11-14). These results in myocytes from genetically engineered mice confirm our previous observations in a gene expression system (i.e. cloned Slo ␣ and ␤ subunits in human embryonic kidney 293 cells) (17). However, wild type and ␤1Ϫ/Ϫ mice are much more valuable models, because they allow the opportunity to test the effects of (xeno)estrogens on blood pressure, vasomotion, gastrointestinal motility, and other smooth muscle activity in the presence or absence of the BK regulatory ␤1 subunit. BK channels of colonic smooth muscle cells from ␤1Ϫ/Ϫ mice were similar to those of wild type with the exception for markedly reduced Ca 2ϩ /voltage-sensitivity and faster activation kinetics. Tamoxifen, 4-OH tamoxifen, and 17␤estradiol activated BK channels and shifted the V1 ⁄2 of activation to more negative potentials only in myocytes from wild type mice. (Xeno)estrogens decreased the conductance of BK channels in myocytes from both wild type and ␤1Ϫ/Ϫ mice. Therefore, whereas the activation of BK channels by (xeno)estrogens depends on the presence of the regulatory ␤1 subunit, the reduction in conductance caused by these compounds does not. These data demonstrate that the regulatory ␤1 subunit is necessary for the stimulatory effects of (xeno)estrogens on native smooth muscle BK channels and extend our understanding of the regulation of BK channels.
Tamoxifen has effects on smooth muscle tissues consistent with the activation of BK channels. In myometrium, tamoxifen inhibits spontaneous and agonist-induced contractions (25). Tamoxifen also has vascular effects, as it relaxes myometrial arteries (26), increases uterine blood flow (26), and hyperpolarizes and relaxes cerebral arteries (27). High extracellular K ϩ , which limits the degree to which openings of BK channels can hyperpolarize smooth muscle, inhibits the relaxing effect of tamoxifen on myometrial smooth muscle (25). In clinical use, tamoxifen chemotherapy commonly causes facial flushing, reflecting a loss of vasomotor tone (28). All of these smooth muscle effects of tamoxifen can be explained in part by the activation of BK channels. However, tamoxifen also affects other ion channels. Tamoxifen inhibits volume-sensitive Cl Ϫ and delayed rectifier K ϩ currents in canine colonic smooth muscle (29). Tamoxifen also inhibits voltage-gated Na ϩ (30), L-type Ca 2ϩ (29,31), and non-selective cation channels (32,33). Tamoxifen effects on ion channels may be responsible for some therapeutic side effects (e.g. for example Q-T interval prolongation and neurological symptoms (34)) as well as smooth muscle relaxation in vitro (25)(26)(27). Most probably, the effects of tamoxifen in vitro and in vivo are multifaceted and difficult to identify specifically.
The mechanism of action of tamoxifen and 17␤-estradiol, determined on recombinant BK channels in heterologous expression systems, involves the regulatory ␤1 subunit (16,17). However, such a mechanism had not been demonstrated in native smooth muscle cells. A wide variety of smooth muscle cells express the ␤1 subunit (18,19), thus it has been difficult to test the importance of the ␤1 subunit in responses of native myocytes to (xeno)estrogens. Genetically engineered mice have provided the opportunity to test a ␤1 subunit-null smooth muscle cell type (13,14). We tested the hypothesis that the regulatory ␤1 subunit was essential for the activation of native smooth muscle BK channels by tamoxifen, 4-OH tamoxifen, and 17␤-estradiol. The data support the hypothesis and strengthen the idea that the regulatory ␤1 subunit functions as a membrane receptor for (xeno)estrogens that transduces the signal to effects on BK channels. The effects of (xeno)estrogens to increase BK channel NP o are non-genomic, as they were rapid, reversible, and occurred in cell-free patches that were removed from nuclear influences. These are well established characteristics of non-genomic mechanism of steroid hormone action (36). Whereas the ␤1 subunit of the BK channel is essential for conferring sensitivity to 17␤-estradiol, tamoxifen, and 4-OH tamoxifen, the exact mechanism of action has yet to be elucidated, but several possibilities can be excluded from our studies and include changes in Ca 2ϩ , nitric-oxide production, and changes in kinase activity.
