Regulation of Phospholipase C- b 3 Activity by Na 1 /H 1 Exchanger Regulatory Factor 2*

Among the phospholipase C that catalyzes the hydrolysis of phosphatidylinositol 4,5-bisphosphate, four mam-malian phospholipase C- b (PLC- b ) isotypes (isotypes 1–4) are activated through G protein-coupled receptors (GPCRs). Although the regulation of the PLC- b s by GPCRs and heterotrimeric G proteins has been extensively studied, little is known about the molecular de-terminants that regulate their activity. The PLC- b isozymes carry a putative PSD-95/Dlg/ZO-1 (PDZ) binding motif ( X (S/T) X (V/L)COOH) at their carboxyl terminus, which is implicated in specific interactions with anchor proteins. Using the yeast two-hybrid system, we identified Na 1 /H 1 exchanger regulatory factor 2 (NHERF2) as a protein that interacted with a C-terminal heptapeptide of PLC- b 3. Immunoprecipitation studies revealed that NHERF2 interacts the N-terminal region was generated by polymerase chain reaction using primers 5 9 -CGGAATTCATGGC-CGCGCCGGAGCCG-3 9 and 5 9 -CCGCTCGAGCCGAGGGCG- CAGCTC-3 9 and digested with Eco RI and Xho I. The C-terminal region was generated by polymerase chain reaction with primers (5 9 - ACGCGTCGACACAGATGAACACTTCAAG-3 9 and 5 9 -CGTCTAGAT-CAGAAGTTGCTGAAGATTTC-3 9 ) and digested with Sal I and Xba I. The fragments were then ligated at their Sal I and Xho I sites, respec-tively, and the resulting construct was inserted into pCDNA3.1-HisC. PLC- b isotypes and their FLAG-tagged forms in pCDNA3.1 and pCMV2 were constructed from the cDNA of each of the previously isolated PLC- b isotype. Each of the individual residues of the PDZ-binding motif (NTQL) of PLC- b 3 were mutated to Ala codons and confirmed by DNA sequencing. Antibodies— For production of NHERF2 antibodies, His 6 fusion protein of the entire NHERF2 was purified and used for immunization of mice. After four immunizations, splenocytes were fused with myeloma cells at a ratio of 10:1 using polyethylene glycol 1500. The hybrids were plated in 96-well plates in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and selected with hypoxanthine/ aminopterin/thymidine. Supernatants of the hybrids were used to screen the pools releasing the antibody for NHERF2. Positive hybrido- mas were cloned twice by limit dilution. Large quantities of specific monoclonal antibodies were produced by preparation of ascites fluid in Balb/c mice. Purified antibodies were used for immunological studies. Antibodies of PLC- b isotypes were made using the C-terminal polypep- tides expressed in E. coli following the above procedure. Anti-HA monoclonal antibody and anti-FLAG

Among the heterotrimeric G proteins, the G␣ q family activates PLC-␤ upon stimulation with ligands (3). Likewise, the ␤␥ subunit, which dissociates from the G␣ i/o family, activates PLC-␤ in a pertussis toxin-sensitive manner (4). Currently, four isotypes of PLC-␤ have been identified in mammals. Each isotype exhibits a different sensitivity to the G␣ q family and ␤␥ subunit. It has been reported that the GTP␥S-activated G␣ q or G␣ 11 subunits stimulated PLC-␤ isotypes with the ranking order of potency: PLC-␤1 Ն PLC-␤3 Ͼ PLC-␤2. PLC-␤4 was also activated by the G␣ q subunit (5). The G␤␥ subunit activated PLC-␤ in vitro according to the following hierarchy: PLC-␤3 Ͼ PLC-␤2 Ͼ PLC-␤1 (6,7). However, several in vivo studies have shown that the receptors of chemoattractants including interleukin-8, C5a, and formyl-Met-Leu-Phe can specifically activate PLC-␤2 via the G␤␥ subunits from G␣ i/o in a pertussis toxin-sensitive manner (8 -10). G␣ q and G␤␥ subunits may thus independently modulate each PLC-␤ isotype in cooperation with other intracellular factors.
The distribution of the PLC-␤ isotypes has been investigated in several tissues and cells. In the brain, PLC-␤1 is highly expressed in the cerebral cortex and hippocampus, and PLC-␤4 is highly expressed in cerebellum. PLC-␤3 is expressed throughout the brain (11)(12)(13). The expression patterns of the PLC-␤ isotypes in different cell lines are also different (14). While PLC-␤3 is expressed in most cells, PLC-␤2 is dominant in hematopoietic cells (14,15). All isotypes except PLC-␤2 are detected in PC12 cells. 2 These findings suggest the possibility that each PLC-␤ isotype plays a distinct role in various GPCR signaling events.
