Polycystin-1 Activates the Calcineurin/NFAT (Nuclear Factor of Activated T-cells) Signaling Pathway*

Regulation of intracellular Ca2+ mobilization has been associated with the functions of polycystin-1 (PC1) and polycystin-2 (PC2), the protein products of the PKD1 and PKD2 genes. We have now demonstrated that PC1 can activate the calcineurin/NFAT (nuclear factor of activated T-cells) signaling pathway through Gαq -mediated activation of phospholipase C (PLC). Transient transfection of HEK293T cells with an NFAT promoter-luciferase reporter demonstrated that membrane-targeted PC1 constructs containing the membrane proximal region of the C-terminal tail, which includes the heterotrimeric G protein binding and activation domain, can stimulate NFAT luciferase activity. Inhibition of glycogen synthase kinase-3β by LiCl treatment further increased PC1-mediated NFAT activity. PC1-mediated activation of NFAT was completely inhibited by the calcineurin inhibitor, cyclosporin A. Cotransfection of a construct expressing the Gαq subunit augmented PC1-mediated NFAT activity, whereas the inhibitors of PLC (U73122) and the inositol trisphosphate and ryanodine receptors (xestospongin and 2-aminophenylborate) and a nonspecific Ca2+ channel blocker (gadolinium) diminished PC1-mediated NFAT activity. PC2 was not able to activate NFAT. An NFAT-green fluorescent protein nuclear localization assay demonstrated that PC1 constructs containing the C-tail only or the entire 11-transmembrane spanning region plus C-tail induced NFAT-green fluorescent protein nuclear translocation. NFAT expression was demonstrated in the M-1 mouse cortical collecting duct cell line and in embryonic and adult mouse kidneys by reverse transcriptase-PCR and immunolocalization. These data suggest a model in which PC1 signaling leads to a sustained elevation of intracellular Ca2+ mediated by PC1 activation of Gαq followed by PLC activation, release of Ca2+ from intracellular stores, and activation of store-operated Ca2+ entry, thus activating calcineurin and NFAT.

Autosomal dominant polycystic kidney disease (PKD) 1 is one of the most common genetic disorders worldwide, affecting 1 in 200 -1,000 individuals (1)(2)(3)(4)(5)(6)(7). The primary pathology in autosomal dominant polycystic kidney disease involves the formation and growth of numerous fluid-filled cysts in the kidney. Other manifestations include hepatic and pancreatic cysts, cardiac valve defects, intracranial and aortic aneurysms, inguinal hernia, and colonic diverticulae (8 -12). Approximately 85% of individuals with autosomal dominant polycystic kidney disease carry mutations in the PKD1 gene, with the remaining cases caused by mutations in the PKD2 gene.
The product of the PKD1 gene, polycystin-1 (PC1), is a plasma membrane protein consisting of a large extracellular N-terminal region followed by 11 membrane-spanning domains and a cytoplasmic C terminus of ϳ200 amino acids (6,13,14). PC1 is thought to be a signaling receptor involved in cell-cell or cell-matrix interactions (15-17); however, its large size and multimembrane-spanning structure would suggest that it has multiple functions. The PKD2 gene product, polycystin-2 (PC2), has been shown to be a Ca 2ϩ -sensitive nonspecific cation channel whose activity may be regulated by direct or indirect interaction with the C-terminal cytoplasmic tail of PC1 (5, 18 -22). Recently, PC1 and PC2 have been shown to have a role in ciliary mechanosensory Ca 2ϩ entry (23).
Numerous studies have suggested that defects in the regulation of epithelial cell growth are involved in the cyst-forming process, because epithelial cells that line cysts are both hyperproliferative and hyperapoptotic (1, 24 -26). As cysts expand, there is also a conversion of tubular epithelial cells from a normal phenotype involving net solute reabsorption to a cystic phenotype involving net solute secretion (27). The secretory phenotype may involve the function of apically localized chloride channels that are regulated by changes in Ca 2ϩ homeostasis (28 -30). Although Pkd1 and Pkd2 null mice develop polycystic kidneys and have cardiac maldevelopment, most die in utero of vascular failure (31)(32)(33)(34)(35). It has been shown recently that vascular smooth muscle cells from Pkd2 heterozygotes have impaired intracellular Ca 2ϩ regulation, which may underlie the vascular phenotype (36).
Studies employing direct pull-downs and transfected cells have shown that the PC1 C-tail binds and activates heterotrimeric G proteins (37), activates c-Jun N-terminal kinase and the AP-1 transcription factor in a G protein-dependent fashion (38 -40), activates protein kinase C (38), and modulates Wnt signaling by inhibiting glycogen synthase kinase-3␤ (GSK-3␤) and stabilizing ␤-catenin (41). PC1 overexpression has also been shown to activate the JAK2/STAT1 pathway leading to the up-regulation of p21(waf1) and potentially to cell cycle arrest (42).
