Polycystin-1 Activation of c-Jun N-terminal Kinase and AP-1 Is Mediated by Heterotrimeric G Proteins*

Functional analysis of polycystin-1, the product of the gene most frequently mutated in autosomal dominant polycystic kidney disease, has revealed that this protein is involved in the regulation of diverse signaling pathways such as the activation of the transcription factor AP-1 and modulation of Wnt signaling. However, the initial steps involved in the activation of such cascades have remained unclear. We demonstrated previously that the C-terminal cytosolic tail of polycystin-1 binds and activates heterotrimeric G proteins in vitro . To test if polycystin-1 can activate cellular signaling cascades via heterotrimeric G protein subunits, polycystin-1 C-terminal tail-mediated c-Jun N-terminal kinase (JNK) and AP-1 activities were assayed in transiently transfected 293T cells in the presence of dominant-negative, G protein inhibiting constructs, and in the presence of cotransfected G (cid:1) subunits. The results showed that polycystin-1-mediated JNK/AP-1 activation is mediated by G (cid:1) and G (cid:2)(cid:3) subunits. Polycystin-1-mediated AP-1 activity could be significantly augmented by cotransfected G (cid:1) i , G (cid:1) q , and G (cid:1) 12/13 subunits,

Autosomal dominant polycystic kidney disease (ADPKD) 1 is a major inherited disorder that is characterized by the growth of large, fluid-filled cysts from the tubules and collecting ducts of affected kidneys, and by a number of extrarenal manifestations including liver and pancreatic cysts, heart valve defects, and cerebral, and aortic aneurysms (1)(2)(3)(4). Approximately 85% of all cases of ADPKD are caused by mutations in the PKD1 gene, with most (if not all) of the remaining cases being caused by defects in the PKD2 gene. The PKD1 gene product, polycys-tin-1, appears to be a large membrane-associated glycoprotein (5)(6)(7)(8)(9) that is thought to play a role in cell-cell and/or cell-matrix interactions (10). Polycystin-1 is composed of a large N-terminal extracellular domain of about 3,000 amino acids, a multimembrane spanning domain of about 1,000 amino acids containing 11 transmembrane segments, and a C-terminal cytosolic domain of about 200 -225 amino acids. Polycystin-2, the protein product of the PKD2 gene, appears to be a Ca 2ϩpermeable, nonselective cation channel whose activity may be regulated by a direct interaction with the C-terminal cytosolic domain of polycystin-1 (11)(12)(13)(14).
Sequence analysis of polycystin-1 has suggested that the C-terminal cytosolic domain of the protein is involved in protein-protein interactions and signal transduction (15). Biochemical analyses support this hypothesis, as transient transfection of the C-tail of polycystin-1 modulates Wnt signaling via stabilization of ␤-catenin (16) and activates the transcription factor AP-1 via c-Jun N-terminal kinase (JNK) and protein kinase C (17). Polycystin signaling has also been implicated in the regulation of a cell growth phenotype, as ADPKD renal epithelial cells are susceptible to abnormal proliferation in response to cAMP (18 -20). Furthermore, stable transfection of full-length polycystin-1 in Madin-Darby canine kidney cells suppresses spontaneous cyst formation and inhibits cellular growth rates and apoptosis (21). Despite these clues, however, the primary cytosolic events that initiate these signaling cascades have remained unclear.
A number of observations have suggested that heterotrimeric G proteins play a role in polycystin-1-mediated signaling. Polycystin-1 has been shown to bind and stabilize RGS7 (regulator of G protein signaling 7) (22), a member of the newly identified family of RGS proteins that are capable of regulating G protein-dependent signaling cascades by accelerating the hydrolysis of GTP bound to G␣ subunits of certain heterotrimeric G proteins (23). Furthermore, RGS7 has been identified as a candidate for a genetic modifier of bpk, a murine model of PKD that is similar to human autosomal recessive PKD (24). We have also shown that polycystin-1 binds and activates heterotrimeric G i /G o proteins in vitro (25). However, a direct link between polycystin-1-mediated signaling pathways and heterotrimeric G proteins has not been established.
