Polycystin 2 Interacts with Type I Inositol 1,4,5-Trisphosphate Receptor to Modulate Intracellular Ca2+ Signaling*

Autosomal dominant polycystic kidney disease, a common cause of renal failure, arises from mutations in either the PKD1 or the PKD2 gene. The precise function of both PKD gene products polycystins (PCs) 1 and 2 remain controversial. PC2 has been localized to numerous cellular compartments, including the endoplasmic reticulum, plasma membrane, and cilia. It is unclear what pools are the most relevant to its physiological function as a putative Ca2+ channel. We employed a Xenopus oocyte Ca2+ imaging system to directly investigate the role of PC2 in inositol 1,4,5-trisphosphate (IP3)-dependent Ca2+ signaling. Cytosolic Ca2+ signals were recorded following UV photolysis of caged IP3 in the absence of extracellular Ca2+. We demonstrated that overexpression of PC2, as well as type I IP3 receptor (IP3R), significantly prolonged the half-decay time (t½) of IP3-induced Ca2+ transients. However, overexpressing the disease-associated PC2 mutants, the point mutation D511V, and the C-terminally truncated mutation R742X did not alter the t½. In addition, we found that D511V overexpression significantly reduced the amplitude of IP3-induced Ca2+ transients. Interestingly, overexpression of the C terminus of PC2 not only significantly reduced the amplitude but also prolonged the t½. Co-immunoprecipitation assays indicated that PC2 physically interacts with IP3R through its C terminus. Taken together, our data suggest that PC2 and IP3R functionally interact and modulate intracellular Ca2+ signaling. Therefore, mutations in either PC1 or PC2 could result in the misregulation of intracellular Ca2+ signaling, which in turn could contribute to the pathology of autosomal dominant polycystic kidney disease.

Autosomal dominant polycystic kidney disease (ADPKD) 2 is one of the most common causes of renal failure (1). It arises from mutations in either the PKD1 or the PKD2 genes (1)(2)(3)(4)(5). The primary phenotype of ADPKD is the presence of numerous fluid-filled cysts in the kidney, liver, pancreas, and intestine (4). Mutations in PKD1 account for the majority (ϳ85%) of ADPKD cases and are associated with more severe clinical presentations and an earlier onset of renal failure (1,2,4,5). Although both PKD genes have been identified, the molecular mechanisms leading to these clinical symptoms are still not fully clear. Hence, it is important to understand the biological roles of the PKD gene products polycystins (PCs) 1 and 2.
PC1 is a 4,302-amino-acid integral membrane protein that has 11 putative transmembrane domains, a long extracellular N terminus, and a short intracellular C terminus (6). The large extracellular portion contains a novel combination of motifs originally found in other proteins that are predicted to be involved in cell-cell and cell-matrix interactions (1-3, 6, 7). Its C terminus contains a coiled-coil domain responsible for interacting with PC2 (8,9). PC2 is a 968-amino-acid protein containing six transmembrane domains (10). Transmembrane domains 5 and 6 share a significant similarity with TRPC1, a member of the transient receptor potential cation channel family (10 -12). Both the N and the C termini are cytosolic. The C terminus of PC2 contains an ER retention signal, an EF-hand motif, and a coiled-coil domain (10,11,13). Electrophysiological studies indicate that PC2 is a Ca 2ϩ -activated, non-selective cation channel with multiple subconductance states and a high permeability to Ca 2ϩ (14 -17). Although it is well accepted that PC1 is expressed in the plasma membrane, the subcellular localization of PC2 has been controversial (11,13,15,18). PC2 has been localized to the ER, the plasma membrane, and the primary cilia of kidney cells (15, 16, 18 -22). On one hand, it has been recently proposed that PC1 and PC2 interact in the primary cilia, in which PC1 acts as a mechanical receptor and PC2 as a Ca 2ϩ influx channel (22). On the other hand, it has also been suggested that PC2 functions as an intracellular Ca 2ϩ channel (16). In addition, evidence suggests that PC1 mutants can cause the mislocalization of PC2 from the cell surface to the ER, which might contribute to the pathogenesis of ADPKD (15). Therefore, it is important to investigate the role of the ER-localized PC2 in regulating intracellular Ca 2ϩ signaling.