Although it is currently thought that the polycystins are involved in regulating intracellular Ca 2ϩ levels (16, 18 -23), neither the ability of the polycystins to lead to sustained Ca 2ϩ increases nor the identity of downstream targets of these Ca 2ϩ signals has been demonstrated. Transient intracellular Ca 2ϩ increases are associated with cellular functions such as muscle contraction, synaptic transmission, or neuroendocrine secretion. In contrast, sustained Ca 2ϩ signals are known to affect transcriptional events leading to adaptive cellular changes and to changes in cell proliferation and cell differentiation (43)(44)(45)(46)(47). A cellular target for sustained increases in Ca 2ϩ is calcineurin, a ubiquitous serine-threonine phosphatase (48). An important intracellular substrate for calcineurin is NFAT (nuclear factor of activated T-cells) (49). In its inactive hyperphosphorylated form, NFAT is sequestered in the cytosol. Signals causing sustained Ca 2ϩ increases result in calcineurin activation, dephosphorylation of NFAT, and translocation of NFAT to the nucleus where it regulates target genes, often at composite NFAT/AP-1 elements (49,50). The termination of NFAT signaling occurs through rephosphorylation of NFAT by GSK-3␤ (51,52), resulting in its return to the cytoplasm. The immunosuppressive drugs, cyclosporin A (CSA) and FK506, inhibit calcineurin and thus nuclear translocation of NFAT (50,53). Five NFAT isoforms have been isolated, of which four (NFAT c1-c4) are calcineurin-sensitive (49,54). Activation of calcineurin/NFAT signaling has been shown to regulate cell differentiation, apoptosis, and cellular adaptation in a wide variety of cell types and tissues (55) and to regulate the noncanonical Wnt/Ca 2ϩ pathway during embryonic development (56). In attempting to identify a connection between polycystin function and Ca 2ϩ signaling, we determined that PC1 is able to activate a pathway that leads to calcineurin activation and translocation of dephosphorylated NFAT to the nucleus. The activation of NFAT by PC1 suggests that polycystin-1 may function to integrate Ca 2ϩ signals at NFAT target genes.

EXPERIMENTAL PROCEDURES
Cell Lines and DNA Constructs-HEK293T cells (ATCC) were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 4.5% glucose and L-glutamine supplemented with 10% heat-inactivated fetal bovine serum (Hyclone) and penicillin/streptomycin. M-1 mouse cortical collecting duct cells (57) were maintained in DMEM/F12 medium supplemented with 5% fetal bovine serum. Mouse PC1 C-tail constructs were subcloned downstream of sIg (CD5 signal peptide, CH2-CH3 domains of human IgG, and CD7 transmembrane domain) in pcDNA1.1 and pcDNA3.1 (Invitrogen) as described earlier (37,39). As shown in Fig. 1, the C-tail constructs included the C-terminal 222 aa of PC1 (PC1-LT), the C-terminal 193 aa of PC1 (PC1-HT), the C-terminal 120 aa of PC1 (PC1-AT), which lacks the heterotrimeric G protein activation domain but contains the coiled-coil region; and the membrane proximal 111-aa C-tail region of PC1 (PC1-LS), which contains the heterotrimeric G protein activation domain but lacks the coiled-coil region. A control construct (sIg-0) consisted of sIg only. The PC1-11TM construct ( Fig. 1) was constructed from the 3Ј region of a mouse PC1 cDNA (13,39,58) containing all of the 11 transmembrane domains and the C-tail (aa 3010 -4293). This region was subcloned downstream of the CD5 signal sequence and the CH2-CH3 IgG domains in pcDNA1.1 and pcDNA3.1 (13,39). As a control, a stop codon was introduced at aa 3092, just past the first transmembrane domain (sIg-stop). HA-tagged and Myc-tagged PC2 constructs (59) were obtained from Dr. L. Tsiokas (University of Oklahoma Health Sciences Center). As a control for HA-PC2, a stop codon was introduced, giving rise to a construct encoding aa 1-379 (PC2-stop). EE-tagged G␣ q in pcDNA3.1 was obtained from Guthrie cDNA Resource Center (39). Cis-acting 4ϫ pNFAT-luciferase and pBlueScript were from Stratagene. The pNFAT-luciferase construct has a TATA box and four 30-bp repeats containing the composite ARRE-2 site from the human interleukin-2 promoter sequence: 5-GGAGGAAA-AACTGTTTCATACAGAAGGCGT-3Ј. The corresponding NFAT (GGA-AAA) and AP-1 (TGTTTCA) elements are underlined (60). The pRL-null construct was from Promega.
NFAT Luciferase Assay-HEK293T cells were plated at a density of ϳ7.5 ϫ 10 5 cells/well of a six-well plastic plate in DMEM plus 10% heat-inactivated fetal bovine serum. 24 h after plating, the cells were transiently transfected with the Ca 2ϩ phosphate precipitation method (61). A total of 3 g of plasmid DNA was used to transfect each well, which contained 50 -500 ng of PC1 DNA or control construct, 100 ng of the ARRE-2 NFAT/AP-1 promoter-reporter construct (firefly luciferase), 1.5-5 ng of Renilla luciferase-null, pBlueScript as filler DNA, and, where indicated, 50 ng of EE-tagged G␣ q . 6 h post-transfection, the medium was replaced with serum-free DMEM and the cultures were incubated for an additional 20 h. Inhibitors were added 4 h before harvesting. The cells were then lysed in Passive lysis buffer (Promega), and 20 l of cell lysate was used with the dual luciferase assay kit (Promega) using an EG&G Berthold 9507 Luminometer. Data were analyzed by one-way ANOVA using GraphPad software (GraphPad Software, Inc., San Diego, CA).
Western Blot Analysis-HEK293T cell lysates were boiled in the presence of 2ϫ sample buffer, fractionated by SDS-PAGE, and transferred to Immobilon-P membranes (Millipore). Western blotting was performed using anti-HA antibody (Roche Applied Science) and anti-PC1 C-terminal peptide (antibody A19) (62). Anti-human IgG Western blots were performed as described earlier (13,39,62). Secondary antibodies conjugated to alkaline phosphatase were used to detect the immobilized antibodies by chemiluminescence with CDP-Star substrate (Amersham Biosciences) according to the manufacturer's instructions.