To determine whether heterotrimeric G proteins mediate polycystin-1 signaling, we tested the effects of expression of the G␤␥ sequestering constructs, dominant-negative G␣ i2 and ␤-adrenergic receptor kinase C-terminal tail (␤ARK-ct), on polycystin-mediated JNK activation. The evidence demonstrated that polycystin-1 activates JNK signaling via G␤␥ subunits of heterotrimeric G proteins. We also showed that a dominant-negative p115RhoGEF construct inhibited AP-1 activation and that wild-type G␣ 12 and G␣ 13 subunits effectively augmented polycystin-1 signaling to AP-1. G␣ i and G␣ q family subunits were also found to be capable of stimulating polycystin-1 signaling to AP-1 but less effectively than G␣ 12 family subunits. These results, taken together with our previous in vitro G protein binding and activation studies (25), indicate that polycystin-1 can couple with and signal via heterotrimeric G proteins and thus is an atypical heterotrimeric G proteincoupled receptor.
JNK Assays-Human embryonic kidney 293T cells were maintained in DMEM (1ϫ MOD) with L-glutamine and 1 g/liter glucose (CellGro) with 500 units/liter penicillin and 0.5 mg/liter streptomycin (Sigma) and 10% fetal calf serum at 37°C in 5% CO 2 (4.5 g/liter glucose and heat-inactivated serum were used for the endogenous JNK and AP-1 assays). Cells were transfected using a modified calcium phosphate protocol as described previously (28). Following the addition of DNA precipitates, cells were then incubated at 37°C in 5% CO 2 for 4 h, after which the medium was replaced with serum-free growth medium or with medium containing 0.5% serum. Pertussis toxin (200 ng/ml) was added either 1 h before or 4 h after transfection. Cells were lysed 16 or 24 -26 h following transfection. They were then washed in phosphatebuffered saline and scraped into 0.5 ml per T25 flask of Tris/Triton lysis buffer (TLB) (20 mM Tris, pH 7.4, 137 mM NaCl, 25 mM ␤-glycerol phosphate, 2 mM EDTA, 1 mM Na 3 VO 4 , 2 mM Na 2 P 2 O 7 , 1% Triton X-100, 10% glycerol) plus 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin, 2 mM benzamidine, and 0.5 mM dithiothreitol (TLBϩ). The cells were vortexed 20 s at 4°C and incubated on ice for 20 min, and the lysates were clarified by centrifugation (14,000 ϫ g) for 10 min. Supernatants were placed at Ϫ80°C until use. Protein concentration was determined using the Bio-Rad Detergent Compatible Assay Kit. Protein A/Gϩ-agarose beads (50 l per reaction; Santa Cruz Biotechnology) were washed twice in TLB and once in TLBϩ and were suspended in 0.5 ml of TLBϩ per 6 reactions plus 0.4 g per reaction of anti-HA probe F-7 (Santa Cruz Biotechnology) for assaying HA-JNK1␤ activity. The beads were rotated at 4°C for ϳ45 min and spun down briefly in a microcentrifuge; the supernatant was decanted, and the beads were resuspended in aliquots of 100 l/reaction. 200 g of cellular lysate and TLBϩ was added to a final volume of 0.6 ml; the reactions were rotated at 4°C for 2 h and then washed 3 times in TLBϩ and twice in kinase buffer (25 mM HEPES, pH 7.4, 25 mM ␤-glycerol phosphate, 25 mM MgCl 2 , 0.1 mM Na 3 VO 4 , 0.5 mM dithiothreitol). The beads were then resuspended in 31 l of kinase buffer containing 1 mM ATP, 8 g of GST-Jun-(1-79) substrate, and [␥-32 P]ATP (3,000 Ci/mmol; PerkinElmer Life Sciences) and were incubated in a room temperature water bath for 5 min. Reactions were terminated by the addition of 20 l of 2ϫ Laemmli loading buffer and were placed on dry ice until loading. Reactions were boiled 5 min and then loaded and electrophoresed, and fusion protein was detected as described previously (29). 32 P incorporation was quantified using a PhosphorImager SI (Molecular Dynamics). Endogenous JNK was assayed with the PathDetect Trans-Reporting System (Stratagene) by following the manufacturer's instructions.
