Polycystin-2 Activity Is Controlled by Transcriptional Coactivator with PDZ Binding Motif and PALS1-associated Tight Junction Protein*

Autosomal dominant polycystic kidney disease (ADPKD) is the most frequent monogenic cause of kidney failure, characterized by the development of renal cysts. ADPKD is caused by mutations of the polycystin-1 (PC1) or polycystin-2 (PC2) genes. PC2 encodes a Ca2+-permeable cation channel, and its dysfunction has been implicated in cyst development. The transcriptional coactivator with PDZ binding motif (TAZ) is required for the integrity of renal cilia. Its absence results in the development of renal cysts in a knock-out mouse model. TAZ directly interacts with PC2, and it has been suggested that another yet unidentified PDZ domain protein may be involved in the TAZ/PC2 interaction. Here we describe a novel interaction of TAZ with the multi-PDZ-containing PALS1-associated tight junction protein (PATJ). TAZ interacts with both the N-terminal PDZ domains 1–3 and the C-terminal PDZ domains 8–10 of PATJ, suggesting two distinct TAZ binding domains. We also show that the C terminus of PC2 strongly interacts with PDZ domains 8–10 and to a weaker extent with PDZ domains 1–3 of PATJ. Finally, we demonstrate that both TAZ and PATJ impair PC2 channel activity when co-expressed with PC2 in oocytes of Xenopus laevis. These results implicate TAZ and PATJ as novel regulatory elements of the PC2 channel and might thus be involved in ADPKD pathology.

Autosomal dominant polycystic kidney disease (ADPKD) is the most frequent monogenic cause of kidney failure, characterized by the development of renal cysts. ADPKD is caused by mutations of the polycystin-1 (PC1) or polycystin-2 (PC2) genes. PC2 encodes a Ca 2؉ -permeable cation channel, and its dysfunction has been implicated in cyst development. The transcriptional coactivator with PDZ binding motif (TAZ) is required for the integrity of renal cilia. Its absence results in the development of renal cysts in a knock-out mouse model. TAZ directly interacts with PC2, and it has been suggested that another yet unidentified PDZ domain protein may be involved in the TAZ/PC2 interaction. Here we describe a novel interaction of TAZ with the multi-PDZ-containing PALS1-associated tight junction protein (PATJ). TAZ interacts with both the N-terminal PDZ domains 1-3 and the C-terminal PDZ domains 8 -10 of PATJ, suggesting two distinct TAZ binding domains. We also show that the C terminus of PC2 strongly interacts with PDZ domains 8 -10 and to a weaker extent with PDZ domains 1-3 of PATJ. Finally, we demonstrate that both TAZ and PATJ impair PC2 channel activity when co-expressed with PC2 in oocytes of Xenopus laevis. These results implicate TAZ and PATJ as novel regulatory elements of the PC2 channel and might thus be involved in ADPKD pathology.
Transcriptional coactivator with PDZ binding motif (TAZ) 4 was initially identified as a 14-3-3 protein and NHERF-2 (Na ϩ /H ϩ exchange regulator factor-2) binding partner (1). The protein contains a conserved WW domain, a coiled-coil domain, a transactivation domain, and a C-terminal PDZ binding motif (2). In human and mouse, TAZ mRNA is expressed in virtually every tissue with the highest expression in the kidneys (1). TAZ serves as a co-regulator for various transcription factors and has been shown to regulate diverse cellular and developmental processes in bone, muscle, fat, lung, heart, and limbs (2). It promotes cell proliferation, induces epithelial-to-mesenchymal transition, and increases cell migration and invasion (3)(4)(5)(6). Regulation of TAZ is mediated by the Hippo pathway, which is crucial for control of organ size and apical cell polarity in Drosophila and mammals (7,8). Phosphorylation of TAZ by large tumor suppressor kinase (LATS) creates a 14-3-3 binding site and translocates it from the nucleus to the cytoplasm, thereby inhibiting the transcriptional and growth-promoting activity (5). Because of its distinct intracellular localizations, TAZ is believed to have ␤-catenin-like functions by integrating extracellular signals from the plasma membrane and the cytoplasm to the nucleus (9). In addition, it has been suggested that TAZ mediates the cross-talk between the Wnt/␤-catenin and Hippo signaling pathways (10). Recently, it has been shown that TAZ-deficient mice and zebrafish develop polycystic kidneys, indicating that TAZ might be a new player in the field of polycystic kidney disease (PKD) (11)(12)(13).