Measurement of Intracellular Ca 2ϩ -HEK293T cells were plated at ϳ5 ϫ 10 4 cells/well/six-well plate on type I collagen-coated glass coverslips 24 h prior to co-transfection with 450 ng of PC1-LT or sIg-0 and 22.5 ng of a cytomegalovirus-enhanced GFP construct (per 3 wells) (Invitrogen) to identify transfected cells plus 9 g of pBlueScript DNA. 4 h post-transfection, the medium was replaced with DMEM plus 0.5% heat-inactivated serum and the cultures were incubated for an additional 20 h. Cells were loaded with 5 M Fura-2/AM (Teflabs) in DMEM/ F12 equilibrated with 5% CO 2 , 95% air for 30 min at 37°C. The cells were rinsed with a HCO 3 -Ringer's solution containing 2 mM CaCl 2 , and the coverslips were mounted in a thermally controlled chamber on the stage of a Nikon inverted microscope equipped with a monochromator. The chamber was continuously perfused with Ringer's solution equilibrated with 5% CO 2 , 95% air at 37°C. Transfected cells were identified by viewing GFP fluorescence with a fluorescein isothiocyanate (FITC) filter set (490-nm excitation and 535-nm emission). Base-line Fura-2 measurements were made with dual excitation wavelengths of 340 and 380 nm. The measurement of emitted light at 510 nm was restricted to GFP-expressing cells using an adjustable iris in front of a digital photomultiplier detection system (Photon Technology International). Felix 32 analysis software (Photon Technology International) controlled the monochromator and data acquisition to generate the 340/380 fluorescence ratio (F 340 /F 380 ). After a steady state (F 340 /F 380 ) was established, 10 mM caffeine was added to release Ca 2ϩ from ryanodine-sensitive stores. At the end of each experiment, cells were permeabilized with 2 M ionomycin in Ringer's solution containing 2 mM Ca 2ϩ to determine the maximum (F 340 /F 380 ) ratio (R max ) and then 10 mM EGTA was added to determine the minimum ratio (R min ). Background correction for the glass coverslips and GFP fluorescence was measured in several cells expressing GFP at the same intensity and was subtracted from the experimental values. F 340 /F 380 ratios were converted to [Ca 2ϩ ] using the equation [Ca 2ϩ ] ϭ K d ϫ ((R Ϫ R min )/R max Ϫ R)) ϫ (S f380 /S b380 ), where the dissociation constant (K d ) of Fura-2 for Ca 2ϩ is 224 nM, R max and R min are F 340 /F 380 ratios for Ca 2ϩ -saturating and Ca 2ϩ -free conditions, and S f380 and S b380 are fluorescence signals at 380 nm for free Ca 2ϩ and bound Ca 2ϩ , respectively (63). A significant difference in intracellular [Ca 2ϩ ] between cells transfected with sIg-0 and PC1-LT was determined using a parametric Student's unpaired t test. Values are represented as the mean Ϯ S.E.
NFAT Nuclear Translocation Assay-HEK293T cells were seeded in Lab-TekII chamber slides (Nunc) and were co-transfected with 100 ng of an HA-NFATc1-GFP expression vector (64) and 200 ng of PC1-LT or PC1-11TM or their respective controls. 20 h post-transfection, the cells were washed three times with PBS and fixed for 10 min with freshly prepared 4% paraformaldehyde at room temperature. The cells were washed with three changes of PBS for 5 min each and counterstained with DAPI for an additional 5 min. The slides were rewashed with PBS, air dried, mounted with antifade (Molecular Probes), and were exam-ined with a Nikon fluorescence microscope equipped with a Spot 32 camera. Where indicated, green (HA-NFATc1-GFP) and blue (DAPI) images were merged to distinguish cytosolic and nuclear NFAT. The phosphorylation states of cytosolic and nuclear HA-NFATc1-GFP were determined by Western blotting using an anti-HA antibody.
Immunofluorescence and Confocal Microscopy-To demonstrate cell surface expression, ϳ3 ϫ 10 5 cells/chamber were plated and grown for 24 h. 24 h after transfection, the cells were washed with PBS, fixed with 2% paraformaldehyde at 4°C for 30 min, and washed with PBS three times on ice. The cells were then incubated with primary antibody (1:50 anti-human IgG-FITC) at 4°C for 1 h, washed with PBS three times, and coverslipped with VectaShield. The cells were analyzed at ϫ40 magnification. As controls (data not shown), cells were permeabilized in 0.1% Triton X-100 and 2% paraformaldehyde prior to incubation with antibody. For the co-expression studies, 1:50 dilutions of anti-human IgG-Texas Red (for PC1-LT) and anti-HA-FITC (for PC2) were used.
To localize endogenous NFAT in kidney epithelial cells, M-1 mouse cortical collecting duct cells were seeded in chamber slides. 2 h before processing, the cells were treated with the Ca 2ϩ ionophore, A23187, or with CSA. The cells were then washed with PBS and fixed with freshly prepared 4% paraformaldehyde for 15 min at room temperature. The cells were washed again with PBS and were permeabilized with freshly prepared 0.2% Triton X-100 in PBS for 10 min. The cells were washed three times with PBS, blocked with 1% bovine serum albumin in PBS for 10 min, and washed with PBS again. The cells were incubated with polyclonal anti-NFATc1 (Santa Cruz Biotechnology) for 2 h in the dark at room temperature. Unbound primary antibody was removed with 4ϫ changes of PBS of 5 min each, and the preparations were incubated with fluorescein-labeled goat anti-rabbit secondary antibody for 1 h, washed with PBS, and counterstained with DAPI. Controls included NFATc1 antibody:blocking peptide (1:5 vol) or secondary antibody only. Controls were negative (data not shown). The slides were visualized with a Zeiss LSM510 confocal microscope.