AP-1 Assays-Endogenous AP-1 was assayed with the PathDetect Cis-Reporting System (Stratagene), which utilizes a 7ϫ AP-1 reporter. 293T cells were plated 24 h prior to transfection at ϳ7.5 ϫ 10 5 cells per well in 6-well plates. A total of 7 g of DNA was used for calcium phosphate precipitation, with 3 g of control or polycystin-1 construct, 1 g of reporter, 5 ng of RL-null, 100 or 500 ng of G␣ construct, 1 g of DN p115RhoGEF (DN p115), and pBS as the filler. The medium, lacking serum, was changed 3.5-4 h after transfection was begun, and cells were incubated for 24 -26 h longer. Cells were lysed in Passive Lysis Buffer (Promega), and 20 l of lysate were used for the Dual-Luciferase assay (Promega) on an EG & G Berthold 9507 luminometer. Expression of the G␣ constructs was confirmed by Western blotting.

Dominant-negative ␤ARK-ct and G␣ i2 Constructs Inhibit JNK Activity Mediated by the C-terminal Cytosolic Tail of Human and Mouse Polycystin-1-Transfection of a human C-terminal cytosolic polycystin-1 construct into 293T cells was
shown to activate c-Jun N-terminal kinase (17). Because JNK is known to be activated by G␤␥ subunits (30,31), 293T cells were cotransfected with cDNAs encoding a C-terminal polycystin-1 fusion protein, HA-JNK1␤, and a G␤␥-sequestering construct, the ␤-adrenergic receptor kinase C-terminal tail (␤ARKct) (32). Following transfection with the human C-tail MN6 construct (see Fig. 1), cells were lysed; HA-JNK1␤ was immunoprecipitated with an anti-HA antibody, and HA-JNK activity was determined by an immune complex kinase assay. JNK activity in MN6-transfected cells was increased Ͼ7-fold over that of control cells transfected with sIg-0 ( Fig. 2A). Cotransfection of ␤ARK-ct reduced JNK activity ϳ30% ( Fig. 2A), suggesting that this JNK activation is mediated by G␤␥ subunits.
Because overexpression of G␣ subunits can also sequester G␤␥ subunits, 293T cells were cotransfected with cDNAs encoding MN6, HA-JNK1␤, and a dominant-negative DN G␣ i2 that remains in an inactive, ␤␥-bound state upon GTP binding (G␣ i2 G204A) (33). Cotransfection of increasing amounts of DN G␣ i2 decreased JNK activation in a dose-dependent fashion by an average of ϳ65% (Fig. 2B). A similar experiment was carried out using a wild-type (WT) G␣ i2 construct. As seen in Fig  2C, cotransfection of the WT construct inhibited JNK activation to nearly the same degree as the DN construct (ϳ60% inhibition). The observation that both DN and WT G␣ i2 subunits reduce polycystin-mediated JNK activation suggests that they inhibit signaling by sequestering G␤␥ subunits.
To investigate the effects of these inhibitors on endogenous JNK activity, 293T cells were cotransfected with a murine polycystin-1 C-tail construct, sIg-PKD222 (see Fig. 1), one of the inhibitor constructs, a c-Jun activation domain/GAL4 DNA binding domain fusion protein, and a GAL4 promoter-reporter construct (Fig. 3). Under these conditions, all three constructs completely inhibited endogenous JNK activation, supporting the idea that G␤␥ subunits mediate polycystin-1-induced JNK activation from the C-tail constructs. The WT and DN inhibitors utilized in these studies were ineffective in inhibiting either endogenous JNK or HA-JNK1␤ activated by constitutively active MEKK, a kinase that acts upstream of JNK (34), demonstrating the specificity of inhibition by ␤ARK-ct and the DN and WT G␣ i2 constructs (data not shown).
