Activation of NFATc3 Down-regulates the β1 Subunit of Large Conductance, Calcium-activated K+ Channels in Arterial Smooth Muscle and Contributes to Hypertension*

Large conductance, Ca2+-activated K+ (BK) channels modulate the excitability and contractile state of arterial smooth muscle. Recently, we demonstrated that during hypertension, expression of the accessory β1 subunit was decreased relative to the pore-forming α subunit of the BK channel. Reduced β1 subunit expression resulted in BK channels with impaired function due to lowered sensitivity to Ca2+. Here, we tested the hypothesis that activation of the calcineurin/NFATc3 signaling pathway down-regulates β1 expression during angiotensin II-induced hypertension. Consistent with this hypothesis, we found that in vivo administration of angiotensin II-activated calcineurin/NFATc3 signaling in arterial smooth muscle. During angiotensin II infusion, arterial smooth muscle BK channel function was decreased in wild type (WT) but not in NFATc3 null (NFATc3-/-) mice. Accordingly, β1 expression was decreased in WT but not in NFATc3-/- arteries. Angiotensin II-induced down-regulation of the β1 subunit required Ca2+ influx via L-type Ca2+ channels. However, in the absence of angiotensin II, moderate elevation of [Ca2+]i alone was not sufficient to activate NFAT transcriptional activity and, thus, decrease β1 subunit expression. Importantly, angiotensin II infusion increased systemic blood pressure to a lower extent in NFATc3-/- than in WT mice, indicating that this transcription factor is required for the development of severe hypertension during chronic angiotensin II signaling activation. We conclude that activation of calcineurin and NFATc3 during sustained angiotensin II signaling down-regulates the expression of the β1 subunit of the BK channel, which in turn contributes to arterial dysfunction and the development of hypertension.

BK 2 channels are abundantly expressed in smooth muscle cells lining the walls of small resistance arteries. In these cells BK channels are composed of pore-forming ␣ and accessory ␤1 subunits (1)(2)(3), the later of which appears to be uniquely expressed in smooth muscle. The ␤1 subunit modulates BK channel function by stabilizing the open conformation and increasing the intrinsic Ca 2ϩ sensitivity of the ␣ subunit (4,5). This enhances the activity of BK channels and strengthens the coupling between these channels and their physiological activators, "Ca 2ϩ sparks" (6).
The functional relevance of the ␤1 subunit of the BK channel was underscored in a recent series of studies, which demonstrated that in ␤1 null mice, arterial smooth muscle BK channels had lower Ca 2ϩ sensitivities than their wild type counterparts. Consequently, these BK channels were decoupled from Ca 2ϩ sparks, which resulted in depolarized arterial smooth muscle, increased vasoconstriction, and ultimately hypertension (3,7).
We recently examined whether decreased ␤1 expression contributed to arterial dysfunction during hypertension (8,9). We found that there was pronounced down-regulation of the ␤1, but not the ␣ subunit, of BK channels in arterial smooth muscle during hypertension (8,9). Accordingly, BK channels from hypertensive arterial myocytes were decoupled from Ca 2ϩ sparks. The ensuing loss of the hyperpolarizing influence of BK channels contributed to increased membrane depolarization, [Ca 2ϩ ] i , and arterial tone, three hallmarks of arterial dysfunction during hypertension (10,11). However, at present the molecular mechanisms underlying these changes in BK channel composition in arterial smooth muscle are unclear.
Our group demonstrated that chronic angiotensin II signaling decreases the expression of voltage-gated K ϩ channels (Kv2.1) in arterial myocytes through the activation of the transcription factor NFATc3 by the Ca 2ϩ -dependent phosphatase calcineurin (12). Here, we tested the hypothesis that angiotensin II selectively down-regulates the BK channel ␤1 subunit in arterial smooth muscle through the calcineurin/ NFATc3 pathway.
We found that in vivo administration of angiotensin II increased calcineurin/NFATc3-mediated transcriptional activity in arterial myocytes. This resulted in decreased expression of the ␤1, but not the ␣, subunit of the BK channel. The effect of angiotensin II on ␤1 expression was independent of changes in blood pressure. We also show that calcineurin/NFATc3 activity was not only necessary for ␤1 subunit down-regulation by angiotensin II but was sufficient in doing so. Importantly, our data indicate that NFATc3 is required for the development of severe hypertension during angiotensin II infusion. Collec-tively, our data suggest that activation of calcineurin and NFATc3 during sustained angiotensin II signaling down-regulates the expression of the ␤1 subunit of the BK channel, which contributes to arterial dysfunction and the development of hypertension.