To date, several studies have attempted to investigate the role of PC2 in Ca 2ϩ signaling pathways based on intracellular Ca 2ϩ measurements (16,23,24). All of these studies relied on plasma membrane G proteincoupled receptors to stimulate the intracellular IP 3 pathway. This could also activate various intracellular signaling pathways and cause multiple effects independent of PC2. To directly investigate the role of PC2 in regulating IP 3 -dependent Ca 2ϩ signaling pathways, we employed a Xenopus oocyte confocal Ca 2ϩ image-recording system. As a well established expression system, Xenopus oocyte also offers four advantages in performing Ca 2ϩ imaging experiments: 1) intracellular Ca 2ϩ release can be initiated by UV photolysis of caged IP 3 to bypass the complicated plasma membrane signaling transduction pathways (25,26); 2) cytosolic Ca 2ϩ signaling can be recorded in the absence of extracellular Ca 2ϩ , which eliminates the contribution of Ca 2ϩ influx; 3) cytosolic Ca 2ϩ signaling can be imaged with a high spatial resolution because of its large size (27)(28)(29)(30)(31)(32)(33)(34); 4) only the type I IP 3 receptor (IP 3 R) Ca 2ϩ release channel is endogenously expressed, because there is no endogenous expression of the ryanodine receptor (35,36). Therefore, this simplified Ca 2ϩ recording system allows us to focus on intracellular Ca 2ϩ signaling com-ponents and to directly investigate the role of PC2 in IP 3 -dependent Ca 2ϩ signaling.
Our studies obtained from Ca 2ϩ imaging assays as well as biochemical assays indicate that PC2 functionally interacts with the IP 3 R to modulate intracellular Ca 2ϩ signaling. IP 3 -mediated intracellular Ca 2ϩ signaling is one of the most important intracellular signaling pathways. Misregulation of intracellular Ca 2ϩ signaling can alter cell proliferation as well as apoptosis. Thus, mutations in PC2 would misregulate IP 3mediated Ca 2ϩ signaling, consequently altering cellular proliferation and/or apoptosis and ultimately contributing to the pathological development of ADPKD.

MATERIALS AND METHODS
Constructs-Myc-PC2, encoding the Myc-tagged wild type PC2, and its Myc-tagged C-terminal truncation, R742X, which lacks the EF-hand, the ER retention signal, and the coiled-coil domain, have been described in Hanaoka et al. (15). Another PC2 C-terminal truncation R872X has been described as PKD2-M1 in Qian et al. (8). Compared with R742X, the R872X mutation contains the coiled-coil domain but lacks the very end of the C-terminal tail of PC2. To generate the Myc-tagged R872X construct, the sequence corresponding to bp 2195-3140 of human PKD2 cDNA (GenBank TM accession number NM_000297) containing this mutation was subcloned into Myc-PC2. To generate the Myctagged D511V construct, a cDNA fragment containing the point mutation D511V was subcloned into Myc-PC2. The Myc-tagged TM-CT construct was generated by subcloning the human PKD2 cDNA sequence corresponding to codons 577-968 into vector CS2-6MT (as described in www.xenbase.org/WWW/Marker_pages/PlasMaps/ CS2ϩMT.html). The resulting TM-CT construct contains the last transmembrane domain and the intact C terminus of PC2. All of the constructs were confirmed for their authenticity by sequencing prior to use. The construct encoding rat type I IP 3 R was obtained from Dr. P. Camacho (University of Texas Health Science Center at San Antonio, TX).
Cell Culture and Transient Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 50 units/ml penicillin and 50 g/ml streptomycin in a humidified atmosphere of 5% CO 2 /95% air. Transient transfection was performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.