PKD is characterized by the development of numerous epithelial cysts in the kidneys (14). Although precise molecular mechanisms underlying cyst development remain unknown, it is thought that changes in cell/matrix, cell/cell interactions, Ca 2ϩ signaling, cell proliferation/differentiation, and cell polarity play critical roles in this process (14). Many genes linked to PKD encode proteins associated with primary cilia of epithelial cells, leading to the hypothesis that defects in the ciliary apparatus possess a central etiologic role for PKD development (14,15). In humans, ADPKD is caused by mutations in genes encoding for the transmembrane proteins PC1 and PC2 (14). Both proteins localize to primary cilia, where they act as a mechanosensitive receptor complex promoting Ca 2ϩ influx into the cell (16,17). PC2 has been shown to function as a nonselective Ca 2ϩ -permeable cation channel protein. Dysfunction of PC2 causes PKD, possibly due to the inability of cells to sense mechanical cues that normally regulate tissue morphogenesis (16,17).
In this study, we identified the multi-PDZ domain protein PALS1-associated tight junction protein (PATJ) as a novel interacting partner of TAZ and PC2. The junction and polarity protein PATJ is known to form a complex with CRB3 and PALS1, which, along with the PAR3-PAR6-aPKC (atypical protein kinase C) complex, regulates the apicobasal polarity of epithelial cells (18,19). PATJ possesses 10 PDZ (for PSD95/discs large/zona occludens 1) domains. We show that the N-and C-terminal PDZ domains of PATJ are able to interact with TAZ and PC2. In addition, we demonstrate that TAZ and PATJ regulate PC2 channel activity, suggesting a potential role of the TAZ-PATJ complex in the pathogenesis of PKD.

EXPERIMENTAL PROCEDURES
Plasmids-TAZ cDNA fragments were PCR-amplified from a pcDNA3.1-mTAZ plasmid (a gift from Dr. M. B. Yaffe, Boston, MA). For bacterial expression of TAZ mutants, amplified products were flanked with EcoRI and XhoI sites to facilitate cloning into pGEX4T-1 plasmid (GE Healthcare). We inserted TAZ PCR fragments into pENTR vector (Gateway, Invitrogen) and subsequently into pcDNA3.1-V5 DEST and pDEST22 via LR Clonase according to the manufacturer's instructions (Gateway system, Invitrogen). For generating N-terminal 3xFLAG-tagged TAZ, we used creator splice vectors V37/V180 (a gift from the Pawson group via addgene) (20). Various PATJ deletion mutants for yeast two-hybrid (Y2H) and in vitro binding assays were generated from pEYFP-PATJ (a gift from Dr. B. Margolis, Ann Arbor, MI) and have been described earlier (21,22). A PATJ deletion mutant encoding amino acids (aa) 1-685 was subcloned into pGEX4T-1 via NcoI/NotI sites. A cDNA fragment encoding the last 278 aa of PC2 was amplified from pGEX-PC2 plasmid (23), cloned into pENTR, and shuttled into the pDEST22 and pcDNA3.1-V5 DEST vectors (Invitrogen). For cRNA synthesis and injection in Xenopus oocytes, fulllength TAZ and PATJ cDNAs were PCR-amplified and subcloned into pTLN vector (24). Plasmids suitable for cRNA synthesis for TRPV4 and PC2 were gifts from Dr. W. Kühn, Freiburg, Germany. 5 Y2H Interaction Assay-A truncated version of TAZ was identified as a PATJ-interacting partner in a Y2H screen that has been carried out earlier (21). Y2H co-transformation assays using Saccharomyces cerevisiae strain Y190 were performed as described previously using plasmids containing different PATJ, TAZ, or PC2 mutants, respectively (21,22).