EXPERIMENTAL PROCEDURES
In Vivo Angiotensin II Infusion and Blood Pressure Monitoring-Wild type and NFATc3 Ϫ/Ϫ mice (kindly provided by Dr. Laurie Glimcher) were handled in strict compliance with the guidelines of the University of Washington Institutional Animal Care and Use Committee. Mice were made hypertensive by subcutaneous administration of angiotensin II (800 ng/kg/min) by means of a mini-osmotic pump (Alzet, Durect Corp., Cupertino, CA). Blood pressure was monitored in conscious adult mice before and after angiotensin II treatment using a telemetry system (Data Science International). Briefly, mice were anesthetized with isoflurane. A ventral midline incision from the lower mandible to the sternum was made to isolate the left common carotid. Two lengths of 7-0 silk suture were thread beneath the vessel for retraction and ligation. The artery was permanently ligated at the level of the bifurcation between the interior and exterior carotid. The second suture was used to occlude blood flow temporally and allow insertion of the catheter. A 25-gauge syringe needle was used to make an incision in the artery through which the tip of the catheter was inserted. The catheter was advanced to the thoracic aorta and tied in place with the suture. A subcutaneous pocket was then made to place the transmitter body. After placement of the transmitter body the incision was closed with 5-0 suture. In experiments in which animals had signs of an incomplete Circle of Willis (e.g. NFATc3 Ϫ/Ϫ ), blood pressure was measured from the femoral artery.
Blood pressure was recorded continuously and stored in the hard drive of a PC running Dataquest software (Data Science International). Control blood pressure measurements began 7 days after surgery to allow animals to recover. After this recovery period, blood pressures were recorded for 3 days before the saline or angiotensin II pump was implanted. Blood pressure recordings continued for up to 7 days after pump implantation.
Isolation of Arterial Myocytes-Mice were euthanized with a lethal dose of sodium pentobarbital (100 mg/kg; intraperitoneal). Myocytes were isolated from cerebral and mesenteric arteries as described elsewhere (9). Briefly, arteries were cleaned from connective tissue and placed in digestion buffer containing 130 mM NaCl, 1 mM KCl, 0.2 mM CaCl 2 , 0.5 mM MgCl 2 , 0.33 mM NaH 2 PO 4 , 3 mM pyruvate, 25 mM HEPES, and 22 mM glucose (adjusted to pH 7.4 with NaOH). Arteries were incubated (8 min at 37°C) in digestion buffer containing papain (0.5 mg/ml; Worthington Biochemical) and dithiothreitol (1.0 mg/ml) followed by a second incubation (15 min at 37°C) in digestion buffer supplemented with collagenase type H (1 mg/ml; Sigma). After enzymatic treatment, tissues were washed three times with enzyme-free digestion buffer. Single cells were obtained by gentle mechanical trituration using a set of fire-polished pipettes with progressively smaller tip diameters. Cells were kept in ice-cold digestion buffer and used on the day of isolation.
Organ Culture-Mesenteric and cerebral arterial segments were cultured as described previously (12).
Electrophysiology and [Ca 2ϩ ] i Imaging-BK single channel activity was recorded from inside-out patches under symmetrical K ϩ (140 mM) conditions. The bath and pipette solutions contained 140 mM KCl, 1 mM EGTA, or 1 mM HEDTA, and 10 HEPES adjusted to pH 7.3. The MAXC software (C. Patton, Stanford University Pacific Grove, CA) was used to determine the amount of CaCl 2 needed to achieve the desire free Ca 2ϩ concentration in the bath solution. Single channel currents were digitized at 5 kHz using pCLAMP 8 software (Axon Instruments Inc.) The data were filtered at 1 kHz using a Bessel filter (8 pole). Clampfit (Axon Instruments Inc.) and Analysis of Single Channel Data (G. Droogmans, University of Leuven, Leuven, Belgium) software were use for single channel data analysis. BK channel number, conductance, and open probability (P o ) were determined from all point amplitude histograms. To estimate the number of BK channels per patch, the patches were held at ϩ80 mV in the presence of 10 M Ca 2ϩ , which maximizes the P o of this channels (5).
For [Ca 2ϩ ] i imaging experiments, cells were loaded with the acetoxymethyl form of the fluorescent Ca 2ϩ indicator fluo-5 (5 M) as described previously (12). Confocal images were acquired using Bio-Rad Radiance 2100 confocal system coupled to an inverted Nikon TE300 microscope equipped with a Nikon PlanApo (60ϫ, NA ϭ 1.4) oil-immersion lens. Images were acquired and analyzed with Lasersharp 4.0 (Bio-Rad) and ImageJ software, respectively. In these experiments background-subtracted fluorescence signals were calibrated using the F max equation (13) using K d and F max /F min (i.e. ratio of the fluo-5 fluorescence under saturating and Ca 2ϩ -free conditions) values of 400 nM and 180, respectively.
