Dual effect of tamoxifen on arterial KCa channels does not depend on the presence of the beta1 subunit.

Tamoxifen has been reported to directly activate large conductance calcium-activated potassium (KCa) channels through the KCa beta1 subunit, suggesting a cardio-protective role of this compound. The present study using knock-out (KO) mice for the KCa channel beta1 subunit was aimed at understanding the molecular mechanisms of the effects of tamoxifen on arterial smooth muscle KCa channels. Single channel studies were conducted in excised patches from cerebral artery myocytes from both wild-type and KO animals. The present data demonstrated that tamoxifen can inhibit arterial KCa channels due to a major decrease in channel open probability (P(o)), a mechanism different from the reduction in single channel amplitude reported previously and also observed in the present work. A tamoxifen-induced decrease in P(o) was present in arterial KCa channels from both wild-type and beta1 KO animals. This inhibition was concentration-dependent and partially reversible with a half-maximal concentration constant IC(50) of 2.6 microm. The effect of tamoxifen was actually dual Single channel kinetic analysis showed that tamoxifen shortens both mean closed time and mean open time; the latter is probably due to an intermediate duration voltage-independent blocking mechanism. Thus, tamoxifen block would predominate when KCa channel P(o) is >0.1-0.2, limiting the maximum P(o), whereas a leftward shift in voltage or Ca(2+) activation curves can be observed for P(o) values lower than those values. This dual effect of tamoxifen appears to be independent of the beta1 subunit. The molecular specificity of tamoxifen, or eventually other xenoestrogen derivatives, for the KCa channel beta1 subunit is uncertain.

KCa 1 channels play a fundamental role in regulating and maintaining arterial tone (1). Extensive work on the molecular biology of these channels has revealed that KCa channels exist in most tissues as heterodimers formed by two different subunits termed ␣ and ␤. The ␣ subunit contains the pore-forming region, the voltage sensor, and the Ca 2ϩ binding site, whereas the ␤ subunit (first discovered in smooth muscle, Ref. 2) has regulatory functions (3,4). Previous works on expressed KCa channels have shown that the smooth muscle-specific ␤1 subunit subtype increases the apparent Ca 2ϩ sensitivity of the channel, slows the activation and deactivation kinetics, and increases the sensitivity of the agonist dehydrosoyasaponin-I (5-7). The KCa channel ␤1 subunit has also been shown to have a crucial role as a molecular tuner for vasoregulation. Indeed, animals with gene-targeted disruption of the ␤1 subunit showed an impaired KCa channel function that leads to an increased arterial tone associated with symptoms of hypertension (8,9). More recently, down-regulation of the KCa channel ␤1 subunit was shown to have a role in a model of acquired hypertension (10) as well as in a model of genetic hypertension (11). Conversely, a recently discovered polymorphism of the ␤1 subunit that promotes a gain of function of the KCa channel is associated with a low prevalence of diastolic hypertension in humans (12). Given the crucial role of the KCa channel ␤1 subunit in arterial smooth muscle, together with its highly smooth muscle-specific expression, the search for a specific KCa channel ␤1 subunit pharmacology has recently increased. Tamoxifen, a nonsteroidal triphenylethylene derivative, is predominantly known as a competitive antagonist at the estrogen receptor and is used as an effective prescribed drug in the treatment of hormone-responsive breast cancer (13). Tamoxifen and other xenoestrogenic and estrogenic compounds have also been reported as direct activators of KCa channels in the micromolar range, suggesting a potential cardio-protective role of these compounds. This nongenomic effect of estrogen-related compounds has been attributed to the presence of the ␤1 subunit of the KCa channel (14 -18). Tamoxifen, however, has also been shown to have inhibitory effects on several potassium conductances in the cardiovascular system, such as IK1, Ito, Isus, and HERG (19,20). In KCa channels from canine colonic smooth muscle cells, tamoxifen and other derivatives were also reported to decrease KCa macroscopic currents, mainly due to a reduction in single channel conductance (15,17). The present study shows that tamoxifen can activate or inhibit native arterial KCa channels, depending on the level of channel activity. This study also demonstrates, with the use of animals with a targeted deletion of the KCa channel ␤1 subunit gene, that the dual effect of tamoxifen on arterial KCa channels is not conferred by the ␤1 subunit in native cerebral artery myocytes.