Immunohistochemistry-Kidneys from late gestation embryos or 3 month-old BALB/c mice were fixed in 4% paraformaldehyde, embedded in paraffin, and prepared for immunohistochemistry with the horseradish peroxidase-labeled streptavidin-biotin method and polyclonal anti-NFATc1. Sections were developed using aminoethyl carbazole substrate and were counterstained with hematoxylin. Controls, which were negative, consisted of either normal rabbit IgG or peptide competition (data not shown).

RESULTS
NFAT Activation Requires the G Protein Binding Region of the PC1 C-tail-To identify potential signaling pathways from polycystin-1 that can modulate intracellular Ca 2ϩ homeostasis, we co-expressed various PC1 C-tail deletion constructs (see Fig. 1) with an NFAT promoter-luciferase reporter construct that has four composite NFAT/AP-1 binding sites from the human interleukin-2 promoter (49). As shown in Fig. 2A, the full-length PC1 C-tail construct (PC1-LT) produced significant activation of the NFAT reporter over the control construct lacking PC1 sequences (sIg-0). Removal of the membrane proximal region, including a portion of the heterotrimeric G protein binding domain (PC1-HT) (37), resulted in only weak activation of the NFAT reporter (ϳ2.5fold over the control), which was considerably less than that seen with PC1-LT. A C-tail fusion protein that completely lacks the G protein binding and activation region (PC1-AT) was not able to activate this promoter. In contrast, a construct that contains the G protein activation region but lacks the coiled-coil domain (PC1-LS) was able to activate NFAT, albeit less so than the full-length C-tail ( Fig. 2A). The expression levels of the various PC1 fusion constructs were monitored by Western blot analysis using anti-human IgG ( Fig.  2A, bottom). These results suggest that PC1-induced NFAT activation is dependent on an intact G protein binding and activation region and does not require the coiled-coil region.
To further test the involvement of G proteins in PC1-mediated NFAT activation, HEK293T cells were co-transfected with a G␣ q expression construct, PC1-LT, and the NFAT reporter (Fig. 2B). Co-expression of the G␣ q expression vector with the PC1 cDNA resulted in further stimulation of PC1-LT-induced NFAT activation. This augmentation of NFAT activity was not observed when G␣ q was co-expressed with the sIg-0 control DNA or when expressed alone. Co-expression of G␣ 12 , but not FIG. 1. PC1 fusion constructs. A, C-tail constructs. All of the constructs have a CD5 signal sequence at the N terminus (not shown) followed by the CH2-CH3 domains of human IgG, hence sIg, and a CD7 transmembrane domain. The control construct, sIg-0, lacks PC1 sequences. PC1-LT encodes the C-terminal 222 amino acids (aa 4072-4293) of mouse PC1 starting with a leucine and ending with a threonine. The N-terminal region of the PC1 sequence in PC1-LT contains the 11th transmembrane domain (aa 4075-4095) (13). The deletion constructs PC1-HT, PC1-AT, and PC1-LS, respectively, contain a portion of the G protein binding region (see Ref. 37) and the coiled-coil region, just the coiled-coil region, or just the G protein binding region (HT, AT, and LS refer to the first and last aa of each PC1 sequence). B, 11TM constructs. PC1-11TM encodes the C-terminal 1284 amino acids (aa 3010 -4293) of mouse PC1 including all 11 transmembrane domains and the C-tail cloned downstream of the CH2-CH3 domains of human IgG. The control construct, sIg-stop, has a stop codon just to the Cterminal side of the first transmembrane domain (13,39) and encodes aa 3010 -3091 of PC1. C, cell surface expression of the PC1 constructs. HEK293T cells were transiently transfected with 470 ng/well/six-well plate of the PC1 C-tail deletion constructs (sIg-0, PC1-LT, PC1-HT, PC1-AT, or PC1-LS) in pcDNA 1.1 or with 250 ng/well/six-well plate of the PC1 C-terminal constructs (sIg-stop, PC1-11TM) in pcDNA 3.1. The final DNA amount was made 3 g/well using pBluescript DNA. Surface expression was detected on nonpermeabilized cells using a FITC-conjugated mouse anti-human IgG monoclonal antibody as described under "Experimental Procedures." G␣ i , also stimulated NFAT activation (data not shown). Previous work from our laboratory demonstrated that the cytosolic C-tail of PC1 is capable of activating c-Jun N-terminal kinase and AP-1 in a heterotrimeric G protein-dependent manner utilizing G␣ i , G␣ q , G␣ 12 , and G␣ 13 (39).