RNA Isolation and RT-PCR-Total RNA was isolated from mice cerebral and mesenteric arteries using the RNAeasy Micro kit (Qiagen, Valencia, CA) as instructed by the manufacturer. To identify the presence of BK channel ␣ and ␤1 subunit transcripts in mice arteries, we designed primers specific for the BK channel ␣ subunit (GenBank TM accession number NM_010610; sense nt 1204 -1227 and antisense nt 1457-1478, amplicon ϭ 275 bp) and ␤1 subunit (GenBank TM accession NM_031169.2; sense 604 -627 and antisense 780 -803, amplicon ϭ 200 bp). We used ␤-actin (GenBank TM accession V01217; sense nt 2384 -2404 and antisense nt 3071-3091, amplicon ϭ 496 bp) transcript levels as an internal control. The ␤-actin primers amplify a region between exons 4 and 6 such that genomic contamination within the RNA preparation is identified by the presence of a 708-bp band in addition to the 496-bp band. Reverse transcription (with random primers) and amplification was performed using the OneStep RT-PCR kit from Qiagen following the manufacturer's instructions. Amplicons were visualized using 2% agarose gel electrophoresis.
Calcineurin and NFAT Activity-To assess NFAT activity, we took advantage of genetically engineered mice (NFAT-luc) in which luciferase expression is transcriptionally regulated by NFAT (14). Arteries from saline-and angiotensin II-treated NFAT-luc mice were collected as described above. Luciferase transcript expression was evaluated by conventional RT-PCR (Photinus pyralis; firefly luciferase, GenBank TM accession M15077; sense nt 520; antisense nt 1009). Calcineurin activity in arterial smooth muscle from control and Ang II-treated mice (NFAT-luc) was assessed using a colorimetric calcineurin cellular activity assay kit (Calbiochem) following the manufacturer's instructions.
Chemicals and Statistics-All chemicals were from Sigma unless stated otherwise. Data are presented as the mean Ϯ S.E. Two-group comparisons were made using Student's t test, and a p value of less than 0.05 was considered significant. A Wilcoxon ranked test was performed for data sets that did not have a normal distribution. The asterisk (*) symbol used in the figures illustrates a statistically significant difference between groups.

In Vivo Administration of Angiotensin II Activates the Calcineurin/NFATc3 Signaling Pathway in Arterial Smooth
Muscle-We measured calcineurin and NFAT activity in mice infused with saline (control) and angiotensin II via subcutaneous osmotic mini-pump. For these experiments, we used a transgenic NFAT reporter mouse (NFAT-luc) in which luciferase expression is controlled by multiple NFAT binding sites (15). Thus, in these mice, luciferase transcript levels can be used as an indicator of NFAT activity.
Tissues (mesenteric and cerebral arteries) were then harvested from control and angiotensin II mice 5 days after the implantation of pumps. As shown in Fig. 1B, calcineurin activity was nearly 5-fold higher in angiotensin II-infused than in saline-infused control arteries (p Ͻ 0.05, n ϭ 5).
Next, we investigated if the observed increase in calcineurin activity induced by angiotensin II translated into increased NFAT-mediated transcriptional activity. We used RT-PCR to detect luciferase transcript expression in arteries from salineand angiotensin II-infused NFAT-luc mice (Fig. 1C). As a control, ␤-actin expression was also measured in these tissues. We found that luciferase, but not ␤-actin, transcript expression was higher in angiotensin II than in control arteries (n ϭ 5; p Ͻ 0.05; Fig. 1). Indeed, it is important to note that luciferase transcript levels could not be detected in any of the control tissues examined, indicating that basal NFAT transcriptional activity in arterial smooth muscle is low. These results corroborate our calcineurin activity data, suggesting that calcineurin function and NFAT transcriptional activity increase in arterial smooth muscle in response to sustained in vivo administration of angiotensin II.
Activation of Calcineurin Is Necessary for Decreased ␤1 Subunit Expression after Angiotensin II Infusion-We investigated if the increase in calcineurin/NFAT activity described above is linked to decreased expression of the ␤1 subunit of BK channels during angiotensin II infusion (8). To begin, we analyzed the promoter region of the mouse BK channel ␣ and ␤1 subunits genes to determine whether they contain putative NFAT binding sites (GGAAA). Our analysis revealed the presence of multiple putative NFAT binding sites in the promoter region of the mouse ␤1 (GenBank TM accession number NM_031169; Ϫ84 to Ϫ1202) but not ␣ subunit genes (GenBank TM accession number NM_010610; Ϫ84 to Ϫ1202).
On the basis of these findings, we examined if activation of the calcineurin/NFATc3 pathway during sustained angiotensin II administration was involved in selective down-regulation of the BK channel ␤1 subunit in arterial smooth muscle cells. First, we tested the hypothesis that calcineurin activity is necessary for down-regulation of BK channel function in arterial smooth muscle. In these experiments, mice were infused with angiotensin II and/or the calcineurin inhibitor FK506 (3 mg/kg twice a day) for 5 days. In control experiments calcineurin activity was measured in arteries isolated from mice infused with saline, angiotensin II, FK506, and angiotensin II ϩ FK506 ( Fig. 2A). We found that calcineurin activity was similar in arteries from control and FK506 mice (n ϭ 5, p Ͼ 0.05). As noted above, calcineurin activity was higher in angiotensin II than in control arteries (n ϭ 5, p Ͻ 0.05). Note, however, that angiotensin II failed to increase calcineurin activity in the presence of FK506 (n ϭ 5, p Ͼ 0.05). These data suggest that basal calcineurin activity is minimal in arterial smooth muscle and that infusion of FK506 prevents the activation of this phosphatase during in vivo angiotensin II administration.