JNK Signaling via the Complete Multimembrane Spanning Region of Polycystin-1-The results shown in Figs. 2 and 3 suggest that the C-terminal tail of polycystin-1 signals in a constitutive fashion, because the C-tail constructs do not contain extracellular polycystin-1 sequences. Because many GPCRs utilize intracellular loops to initiate signaling, we wanted to determine whether a larger polycystin-1 construct containing all of the putative intra-and extracellular loops would behave in a manner similar to that of the human sIg-PKD-MN6 and the murine sIg-PKD222 constructs. Thus, a cDNA encoding the C-terminal 1,284 amino acids of murine polycystin-1 was fused to the extracellular portion of the sIg.7 membrane targeting cassette (lacking the CD7 transmembrane domain). This construct, sIg-PKD1284 (see Fig. 1), encodes the complete, predicted multimembrane spanning and intra-and extracellular loops of polycystin-1 (8) and might be expected to function similarly to native polycystin-1.
Consistent with observations using the shorter C-terminal tail constructs, sIg-PKD1284 also activated HA-JNK1␤. ␤ARKct (Fig. 4A) and DN G␣ i2 (data not shown) were both effective at blocking JNK activity, reducing this activation ϳ65 and ϳ35%, respectively. However, in contrast to our observations with the shorter polycystin-1 constructs, WT G␣ i2 caused a stimulation in JNK activity, which was more effective at lower amounts of the transfected WT G␣ i2 (Fig. 4B). The stimulation of JNK at low concentrations of WT G␣ i2 suggested that polycystin-1 may be capable of utilizing G␣ i2 for signaling and that the reduced JNK activation at higher concentrations of WT G␣ i2 may be due to sequestration of G␤␥ by overwhelming amounts of the G␣ subunits. Endogenous JNK was also completely inhibited by ␤ARK-ct and DN G␣ i2 but showed some stimulation with WT G␣ i2 when activated by the sIg-PKD1284 polycystin-1 construct (data not shown). Despite this stimulation of JNK activity by exogenous WT G␣ i2 , we could only partially block polycystin-induced JNK activation with pertussis toxin. In four experiments in which inhibition was seen, HA-JNK1␤ activation by the mouse and human C-tail constructs was inhibited from 9 to 43% (mean ϭ 25%). These results suggest that polycystin-induced JNK signaling is only partially mediated by G i and may therefore involve other G protein families.
AP-1 Signaling Is Mediated by the G␣ Subunits of Heterotrimeric G Proteins-AP-1 can be activated by a number of signaling pathways, including the JNK pathway (35)(36)(37). Previous work (17) had shown that the human polycystin-1 C-tail can activate AP-1 in a JNK-and PKC-dependent fashion. We have seen that the sIg-PKD1284 construct (data not shown), as well as the sIg-PKD222, can activate AP-1 when assayed using a 7ϫ AP-1 promoter. To identify the G protein families that can be activated by polycystin-1, wild-type G i , G q , and G 12 family ␣ subunits were cotransfected into 293T cells with the sIg-PKD222 construct, and AP-1 was assayed. At 100 ng of transfected G␣ construct, G␣ 12 and G␣ 13 were found to be quite effective at stimulating polycystin-1-induced AP-1 activity (Fig.  5A). G␣ i1 , G␣ i2 , G␣ i3 , and G␣ q were much less effective com- FIG. 1. Polycystin-1 fusion constructs. A, models for full-length polycystin-1 (left), the C-terminal tail constructs (middle), and the multimembrane spanning construct (right). B, structures of the cDNAs encoding the human and mouse polycystin-1 fusion proteins. All of the constructs have a CD5 signal sequence followed by the CH2-CH3 domains of human IgG (sIg). The short C-tail constructs also have the CD7 transmembrane domain. sIg-PKD-MN6, which encodes amino acids 4,077-4,241 of human polycystin-1, and its control construct, sIg-0-12 lacking the C-tail, are cloned in pcDM12; sIg-PKD222, which encodes amino acids 4,072-4,293 of mouse polycystin-1, and its control construct, sIg-0-1.1 lacking the C-tail, are cloned in