NFAT Activation by the PC1 C-tail Involves PLC Activation, Ca 2ϩ Release from Internal Stores, and Ca 2ϩ Entry-Numerous studies have demonstrated that ligand-dependent activation of G q -coupled receptors results in the activation of PLC (66 -68), which hydrolyzes membrane phospholipids to give rise to two intracellular signaling molecules, inositol trisphos-phate (IP3) and diacylglycerol (69), which in turn stimulate Ca 2ϩ mobilization from intracellular stores and Ca 2ϩ entry from the extracellular pool (70). To address whether PC1-mediated NFAT activation involves PLC activation, we made use of a known inhibitor of PLC, U73122 (71). As shown in Fig. 3A, the addition of 50 M U73122 at 4 h before harvesting cells caused a significant decrease in PC1-mediated NFAT activation. Higher concentrations of U73122 were toxic to the cells (data not shown). We then tested the purported inhibitor of IP3 receptors, xestospongin (72). As shown in Fig. 3B, xestospongin partially, although significantly, decreased PC1-LT-mediated NFAT activation. The treatment of HEK293T cells with 2-aminophenylborate (2-APB), which inhibits both the IP3 receptor and the ryanodine receptor (73), almost completely abolished PC1-mediated NFAT activation (Fig. 3C).
These results suggest that the initial events resulting in PC1-mediated NFAT activation may be, at least in part, dependent on G q signaling to PLC followed by Ca 2ϩ release from intracellular stores. NFAT activation has been shown to require sustained increases in intracellular Ca 2ϩ that are dependent on Ca 2ϩ entry via store-operated calcium release-activated Ca 2ϩ channels (74). To determine whether the PC1-mediated NFAT activation is dependent on Ca 2ϩ entry, we treated PC1-LT and NFAT luciferase-transfected HEK293T cells with gadolinium hydrochloride (GAD) at 4 h before harvesting. As shown in Fig. 3D, GAD treatment completely abolished PC1-mediated NFAT activation. Similar results were obtained when PC1-transfected cells were treated with the extracellular Ca 2ϩ chelator, EGTA (data not shown). These results suggest that NFAT activation by PC1 is dependent both on Ca 2ϩ release from internal stores and on Ca 2ϩ entry from the extracellular pool.
The PC1 C-tail Causes Sustained Ca 2ϩ Increases and Calcineurin-dependent NFAT Activation-Activation and nuclear translocation of NFAT is known to depend on the activity of the Ca 2ϩ /calmodulin-dependent protein phosphatase, calcineurin, which requires sustained increases in intracellular Ca 2ϩ levels for NFAT activation (74). To test whether PC1mediated signal transduction can lead to a sustained elevation in intracellular Ca 2ϩ , HEK293T cells were co-transfected with PC1-LT or the control construct, sIg-0, and a GFP expression construct to identify transfected cells. After 24 h, the cells were loaded with Fura-2/AM and both basal and caffeine-stimulated intracellular Ca 2ϩ levels were determined. As shown in Fig.  4A, both steady-state intracellular Ca 2ϩ levels (76.3 Ϯ 17.8 nM (ϮS.E., n ϭ 7) for sIg-0 versus 137.1 Ϯ 19.5 nM (ϮS.E., n ϭ 8) for PC1-LT (p Ͻ 0.05)) and peak caffeine-induced Ca 2ϩ release were higher in PC1-LT-transfected cells. Thus, it appears that the PC1-LT construct can give rise to significant long-term elevations in basal intracellular Ca 2ϩ .
To test whether this PC1-mediated elevation in intracellular Ca 2ϩ can activate calcineurin, we used the calcineurin inhibitor CSA. HEK293T cells transfected with PC1-LT and the NFAT reporter construct were treated with 100 ng/ml CSA 4 h before harvesting. Fig. 4B shows that CSA treatment completely abolished the PC1-mediated NFAT activation. As a control (data not shown), we found that PC1-mediated AP-1 (7ϫ AP-1, Stratagene) activation was not inhibited by CSA treatment. The inhibition by CSA suggests that the PC1-mediated activation of the composite ARRE-2 NFAT/AP-1 promoter is strongly dependent on calcineurin activation and therefore is an NFAT-dependent process.
PC2 Does Not Activate NFAT-The co-assembly of PC1 and PC2 through an interaction involving the C-tails of both proteins has been shown to create a novel Ca 2ϩ current (4, 5, 18, 19, 75). To determine whether PC1 activation of NFAT involves PC2, cells were transfected with a full-length PC2 construct.

FIG. 2. The G protein binding region of the PC1 C-tail is required to activate NFAT. A, effects of C-tail deletion constructs.
HEK293T cells were transiently co-transfected with an NFAT-responsive promoter-firefly luciferase reporter construct (100 ng/well/six-well plate) and a control Renilla luciferase construct (1.5 ng/well) together with 500 ng/well of either the control sIg-0 or one of various PC1 C-tail deletion constructs (PC1-LT, PC1-HT, PC1-AT, or PC1-LS). The final DNA amount was made 3 g/well using pBluescript DNA. The cells were harvested and lysed 24 h post-transfection using Promega 1ϫ passive lysis buffer. Luciferase activity was measured using the dual luciferase reporter assay system (Promega). The Renilla luciferase values for the wells within an individual 3-well experiment were used to correct firefly luciferase values. A promoterless pCis-CK plasmid showed no activity (data not shown). The bars represent average firefly luciferase Ϯ S.D. of a representative experiment of three individual experiments, each done in triplicate. Asterisk represents data significant at p Ͻ 0.01 compared with control (sIg-0). Below are anti-human IgG Western blots demonstrating expression levels of the various fusion proteins for two (of the three) samples in the experiment. B, augmentation of PC1 C-tail-mediated NFAT activation by G␣ q . HEK293T cells were transiently co-transfected as above with NFAT luciferase, Renilla luciferase, and 200 ng/well of either sIg-0 or the PC1 C-tail construct PC1-LT together with 50 ng/well G␣ q or pcDNA3.1 (vector control for G␣ q ) and were harvested 24 h post-transfection. The values represent mean Ϯ S.D. of a representative experiment of three individual experiments, each done in duplicate or triplicate. Asterisk represents data significant at p Ͻ 0.01 compared with PC1-LT. Below are anti-human IgG Western blots demonstrating expression levels of the fusion proteins for two samples in the experiment. Fig. 5 shows that transfection of 50 ng/well of a full-length HA-tagged PC2 construct did not activate NFAT (amounts of up to 2 g/well did not significantly activate NFAT; data not shown). Furthermore, PC1-mediated NFAT activation was not augmented by co-transfection with 50 ng/well HA-tagged PC2 construct but instead it appeared to be somewhat inhibited. These results suggest that it is PC1 and not PC2 that is responsible for the PC1-LT activation of NFAT.