If calcineurin activation is necessary for decreasing ␤1 subunit expression during sustained angiotensin II infusion, then FK506 should prevent such down-regulation from occurring. Because the ␤1 subunit enhances the Ca 2ϩ sensitivity of BK channels, we examined the open probability (P o ) of BK channels from control, angiotensin II, FK506, and angiotensin II ϩ FK506 myocytes at 1 and 10 M Ca 2ϩ . Single BK channel cur- rents were recorded using the inside-out configuration of the patch clamp technique at Ϫ40 mV in the presence of 1 and 10 M Ca 2ϩ in the cytosolic aspect of the channel. Fig. 2B shows representative traces of BK channels from control, angiotensin II, FK506, and angiotensin II ϩ FK506 myocytes.
As expected, the P o of BK channels in control cells increased as Ca 2ϩ was elevated from 1 to 10 M (Fig. 2, B-C). Consistent with our previous study (8), we found that the P o of BK channels in angiotensin II myocytes was lower than in control cells at each Ca 2ϩ concentration examined (n ϭ 9 patches, p Ͻ 0.05). In contrast, the P o values of BK channels from FK506 and control cells were similar, indicating that basal calcineurin activity does not influence ␤1 function in arterial myocytes (p Ͼ 0.05). Note, however, that the P o of BK channels from angio-tensin II ϩ FK506 myocytes at 1 and 10 M Ca 2ϩ was similar to that of control channels (p Ͼ 0.05; Fig.  2C). These data indicate that coinfusion of the calcineurin inhibitor FK506 prevents the decrease in BK channel Ca 2ϩ sensitivity associated with chronic angiotensin II administration.
In addition to increasing the intrinsic Ca 2ϩ sensitivity of the BK channel ␣ subunit, the ␤1 subunit also increases the open time of these channels. Thus, we examined the open times of BK channels in control, angiotensin II, FK506, and FK506 ϩ angiotensin II myocytes by constructing open time histograms at a membrane potential of ϩ40 mV with a free [Ca 2ϩ ] of 1 M. Fig. 2D shows representative single-channel recordings and open time histograms for all the experimental conditions examined. As reported previously (8), the open dwell times of BK channels in myocytes from angiotensin II-infused mice are decreased relative to those of BK channels from control cells (4.3 Ϯ 0.5 versus 12.0 Ϯ 1.8 ms, p Ͻ 0.05, n ϭ 9 patches each), which is consistent with decreased ␤1 subunit function in angiotensin II-exposed cells. Note, however, that control, FK506, and FK506 ϩ angiotensin II BK channels had similar open times. From these results we conclude that FK506-sensitive activation of calcineurin is required for decreased BK channel open times after angiotensin II infusion.
One potential mechanism by which FK506 prevented the decrease in BK channel function (i.e. shorter open times and lowered Ca 2ϩ -sensitivity) after angiotensin II infusion is by preventing down-regulation of the ␤1 subunit. To investigate this possibility, we measured ␣ and ␤1 transcript levels in control, angiotensin II, FK506, and angiotensin II ϩ FK506 arteries using RT-PCR (Fig. 3). We found that expression of ␣ transcript was similar in control, angiotensin II, FK506, and angiotensin II ϩ FK506 arteries. As reported previously (8), ␤1 transcript expression was about 60 Ϯ 5% lower in angiotensin II than in control arteries (n ϭ 6 mice; p Ͻ 0.05). Note, that the calcineurin inhibitor FK506 prevented angiotensin II infusionmediated down-regulation of the BK channel ␤1 subunit. In light of our electrophysiological data, these findings suggest that calcineurin activity is necessary for down-regulation of ␤1 subunit expression and, hence, decreased BK channel function during angiotensin II-induced hypertension.
NFATc3 Is Necessary for Decreased ␤1 Subunit Expression after Angiotensin II Infusion-Calcineurin dephosphorylates and, thus, activates the transcription factor NFATc3 in arterial myocytes (12,16). Because we have shown that calcineurin activity is necessary for ␤1 subunit down-regulation by angio-  tensin II, we hypothesized that NFATc3 may also be necessary. A testable prediction of this hypothesis is that angiotensin II infusion will not decrease ␤1 subunit function in arteries from NFATc3 null mice (NFATc3 Ϫ/Ϫ ). Accordingly, we implemented an experimental strategy similar to the one described above to assess the role of NFATc3 on ␤1 subunit function. We examined the apparent Ca 2ϩ sensitivity and open dwell times of BK channels in excised patches from myocytes isolated from wild type (WT) and NFATc3 Ϫ/Ϫ mice infused with saline or angiotensin II (see Fig. 4). We found that the Ca 2ϩ sensitivity and open dwell times of BK channels in control (i.e. salineinfused WT and NFATc3 Ϫ/Ϫ mice) were not different (panels A and B in Fig. 4; p Ͼ 0.05). This indicates that that basal NFATc3 activity is not necessary for normal BK channel function in arterial myocytes.