pcDNA1.1. sIg-PKD222 encodes the C-terminal 222 amino acids of polycystin-1, and MN6 encodes a shorter C-terminal domain construct that is truncated in the coiled-coil domain (68). sIg-PKD1284, which encodes amino acids 3,010 -4,293 of mouse polycystin-1, and its control construct sIg-stop, which has a stop codon just to the 3Ј side of the first putative transmembrane domain (8) and therefore encodes amino acids 3,010 -3,091 of polycystin-1, are cloned in pcDNA1.1. The sIg-PKD1284 construct encodes the complete multimembrane spanning and intra-and extracellular loops of polycystin-1 and is expected to adopt a structure more representative of native polycystin-1. pared with G␣ 12 and G␣ 13 (Fig. 5A and data not shown). At 500 ng of transfected G␣ construct, G␣ i1 , G␣ i2 , and G␣ i3 were found to be more effective, and G␣ q was comparable with G␣ 12 in stimulating AP-1 ( Fig. 5B and data not shown). To determine the contribution of endogenous G␣ 12 family subunits to polycystin-induced activation of AP-1, we tested the effect of the dominant-negative inhibitor of G␣ 12 family subunits, DN p115RhoGEF (DN p115). The DN p115 construct lacks the RhoGEF domain of p115RhoGEF, but contains the RGS domain, and therefore can bind to and inhibit G␣ 12 and G␣ 13 subunits (27,38,39). Fig. 6 shows two representative experiments (left and right) in which AP-1 activity in polycystintransfected cells was inhibited ϳ24 and 38% by the DN p115 construct, confirming that polycystin-induced AP-1 activity is mediated in part by endogenous G 12 family heterotrimeric G proteins.

DISCUSSION
This study has demonstrated that polycystin-1 activates JNK and AP-1 signaling via G␣ and G␤␥ subunits of heterotrimeric G proteins and that polycystin-1 can couple with G i , G q , and G 12 family proteins. Polycystin-1 contains sequence elements found in a number of GPCRs, including a latrophilin/ CL-1-like GPCR proteolytic site (40). This GPCR proteolytic site domain has been identified in CIRL-l, a member of a recently identified subfamily of large orphan receptors that contains structural features typical of cell adhesion proteins and GPCRs (41). We showed previously that polycystin-1 also contains a polybasic domain in its C-terminal tail that can activate G i /G o proteins in vitro (25). This domain lies in the most highly conserved region of polycystin-1 and is part of a sequence capable of forming stable binding interactions with heterotrimeric G i /G o proteins in vitro (25). Polycystin-1 has an unusual structure for a GPCR, as it is thought to have 11 transmembrane domains rather than 7 and thus would have to

FIG. 2. Inhibition of human polycystin-1 C-tail-mediated HA-JNK1␤ activation by dominant-negative interfering constructs.
293T cells were cotransfected with sIg-PKD-MN6 or sIg-0, HA-JNK1␤, and increasing amounts (shown in g transfected DNA) of ␤ARK-ct (A), DN G␣ i2 (B), or WT G␣ i2 (C). Cells were plated at 4 ϫ 10 5 cells per T25 flask, grown for 2 days, and transfected for 16 h. Each flask was transfected with a total of 7 g of DNA, which included the indicated amounts of ␤ARK-ct (A) brought to 3 g of DNA with pRK5, or with a total of 10 g of DNA, which included the indicated amounts of DN G␣ i2 (B) or WT G␣ i2 (C) brought to 6 g of DNA with pRK5, plus 2 g of HA-JNK1␤, and 2 g of either sIg-MN6 or sIg-0. Cells were lysed, and protein concentrations were determined. Equal amounts of protein were used to assay HA-JNK1␤ activity, as described under "Experimental Procedures." Data are expressed as fold stimulation relative to the value for sIg-0 without inhibitors, which was set at one. Independent experiments were carried out 3 (n ϭ 3) or 4 (n ϭ 4) times. Error bars represent S.D. Significant differences between treated and control samples are indicated by one asterisk (p Ͻ 0.05), as determined by one-way ANOVA. Western blots demonstrate expression of the various constructs from one of the experiments; endogenous JNK (endg JNK).