MATERIALS AND METHODS
Cell Isolation-Murine arterial smooth muscle cells were isolated as previously described (8). Briefly, wild-type C57Bl/6 (Black 6) mice and ␤1 subunit KO mice were used in the experiments. Adult mice (25-35 g; 3-8 months old) of either sex were euthanized by peritoneal injection of pentobarbital solution (150 mg/kg). The genotype of the mice was confirmed by PCR analysis of DNA obtained from mouse tails or blood samples following euthanization of the animals. Cerebral arteries were carefully dissected on ice and then digested with papain (0.3 mg/ml papain and 1 mg/ml dithioerythritol for 8 min at 37°C) and collagenase (1 mg/ml collagenase type F and type H in a 70%:30% mixture, respectively, incubated for 7 min at 37°C). The digested tissue was triturated with a fire polished glass Pasteur pipette to yield single smooth muscle cells. Cells were kept on ice until use.
Single Channel Recordings-Single channel currents were recorded from excised membrane patches of isolated arterial myocytes in symmetrical solutions containing 140 mM KCl, 10 mM HEPES (pH 7.2), 1 mM Mg 2ϩ , 5 mM EGTA or HEDTA, and Ca 2ϩ free concentration of 10 M. For Ca 2ϩ activation curves, bath solutions contained 140 mM KCl, 1 mM EGTA (or HEDTA pCa Ͼ 6), 10 mM HEPES, and 5 mM Tris base. Free Ca 2ϩ concentration was adjusted with different amounts of CaCl 2 calculated with MaxChelator software (C. Patton, Stanford University). Final free Ca 2ϩ concentration was measured with a Ca 2ϩ -sensitive electrode calibrated according to the manufacturer's instructions (World Precision Instruments, Sarasota, FL). Single channel activity was recorded at steady potentials indicated in the text with an Axopatch 200B amplifier (Axon Instruments, Union City, CA). Currents were filtered at 2-5 kHz and digitized at 20 kHz.
Solution Exchange-Fast solution exchange was performed with a multibarreled application pipette positioned in front of the patch pipette and switched with a remote-controlled piezoelectric device. Subsecond exchange time is normally achieved using this system (SF-77B Perfusion Fast Step; Warner Instruments, Hamden, CT).
Chemicals-All chemicals were obtained from Sigma and Calbiochem. All experiments were conducted at room temperature (20 -22°C). Tamoxifen was dissolved in Me 2 SO. Control solutions contained this vehicle (1:10,000 Me 2 SO).
Data Analysis and Statistics-Single channel data analysis was performed with Clampfit 9.0 (Axon Instruments). Curve fitting and linear regression analyses were performed with Origin 6.1 software (Origin-Lab Corp., Northampton, MA). Concentration-response relationship for tamoxifen inhibition was fitted to a Hill equation of the following form: Inhibition ϭ Inhibition max /(1 ϩ (IC 50 /[Tamoxifen]) NHill ), where Inhibition is 1 Ϫ (P o tamoxifen /P o control ), P o is the open probability of the channel in control conditions or in the presence of tamoxifen as indicated, Inhibition max is the maximal inhibition obtained, IC 50 is the concentration of tamoxifen needed to obtain half-maximal inhibition, and [Tamoxifen] is the tamoxifen concentration used. Calcium activation curves were also fitted to a Hill equation of the following form: P o ϭ P o max /(1 ϩ (IC 50 /[Ca 2ϩ ]) NHill ), where P o is the open probability of the channel, P o max is the maximal P o , IC 50 is the concentration of Ca 2ϩ needed to obtain half-maximal P o , and [Ca 2ϩ ] is the Ca 2ϩ concentration used. N Hill is the Hill coefficient and represents the slope factor of the Hill equation. Voltage activation curves were fitted with a Boltzmann equation of following form: P o ϭ P o max /(1 ϩ exp(V Ϫ V1 ⁄2 )/dV), where P o and P o max have the meaning described above, V is the holding potential, V1 ⁄2 is the voltage for half-maximal activation, and dV represents the slope factor. Unless indicated, data are presented by mean Ϯ S.E. from n number of cells. Statistical analyses were made with paired and unpaired Student's t test, as appropriate. The threshold for statistical significance was set at p Ͻ 0.05. Statistical tests were run in Origin 6.1 and SigmaStat 2.03 (Systat Software Inc., Point