PC1-mediated Dephosphorylation and Nuclear Translocation of NFAT-NFAT requires dephosphorylation by calcineurin for its nuclear translocation (74). To determine whether NFAT is dephosphorylated following transfection of PC1-LT, we assayed the relative amounts of phosphorylated and dephosphorylated forms of a co-transfected HA-tagged NFATc1-GFP fusion protein by Western blot analysis using an anti-HA antibody to determine their relative electrophoretic mobilities (64). As shown in Fig. 6A, the predominant form of HA-NFATc1-GFP in sIg-0 control-transfected cells was the slower migrating phosphorylated (cytosolic) band (lanes 1 and  2). The treatment of these cells with the Ca 2ϩ ionophore A23187 shifted most of the HA-NFATc1-GFP to the faster migrating dephosphorylated (nuclear) form (lanes 3 and 4). Co-transfection with PC1-LT also resulted in a shift to the faster migrating dephosphorylated form (compare lanes 1 and 2 with lanes 5 and 6), which was prevented by CSA treatment (lanes 7 and 8). Intracellular localization of HA-NFATc1-GFP was assessed by visualization of the GFP fluorescence. Cotransfection with sIg-0 resulted in predominantly cytoplasmic localization of HA-NFATc1-GFP (Fig. 6B, upper left), whereas co-transfection with PC1-LT resulted in significant nuclear localization of HA-NFATc1-GFP (upper right), in many cases in a characteristic punctate pattern (76). Co-transfection with HA-NFATc1-GFP and PC1-11TM, which contains all of the 11 transmembrane domains as well as the C-tail (13) (see Fig. 1), also resulted in nuclear translocation of HA-NFATc1-GFP (Fig.  6B, lower left versus lower right). The 11TM construct was also effective in activating the NFAT luciferase reporter (Fig. 6C).
PC1-LS-mediated NFAT Activation Is Enhanced by LiCl-PC1 has been shown to inhibit GSK-3␤, leading to the stabilization of ␤-catenin and activation the ␤-catenin target gene, siamois (41). The distal C-tail region of PC1, which contains the coiled-coil domain, was shown to have this activity. This region of the PC1 C-tail corresponds to the PC1-AT construct used in our experiments (see Figs. 1 and 2). Because GSK-3␤ is known to phosphorylate nuclear NFAT, thus causing it to translocate back to the cytoplasm (52), it is possible that there are two activities in the PC1 C-tail that can enhance NFAT activity, one mediated by the proximal C-tail leading to a sustained elevation in intracellular Ca 2ϩ and the other mediated by the distal C-tail leading to an inhibition of GSK-3␤ and thus to the nuclear retention of NFAT. Consistent with this finding, the PC1-LT construct (which contains both domains) gave rise to a higher level of NFAT activation than did the PC1-LS construct, which lacks the GSK-3␤ inhibitory domain (see Figs.  1 and 2A). If this difference is the due to the inability of PC1-LS to inhibit GSK-3␤, it should be possible to increase PC1-LSmediated NFAT activity to the same level achieved by PC1-LT by inhibiting GSK-3␤ with LiCl (77). As shown in Fig. 7, NFAT activity was significantly lower with PC1-LS than with PC1-LT in the absence of LiCl but was increased to the same level in the presence of LiCl.
NFAT Localization in the M-1 Mouse Cortical Collecting Duct Cell Line and in Renal Tubules-An analysis of the temporal and spatial patterns of PC1 expression in normal and polycystic kidney tissues has suggested that PC1 functions in epithelial differentiation and maturation (78 -81). To understand the role of PC1-mediated NFAT signaling in the kidney, it will be important to establish that NFAT is expressed in renal epithelial cells. To do this, M-1 mouse cortical collecting duct cells were examined by immunofluorescence using an anti-NFATc1 antibody that recognizes all of the NFAT family members. As shown in Fig. 8A (upper panels), untreated M-1 cells express NFAT, which appears to be largely cytoplasmic but also nuclear under basal cell growth conditions (green fluorescence). Treatment of the cells with the Ca 2ϩ ionophore A23187 caused nuclear accumulation of NFAT (Fig. 8A, middle panels), whereas treatment of the cells with CSA resulted in cytoplasmic retention (Fig. 8A, lower panels). RT-PCR (Fig. 8B) showed that the NFAT c1-c4 genes are expressed in M-1 cells.