As expected from the data presented above and our previous study (8), we found that angiotensin II infusion decreased the Ca 2ϩ sensitivity and open dwell times of WT BK channels. However, note that the Ca 2ϩ sensitivity and open dwell times of BK channels of cells from NFATc3 Ϫ/Ϫ mice infused with angiotensin II were similar to those from control WT and NFATc3 Ϫ/Ϫ cells ( Fig. 4; p Ͼ 0.05). These data indicate that the calcineurin/ NFATc3 signaling pathway is necessary for angiotensin II-mediated decreases in BK channel Ca 2ϩ sensitivity and open times.
Next, we investigated if NFATc3 activity is involved with decreased BK channel ␤1 expression after angiotensin II infusion. RT-PCR was used to assess ␣ and ␤1 transcript expression in WT, NFATc3 Ϫ/Ϫ , WT ϩ angiotensin II, and NFATc3 Ϫ/Ϫ ϩ angiotensin II arteries (Fig. 5). Our data indicate that ␣ subunit transcript expression was similar under all experimental conditions. As expected, angiotensin II infusion decreased ␤1 transcripts in WT arteries but failed to do so in NFATc3 Ϫ/Ϫ arteries (Fig. 5). ␤1 subunit transcript expression was similar in control (i.e. no angiotensin II), WT, and NFATc3 Ϫ/Ϫ arteries (n ϭ 6 animals; p Ͼ 0.05). Taken together, our findings indicate that sustained angiotensin II infusion activates the calcineurin/NFATc3 signaling pathway, which results in down-regulation of ␤1 subunit and, thus, decreases BK channel function.
Angiotensin II-induced Down-regulation of the BK Channel ␤1 Subunit Is Independent of Changes in Blood Pressure-In the next series of experiments, we investigated if the effects of systemic angiotensin II administration on arterial smooth muscle BK channels were dependent on changes in arterial pressure. To examine this issue, we placed WT cerebral arterial segments in organ culture (12,17) for 48 h under control conditions (i.e. medium only) and in the presence of angiotensin II (100 nM). At the end of this incubation period, we dissociated cells from these arteries and examined the activity (P o ) of BK channels in inside-out patches at 1 and 10 M Ca 2ϩ at Ϫ40 mV as described above. Note that all experiments were performed in the absence of angiotensin.
As observed during in vivo administration of angiotensin II (see above), the P o of BK channels from control WT myocytes was higher than those from angiotensin II-treated WT cells at 1 and 10 M Ca 2ϩ (Fig. 6, A and B). This indicates angiotensin II signaling is sufficient to decrease 〉⌲ channel function in arterial myocytes. To determine whether, as with in vivo administration, angiotensin II decreases BK channel Ca 2ϩ sensitivity through the activation of calcineurin, we cultured WT arteries   We found that the P o of BK channels from control (n ϭ 5) and from cells in which calcineurin activity was inhibited, CsA (n ϭ 6) and angiotensin II ϩ CsA (n ϭ 8), were not different ( Fig. 6B; p Ͼ 0.05). This indicates that calcineurin activity is necessary for down-regulation of BK channel function during sustained angiotensin II signaling.

Transcriptional Control of BK Channel Function
We also examined the effects of sustained angiotensin II exposure on the activity of BK channels from NFATc3 Ϫ/Ϫ myo-cytes (Fig. 6A). We found that, as with in vivo administration of angiotensin II (see Figs. 4 and 5 above), the P o of BK channels in control WT and NFATc3 Ϫ/Ϫ myocytes was similar. Importantly, note that unlike channels from WT cells, sustained exposure to angiotensin II had no effect on the activity of BK channels from NFATc3 Ϫ/Ϫ myocytes (p Ͼ 0.05). Collectively, our data suggest that sustained angiotensin II signaling decreases the Ca 2ϩ sensitivity of BK channels in arterial myocytes via activation of the calcineurin/NFATc3 signaling pathway independent of changes in blood pressure.
To complement these experiments, we again used RT-PCR to determine BK channel ␣ and ␤1 subunit transcript expression in WT, NFATc3 Ϫ/Ϫ , WT ϩ CsA, WT ϩ angiotensin II, WT ϩ angiotensin II ϩ CsA, and NFATc3 Ϫ/Ϫ ϩ angiotensin II arteries (Fig. 7). As before, we found that ␣ subunit transcript levels were similar under all experiments conditions (n ϭ 5; p Ͼ 0.05). Consistent with the electrophysiological data described above, ␤1 transcript expression was lower in WT arteries cultured in the presence of angiotensin II than in arteries exposed to medium only. However, we found that ␤1 subunit transcript levels were similar in WT, NFATc3 Ϫ/Ϫ , WT ϩ CsA, WT ϩ angiotensin II ϩ CsA, and NFATc3 Ϫ/Ϫ ϩ angiotensin II arteries (n ϭ 5, p Ͼ 0.05). It is important to note that similar results were obtained in experiments in which the highly specific calcineurin autoinhibitory (100 M) peptide was used instead of CsA to inhibit calcineurin (data not shown). These data suggest that sustained angiotensin II signaling in arterial myocytes is sufficient to decrease ␤1 subunit transcript expression via calcineurin/NFAT signaling.