PLC-␤ isotypes have a long C-terminal regions (ϳ400 residues) that have relatively low homology among them (7). In studies using mutants with deletions in the C-terminal region of PLC-␤1, it was determined that the segment (residues 903-1142) was required for binding and stimulation by G␣ q (17). Moreover, there are short consensus sequences known as postsynaptic density-95/discs large/ZO-1 (PDZ)-binding motifs that consist of the amino acids -X(S/T)X(V/L)-COOH at the immediate C terminus of the PLC-␤3 isotype (18). PDZ domains exist in a large number of multifunctional proteins that mediate protein-protein interactions at the postsynaptic density in neurons and junctional complexes in epithelia (19,20). PSD-95/SAP90, which contains three PDZ domains, seems to assemble a receptor and channel complex by interacting with a  1 The abbreviations used are: GPCR, G protein-coupled receptor; PLC, phospholipase C; GTP␥S, guanosine 5Ј-3-O-(thio)triphosphate; PDZ, PSD-95/Dlg/ZO-1; PSD-95, postsynaptic density-95; HA, hemagglutinin; NHERF, Na ϩ /H ϩ exchanger-regulating factor; GST, glutathione S-transferase; X-gal, 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside. type of glutamate receptor (N-methyl-D-aspartate receptor), and a shaker type K ϩ channel and neuronal nitric-oxide synthase (21,22). GRIP has seven PDZ domains and interacts with the ␣-amino-3-hydroxy-5-methyl-4-isoxazole propionate receptor, another type of glutamate receptor potential through the fourth and fifth PDZ domains (23). The phototransduction pathway in the compound eye of Drosophila is spatially organized by inactivation no afterpotential D, a five-PDZ-containing protein. By interacting with PLC-␤, eye-specific protein kinase C and light-activated Ca 2ϩ channel transient receptor potential through its PDZ domains, INAD enables the visual signaling cascade through rhodopsin to be efficiently regulated (24,25). These reports suggest that the PDZ-containing proteins may involve in various G protein-coupled receptor-mediated signaling pathways. Therefore, it is reasonable to assume that this motif may provide the PLC-␤ isotypes with specificity in the signal transduction pathway by allowing them to interact with various PDZ-containing proteins.
In this study, we identify some of the proteins interacting with PLC-␤3 using the yeast two-hybrid system. Here, we show that NHERF2 (Na ϩ /H ϩ exchanger-regulating factor 2) interacts with the C terminus of PLC-␤3 via the second PDZ domain. This interaction is isotype-specific in that the PLC-␤3 binds specifically to NHERF2 rather than other PLC isotypes. Furthermore, NHERF2 enhances the activation of PLC-␤ induced by carbachol treatment. These results suggest that NHERF2 may play a pivotal role in the organization and modulation of the PLC-␤3-mediated signaling.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid System-Two-hybrid screening was performed using the L40 yeast strain harboring His3 and ␤-galactosidase as reporter genes as described previously (26,27). DNA oligomers encoding the C-terminal heptapeptides of PLC-␤ isotypes were inserted into pBHA (LexA fusion vector) and used as bait to screen a human fetal liver cDNA library inserted into the activation domain of GAL4 in pGAD10 (CLONTECH, Palo Alto, CA). We screened ϳ1 ϫ 10 6 primary transformants for interacting proteins. Clones specifically interacting with the bait were identified by His growth and X-gal activity assay. The DNAs of positive clones were sequenced, and the sequences were compared with sequences in the data bank using the National Center for Biotechnology Information (NCBI) BLAST.
Plasmid Constructs, Site-directed Mutagenesis-In order to express NHERF2 in Escherichia coli, the cDNA of human NHERF2 was inserted into the glutathione S-transferase (GST) fusion protein vector, pGEX4T-1 (Amersham Pharmacia Biotech). Various constructs of NHERF2 in pET30a were provided by Dr. C. H. Yun from Johns Hopkins University (28), and HA-tagged NHERF1 was kindly provided by Dr. Robert Lefkowitz from Duke University (29). GST or His 6 fusion proteins were expressed in E. coli BL21 by induction with 0.5 mM isopropyl-␤-D-thiogalactopyranoside for 4 h at 27°C. Bacterial lysates were prepared by sonication in ice-cold PBS in the presence of protease inhibitors. Proteins were purified from the soluble fraction on glutathione-Sepharose 4B (Amersham Pharmacia Biotech) for GST fusion proteins or Ni 2ϩ -nitrilotriacetic acid resin (Invitrogen) for His 6 fusion proteins.