Oocyte Methods and Confocal Ca 2ϩ Imaging with Oocytes-Synthetic mRNAs were prepared using the SP6 MEGASCript kit according to the manufacturer's protocol (Ambion, Austin, TX). Stages V-VI defolliculated albino Xenopus oocytes were injected with 25-50 ng of mRNA of wild type PC2 or its mutants and cultured in 50% L-15 medium for 4 -9 days at 16°C (33). To perform the Ca 2ϩ imaging assay, oocytes were first co-injected with fluorescent Ca 2ϩ indicator Oregon green II (OG2, 12.5 M final; Invitrogen) and NPE-caged IP 3 (5 M final; Invitrogen) 60 min before imaging. Ca 2ϩ release was initiated by photolysis of caged IP 3 using a UV laser (355/363 nm). Adjusting the power level and/or the duration of exposure time generated different doses of IP 3 . Ca 2ϩ images were acquired at a rate of 0.5 s/frame on a confocal laser scanning microscope (model LSM510, Zeiss) using a 10ϫ objective (numerical aperture ϭ 0.50) at zoom 1. Imaging was performed in recording buffer containing 96 mM NaCl, 2 mM KCl, 2 mM MgCl 2 , 5 mM HEPES, pH 7.5, 1 mM EGTA.
Western Blot-Oocytes were harvested and processed as previously described (33). Oocytes were incubated on ice for 30 min with lysis buffer containing 20 mM Tris-HCl, pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 2% Triton X-100 supplemented with a protease inhibitor mixture (Roche Applied Science). After homogenization, the samples were centrifuged for 15 min at 4,500 ϫ g, and the supernatants were collected. COS-7 cells were harvested and processed as previously described (37). Briefly, the cells were solubilized in lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, and complete protease inhibitor). The cell lysates were spun at 14,000 ϫ g for 15 min at 4°C to pellet insoluble material.
The protein concentrations of the supernatants from both cell types were quantified with a bicinchoninic acid protein assay kit (Pierce). After incubation in Laemmli buffer at 42°C for 30 min, the protein samples were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane. Mouse anti-Myc monoclonal antibody (in a dilution of 1:1,600; Santa Cruz Biotechnology) was used for Myc-polycystin 2 detection, rabbit anti-IP 3 R-1 polyclonal antibody (in a dilution of 1:1,000, Abcam, Cambridge, MA) for endogenous and exogenous expression of type I IP 3 R, and mouse anti-actin monoclonal antibody (in a dilution of 1:2,000, Chemicon, Temecula, CA) for actin detection. After probing with horseradish peroxidase-conjugated sheep antimouse IgG secondary antibody (in a dilution of 1:10,000; Amersham Biosciences), Western blots were visualized by ECL detection reagents according to the manufacturer's protocol (Amersham Biosciences).
Co-immunoprecipitation Assay-Total protein extracts from oocytes or COS-7 cells overexpressing PC2 or its mutants were obtained as described under "Western Blot." 1 ⁄ 20 of the extraction samples were saved to detect expression levels of PC2 mutants. Primary antibody against type I IP 3 R (Abcam) was added. Samples were rotated overnight at 4°C and followed by the addition of 20 l of protein A/G-agarose beads (Santa Cruz Biotechnology). After rotating for 2 h, the beads were spun down and washed three times with lysis buffer. The immunocomplexes were resolved by SDS-PAGE. After being transferred to a polyvinylidene difluoride membrane, Myc monoclonal antibody was used to detect the PC2. Alternatively, Myc monoclonal antibody was used to pull down immunocomplexes, and the endogenous IP 3 R was detected by anti-IP 3 R antibody.
Imaging Analysis-Ca 2ϩ images were analyzed using the public domain National Institutes of Health ImageJ program (available at rsb.info.nih.gov/ij.).
Statistics Analysis-Statistical significance was determined by Student's t test or one-way analysis of variance as appropriate and accepted as p Ͻ 0.05.