Ca 2ϩ Influx through L-type Ca 2ϩ Channels Is Necessary for Angiotensin II-mediated Down-regulation of the ␤1 Subunit-Angiotensin II increases intracellular Ca 2ϩ ([Ca 2ϩ ] i ) at least in part by increasing the activity of L-type Ca 2ϩ channels. Indeed, we (12) and others (16,18,19) have linked increased Ca 2ϩ influx via L-type Ca 2ϩ channels to calcineurin/NFAT activation. Thus, we tested if Ca 2ϩ entry via L-type Ca 2ϩ channels is necessary for decreased ␤1 subunit expression and BK channel function after angiotensin II exposure. To do this we compared the P o of BK channels (at 1 and 10 M Ca 2ϩ as above) from smooth muscle cells dissociated from WT arteries cultured in the presence of the benzothiazepine L-type Ca 2ϩ channel blocker diltiazem (10 M;) and angiotensin II ϩ diltiazem (Fig.  8, A and B). Our data indicate that after L-type Ca 2ϩ channel blockade, angiotensin II did not change the P o of BK channels in WT arteries in culture (n ϭ 6; p Ͻ 0.05).  We used RT-PCR to determine ␣ and ␤1 subunit transcript expression in WT (n ϭ 5), WT ϩ diltiazem (n ϭ 5), WT ϩ angiotensin II (n ϭ 5), and WT ϩ angiotensin II ϩ diltiazem (n ϭ 5; Fig. 8C). As above, we found that ␣ transcript levels were similar under all experimental conditions (p Ͼ 0.05). Again, as suggested by our single channel data presented above, ␤1 subunit transcript expression was lower in WT arteries exposed to angiotensin II than in control WT arteries and that diltiazem prevented a decrease in ␤1 transcript expression produced by angiotensin II. Indeed, ␤1 transcript expression was similar in WT, WT ϩ diltiazem, and WT ϩ angiotensin II ϩ diltiazem. From this, we conclude that Ca 2ϩ influx through L-type Ca 2ϩ channels is necessary for angiotensin II-induced decreases in ␤1 transcript expression and BK channel activity.

Elevated [Ca 2ϩ ] i Is Not Sufficient in Increasing Transcriptional NFAT Activity and Down-regulating ␤1
Subunit Expression-We investigated if increasing [Ca 2ϩ ] i can induce NFAT transcriptional activity and thereby decrease ␤1 subunit expression. To do this, we examined the relationship between [Ca 2ϩ ] i and NFAT activity (Fig. 9A). For these experiments, [Ca 2ϩ ] i was modified by increasing external Ca 2ϩ from 2 to 5 mM.
First, we determined the relationship between external Ca 2ϩ and [Ca 2ϩ ] i in arterial myocytes at these two external Ca 2ϩ concentrations. As expected, increasing external Ca 2ϩ from 2 to 5 mM raised [Ca 2ϩ ] i from about 190 to 225 nM, respectively (n ϭ 5; Fig. 9A). Note that application of 100 nM angiotensin II in the sustained presence of 2 mM external Ca 2ϩ elevated [Ca 2ϩ ] i from 190 to 230 nM. Thus, increasing external Ca 2ϩ from 2 to 5 mM increased [Ca 2ϩ ] i to a similar extent than angiotensin II (100 nM) with 2 mM Ca 2ϩ in arterial myocytes.
We used NFAT-luc arteries to determine the relationship between [Ca 2ϩ ] i and transcriptional NFAT activity. In these experiments luciferase transcript expression was measured in arteries cultured for 48 h in 2 and 5 mM external Ca 2ϩ (Fig. 9A). For comparison, we plotted in the same graph NFAT activity at 2 mM external Ca 2ϩ in the presence of angiotensin II (100 nM).
We found that, unlike what we observed in arteries exposed to 100 nM angiotensin II and 2 mM external Ca 2ϩ , NFAT activity (i.e. luciferase expression) did not change when external Ca 2ϩ was increased from 2 to 5 mM in the absence of angiotensin II (n ϭ 6, p Ͼ 0.05; Fig.  6A). Consistent with this, Fig. 9B shows that, unlike angiotensin II-treated smooth muscle, ␣ and ␤1 subunit transcript expression was not different at 2 and 5 mM external Ca 2ϩ concentration in the absence of angiotensin II (p Ͼ 0.05; Fig. 9B). Taken together, these data suggest that whereas angiotensin II-mediated increases in Ca 2ϩ entry through L-type Ca 2ϩ channels are necessary for NFAT activation and ␤1 subunit down-regulation, an increase in [Ca 2ϩ ] i by itself is not a sufficient stimulus to decrease expression of this BK channel subunit.  NFATc3 Is Required for Angiotensin II-induced Hypertension-Having established that NFATc3 is required for ␤1 subunit down-regulation in arterial smooth muscle during in vivo angiotensin II administration, we examined the effects of this vasoactive peptide on blood pressure in NFATc3 Ϫ/Ϫ and WT mice. Two recent studies have demonstrated that loss of BK channel ␤1 subunit causes hypertension (3,7). Thus, it is reasonable to hypothesize that angiotensin II infusion into NFATc3 Ϫ/Ϫ mice would either not increase blood pressure or increase it to a lesser extent than in WT mice (Fig. 10A).