Bacterial Protein Expression and in Vitro Binding Assays-Synthesis of recombinant GST fusion proteins in Escherichia coli strain BL21 was performed as described (21). For GST binding assays, equal amounts of GST fusion proteins were immobilized on GST-Sepharose beads (GE Healthcare) for 3 h at 4°C. Beads were washed five times with PBS and incubated with HEK293T lysates as indicated at 4°C overnight on a rotating platform. After washing (1ϫ PBS), bound proteins were analyzed by SDS-PAGE and Western blot as described previously (21).
Co-immunoprecipitation Assays-Co-IPs were performed as described (21). Briefly, HEK293T cells were transiently transfected using the calcium phosphate method. After 24 h of incubation, cells were washed twice in PBS and lysed in a 1% Triton X-100 lysis buffer. Lysates containing equal amounts of proteins were incubated with anti-FLAG affinity gel (Sigma) overnight on a rotating platform. Beads were washed extensively with lysis buffer and boiled for 5 min at 95°C, and bound proteins were resolved by 12% SDS-PAGE and analyzed by Western blot as already described.
cRNAs and Double Electrode Voltage Clamp-Oocytes were obtained from anesthetized Xenopus laevis frogs as reported earlier (25). They were injected separately or in combination with cRNA (10 ng, 47 nl of double-distilled water) encoding PC2, TRPV4, TAZ, and PATJ. Water-injected oocytes served as controls. 2-4 days after injection, oocytes were impaled with two electrodes (Clark Instruments Ltd., Salisbury, UK), which had a resistances of Ͻ1 megaohms when filled with 2.7 mol/ liter KCl. Using two bath electrodes and a virtual ground head stage, the voltage drop across the serial resistance was effectively zero. Membrane currents were measured by voltage clamping (oocyte clamp amplifier, Warner Instruments, Hamden, CT) in intervals from Ϫ60 to ϩ40 mV, in steps of 10 mV, each 1 s. The bath was continuously perfused at a rate of 5 ml/min. All experiments were conducted at room temperature (22°C).
Materials and Statistical Analysis-All compounds used were of highest available grade of purity and were purchased from Sigma or Calbiochem. Student's t test was used for statistical analysis. p value of Ͻ0.05 was regarded as significant.

RESULTS AND DISCUSSION
We performed a Y2H screen using the last four PDZ domains of PATJ (aa 1239 -1801) as a bait. The same strategy previously identified kidney brain protein (KIBRA) as a novel PATJ interaction partner (21). Another yeast clone identified in this screen encoded for TAZ protein.
PATJ contains 10 PDZ domains in addition to its N-terminal Lin-2 and Lin-7 (L27) domain (Fig. 1D). PDZ domains are modular protein interaction domains comprising 80 -100 aa, usually associating with the last four residues of bound proteins. Corresponding binding motifs can be subdivided into three binding classes (26). At its extreme C terminus, TAZ contains a class I PDZ binding motif Leu-Thr-Trp-Leu (LTWL). To find out whether the PATJ/TAZ interaction is mediated by this motif, we performed Co-IP assays using cell extracts expressing 3xFLAG TAZ full length (FLAG-TAZ fl) and a 3xFLAG TAZ mutant lacking the LTWL (FLAG-TAZ-LTWL) motif (Fig.  1A). Our results show that full-length TAZ co-precipitates with Myc-PATJ, suggesting that both are part of the same protein complex, whereas TAZ lacking the LTWL motif did not precipitate PATJ. These data have been confirmed by GST pulldown (Fig. 1B) as well as Y2H co-transformation assays (Fig. 1C).