RESULTS
Overexpression of PC2 Enhances IP 3 -dependent Ca 2ϩ Release-We employed a Xenopus oocyte Ca 2ϩ imaging system to record the cytosolic Ca 2ϩ signaling following UV photolysis of caged IP 3 in the absence of extracellular Ca 2ϩ . We first overexpressed Myc-tagged PC2 in this system to see whether PC2 overexpression altered the IP 3 -dependent Ca 2ϩ signaling pathway. We demonstrated that overexpression of Myctagged PC2 significantly prolonged the half-decay time (t1 ⁄ 2 ) of IP 3 -in-duced Ca 2ϩ transients compared with H 2 O-injected control oocytes without changing the initial amplitude (Fig. 1A). In this recording system, Ca 2ϩ transients are initiated by the opening of IP 3 R. t1 ⁄ 2 could be influenced by both Ca 2ϩ release from the ER and Ca 2ϩ clearance from the cytosol. Longer t1 ⁄ 2 may indicate enhanced Ca 2ϩ release and/or reduced Ca 2ϩ clearance. In oocytes, Ca 2ϩ clearance is mainly regulated through the activity of Sarco ER Ca 2ϩ ATPases (SERCAs) (26,27,38,39). To examine whether the activity of endogenous SERCAs were affected by overexpression of PC2, an in vivo Ca 2ϩ imaging method was used to indirectly measure the activity of SERCAs. In this assay, we elevated cytosolic Ca 2ϩ by UV photolysis of the caged Ca 2ϩ reagent and recorded the decay of the Ca 2ϩ signal. The t1 ⁄ 2 under this circumstance should primarily be determined by SERCA activity. We found that the t1 ⁄ 2 was not significantly different between the control and PC2-overexpressed oocytes, indicating a similar SERCA activity in these two groups (Fig. 1B). In oocytes, overexpressed PC2 was predominantly localized in the ER, as detected by endoglycosidase H (Endo H) sensitivity analysis (Fig. 1, C and D).
Careful analysis of IP 3 -induced Ca 2ϩ transients revealed two components: a rapid Ca 2ϩ response and a delayed Ca 2ϩ response. The rapid Ca 2ϩ response represents the initial opening of IP 3 R in response to IP 3 stimuli, whereas the delayed Ca 2ϩ response might represent the subsequent Ca 2ϩ -activated Ca 2ϩ release from intracellular stores. An unchanged initial amplitude and initial decay slope following IP 3 stimulation indicate the similar rapid Ca 2ϩ response between H 2 O-and PC2-injected oocytes. To further examine whether the prolonged t1 ⁄2 associated with PC2 overexpression is attributed from the delayed Ca 2ϩ response, we isolated and compared the delayed Ca 2ϩ response between H 2 O control and PC2-overexpressed oocytes as detailed below. Ca 2ϩ transients from the H 2 O control group were subdivided into two subgroups ( Fig. 2A). Approximately 43% of the oocytes (subgroup I) exhibited predominantly the rapid Ca 2ϩ response, whereas the delayed component was too small to be reliably analyzed. The decay of these curves fit well with the first order exponential decay equation (Y ϭ A 1 ϫ Exp(ϪX/T 1 ) ϩ Y 0 , R 2 Ն 0.99) ( Fig. 2A). Ca 2ϩ transients from the remaining oocytes (subgroup II) containing the delayed response with larger amplitude thus could not be fit well with a single exponential (0.9 Ͻ R 2 Յ 0.98). To isolate the delayed Ca 2ϩ  response component, we first obtained an average decay equation from subgroup I. We then subtracted this component from each recording curve in subgroup II after normalizing the peak values. As shown in Fig.  2A, this isolated component appears as a small wave and could be considered as an endogenous delayed Ca 2ϩ release from intracellular stores. We averaged all of the delayed Ca 2ϩ response curves from subgroup II in the H 2 O control (Fig. 2B). The rising curve was well fit with a first order exponential activation equation (R 2 ϭ 0.97); A 1 , T 1 , and Y 0 were Ϫ0.285 Ϯϩ0.007, 34.8 Ϯ 2.3, and 0.305 Ϯ 0.008 respectively.