NFAT was also localized in embryonic and adult mouse kidneys. Kidneys from late gestation embryos or 3-month-old BALB/c mice were prepared for immunohistochemistry with the NFATc1 pan-antibody. In embryonic kidneys (Fig. 8C, top  and middle panels), the most prominent immunostaining was found in the nephrogenic zone in vesicles (V) and S-shaped bodies with less staining in the ureteric buds (UB) except at the ureteric bud tips (arrows). NFAT staining was retained in early tubular (T) and glomerular structures in the embryonic kidney. In adult kidney (Fig. 8C, bottom panel), NFAT staining was most prominent in tubular epithelial cells within the cortex. DISCUSSION This report demonstrates that the C-tail of PC1 can activate calcineurin/NFAT signaling, a process known to be dependent on sustained elevations in intracellular Ca 2ϩ . Unlike transient increases in Ca 2ϩ , which are responsible for short-term events such as muscle contraction or synaptic transmission, sustained increases in cytosolic and nuclear Ca 2ϩ can give rise to longlasting outcomes such as adaptive cellular responses involving changes in cell differentiation and cell death, which are dependent on changes in gene expression (44,82). Based on our results, we propose that PC1-mediated heterotrimeric G protein signaling triggers PLC activation followed by an initial IP3-dependent release of Ca 2ϩ from internal stores, which then causes Ca 2ϩ entry through store-operated Ca 2ϩ release-activated Ca 2ϩ channels, thus activating calcineurin and NFAT. Support for the idea that PC1-mediated NFAT activation involves G protein signaling comes from experiments using C-tail deletion constructs with complete or partial G protein binding ability (37). The PC1 construct that completely lacks the G protein binding domain (PC1-AT) was not able to activate NFAT, and the construct that partially lacks the G protein binding domain (PC1-HT) had diminished ability to activate NFAT. In addition, the G␣ q gain-of-function studies showed that co-transfection of G␣ q augmented PC1-mediated NFAT activation. These results suggest that the PC1 C-tail can couple with G q to activate NFAT. Previous studies from our laboratory showed that the PC1 C-tail can couple with G q as well as with G i/o and G 12/13 to activate c-Jun N-terminal kinase and AP-1 (39). Others (83) have corroborated the observation that PC1 can activate G i/o .
It is widely known that stimulation of G q -coupled receptors activates PLC␤ (66 -68). Thus, the possibility exists that PC1 activates the ␤ isoforms of PLC in HEK293T cells, which can then initiate the release of Ca 2ϩ to activate NFAT. Indeed, we showed that a PLC inhibitor, U73122, significantly decreased PC1-mediated NFAT activation. It has been demonstrated that In the upper panels, the green (HA-NFATc1-GFP) and blue (DAPI) images were merged (in the lower panels, the cells were not stained with DAPI). Arrows point to nuclei. C, HEK293T cells were co-transfected as described in Fig. 2 with the NFAT-responsive promoter-firefly luciferase reporter construct (100 ng/well/six-well plate) and control Renilla luciferase construct (1.5 ng/well) together with 300 ng/well of either sIg-stop or PC1-11TM (see Fig. 1). The bars represent average firefly luciferase Ϯ S.D. for two experiments (each done in duplicate) carried out with different DNA preparations. Below are Western blots demonstrating expression levels of the fusion proteins using anti-PC1 and anti-IgG for PC1-11TM and anti-IgG for sIg-stop. HEK293 cells express all of the phosphoinositide-specific PLC␤ isoforms (PLC␤1-␤4) (84). The sensitivity of these isoforms to G q activation differs with the order ␤1 Ͼ ␤3 Ͼ ␤4 Ͼ ␤2 (85). As with the G␣ q subunits, co-transfected G␣ 12 subunits were also able to augment PC1-mediated NFAT activation (data not shown). G␣ 12 is known to activate PLC⑀ (86), which can also initiate Ca 2ϩ release and subsequent NFAT activation. In two of the well characterized animal models of PKD (87,88), the activity of PLC␥ has been shown to be increased (89). These observations, although not directly involving PC1, suggest the possible importance of signaling events leading to PLC activa-tion in the pathophysiology of PKD.
Inhibition by 2-APB and xestospongin suggests that Ca 2ϩ release is involved in PC1-mediated NFAT activation. The complete inhibition of NFAT activation by GAD, which blocks Ca 2ϩ entry (90), suggests that PC1-mediated Ca 2ϩ release triggers Ca 2ϩ entry through store-operated or Ca 2ϩ release-activated Ca 2ϩ channels. The addition of EGTA to the extracellular medium also abolished PC1-mediated NFAT activation (data not shown), supporting this contention. Although the identity of plasma membrane Ca 2ϩ release-activated Ca 2ϩ channels has not yet been established, their role has been postulated in sustained Ca 2ϩ elevations (90 -92). Other mechanisms for Ca 2ϩ entry may also be possible, potentially involving PC2. However, the expression of PC2 alone did not result in NFAT activation and co-expression of PC2 with PC1-LT did not augment PC1-mediated NFAT activation. PC2 has been demonstrated to be constitutively active when co-assembled with PC1 and targeted to the plasma membrane (19). These observations suggest that PC2 is not involved in NFAT activation and further suggest that PC1 alone is capable of generating Ca 2ϩ signals.