To map the TAZ binding site within the PATJ protein, we performed Y2H (Fig. 1D) and GST pulldown assays (Fig. 1E) using various truncated PATJ mutants. We could show that TAZ associates with both the N-terminal PDZ domains 1-3 and the C-terminal PDZ domains 8 -10 of PATJ, suggesting two distinct TAZ binding domains. Binding of PDZ domain proteins to their target molecules may play a role in clustering protein complexes at the plasma membrane. Furthermore, it has been demonstrated that PDZ domain proteins contribute to the formation of macromolecular complexes but may also functionally affect their binding partners (27). TAZ is regulated throughout its different localizations within the cell, e.g. the nucleus, cytoplasm, and plasma membrane, through association with different protein partners. Most of the identified TAZ binding partners are transcription factors at the nucleus, but its binding to 14-3-3 protein has been shown to result in a cytosolic translocation (2). So far, only one protein, the PDZ domaincontaining protein NHERF-2, has been described to link TAZ to the plasma membrane (1). Its novel interaction with the polarity protein PATJ described in this study might link TAZ to apical cell-cell junctions, suggesting a putative role in the establishment or maintenance of epithelial cell polarity. This is of particular interest with regard to the participation of TAZ in polycystic kidney disease because it has been hypothesized that tight junction and polarity proteins might also be involved in PKD (28). A well maintained apicobasal polarity is fundamental to renal tubule development and maturation because it coordinates the proper distribution of receptors, transporters, and channels (29,30). Cell polarity defects may result in aberrant renal epithelial organization and tubulogenesis and consequently induce cystogenesis in the epithelial tissues. Several abnormalities in the polarized distribution of membrane proteins and cell-cell contact proteins such as E-cadherin have been reported in ADPKD in cyst-lining epithelia (29). Recently, the polycystin complex has been implicated in junction formation and establishment of apicobasal polarity (31). In addition, it has been shown that PATJ, among other polarity proteins such as PALS1 and PAR6, interacts with nephrocystins, suggesting a link between tight junction formation and cystic kidney disease (32). The function of the epithelial primary cilium is a key factor in the etiology of cystic kidney disease (15). Interestingly, polarity proteins such as CRB3 and members of the PAR complex are localized not only at the tight junction but also at the primary cilium of Madin-Darby canine kidney cells controlling ciliogenesis (33,34). TAZ knockdown cells present shortened cilia and display a similar phenotype as described for CRB3 and PAR3 knockdown cells, suggesting that TAZ may also participate in ciliogenesis (11,33,35). The asso-ciation between TAZ and PATJ indicates that TAZ, polarity proteins, and other proteins involved in PKD share a common pathway participating in regular epithelial function and structural integrity.
Mutation and dysfunction of PC2 causes ADPKD. PC2 is a member of the transient receptor potential (TRP) family and functions as a Ca 2ϩ -permeable cation channel (16). Together with PC1, it acts as a mechanosensor at the primary cilium mediating extracellular calcium influx in response to fluid flow (16,17). The flow-mediated Ca 2ϩ transient requires the presence of functional PC2 and has been implicated in the regulation of tubular polarity and morphology (16,17).
Tian et al. (12) showed that TAZ interacts with the C terminus of PC2 and mediates PC2 degradation via a SCF ␤ -TRP E3 ubiquitin ligase pathway. However, deletion of the last 5 amino acids representing a PDZ binding motif in PC2 diminished TAZ association with PC2, indicating that another PDZ domain-containing protein may be involved in the TAZ/PC2 interaction.
Therefore, we speculate that the multi-PDZ protein PATJ interacts not only with TAZ but also with PC2. Because PATJ is an intracellular membrane-associated protein, the interactions must involve the cytosolic tail of PC2. Thus, we investigated whether PATJ directly associates with the cytosolic C terminus of PC2, containing the putative class I PDZ binding domain Asn-Val-His-Ala (NVHA). We demonstrated in Y2H assays that the cytosolic tail of PC2 shows a strong interaction with the C-terminal PDZ domains 8 -10 and also a weak binding to the N-terminal PDZ domains 1-3 of PATJ ( Fig. 2A). Our data were confirmed by Co-IPs using HEK293T lysates expressing 3xFLAG-tagged PATJ N-(FLAG-PATJ aa 1-554) and C-terminal (FLAG-PATJ aa 1239 -1801) deletion mutants, respectively. The assays revealed that the C terminus of PC2 precipitates both the PDZ domains 1-3 as well as the PDZ domains 7-10 of PATJ, whereas the control did not (Fig. 2B). These data have been confirmed by GST pulldown assays (Fig. 2C).