We then applied this procedure to Ca 2ϩ transients from PC2-overexpressed oocytes. To facilitate analysis, the Ca 2ϩ transients were subdivided into three subgroups ( Fig. 2A, I-III) according to the value of R 2 (I, R 2 Ն 0.99; II, 0.9 Ͻ R 2 Յ 0.98; and III, R 2 Ͻ 0.9). The percentages of distribution are: I, ϳ18%; II, ϳ26%; and III, ϳ56%, respectively. It has been reported that there is a variance of expression efficiency among oocytes (40), and we are confident that the variance in the PC2-overexpressed group is largely due to the different expression efficiency among individual oocytes. In subgroup I from PC2-injected oocytes, the parameters of the first order exponential decay equation were not significantly different from the H 2 O-injected control; A 1 , T 1 , and Y 0 are 2.2 Ϯ 0.1, 41.3 Ϯ 4.0, 0.13 Ϯ 0.06 (H 2 0 subgroup I) and 2.1 Ϯ 0.2, 37.4 Ϯ 3.1, 0.28 Ϯ 0.10 (PC2 subgroup I), respectively. We isolated and averaged all of the delayed Ca 2ϩ response curves from subgroups II and III in the PC2-overexpressed group (Fig. 2B). The activation kinetics and peak values of the delayed Ca 2ϩ response curves were significantly different from those of the H 2 O-injected oocytes. In the PC2-overexpressed group, A 1 , T 1 , and Y 0 from the activation equation (R 2 ϭ 0.99) of the rising curve were Ϫ0.349 Ϯ 0.002, 14.1 Ϯ 0.2, and 0.343 Ϯ 0.001 respectively. The delayed Ca 2ϩ release from the PC2-overexpressed group reached higher levels more rapidly than the H 2 O-injected oocytes. Taken together, these results suggested that overexpression of PC2 modulated intracellular IP 3 -dependent intracellular Ca 2ϩ signals and enhanced Ca 2ϩ release upon IP 3 stimulation.
Overexpression of D511V Reduces IP 3 -dependent Ca 2ϩ Release-To further test whether the enhanced Ca 2ϩ release was associated with the channel activity of PC2, Ca 2ϩ imaging experiments were carried out in oocytes overexpressing mutant forms of PC2. Two disease-associated mutations of PC2 were used in this experiment. D511V, a point mutation in the third transmembrane domain, has previously been shown to be a dead channel mutation by electrophysiological studies (16). The  DECEMBER 16, 2005 • VOLUME 280 • NUMBER 50 second mutant, R742X, is a C terminus truncation that lacks the EF-hand, the ER retention signal, and the coiled-coil domain. Although its channel activity has been controversial, previous studies suggest that overexpressed R742X is predominantly localized near the plasma membrane, whereas the wild type PC2 is in the ER (15,19,41). We found that neither of the PC2 mutants affect the t1 ⁄ 2 of IP 3 -induced Ca 2ϩ transients compared with the H 2 O-injected control (Fig. 3, A and B). These results would support the hypothesis that PC2 itself functions as an intracellular Ca 2ϩ release channel.

PC2 Interacts with IP 3 R
Interestingly, we found that the amplitude of IP 3 -induced Ca 2ϩ transients were significantly reduced in D511V-overexpressed oocytes (Fig.  3B). We further recorded IP 3 -induced Ca 2ϩ transients under different doses of IP 3 . Again, we observed that D511V mediated a low percentage of detectable Ca 2ϩ response as well as the reduced amplitude following each applied IP 3 concentration (Fig. 3C). To demonstrate whether the endogenous expression of IP 3 R was changed by overexpression of D511V, we measured the endogenous expression of IP 3 R, and we found that the endogenous expression level of IP 3 R in the oocyte was not affected by overexpression of any of the constructs we tested (Fig. 3, D  and E). These observations indicated that overexpression of D511V might functionally modulate the activity of IP 3 R.