Richmond, CA). Fig. 1 shows KCa single channel activity recorded in excised patches at 10 M free Ca 2ϩ concentration. Application of 5 M tamoxifen to the bath promoted two types of changes in KCa channels from WT animals that are illustrated in Fig. 1A: a) a pronounced decrease in channel activity where NP o changes from 0.75 to 0.3; and b) a moderate reduction in single channel amplitude (from 10 to 9.3 pA) that was previously reported (15,17). Corresponding all point histograms on the right illustrate both the change in activity and the small change in channel amplitude. These types of changes can also be observed in the absence of the ␤1 subunit (KO). Fig. 1B shows KCa channel activity from KO animals recorded, in this case, at ϩ40 mV. Tamoxifen (5 M) induced a decline in NP o (from 1.8 to 0.32) with a concomitant slight reduction in single channel amplitude (from 9.25 to 8.8 pA) that can be also observed in the corresponding all point histogram on the right. Fig. 1C summarizes the inhibitory effect of 5 M tamoxifen on arterial KCa channel activity from both WT and ␤1 KO animals. Tamoxifen promoted a significant decrease in NP o (Fig. 1C, left panel) with respect to control conditions in WT KCa channels (46 Ϯ 13% reduction, p Ͻ 0.05, n ϭ 5, paired Student's t test) and also in ␤1 KO KCa channels (49 Ϯ 8% reduction, p Ͻ 0.05, n ϭ 9, paired Student's t test). Tamoxifen also induced a small reduction in the absolute single channel amplitude (Fig. 1C, right panel) from 9.97 Ϯ 0.26 to 8.81 Ϯ 0.26 pA in WT (control versus tamoxifen, respectively, p Ͻ 0.05, n ϭ 5, paired Student's t test) and from 9.86 Ϯ 0.22 to 8.61 Ϯ 0.34 pA in KO (control versus tamoxifen, respectively, p Ͻ 0.05, n ϭ 6, paired Student's t test). These findings showed that, besides a small reduction in channel amplitude, tamoxifen inhibition of KCa channel activity is the main inhibitory mechanism and that this inhibition is independent of the presence of the ␤1 subunit.

Tamoxifen Inhibits Arterial Smooth Muscle KCa Channels in the Presence or Absence of the ␤1 Subunit-
Tamoxifen Inhibition of KCa Channels Is Concentration-dependent and Partially Reversible- Fig. 2A illustrates single channel activity at 10 M Ca 2ϩ and Ϫ40 mV holding potential from an inside-out patch containing three active channels from cerebral artery isolated myocytes from WT animals. KCa channel activity decreased with increasing concentrations of tamoxifen applied to the bath. P o goes from 0.89 in control to 0.87, 0.71, 0.56, 0.40, 0.20, and 0.17 with increasing tamoxifen concentrations of 0.1, 1, 3, 5, 10, and 20 M, respectively. Upon washout, the P o is partially restored from 0.17 at 20 M tamoxifen to 0.47 in washout. This reversibility was observed in three more independent experiments. Fig. 2B shows the concentration-dependent inhibition curve of WT KCa channels obtained from eight different patches at Ϫ40 mV and 10 M Ca 2ϩ . Patches included in this curve lasted long enough to be assayed at least for three different concentrations of tamoxifen. The experimental points were fitted to a Hill function (continuous line) with an N Hill of 1.47, suggesting more than one site for interaction between tamoxifen and the KCa channel protein. The half-maximal concentration constant (IC 50 ) obtained from the fit was 2.6 M. Interestingly, analysis of the inhibition of single channel amplitude in the traces presented in Fig. 2A also shows a concentration-dependent behavior. The decay in single channel amplitude can be well fitted by a Hill function (data not shown). The values obtained for N Hill and IC 50 were 1.26 and 2.5 M, respectively, very similar to the values obtained for the inhibition of the P o (1.47 and 2.6 M, respectively, Fig. 2B). However, the maximum inhibition can only reach 20% compared with the 76% obtained for the P o (Fig. 2B). These data indicate that tamoxifen inhibition of arterial KCa channels is largely due to a novel mechanism that is concentrationdependent and reversible and decreases the activity of the channel. In addition, a concentration-dependent small reduction of single channel amplitude can also be observed.