We found that mean arterial pressure in control WT (114.2 Ϯ 5 mm Hg, n ϭ 4) and NFATc3 Ϫ/Ϫ mice (114.6 Ϯ 3 mm Hg, n ϭ 5) was similar (p Ͼ 0.05). Consistent with our hypothesis, angiotensin II infusion produced a change in mean arterial pressure (⌬MAP) in NFATc3 Ϫ/Ϫ mice of 18 Ϯ 4 mm Hg (n ϭ 5) versus 35 Ϯ 4 mm Hg in WT mice (n ϭ 5; p Ͻ 0.05) (Fig. 10B). Thus, angiotensin II infusion evoked an Ϸ2.0-fold lower increase in blood pressure in NFATc3 Ϫ/Ϫ than in WT mice. These data are in concordance with the biochemical, molecular, and functional data described above and suggest for the first time that NFATc3 plays a crucial role in the development of severe hypertension during chronic angiotensin II infusion.

DISCUSSION
In this study we provide the first demonstration that the calcineurin/NFATc3 signaling pathway regulates the expression of the ␤1 subunit of the BK channel in arterial smooth muscle and plays a critical role in the development of angiotensin IIinduced hypertension. Our data indicate that sustained angiotensin II signaling activates the Ca 2ϩ -sensitive phosphatase calcineurin and transcription factor NFATc3 in arterial myocytes. NFATc3 selectively down-regulates the ␤1 subunit but not the pore-forming ␣ subunit of the BK channel. Down-regulation of ␤1 expression reduces the activity of BK channels by decreasing the Ca 2ϩ sensitivity of these channels. Furthermore, we demonstrate that NFATc3 Ϫ/Ϫ mice undergo a significantly smaller (Ϸ2.0-fold) increase in blood pressure than WT mice after Ang II infusion. These data support the novel concept that activation of calcineurin/NFATc3 signaling down-regulates the ␤1 subunit of BK channels, hence decreasing the function of these channels, which contributes to vascular dysfunction and to the development of hypertension.
Previous studies provide insight into the mechanisms by which sustained angiotensin II signaling activates calcineurin and NFATc3 in cerebral arterial myocytes. In these cells angiotensin II elicits a small but sustained increase in global [Ca 2ϩ ] i (12). A generally accepted model of angiotensin II signaling proposes that stimulation of angiotensin type 1 receptors activate heterotrimeric G q proteins positively coupled to protein kinase C and IP 3 signaling. In addition, a recent paper (20) suggested that angiotensin II also stimulates calcineurin activity indirectly by down-regulation of endogenous calcineurin inhibitory peptides secondary to inhibition of Foxo transcription factors.
Activation of either protein kinase C or IP 3 pathways could potentially contribute to increased global [Ca 2ϩ ] i . Increased IP 3 levels promote Ca 2ϩ release via IP 3 -sensitive Ca 2ϩ channels in the sarcoplasmic reticulum of arterial myocytes. By activating protein kinase C, angiotensin II could also increase Ca 2ϩ influx via L-type Ca 2ϩ channels through multiple direct and indirect mechanisms. For example, phosphorylation of L-type Ca 2ϩ channels by protein kinase C (21) induces these channels to operate in a high open probability gating mode that create zones of persistent Ca 2ϩ influx (i.e."persistent Ca 2ϩ sparklets") (22,23). Protein kinase C could also increase Ca 2ϩ influx via L-type Ca 2ϩ channels indirectly by causing membrane depolarization due to decreased BK (8, 24 -26) and voltage-gated, Ca 2ϩ -independent (Kv) K ϩ currents (27,28). However, in arterial smooth muscle, global increases in [Ca 2ϩ ] i produced by angiotensin II were abrogated by inhibiting L-type Ca 2ϩ channels. Thus, in this tissue Ca 2ϩ influx through L-type Ca 2ϩ channels appears to be the predominant mechanism for increased [Ca 2ϩ ] i after angiotensin II exposure.
At present details of the Ca 2ϩ signal(s) responsible for the activation of calcineurin in arterial myocytes is unclear. Recent work (29) by our group in cardiac myocytes suggests that a sustained increase in global [Ca 2ϩ ] i is sufficient to activate calcineurin/NFATc3 signaling in these cells. Accordingly, we (12) and others (16) have demonstrated that Ca 2ϩ influx through L-type Ca 2ϩ channels increases global [Ca 2ϩ ] i and is required for activation of calcineurin/NFATc3 signaling in arterial myocytes during G q/11 -mediated signaling.