PLC-␤ isotypes and their FLAG-tagged forms in pCDNA3.1 and pCMV2 were constructed from the cDNA of each of the previously isolated PLC-␤ isotype. Each of the individual residues of the PDZbinding motif (NTQL) of PLC-␤3 were mutated to Ala codons and confirmed by DNA sequencing.
Antibodies-For production of NHERF2 antibodies, His 6 fusion protein of the entire NHERF2 was purified and used for immunization of mice. After four immunizations, splenocytes were fused with myeloma cells at a ratio of 10:1 using polyethylene glycol 1500. The hybrids were plated in 96-well plates in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and selected with hypoxanthine/ aminopterin/thymidine. Supernatants of the hybrids were used to screen the pools releasing the antibody for NHERF2. Positive hybridomas were cloned twice by limit dilution. Large quantities of specific monoclonal antibodies were produced by preparation of ascites fluid in Balb/c mice. Purified antibodies were used for immunological studies. Antibodies of PLC-␤ isotypes were made using the C-terminal polypeptides expressed in E. coli following the above procedure. Anti-HA monoclonal antibody and anti-FLAG monoclonal antibody were purchased from Roche Molecular Biochemicals and Sigma. Anti-His 6 polyclonal antibody was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Pull-down Assay-COS7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 50 IU/ml penicillin, 50 g/ml streptomycin, and 10% fetal bovine serum in a 5% CO 2 , 95% air incubator at 37°C. The cells were lysed in lysis buffer (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 1 M pepstatin, and 1% Triton X-100). The lysate was centrifuged at 15,000 ϫ g for 15 min at 4°C. The supernatant was incubated with 10 g of whole NHERF2 or GST fusion fragments thereof immobilized on glutathione-Sepharose 4B or Ni 2ϩ -nitrilotriacetic acid resin. The beads were washed four times with lysis buffer, followed by boiling in SDS sample buffer. The precipitates were resolved by SDS-polyacrylamide gel electrophoresis and were immunoanalyzed.
COS7 Cell Transfection and Immunoprecipitation-COS7 cells were transfected using Fugene 6 (Roche Molecular Biochemicals) at a ratio of 1.0 g of total plasmid DNA and 3 l of Fugene in 100 l of medium. After 36 h, the transfected cells were harvested, lysed in lysis buffer, and used for immunoprecipitation with appropriate antibodies as indicated in each figure. Immunocomplexes were collected by binding to protein A-Sepharose and washed four times with lysis buffer prior to electrophoresis. The proteins were eluted with SDS-sample buffer, analyzed by SDS-polyacrylamide gel electrophoresis, and immunoblotted.
Measurement of Total Inositol Phosphates-COS7 cells were seeded at 2 ϫ 10 5 cells/well in a six-well plate and transfected with 0.1 g of muscarinic receptor type 1 plasmid and 0.9 g of wild type or deletion mutant plasmid of NHERF2 using Fugene 6. After 24 h, the transfected cells were incubated with 2 Ci/well [ 3 H]inositol (NEN Life Science Products) in 2 ml of culture medium for 12 h at 37°C. After incubation with 10 mM LiCl for 10 min, the cells were treated with 10 M carbachol for 30 min and lysed by the addition of ice-cold 5% perchloric acid. Accumulated [ 3 H]inositol phosphates were determined as described previously (30). Briefly, the extracts were centrifuged, diluted 1:5 with distilled water, and applied to AG1-X8 anion exchange resin (Bio-Rad). The column was washed extensively with 4 ml of distilled water and with 10 ml of 0.06 M ammonium formate containing 5 mM sodium tetraborate. Inositol phosphates were eluted with 2 ml of 1 M ammonium formate and 0.1 M formic acid. The radioactivity in eluate was measured using a ␤-scintillation counter (Packard).
Cell Lines Expressing the Recombinant NHERF2-To characterize the effects of NHERF2 in the whole cell, we established HeLa cell lines expressing deletion mutants of NHERF2 on a background of endogenous NHERF2, PLC-␤3, and muscarinic receptor expression. Briefly, each construct in pCDNA3.1, a plasmid that confers resistance to G418, were transfected into HeLa cells using Fugene 6. After cells were selected in the presence of 800 g/ml G418, the expression level of each mutant was checked with anti-NHERF2 antibody.