Tamoxifen Can Have a Dual Effect, Depending on the Level of KCa Channel Activity-To confront the present observations that tamoxifen reversibly inhibits KCa channels in a concentration-dependent manner with its well-documented activator effect on KCa channels (14,15,17), a series of experiments varying either voltage or free Ca 2ϩ concentration were performed using arterial myocytes from WT animals. is Ϫ39 Ϯ 2 mV, and the slope factor (dV) of the fitted curve is 23 Ϯ 2 mV. Under these conditions, tamoxifen did not significantly change the V1 ⁄2 (Ϫ38 Ϯ 4 mV, p Ͼ 0.7, unpaired Student's t test) or the slope factor of the curve (25 Ϯ 4 mV, p Ͼ 0.5, unpaired Student's t test), but it did significantly reduce the P o max of the WT KCa channels from 0.82 Ϯ 0.01 to 0.49 Ϯ 0.02 (control versus tamoxifen, respectively, p Ͻ 0.001, unpaired Student's t test). If external Ca 2ϩ was reduced to 100 nM, KCa channel activity decreased dramatically, and the number of events to determine the P o of the channel with accuracy at hyperpolarized voltages was very low. Nonetheless, under these conditions, the tamoxifen activator effect can be detected at a holding potential of ϩ80 mV, which is the maximum depolarized voltage tested without compromising the stability of the membrane patch. Under these conditions, the P o of the KCa channel increases from 0.038 Ϯ 0.01 to 0.11 Ϯ 0.3 (control versus tamoxifen, respectively, n ϭ 7, p Ͻ 0.05, paired Student's t test). The dual effect of tamoxifen was illustrated for comparison at ϩ80 mV in the paired experiment shown in Fig.  3A, in which KCa channel activity was inhibited by 5 M tamoxifen at 10 M Ca 2ϩ concentration (from 0.6 to 0.43, control versus tamoxifen, respectively; 10 M Ca 2ϩ ) and activated by 5 M tamoxifen at 100 nM Ca 2ϩ concentration (from 0.016 to 0.075, control versus tamoxifen, respectively; 100 nM Ca 2ϩ ). Single channel amplitude also diminished in the presence of tamoxifen (from 17.3 to 15.5 pA, control versus tamoxifen, respectively; 10 M Ca 2ϩ ). Similarly, at low Ca 2ϩ concentrations, single channel amplitude was also diminished by the presence of tamoxifen (from 20.5 to 17.3 pA, control versus tamoxifen, respectively; 100 nM Ca 2ϩ ).
The dual effect of tamoxifen on WT KCa channels was also studied in an experimental series involving inside-out patches held at a single membrane potential (ϩ40 mV) and tested at different intracellular Ca 2ϩ concentrations. Rapid solution exchange (see "Materials and Methods") was used to expose the cytosolic side of the channels to several Ca 2ϩ concentrations in the presence and absence of 5 M tamoxifen. Fig. 4A Fig. 4B shows the same plot in a log-log scale to expand the changes observed at low Ca 2ϩ and low P o values. These data indicated that tamoxifen reduces the maximum P o of the KCa channels, probably due to a channel block. In addition, tamoxifen also promoted a leftward shift in the Ca 2ϩ activation curve of the channels.

Single Channel Analysis Reveals That Tamoxifen Abbreviated Both Mean Closed Time ( c ) and Mean Open Time ( o ) in WT KCa
Channels-Kinetic analysis of the effect of tamoxifen on WT KCa channels was performed in a subset of five experiments from Fig. 3B in which only one active channel was present in the membrane patches. Patches were studied at a Ca 2ϩ concentration of 10 M at several holding potentials (from Ϫ80 to ϩ 80 mV). Fig. 5A (top panel) shows that in control conditions, single channel activity was evident at all the voltages studied. The addition of 5 M tamoxifen to the bath solution promoted a decrease in single channel amplitude together with a decrease in single channel activity at voltages more positive than Ϫ40 mV. At more negative potentials (i.e. Ϫ60 and Ϫ80 mV), single channel openings appear to be more frequent. This change in activity is summarized in Fig. 5B (top  right panel). Experimental points obtained in these five experiments were fitted to a Boltzmann function (solid lines). The plot shows that tamoxifen promotes a reduction in P o max (from 0.84 Ϯ 0.02 to 0.53 Ϯ 0.01, control versus tamoxifen, respectively, p Ͻ 0.001, unpaired Student's