PC2 Interacts with IP 3 R-By performing a biochemical co-immunoprecipitation assay with PC2-overexpressed oocytes, a novel association between PC2 and the endogenous type 1 IP 3 R was observed (Fig. 4A). This is not unique to the oocyte overexpression system, because we can detect this interaction in the mammalian cell expression system as well (Fig. 4B). The PC2 C terminus-truncated mutation R742X cannot be co-immunoprecipitated with IP 3 R (Fig. 5A). Because this might be due to the mislocation of R742X from the ER into the plasma membrane, we performed an Endo H sensitivity analysis and found that overexpressed R742X is sensitive to Endo H treatment, indicating its predominant ER localization (Fig. 5B). To further identify the domain in PC2 that is essential for the association, co-immunoprecipitation experiments were carried out in oocytes expressing various PC2 mutants, including D511V, R742X, R872X, and the PC2 C terminus (TM-CT). We found that TM-CT, R872X, and D511V still associated with IP 3 R (Fig. 5C). Compared with the wild type PC2, both TM-CT and R872X were weakly associated with IP 3 R, whereas D511V was more strongly associated. These results indicate that PC2 interacts with IP 3 R through its C terminus, most likely via its coiled-coil domain.
Overexpression of Type I IP 3 R Associates with Prolonged t 1/2 and Reduced Amplitude-We then investigated whether overexpression of type I IP 3 R could alter the IP 3 -dependent Ca 2ϩ signaling in our recording system. We found that overexpression of IP 3 R significantly prolonged the t1 ⁄ 2 of IP 3 -induced Ca 2ϩ transients compared with H 2 O-injected control oocytes (Fig. 6A). This phenotype started to appear on day 7 of post-injection (Fig. 6B). The initial amplitudes, however, were not significantly different when we employed a high affinity Ca 2ϩ indicator, Oregon Green 2 (OG2, K d ϳ 580 nM). To eliminate the possibility that the high affinity Ca 2ϩ dye could saturate and/or disturb the intracellular Ca 2ϩ signaling, we repeated the experiments using a low affinity Ca 2ϩ indicator, OGB-5N (K d ϳ 25 M). Under this circumstance, we found that IP 3 R-overexpressed oocytes exhibited a significantly lower initial amplitude also starting on day 7 of post-injection (Fig. 6, A and B). By Western blot, we demonstrated that the expression of exogenous IP 3 R reached to its peak on day 7 of post-injection (Fig. 6C). Taken together, these observations demonstrated overexpression of IP 3 R facilitated IP 3 -dependent Ca 2ϩ release from intracellular stores. . Polycystin 2 physically associates with IP 3 R. A, co-immunoprecipitation of PC2 and endogenous IP 3 R in oocytes overexpressed Myc-PC2. Endogenous IP 3 R was immunoprecipitated (IP) with a polyclonal rabbit anti-IP 3 R antibody (Abcam), and immunocomplexes were analyzed by Western blots (IB) with either mouse anti-Myc monoclonal antibody (Santa Cruz Biotechnology) or anti-IP 3 R antibody (left panels). Alternatively, Myc-PC2 was immunoprecipitated with anti-Myc antibody, and immunocomplexes were analyzed with Western blots (right panels). Expression levels of Myc-PC2 and endogenous IP 3 R before immunoprecipitation are shown labeled as Input. B, co-immunoprecipitation (Co-IPs) of PC2 and endogenous IP 3 R in COS-7 cells overexpressed Myc-PC2. Two detergents were used, 1% Triton X-100 (TNX) or 1% of Nonidet P-40 (NP40). IB, immunoblot. Un, untransfected control.

FIGURE 5. C terminus of PC2 is essential for the interaction between PC2 and IP 3 R.
A, co-immunoprecipitation of endogenous IP 3 R with wild type PC2 or its truncation, R742X, in the oocyte overexpression system. B, overexpressed R742X in oocyte is also sensitive to Endo H treatment, indicating its ER localization. C, co-immunoprecipitation of endogenous IP 3 R with wild type PC2 or its mutants, D511V, TM-CT, R742X, and R872X, in the oocyte overexpressing system.