FIG. 3. Inhibition of mouse polycystin-1 C-tail-mediated endogenous JNK activation by dominant-negative interfering constructs.
293T cells were cotransfected with sIg-PKD222 or sIg-0 and the inhibitors ␤ARK-ct, DN G␣ i2 , WT G␣ i2 , or pRK5 alone (vector). Cells were plated at ϳ7.5 ϫ 10 5 cells per well of a 6-well plate, grown for 1 day, and transfected for 24 -26 h. Each 3-well triplicate sample was transfected with a total of 7 g of DNA, which included 2 g of the inhibitor constructs or pRK5, 2 g of sIg-PKD222 or sIg-0, 1 g of pFR-Luc, 50 ng of pFA2-cJun, 10 ng of pRL-null, and pBS as the filler.
To determine the level of activation of endogenous JNKs, cells were cotransfected with the c-Jun activation domain/GAL4 DNA binding domain construct (pFA2-cJun) and the luciferase reporter gene under the control of a GAL4 promoter (pFR-Luc). Following transfection, cells were lysed and JNK activation was determined by assaying firefly luciferase activity, and the values were normalized to Renilla luciferase activity. Data are expressed as relative luciferase units (RLUs). Independent experiments were carried out in triplicate (n ϭ 3). Error bars represent S.D. In all cases, the inhibitors significantly decreased JNK activation (p Ͻ 0.001), as determined by one-way ANOVA. The Western blot demonstrates expression of the polycystin constructs from one of the experiments. be classified as an atypical GPCR.
GPCRs can signal through a number of pathways that lead to the activation of JNK and AP-1 (42). Pathways that activate Rac and/or Cdc42 and JNK via heterotrimeric G proteins have been described, including those activated by G␣ 12 and G␣ 13 subunits (43,44), ␤␥ subunits (30,31), G␣ i subunits (45), and possibly G q (46). In addition, AP-1 can be activated by JNKindependent pathways involving G q and G 12 /G 13 activation of Rho, the p38 MAPKs, and BMK1/ERK5 (47,48). Work by others (17) has shown that the polycystin-1 C-terminal cytosolic domain can activate JNK via the small G proteins Rac-1 and Cdc42. Here we demonstrate the first evidence that polycystin-1 mediated signaling to JNK and AP-1 is regulated by heterotrimeric G protein subunits, possibly through a number of pathways.
To test the hypothesis that polycystin-1 modulates JNK and AP-1 activity via heterotrimeric G proteins, we assessed the ability of various polycystin-1 C-terminal constructs to activate JNK and AP-1 in the presence of specific inhibitors of G protein signaling. JNK activation mediated by C-terminal tail polycystin-1 constructs could be inhibited in some experiments up to 100% by dominant-negative inhibitors. Because transfection of a WT G␣ i2 subunit also inhibited JNK activation, it was thought that overexpression of WT and DN G␣ subunits might be inhibiting JNK activation by competing for G␤␥ subunits. Similar observations have been made in the yeast Saccharomyces cerevisiae, where overexpression of the G␣ subunit Gpa1 inhibits G␤␥ signaling (49), and in COS-7 cells, where transient transfection of G␣ i subunits interferes with G␣ q/11 -and G␣ 16mediated signaling via competition for G␤␥ complexes (50). Consistent with this idea, cotransfection of ␤ARK-ct, a specific ␤␥ scavenger, inhibited exogenously expressed HA-JNK1␤ by about 30% (Fig. 2) and completely inhibited endogenous JNK (Fig. 3).