t test) but does not significantly change V1 ⁄2 (Ϫ40 Ϯ 2 versus Ϫ30.5 Ϯ 8 mV, control versus tamoxifen, respectively, p Ͼ 0.2, unpaired Student's t test). However, a decrease in the slope of the Boltzmann function could be detected with tamoxifen in this subset of experiments. Slope factor changed from 17 Ϯ 2 to 33 Ϯ 7 mV (control versus tamoxifen, respectively, p Ͻ 0.05, unpaired Student's t test). This change most likely reflects the combined block and activation mechanisms of tamoxifen. Single channel amplitude was also reduced by tamoxifen (Fig. 5B, top left panel). Slope conductance changed from 227 Ϯ 0.001 to 194 Ϯ 0.002 picosiemens (15% reduction, control versus tamoxifen, respectively, p Ͻ 0.001, unpaired Student's t test). Under control conditions, both c and o were found to vary with voltage. Fig. 5B (bottom  left panel) shows that c can be well described by the exponential function of the following form: c ϭ c (0)exp(ϪV/C) (solid lines), where c (0) represents the mean closed time constant at 0 mV, and C is a slope factor (in mV). Under control conditions, c decreased e-fold by 26 mV, whereas in the presence of tamoxifen, c decreased e-fold by 40 mV. Analysis of the kinetic changes promoted by tamoxifen in this experimental series revealed that tamoxifen-induced reduction in c becomes significant at Ϫ80 mV (from 29.6 Ϯ 5 to 11.6 Ϯ 3 ms, control versus tamoxifen, respectively, p Ͻ 0.05, unpaired Student's t test). Addition of tamoxifen almost abolished the voltage dependence of o that remained around 1.6 ms at every potential tested, ranging from a minimum of 1.2 ms to a maximum of 1.9 ms (Fig. 5B, bottom right panel). The linear regression after the addition of tamoxifen (solid line, r ϭ 0.8, p Ͻ 0.02) showed a 7-fold decrease in the slope from 0.015 to 0.002 ms/mV (p Ͻ 0.03, unpaired Student's t test) These data indicated that tamoxifen induces a voltage-independent block of the channel, together with a destabilization of long-lived closed states present at hyperpolarized voltages. The Dual Effect of Tamoxifen Is Also Present in ␤1 KO KCa Channels-The absence of the ␤1 subunit in arterial KCa channels does not prevent tamoxifen inhibition of KCa channels as demonstrated in Fig. 1. The question remains whether the activator effect of tamoxifen can also be present in this preparation. To further explore the effect of tamoxifen, in the absence of the ␤1 subunit, a series of nine concentration-response experiments were conducted with excised membrane patches from ␤1 KO arterial myocytes at a single Ca 2ϩ concentration (1 M) and a single voltage (ϩ40 mV). Under these conditions, individual variations in initial P o of the ␤1 KO KCa channel were observed. In four membrane patches with an initial P o of Ͼ0.3, application of increasing tamoxifen concentrations promoted a decrease in channel activity consistent with the initial observations made in Fig. 1 at a high Ca 2ϩ concentration (10 M). This tamoxifen concentration-dependent inhibition paralleled that observed in WT KCa channels. Fig. 6A shows that tamoxifen concentration-dependent inhibition of ␤1 KO KCa channels (black circles) closely matches the inhibition obtained with WT KCa channels (gray circles, same data points as in Fig. 2B). However, in the five remaining experiments in which the initial P o was Ͻ0.3, tamoxifen concentration response showed a biphasic pattern (black diamonds) with an initial increase in P o at low concentrations, followed by a decrease in activity that parallels the rest of the experimental points at higher tamoxifen concentrations. In three of those five experiments, the increase in activity peaked at 3 M tamoxifen, whereas in the remaining two experiments, a peak in activity is observed at 5 M tamoxifen. The overall average of these five remaining experiments did not show a significant change in P o ; however, individual experiments showed biphasic behavior with tamoxifen. These data suggest that the concentration-dependent inhibition of KCa channels is conserved, even in the absence of the ␤1 subunit, and that at low channel activity, tamoxifen can produce a biphasic pattern.