Overexpression of the C-terminal Tail of PC2 Exhibits Prolonged t 1/2 and Reduced Amplitude-Because the C terminus of PC2 (TM-CT) itself can associate with IP 3 R, we further performed a Ca 2ϩ imaging assay to determine whether overexpression of TM-CT would modulate IP 3 -dependent Ca 2ϩ signaling. We found that the decay curves, obtained from TM-CT-overexpressed oocytes, could be divided into four subgroups, I-IV (Fig. 7, A and B). The first three subgroups (I, ϳ27%; II, ϳ15%; and III, ϳ27%) exhibited the same phenotypes as the subgroups in PC2-overexpressed oocytes. Interestingly, a significant fraction of oocytes overexpressing TM-CT (subgroup IV, ϳ31%) displayed a Ca 2ϩ plateau upon IP 3 stimulation. The initial amplitude is significantly lower in the TM-CT overexpression group (Fig. 7C). We also measured the endogenous activities of SERCAs in these two groups and found that there was no significant difference (data not shown). These results indicated that the endogenous IP 3 R activity could be modulated by overexpression of TM-CT. Taken together with previous observations that overexpression of IP 3 R also associated with prolonged t1 ⁄ 2 and reduced amplitude, the phenotype associated with TM-CT overexpression suggested an enhanced activity of IP 3 R.

DISCUSSION
PC2 has been characterized as a Ca 2ϩ -permeable cation channel, although there is still a controversy over whether PC2 acts as an intracellular Ca 2ϩ release channel or a cilia/plasma membrane Ca 2ϩ channel  (15,16,22). Here, using a Xenopus oocyte Ca 2ϩ imaging system, we demonstrated that overexpression of PC2 significantly prolonged the t1 ⁄ 2 of IP 3 -dependent Ca 2ϩ release from intracellular stores. Careful analysis of Ca 2ϩ transients further revealed that the prolonged decay was mainly attributed to the enhanced delayed component of the Ca 2ϩ response, which might represent the Ca 2ϩ -activated Ca 2ϩ release. This phenomenon is associated with the channel activity of PC2, because neither of the disease-associated PC2 mutants (D511V nor R742X) alter the t1 ⁄ 2 of IP 3 -induced Ca 2ϩ transients. On one hand, these observations could support the hypothesis that PC2 functions as an intracellular Ca 2ϩ release channel, consistent with the previous report from Koulen et al. (16). Moreover, we demonstrated the physical interaction between PC2 and IP 3 R. This observation provides additional evidence and rationale to support the model that ER-localized PC2 might need to be physically associated with IP 3 R to get close enough to be activated by the IP 3 pathway.
On the other hand, PC2 could functionally modulate IP 3 R channel activity as well. Several observations from the present study support this hypothesis. 1) Overexpression of the C terminus of PC2 (TM-CT) significantly prolonged the t1 ⁄ 2 of IP 3 -induced Ca 2ϩ response. Approximately 31% of TM-CT-overexpressed oocytes exhibited a striking phenotype of Ca 2ϩ plateau upon IP 3 application, indicating a somewhat dysfunctional closure of the IP 3 R. This is the most reliable explanation because of the following reasons: first, TM-CT itself is unlikely to carry a channel activity; second, the activity of the SERCAs for Ca 2ϩ clearance was not changed; and third, TM-CT was able to physically interact with IP 3 R. 2) Overexpression of TM-CT significantly reduced the initial amplitude of IP 3 -induced Ca 2ϩ transient. Keep in mind that overexpression of IP 3 R itself also exhibits reduced amplitude; this could be explained by the increased basal level activity of IP 3 R and the subsequently depleted ER Ca 2ϩ stores. As a consequence of the pre-emptied ER Ca 2ϩ stores, oocytes overexpressing IP 3 R or TM-CT exhibit a reduced rapid Ca 2ϩ response upon IP 3 stimulation. 3) Overexpression of D511V also mediated reduced amplitude upon IP 3 stimulation. Because D511V could physically interact with IP 3 R, one of the reasonable explanations is that D511V might also enhance the basal level of IP 3 R activity to deplete the Ca 2ϩ stores. 4) Finally, R742X, the construct that could not interact with IP 3 R, does not affect either the amplitude or the t1 ⁄ 2 of the IP 3 -induced Ca 2ϩ transients.