The polycystin-1 assays made use of C-tail constructs that apparently function in a constitutively active fashion, because they lack extracellular polycystin-1 sequences. Constitutive activity of GPCRs has been observed previously (51). We also tested a larger fusion protein construct encoding amino acids 3,010 -4,293 of murine polycystin-1. This construct, sIg- Each flask was transfected with a total of 10 g of DNA, which included the indicated amounts of ␤ARK-ct (A) brought to 5 g of DNA with pRK5, or with a total of 11 g of DNA, which included the indicated amounts of WT G␣ i2 (B) brought to 6 g of DNA with pRK5, plus 2 g of HA-JNK1␤, and 3 g of either sIg-PKD1284 or sIg-stop. Cells were lysed and protein concentrations were determined. Equal amounts of protein were used to assay HA-JNK1␤ activity, as described under "Experimental Procedures." Data are expressed as fold stimulation relative to the value for sIg-stop without inhibitors, which was set at one. Independent experiments were carried out 3 (n ϭ 3) or 4 (n ϭ 4) times. Error bars represent S.D. Significant differences between treated and control samples are indicated by one asterisk (p Ͻ 0.05), two asterisks (p Ͻ 0.01), or three asterisks (p Ͻ 0.005), as determined by one-way ANOVA. Western blots demonstrate expression of the various constructs from one of the experiments, endogenous JNK (endg JNK).

FIG. 5. Stimulation of mouse polycystin-1-induced AP-1 activity by G␣ family subunits.
Endogenous AP-1 was assayed in 293T cells transfected as described under "Experimental Procedures," with a total of 7 g of DNA which included 3 g of control or polycystin-1 construct, 1 g of AP-1 reporter, 5 ng of RL-null, 100 ng or 500 ng of G␣ construct, and pBS as the filler. A, 100 ng of G␣ i1 , G␣ i2 , G␣ i3 , G␣ 12 , and G␣ 13 . B, 500 ng of G␣ i1 , G␣ i3 , G␣ q , and G␣ 12 . Data are expressed as relative luciferase units (RLUs). Independent experiments were carried out in triplicate (n ϭ 3). Error bars represent S.D. Significant differences between G␣-stimulated and -unstimulated control samples are indicated by one asterisk (p Ͻ 0.05), three asterisks (p Ͻ 0.005), or four asterisks (p Ͻ 0.001), as determined by one-way ANOVA. Western blots demonstrate expression of the polycystin-1 (Ig-PKD222) and control (Ig-0) constructs from one of each of the experiments. PKD1284, only lacks the N-terminal extracellular domain of polycystin-1 but contains all of the putative transmembrane domain segments (8) and all of the extracellular and intracellular loops, as well as the C-tail. Thus, the product encoded by this construct would be expected to adopt the native polycystin-1 structure over this region of the protein. As with the experiments involving the smaller C-tail constructs, sIg-PKD1284 was also capable of stimulating JNK and AP-1. Furthermore, the sIg-PKD1284-mediated JNK activation was effectively inhibited by ␤ARK-ct and by DN G␣ i2 . However, contrary to our observations utilizing the smaller C-tail constructs, the sIg-PKD1284-mediated JNK activation could be stimulated somewhat by WT G␣ i2 at low concentrations (Fig.  4B). Despite this stimulation, we could only partially inhibit polycystin-1 JNK activation with pertussis toxin. Thus, to identify other G protein families that may be activated by polycystin-1, we transfected cells with three different G␣ i subunits, with G␣ q and with G␣ 12 and G␣ 13 subunits. Although all three G protein families were found to be capable of stimulating polycystin-1 activation of AP-1, G␣ 12 and G␣ 13 were the most effective. Furthermore, polycystin-1-induced AP-1 activity was inhibited with a DN p155RhoGEF construct that specifically inhibits G␣ 12 and G␣ 13 . Our results demonstrate that polycystin-1 is able to couple with G 12 class heterotrimeric G proteins to stimulate AP-1, and they suggest the involvement of other G protein families in JNK/AP-1 signaling. Apparently, the C-tail alone is sufficient for coupling with and activating G protein signaling.