To further explore the observation about biphasic response to tamoxifen, an experimental series at nonsaturating cytosolic Ca 2ϩ (1 M) was performed at varying voltages and tamoxifen concentrations to expand the range at which tamoxifen-induced changes in channel activity can be detected. Original ␤1 KO KCa channel traces shown in Fig. 6C demonstrate that the dual effect of tamoxifen can also be detected in the absence of the ␤1 subunit. Channel activity recorded at ϩ20 mV show that 1 M tamoxifen can increase ␤1 KO KCa channel P o from 0.02 to 0.12 (control versus tamoxifen, respectively). Fig. 6B shows a family of voltage activation curves obtained from five different experiments at increasing tamoxifen concentrations. At a low tamoxifen concentration (1 M), it can be clearly appreciated that, even in the absence of the ␤1 subunit, tamoxifen promotes both a leftward shift of 14 mV (V1 ⁄2 Control ϭ 36 Ϯ 2 mV versus V1 ⁄2 tamoxifen ϭ 22 Ϯ 3 mV, p Ͻ 0.005) and a reduction of P o max (from 0.58 Ϯ 0.02 to 0.39 Ϯ 0.02, control versus tamoxifen, respectively, p Ͻ 0.001, unpaired Student's t test). A change in the slope factor of the Boltzmann curve was also observed (from 18 Ϯ 1 to 24 Ϯ 2 mV, control versus tamoxifen, respectively, p Ͻ 0.05, unpaired Student's t test). The observed change in the slope factor of the Boltzmann curves probably reflects a combined effect between block and activation. Consistent with the initial observations in Fig. 1, further increases in tamoxifen concentration markedly reduce ␤1 KO KCa channel activity, as can be seen in the voltage activation curves obtained with 5 and 10 M tamoxifen, in which P o max drops to 0.29 and 0.18, respectively, but a leftward shift can no longer be detected.

DISCUSSION
Tamoxifen Inhibition of KCa Channels-Despite recent reports suggesting the potential vasodilatory effect of tamoxifen and other xenoestrogen-related compounds acting through activation of KCa channels, the present findings revealed that tamoxifen instead can be an effective blocker of the KCa channels present in arterial smooth muscle. Tamoxifen inhibition of KCa channel was observed for other smooth muscle preparations and attributed to a reduction in single channel conductance produced by tamoxifen. In the present work, the change in unitary current observed in arterial KCa channels accounts only for a marginal reduction of the current (10 -15%). However, a substantial decrease in channel P o is observed in the presence of tamoxifen. The analysis of single channel P o provided here allows a close discrimination between changes in unitary conductance and changes in channel activity. Tamoxifen inhibition of arterial KCa channel P o is concentration-dependent and also partially reversible. KCa channel P o can drop to 50% of the control values at a tamoxifen concentration of ϳ3 M. This inhibition in channel activity is probably also responsible for the reduction in macroscopic KCa currents observed in other smooth muscle preparations (15,17). However, this pronounced decrease in macrospopic currents was attributed to a decrease in channel conductance rather than a change in P o . The evidence provided here suggests that tamoxifen inhibition of KCa channels can be better described by a decrease in channel P o probably due to an intermediate block that was entirely overlooked in the past.
The Mechanisms of the Effect of Tamoxifen on KCa Channels-The fact that tamoxifen concentration-dependent inhibition affinity (IC 50 ϭ 2.6 M) is very close to the reported affinity for dose-dependent activation (ϳ1 M, Ref. 14) emphasizes the dual nature of tamoxifen action in the same concentration range and the need for a very detailed analysis to account for the possible mechanisms. The dual nature of the effects of tamoxifen produces, under certain conditions, a clear crossover of the voltage activation curves or the Ca 2ϩ activation curves (Figs. 4A and 6B). Single channel kinetics revealed that tamoxifen operates through a reduction in the mean open time, but also a through reduction in the mean close time (Fig. 5B). A simple explanation is that tamoxifen can be well described as an intermediate channel blocker of KCa channels (as opposed to a fast or flickery block). However, single channel kinetics simulations (data not shown) have thus far excluded the possibility of tamoxifen behaving as an open channel blocker or a closed channel blocker, at least for very simplified models of three to five states (closed-open-block, closed-blocked-open, or closed 1 -closed 2 -closed 3 -open-blocked). Thus, it appears that the activator effect of tamoxifen cannot be simply described as a release from blocked states. In addition, the Hill coefficient obtained in the present study for the blocking reaction (N Hill ϭ 1.47 Ͼ 1) also suggests that tamoxifen inhibition cannot be taken as a simple bimolecular reaction. Tamoxifen appears to act by destabilizing long-lived closed states while, at the same time, producing a block state of intermediate duration not previously described. Both processes appear to have a very similar tamoxifen affinity, which further complicates the analysis. It is not completely clear at this point whether the intermediate blocking site is physically separated from the activating site in the KCa channel protein. Moreover, it is not clear whether tamoxifen inhibition takes place as a true physical occlusion of the permeation pathway in the mouth of the channel, or whether tamoxifen is acting by shifting the equilibrium between normal gating states of KCa channels. Nonetheless, competition experiments have shown in the past that tamoxifen can reduce the inhibition produced by known KCa channel pore blockers such as tetraethylammonium, iberiotoxin, and charybdotoxin (17), favoring the hypothesis that tamoxifen inhibition can take place within the mouth of the channel. The presence of several channels in the patch pipette in the present work precludes any detailed analysis, which, in combination with a dwell time of the intermediate block in the range of the pre-existing dwell times in normal gating of the channel, further complicates the discrimination between a pore occlusion of the channel or a modification in the gating process of the channel. Competition experiments in outside-out configuration patches, a configuration difficult to obtain under the present recording conditions, will be needed to determine whether tetraethylammonium or iberiotoxin can effectively reduce the tamoxifen inhibition in P o described here. This will help to distinguish between an occlusion mechanism or a gating modifier mechanism.