In contrast to cardiac myocytes (29,30), however, we found that elevation of [Ca 2ϩ ] i in arterial myocytes is not sufficient to increase NFAT transcriptional activity. Indeed, these findings are consistent with recent work (16,19,31) suggesting that isolated elevation of [Ca 2ϩ ] i is not sufficient in inducing nuclear accumulation of NFATc3 in arterial myocytes. These studies suggest that although G q/11 -coupled receptor agonists such as angiotensin II necessarily elevate [Ca 2ϩ ] i , thus promoting nuclear NFATc3 accumulation, concomitant inhibition of Crm-1/JNK-dependent NFATc3 export is also required. In combination with our data, these findings suggest that angiotensin II (and presumably other G q/11 -coupled receptor agonists) increases the nuclear accumulation of NFAT not only by promoting nuclear import but presumably by also decreasing nuclear export. Acting synergistically, these two mechanisms could allow for sufficient nuclear accumulation of NFAT to modulate the expression of multiple genes including the BK channel ␤1 subunit gene. A, representative pressure waveforms of conscious WT and NFATc3 Ϫ/Ϫ mice before and during Ang II administration. B, bar plot of the change in mean arterial pressure (⌬MAP) induced by Ang II infusion in WT and NFATc3 Ϫ/Ϫ mice.
As noted above, we found multiple NFAT binding elements in the promoter region of the mouse BK channel ␤1 but not in the ␣ subunit gene. This is consistent with our finding that NFAT activation decreases ␤1 but not ␣ subunit expression in mouse arterial smooth muscle. As with the mouse, angiotensin II increases calcineurin/NFAT activity (12,18) and selectively down-regulates the BK channel ␤1 subunit in rat arterial smooth muscle (8,9). Thus, it is intriguing to speculate that calcineurin/NFAT signaling is a conserved mechanism by which angiotensin II signaling modulates the function of BK channels in arterial myocytes via changes in ␤1 subunit expression.
Decreasing BK channel ␤1 subunit expression has important functional consequences. By decreasing the Ca 2ϩ sensitivity of the BK channels, down-regulation of ␤1 decouples these channels from activating Ca 2ϩ sparks (3,(7)(8)(9). Decreased BK channel activity depolarizes arterial myocytes. Membrane depolarization in turn increases Ca 2ϩ influx through L-type Ca 2ϩ channels and, hence, raises global [Ca 2ϩ ] i causing smooth muscle cell contraction and arterial constriction. Indeed, mice lacking ␤1 expression have BK channels, which are completely decoupled from Ca 2ϩ sparks (3,7), resulting in arterial hyperconstriction and systemic hypertension. Consistent with this, we found that down-regulation of the BK channel ␤1 subunit in arterial smooth muscle is an early event in the development of hypertension (8,9).
Our data suggest that basal NFAT transcriptional activity is low and insufficient to down-regulate BK channel ␤1 subunit expression in arterial myocytes. It is intriguing to speculate that this contributes at least in part to the relatively high expression of ␤1 subunit in these cells under control conditions. As noted above, high ␤1 subunit expression increases the Ca 2ϩ sensitivity of BK channels and maintains the coupling between these channels and Ca 2ϩ sparks tight.
An important observation in this study is that angiotensin II infusion evoked a significantly smaller increase in blood pressure in NFATc3 Ϫ/Ϫ than in WT mice. The work presented here in combination with recent work by our group (12,29,30) provide insight into the mechanisms underlying decreased blood pressure changes in NFATc3 Ϫ/Ϫ . Note that in addition to modulating BK channel ␤1 subunit expression, calcineurin/ NFATc3 signaling decreases the function of Kv currents in arterial smooth muscle by specifically decreasing Kv2.1 expression (12). Thus, it is intriguing to speculate that NFATc3 Ϫ/Ϫ mice undergo a smaller increase in blood pressure during angiotensin II infusion at least in part because arterial smooth muscle in these mice does not down-regulate ␤1 and Kv2.1 subunits during chronic angiotensin II signaling. The observation that loss of ␤1 subunit in arterial myocytes is sufficient to induce hypertension gives credence to this view (3,7). However, we cannot rule out the possibility that NFATc3 can modulate arterial tone by regulating the expression of other proteins.
Currently, the mechanisms underlying the relatively small increase in blood pressure after Ang II infusion in NFATc3 Ϫ/Ϫ mice are unclear but may include NFATc3-independent transcriptional and non-transcriptional changes in vascular and kidney (32) function. As described above, in arterial smooth muscle these may include activation of protein kinase C and IP 3 pathways. On the basis of these findings and the BK channel ␤1 subunit data presented here, we propose a model in which calcineurin/NFATc3 signaling constitutes a Ca 2ϩ -dependent signaling pathway that tunes arterial smooth muscle excitability and, hence, arterial blood pressure by regulating the expression of Kv and BK channels.