Subcellular localization of PC2 remains controversial and has been described at the endoplasmic reticulum, primary cilia, and plasma membrane (basolateral and apical), regulated by interaction with diverse adaptor proteins (14,36). We speculate that the interaction with PATJ might mediate the translocation to the plasma membrane and the primary cilium, similar to the mechanism suggested for polycystin-1 (37) or TRPC1 (38). This will have to be elucidated in further studies.
As a next important step, we investigated whether the PC2 interaction with TAZ and/or PATJ has an impact on the PC2 FIGURE 1. TAZ interacts with polarity protein PATJ. A, Co-IP revealed that TAZ interacts through the last 4 aa (LTWL) with PATJ. HEK293T cells were co-transfected with expression plasmids for Myc-PATJ and 3xFLAG-TAZ (FLAG-TAZ (F.TAZ)) as indicated. After immunoprecipitation with ␣-FLAG, PATJ is detectable in the precipitate containing 3xFLAG TAZ full length (FLAG-TAZ fl (F.TAZ fl)) but not TAZ without PDZ binding motif (FLAG-TAZ-LTWL F.TAZ-LTWL). B, GST pulldown experiments. HEK293T lysates containing enhanced YFP (EYFP)-tagged PATJ were incubated with equal amounts of different GST-TAZ fusion proteins. GST was used as a control. C, Y2H assay using PATJ PDZ7-10 (aa 1239 -1801) as bait and TAZ full length or TAZ-LTWL as prey confirms the binding of PATJ to TAZ. Interaction was scored by His auxotrophy (ϪHIS) and ␤-galactosidase activity (LacZ). D, Y2H assays using PATJ deletion mutants revealed TAZ binding sites within the N-terminal PDZ domains 1-3 and C-terminal PDZ domains 8 -10 of PATJ. Green box, PDZ domain; blue box, L27 domain. E, GST pulldown assay using V5-tagged TAZ incubated with equal GST-PATJ fusion proteins comprising aa 1-685 (L27-PDZ4) or aa 1239 -1801 (PDZ7-10) supports the Y2H data, suggesting two distinct TAZ binding sites within PATJ. channel activity, which is pathologically disturbed in PKD (16). As shown earlier, PC2 utilizes TRPV4 to form a mechano-and thermosensitive molecular sensor at the cilium of vertebrate cells (39). The fact that TRPV4 and PC2 are able to form heterooligomeric channels was used in the present study to analyze PC2 activity. Although it has been demonstrated earlier that heterooligomeric PC2-TRPV4 channels can be activated by high (18 mM) extracellular Ca 2ϩ or by hypotonic bath solution (39), we frequently observed that endogenous Ca 2ϩ /Na ϩ currents were activated by high extracellular Ca 2ϩ or hypotonic bath solution, probably activating endogenous TRP channels (40). Therefore, we made use of GSK1016790A (41), which reliably activates whole cell currents only in TRPV4 or PC2-TRPV4 co-expressing oocytes (Fig. 2D).
TRPV4-expressing or PC2-TRPV4 co-expressing oocytes responded to stimulation with low concentrations (50 nM) of GSK1016790A with an activation of a large negative current that was due to an influx of Na ϩ / Ca 2ϩ into the oocytes (Fig. 2E). In contrast, GSK1016790A was unable to activate a significant whole cell conductance in noninjected/ water-injected (con) or PC2-injected oocytes (Fig. 2E). Additional co-expression of PATJ and TAZ reduced the current activation by GSK1016790A (Fig. 2E), and the highest reduction was observed after co-expression of both PATJ and TAZ. As a control, these inhibitory effects of PATJ, TAZ, or PATJ-TAZ on TRPV4 currents could not be observed in the absence of PC2, suggesting that PATJ-TAZ interacts with PC2 rather than TRPV4 (Fig. 2F).