The Role of the ␤1 Subunit-Tamoxifen inhibition of KCa channels appears to reside only in the ␣ subunit. As stated above, this inhibition is mainly due to an intermediate duration block that results in a dramatic reduction of P o , which differs from the reduction in single channel conductance also observed here and elsewhere (15,17). Dick et al. (17) demonstrated that ethylbromide tamoxifen (a membrane-impermeant tamoxifen analog) can reduce single channel amplitude from both the extracellular side and (to a minor extent) the intracellular side. They also showed that ethylbromide tamoxifen can reduce KCa macroscopic currents in outside-out patches. Although the present work does not address which side applies to the tamoxifen effect, it is conceivable that the tamoxifen intermediate blocking site can only be reached from the extracellular side of the channel protein. Regardless of its side, this intermediate blocking site is most likely located only in the ␣ subunit, due to the fact that tamoxifen inhibition takes place in both WT and ␤1 KO KCa channels with very similar affinities and partial reversibility. The present study also demonstrates that the ␤1 subunit is not required to allow tamoxifen-mediated activation of arterial KCa channels. This observation suggests that tamoxifen binding sites are probably located in the ␣ subunit, opening the possibility that both activation and inhibition are physically related, if not the same, sites. It is worth noting, however, that in the absence of the ␤1 subunit, the activator effects of tamoxifen are less pronounced. The leftward shift shown in Fig. 6B for KO conditions is only 14 mV compared with the 24-mV shift reported for colonic smooth muscle under WT conditions (15), suggesting that perhaps the role of the ␤1 subunit is to stabilize or facilitate tamoxifen-bound states rather than being the actual forming subunit of the tamoxifen binding sites as previously suggested (14,15). Indeed, the receptor nature of the ␤1 subunit for different KCa channel regulators remains elusive. The receptor for charybdotoxin (or iberiotoxin) (21,22), Ca 2ϩ (23), dehydrosoyasaponin-I (7), or, in this case, tamoxifen (14) appears to be modulated directly or indirectly by the ␤1 subunit, rather than the ␤1 subunit being part of those receptors in the channel protein (5,24,25). The weaker effect for tamoxifen activation obtained in the absence of the ␤1 subunit, in combination with the inhibitory properties of tamoxifen, can probably explain the differences between the present work and the previously reported lack of effect in KCa channels also from ␤1 KO animals, although colonic smooth muscle cells were used in that study (15). Alternatively, another auxiliary ␤ subunit could have emerged in arterial smooth muscle as a compensatory mechanism in ␤1 KO animals. In this case, this compensatory ␤ subunit should also preserve similar pharmacological characteristics. To date, there is no information about the role of other KCa channel ␤ subunits in tamoxifen-mediated KCa channel activation. In either case, the exclusiveness of the ␤1 subunit in tamoxifen activation becomes questionable, as does the use of tamoxifen as a pharmacological probe for the presence of ␤1 subunits in KCa channels (10,11).
In conclusion, tamoxifen interaction with arterial KCa channels appears to be rather complex, promoting two types of blocks: a previously described fast flickery block, also present here, and an intermediate block, first characterized here, which accounts for a great deal of the inhibition observed in arterial KCa channels in the present work and probably in other studies elsewhere. These blocking effects co-exist with an activator effect through the destabilization of long-lived closed states. Together, these effects appear to reside exclusively on the ␣ subunit or, at minimum, do not appear to depend on the presence of the ␤1 subunit of KCa channels.