Our observation of the reduced Ca 2ϩ response mediated by overexpression of D511V is in contrast with the previous report from Koulen et al. (16), where they did not demonstrate any difference in amplitude of vasopressin-induced Ca 2ϩ transients between control and D511Voverexpressed LLC-PK 1 cells. We would argue that our recording system using photolysis of caged IP 3 eliminates the contribution from plasma membrane receptor signaling. This is beneficial, because the expression of the plasma membrane receptors could be subsequently changed by overexpression of D511V.
The physiological function of PC1 remains elusive. PC2 has been previously reported to interact with PC1 through its coiled-coil domain (8,9). Our data indicated that PC2 might interact with IP 3 R also through its coiled-coil domain. It is quite possible that PC1 and IP 3 R competitively bind to PC2. Therefore, PC1 has the potential to modulate Ca 2ϩ signaling by competitively binding to PC2 and modulating the interaction between PC2 and IP 3 R. Both overexpression and deficiency of PC1 are reported to cause cysts in the kidney (42)(43)(44). This paradox could be explained by the misregulation of the interaction between PC2 and IP 3 R and subsequently alteration of IP 3 -mediated Ca 2ϩ signaling. Recently it was reported that PC1 stably expressed Madin-Darby canine kidney cells exhibit an accelerated rate of decay in ATP-induced Ca 2ϩ transients compared with nontransfected Madin-Darby canine kidney cells (45). This observation is consistent with our model that PC1 overexpression could disrupt the endogenous interaction between PC2 and IP 3 R. The disturbance of their interaction might reduce the activity of endogenous PC2 and/or IP 3 R. Thus, the accelerated decay associated with PC1 overexpression could be due to the reduced IP 3 -dependent Ca 2ϩ release from ER.
Nauli et al. (22) demonstrated that PC1 and PC2 localized on cilia to mediate fluid flow sensing, and IP 3 R was not involved in the downstream pathway. Although their evidence indicates that ER-localized PC2 does not participate in cilia mechanosensation (22), it does not necessarily exclude the importance of a role for the IP 3 -dependent pathway in the pathogenesis of ADPKD. Previous publications report that a significant amount of ATP, as well as the purinergic signal transduction components, are resident in a cystic environment (46). As one of the downstream pathways of ATP stimulation, a dysregulated IP 3 pathway could change ER and cytosolic Ca 2ϩ homeostasis. Inconsistent with this model, it has been reported that in pkd2 ϩ/Ϫ mice, the basal Ca 2ϩ levels in both cytosol and sarcoplasmic reticulum stores are significantly reduced in freshly dissociated smooth muscle cells (47). In addition, low levels of cytosolic Ca 2ϩ were associated with an elevated level of cAMP, which in turn contributes to abnormal cell proliferation and apoptosis in vascular smooth muscle cells (48).
ADPKD patients also suffer from cardiovascular complications, including intracranial aneurysms/aneurysmal ruptures and thoracic aortic dissections (47,48). Type I IP 3 R appears to be the predominant subtype expressed in the aorta and basilar and mesenteric arteries (49). The IP 3 signaling pathway plays an important role in controlling vascular tone by participating in the contractile response of smooth muscle cells and synthesis of nitric oxide from endothelial cells in blood vessels (50 -52).
In conclusion, we suggest that PC2 and IP 3 R could functionally interact with each other to modulate IP 3 -dependent intracellular Ca 2ϩ signaling. On one hand, PC2 itself may function as a Ca 2ϩ channel and can be subsequently activated by the IP 3 signaling pathway. On the other hand, PC2 may be able to modulate the channel activity of IP 3 R. Therefore, mutations in either PC1 or PC2 could result in dysregulation of intracellular Ca 2ϩ signaling and contribute to the pathology of ADPKD, not only in cyst formation but also in the development of cardiovascular complications associated with ADPKD.