Interestingly, we observed that PATJ significantly impairs the channel activity of PC2, indicating PATJ as a novel regulatory subunit of PC2. This is in line with published data for the Drosophila homolog of PATJ called inactivation no after potential D (INAD), which interacts and negatively regulates a TRP cation channel in fly photoreceptor cells (42)(43)(44)(45). In addition, we could determine that TAZ alone is also able to impair PC2 channel activity, although to a lesser extent than PATJ. The strongest reduction of current activity was observed after co-expression of TAZ with PATJ. For Drosophila PATJ, it has already been shown that it is required for localization as well as promoting the stability of interacting proteins regulating the TRP channel activity (43). PATJ might also promote the interaction of TAZ and PC2. In our studies (Fig. 2G, model), PATJ interacts with TAZ as well as PC2. In accordance with the observation that TAZ interacts with two distinct parts of the PATJ molecule, the negative regulatory effect on PC2 channel activity through TAZ might be increased if both PATJ and TAZ are present. TAZdeficient mice develop kidney cysts at least in part due to impaired PC2 degradation, resulting in PC2 accumulation (12). Our data lead us to speculate that there might be an additional mechanism, where TAZ functions as a negative regulator of PC2 channel activity. Loss of the mechanosensory function of PC2 is thought to result in altered cellular signaling, subsequently leading to cyst formation (46). Obviously, a channel FIGURE 2. TAZ and PATJ regulate PC2 channel activity. A, Y2H assay using various PATJ mutants as bait and the C terminus (C-term) of PC2 as prey. Interaction scored by His auxotrophy (ϪHIS) and ␤-galactosidase activity (LacZ) reveals an association of PC2 with the PDZ domains 1-3 and 8 -10. B, Co-IP assays showed interaction of V5-PC2/C-term with both 3xFLAG-tagged PDZ domains 1-3 (FLAG-PATJ aa 1-554 (F.PATJ aa 1-554)) and the PDZ domains 7-10 (FLAG-PATJ aa 1239 -1801 (F.PATJ aa 1239 -1801)) of PATJ. C, GST pulldown assays using the same 3xFLAG-tagged PATJ constructs confirmed that both the PDZ domains 1-3 (FLAG-PATJ aa 1-554) and the PDZ 7-10 (FLAG-PATJ aa 1239 -1801) independently interact with GST-PC2 fusion protein (aa 688 -966) but not with GST alone. D, activation of whole cell currents by the TRPV4 activator GSK1016790A in a Xenopus oocyte injected with cRNA, encoding TRPV4 and PC2. E, functional relevance of TAZ and PATJ interaction with PC2. Summary of GSK1016790A (50 nM)-activated whole cell currents at the negative clamp voltage of Ϫ60 mV is shown. Xenopus oocytes were activated as water-injected control oocytes (con) or PC2-or TRPV4-injected oocytes or after co-expressing TRPV4, PC2, TAZ, and PATJ, as indicated. Co-expression of PATJ and TAZ significantly reduced current activation, and the strongest reduction occurred when both PATJ and TAZ were co-expressed. F, the lack of effect of PATJ-TAZ on GSK1016790A-activated currents (V c ϭ Ϫ60 mV) in TRPV4-expressing oocytes reveals that the observed inactivation is based on PC2 regulation. Mean Ϯ S.E., n ϭ number of cells measured. #, significant increase in whole cell conductance by GSK1016790A (unpaired t test). G, model of the PC2/PATJ/TAZ interaction. PATJ mediates interaction of two TAZ proteins with PC2, resulting in negative regulation of PC2 channel activity.
using Ca 2ϩ as a ubiquitous second messenger requires tight control to prevent detrimental effects on the cellular homeostasis. The absence of TAZ as a regulatory element of PC2 channel activity might result in a dysregulation of intracellular Ca 2ϩ signaling, contributing to the development of ADPKD in TAZ knock-out mice. In addition, we identified the multi-PDZ domain-containing protein PATJ with an established role in apicobasal polarity as a new interaction partner of both TAZ and PC2, functionally regulating PC2 activity. It is a tempting speculation that regulation of apicobasal polarity, in addition to planar cell polarity, might be a new pathway involved in the pathogenesis of PKD.