Polycystin-1 has been shown to regulate Ca 2ϩ flux through an interaction with polycystin-2 (12,13). Polycystin-2 has similarity to the ␣1E-1 subunit of a voltage-activated calcium channel, a class of calcium channels whose activity is potentially regulated by G␤␥ subunits (52,53). Binding of G␤␥ to the ␣1 subunit of these channels results in kinetic slowing and steady-state inhibition of the current. This inhibition may be due to competition (or in some cases cooperative interactions) between G␤␥ and a channel ␤ subunit that regulates membrane targeting and other biophysical properties of ␣1 channels (54,55). Interestingly, antibodies directed against the channel ␤ subunit block the GTPase activity of G␣ o in neuronal membranes, suggesting that the channel ␤ subunit has GAP activity (56). GAP activity may be important in channel function as it promotes the reassociation of G␤␥ and G␣GDP in an inactive heterotrimeric state, thereby preventing binding of G␤␥ to the ␣1 channel subunit. Polycystin-2 is also capable of activating JNK and AP-1 via a pathway dependent on RhoA/Rac-1/Cdc42 (57). The ability of polycystin-2 to regulate Ca 2ϩ flux may account for its ability to activate JNK. Calcium mobilization has been implicated in JNK signaling, as chelation of intracellular and/or extracellular Ca 2ϩ is capable of inhibiting JNK activation via the m2 (58) or m3 (59) muscarinic acetylcholine receptors. Indeed, a recent paper (60) has demonstrated that polycystin-1 can regulate Ca 2ϩ and K ϩ channels in a pertussis toxin-sensitive, G␤␥-dependent fashion, thus supporting this idea.
JNK and AP-1 play critical roles in numerous cellular processes, including cell cycle regulation, cell growth, differentiation, apoptosis, and inflammation (61). JNK signaling has been characterized extensively in Drosophila, being involved in developmental pathways that control planar polarity (62), epidermal adhesion and integrity (63), and dorsal closure (64). Polycystic kidney disease is characterized by epithelial cell proliferation, a dedifferentiated epithelial phenotype, and abnormal epithelial polarity (4,15,65). Polycystin-1 has also been shown to have a direct role in the maintenance of the vasculature and of epithelial integrity (66) and in mediating intercellular adhesion (67). Thus, it is reasonable to envision that JNK and AP-1 signaling could play key roles in polycystin-1 function and cystic disease progression. Our evidence supports the idea that polycystin-1 is a GPCR (Fig. 7). Binding to an unidentified extracellular ligand (or to itself through homotypic interactions) could regulate heterotrimeric G protein signaling that would in turn lead to the control of diverse cellular processes that are misregulated in PKD, such as cell proliferation, differentiation, polarity, and fluid secretion. Endogenous AP-1 was assayed in 293T cells transfected as described under "Experimental Procedures," with a total of 7 g of DNA which included 3 g of control or polycystin-1 construct, 1 g of DN p115 or control plasmid, 1 g of AP-1 reporter, 5 ng of RL-null, and pBS as the filler. Data for two independent experiments, each carried out in triplicate (n ϭ 3), are expressed as direct firefly luciferase activity. AP-1 activity was inhibited by DN p115 ϳ24 and ϳ38%, representing 42 and 30% drops in fold increase, respectively. Error bars represent S.D. Significant differences between DN p115-inhibited and control samples are indicated by one asterisk (p Ͻ 0.05) or two asterisks (p Ͻ 0.01), as determined by one-way ANOVA. Western blots demonstrate expression of the polycystin-1 (Ig-PKD222) and control (Ig-0) constructs. FIG. 7. Model of G protein-coupled polycystin-1 function. Our data demonstrate that polycystin-1-mediated JNK and AP-1 activation are regulated by the G␣ and G␤␥ subunits of heterotrimeric G proteins. G␤␥ subunits could potentially be involved in the regulation of polycystin-2 channel activity. The extracellular event that initiates polycystin-1-mediated G protein signaling remains unknown.