The NFAT-responsive reporter used in these studies contains four copies of a composite NFAT/AP-1 response element that requires cooperative binding of both NFAT and AP-1 (60). This ARRE-2 element has a weak AP-1 site (TGTTTCA), which binds Fos-Jun dimers, such that c-Jun makes direct contact with NFAT bound to its adjacent DNA element. Previously, we demonstrated that the PC1-LT C-tail construct can activate another Fos-Jun-responsive AP-1 reporter, the 7ϫ AP-1 promoter-reporter (pAP-1-luciferase; Stratagene), which contains the AP-1 site, TGACTAA (39). We and others (40) also found that the PC1-HT construct only weakly (if at all) activates the 7ϫ AP-1 promoter. Thus, the very poor activation of the ARRE-2 NFAT/AP-1 reporter by PC1-HT ( Fig. 2A) may reflect an inability of this truncated C-tail construct to activate G protein signaling to both NFAT and AP-1.
The strong response of the NFAT reporter to PC1-LT relative to PC1-LS (see Fig. 2A) may be due to the ability of the distal C-tail region to inhibit GSK-3␤. In fact, the distal C-tail region of PC1 has been implicated as a positive effector of the Wnt/␤catenin pathway by its inhibition of GSK-3␤, stabilization of ␤-catenin, and activation of ␤-catenin target genes (41). It has also been shown that GSK-3␤ can phosphorylate NFAT, causing its nuclear export and thus its inactivation (51,93), and that inhibition of GSK-3␤ with LiCl can increase NFAT activity (77,94). As such, the inhibition of GSK-3␤ by PC1-LT could promote the nuclear retention of NFAT. The observation (see Fig. 7) that LiCl treatment further increased the responses of both PC1-LT and PC1-LS suggests that GSK-3␤ is active in HEK293T cells and can attenuate the NFAT response to the degree to which it is inhibited by the PC1 distal C-tail region. Taken together, these results suggest that the PC1 C-tail can serve to integrate both Ca 2ϩ -dependent and Ca 2ϩ -independent pathways, leading to the activation of NFAT targets.
The temporal and spatial patterns of its kidney expression have suggested that PC1 functions in renal epithelial differentiation and maturation (78 -81). To understand the role of PC1-mediated NFAT signaling in the kidney, it will be important to establish that NFAT is expressed in renal epithelial cells. We have now shown that NFAT is expressed in HEK293T cells (data not shown) and in mouse M-1 cortical collecting duct cells (Fig. 8, A and  B). We have also demonstrated that NFAT is expressed in tubular epithelial cells of the developing and adult mouse kidney (Fig. 8C), which corresponds temporally with PC1 expression (78-81), suggesting the possibility that PC1 and NFAT function together during renal development, perhaps in differentiating tubules, and in FIG. 8. Expression of endogenous NFAT. A, NFAT in M-1 mouse cortical collecting duct cells. M-1 cells were seeded and grown in chamber slides. 2 h before processing, the wells were treated with the Ca 2ϩ ionophore, A23187, or with CSA. The slides were processed as described under "Experimental Procedures." Nuclear and cytosolic localization of NFAT, as indicated by green fluorescence staining, was visualized by confocal microscopy. The nuclei were counterstained with DAPI (blue). Upper panels, untreated M-1 cells (basal); middle panels, M-1 cells treated with 10 M A23187; and lower panels, M-1 cells treated with 100 ng/ml CSA. The left panels show NFAT localization, the middle panels (vertically) show DAPI-stained nuclei merged with NFAT, and the right panels show the cells by phase-contrast microscopy. B, RT-PCR. Total RNA was isolated from M-1 mouse cells as described under "Experimental Procedures." A total of 1 g of RNA was used for RT-PCR analysis using oligonucleotides specific for NFATc1 (c1), NFATc2 (c2), NFATc3 (c3), and NFATc4 (c4). The expected sizes of the products were 344, 403, 580, and 240 bp, respectively. A 100-bp ladder is shown in the marker lane (M). C, NFAT in mouse kidneys. kidneys from late gestation embryos or 3 month-old BALB/c mice were fixed in 4% paraformaldehyde, embedded in paraffin, and processed as described under "Experimental Procedures." Controls (data not shown) consisted of either normal rabbit IgG or peptide competition. In embryonic kidneys (top and middle), the most prominent immunostaining was found in the nephrogenic zone in vesicles (V) and S-shaped bodies (S) with less staining in the ureteric buds (UB) except at the ureteric bud tips (arrows). In adult kidney (bottom), NFAT staining was most prominent in tubular (T) epithelial cells within the cortex. the adult kidney, perhaps in cellular adaptive responses. Recent observations have demonstrated the presence of NFATc1 in tubular epithelial cells of the rat kidney (95), which was localized in nuclei following streptozotocin-induced diabetic renal glomerular and tubular hypertrophy (94). CSA treatment abolished the renal hypertrophy in parallel with a reduction in calcineurin activity and NFAT relocalization to the cytoplasm (95). As such, NFAT activation may be associated with adaptive cellular hypertrophy in the kidney (95) as it is in the heart (96). It has also been reported that NFAT5 is highly expressed in the renal medulla and that NFAT5 knock-out mice have severe renal dysfunction with impaired activation of osmoprotective genes and microcystic tubular dilation (97).
Most mutations in the PKD1 gene are loss-of-function mutations such as nonsense, splicing, and frameshift mutations or deletions, most of which would be expected to cause the loss of expression of the PC1 C-tail. As such, our tail-less control construct, sIg-0, should mimic most of these disease-causing mutations. Thus, if PC1 normally functions in regulating NFAT target genes (60), a loss of PC1 expression by mutation would be expected to cause the abnormal (decreased or increased) expression of these target genes during renal development or in the adult kidney in response to adaptive signals, thus causing or contributing to cyst growth in autosomal dominant polycystic kidney disease.