Cellular signaling mechanisms are known to influence BK channel activity. Nitric-oxide activates BK channels directly (37) and indirectly via protein kinase G (38). Protein kinase A has also been shown to regulate the NP o of BK channels (39,40). In this study, excised patch recordings were performed in solutions with clamped Ca 2ϩ and in the nominal absence of substrates for cellular regulatory mechanisms that affect BK channel NP o (e.g. nitric-oxide production and phosphorylation via cGMP-and cAMP-dependent kinases). Therefore, (xeno) estrogens appear to directly activate BK channels. Our studies, however, do not rule out the possibility that these other mechanisms may contribute to the effects of (xeno)estrogens in intact myocytes. If some of these other cellular signaling mechanisms contribute to the increase in NP o in intact myocytes, then our experiments have underestimated the full effect of (xeno)estrogens to increase BK NP o , as the total cellular response would be a summation of direct and indirect effects. Another critical reason why phosphorylation by protein kinase G is an improbable mechanism underlying BK activation by tamoxifen, 4-OH tamoxifen, and 17␤-estradiol in the excised patches in our study is that a stimulatory protein kinase G phosphorylation site exists on the pore-forming ␣ subunit (41), but our study shows that (xeno)estrogens have no effects on NP o in ␤1Ϫ/Ϫ mice. Additionally, the effects of tamoxifen are also not attributed to periodic fluctuations of NP o or "wanderlust" kinetics (24), because the responses to tamoxifen are concentration-dependent (17) and, as we have shown here, completely reversible. Regardless of any other potential influence or signaling mechanism, the ␤1 subunit is crucial for the effects of tamoxifen, 4-OH tamoxifen, and 17␤-estradiol on BK NP o .
Tamoxifen, its 4-hydroxy metabolite, and 17␤-estradiol produce effects on BK channels directly at the cell membrane, probably by binding the regulatory ␤1 subunit itself, however, this is yet to be confirmed. A previous study (16) has indicated that a membrane-impermeant conjugate of 17␤-estradiol acts only from the extracellular surface, giving a sidedness to the putative binding site. We have obtained preliminary results FIG. 6. 17␤-Estradiol (17␤E) activates BK channels in mouse colonic myocytes only in the presence of the regulatory ␤1 subunit. A, group data showing that 17␤E (10 M) increases the NP o of BK channels in myocytes from wild type (n ϭ 7) but not ␤1Ϫ/Ϫ (n ϭ 9) mice. Currents were recorded from inside-out patches of membrane as in Fig.  2 (symmetrical 140 mM K ϩ , 100 nM free Ca 2ϩ ). NP o was measured before and 5 min after the application of 10 M 17␤E. Asterisk indicates a 17␤E-induced difference from control NP o , whereas the dagger indicates a significant difference in normalized NP o between wild type and ␤1Ϫ/Ϫ mice in the presence of 17␤E. B, group data demonstrating the reduction in ␥ by 17␤E (10 M, same cells as in A). Asterisks indicate a significant difference between control ␥ and that in the presence of 10 M 17␤E.
indicating that a membrane-impermeant form of tamoxifen, i.e. ethylbromide tamoxifen, activates BK channels only from the extracellular surface. 2 It is unknown whether the 17␤-estradiol and xenoestrogen binding sites are the same, however, pilot data from our lab indicates that they are the same as 17␣estradiol, an isomer without estrogenic properties, antagonizes the effects of both tamoxifen and 17␤-estradiol. 2 Furthermore, it is unknown whether the binding site is actually on the ␤1 subunit or on another unidentified effector molecule that acts through the ␤1 subunit. Neither this study, our previous work (17), nor the investigation of Valverde et al. (16) has determined the location of the binding site for (xeno)estrogens. Regardless, the effects of 17␤-estradiol, tamoxifen, and 4-OH tamoxifen to increase BK NP o represent a non-genomic mechanism of ion channel activation involving a regulatory subunit. Furthermore, our data suggest that wild type and ␤1Ϫ/Ϫ mice could be used to test the effects of (xeno)estrogens on smooth muscle function in vivo and in vitro in the presence or absence of the BK channel regulatory ␤1 subunit. Thus, the contribution of BK channels to the effects of tamoxifen on smooth muscle reported previously (25-28) might now be determined and separated from the effects on other ion channels (29 -32).