Subcellular Fractionation of HeLa Cells-HeLa cells expressing deletion mutants of NHERF2 were fractionated as described previously (31). Cells grown to 70 -80% confluency in a 100-mm culture dish were collected and homogenized in 300 l of 20 mM Hepes/NaOH, pH 7.4, by sonication. 80 l of the homogenate was saved for immunoblot analysis, and the remainder was centrifuged at 100,000 ϫ g for 30 min to separate the cytosol from membrane fraction. The pellet was homogenized in 220 l of 20 mM Hepes/NaOH, pH 7.4, containing 1% Triton X-100 and centrifuged at 15,000 ϫ g for 15 min to separate Triton X-100-soluble and -insoluble fractions.

RESULTS
Identification of the PLC-␤3-interacting Protein-The C-terminal heptapeptide (QEENTQL-COOH) of PLC-␤3 was used as NHERF2 Binds PLC-␤3 the bait in the yeast two-hybrid assay screening for proteins interacting with PLC-␤3. Two positive clones from a human liver cDNA library were obtained and sequenced. They were found to contain the fragment of NHERF2 extending from the second PDZ domain to the C-terminal region. NHERF2 initially cloned by Yun et al. has two highly conserved PDZ domains, PDZ1 and PDZ2. These domains, which have been described as modules mediating protein-protein interaction in the submembranous region, recognize the C-terminal consensus sequence (X(S/T)X(V/L)-COOH) of target molecules. The binding specificity of the C-terminal motif of PLC-␤3 for various different PDZ domains was examined with a yeast two-hybrid assay ( Table I). The PLC-␤3 motif interacted specifically with the PDZ2 of NHERF2 but not with the motif in PSD-95 and NHERF1, suggesting that the PLC-␤3 binding is specific for NHERF2.
PLC-␤3 Specifically Interacted with NHERF2-To confirm the association of PLC-␤3 with NHERF2, GST-NHERF2 expressed in E. coli strain BL21 was used in a pull-down assay of COS7 cell lysates. Immunoblot analysis using an anti-PLC-␤3 antibody showed that PLC-␤3 precipitated with GST-NHERF2 but not with GST alone (Fig. 1A). The extracts of COS7 cells expressing FLAG-PLC-␤3 or FLAG-PLC-␤3⌬C (C-terminal four amino acids deleted) were incubated with GST-NHERF2. Whole PLC-␤3 was detected by the anti-FLAG antibody in the precipitate, whereas the C-terminally deleted form was not. This suggests that PLC-␤3 interacted with NHERF2 via the C-terminal PDZ binding motif (Fig. 1B). NHERF1/EBP50 cloned by Reczek et al. shares high amino acid sequence homology with NHERF2 (32). Both isotypes of NHERF regulate the activity of NHE3 and the cystic fibrosis transmembrane conductance regulator in membranes. We examined whether NHERF1 would also interact with PLC-␤3. Extracts of COS7 cells co-expressing FLAG-PLC-␤3 together with HA-tagged NHERF1 or His 6 -tagged NHERF2 were allowed to immunoreact with anti-FLAG antibody. Subsequently, the immunocomplexes were analyzed with appropriate antibodies. As shown in Fig. 2, only NHERF2 was coprecipitated with PLC-␤3, suggesting that PLC-␤3 interacted specifically with NHERF2.
All PLC-␤ isotypes have the consensus sequence of the PDZ binding motif at their C terminus, which enables the PLC-␤s to associate with PDZ domain-containing proteins. We tested the interaction of NHERF2 with each PLC-␤ isotype by co-expressing NHERF2 with each isotype in COS7 cells (Fig. 3). When NHERF2 was precipitated from cellular extracts using an anti-NHERF2 antibody (antibody 2570), only PLC-␤3 was detected in the precipitates (Fig. 3A). In parallel, immunoprecipitation assays using anti-FLAG antibody were performed on extracts of COS7 cells expressing NHERF2 and one of each FLAGtagged PLC-␤ isotype. NHERF2 was detected in the immunoprecipitate together with FLAG-PLC-␤3, but not with the other isotypes (Fig. 3B). These data suggest that NHERF2 specifically interacts with PLC-␤3, although a weak association with PLC-␤2 was also detected (Fig. 3, A and B).
The C-terminal Thr and Leu Residues of PLC-␤3 Are Essential for the Interaction with NHERF2-It is known that PDZ domains bind to C-terminal four amino acids of target proteins. Ser or Thr at the Ϫ2-position and Leu or Val at the terminal position of the PDZ binding motif are essential for the interaction with PDZ domains, while other residues are less effective (33). The binding preferences of NHERF2 were determined by mutating in turn the last four amino acids (NTQL-COOH) of PLC-␤3 to Ala. The extracts of COS7 cells co-expressing each FLAG-tagged, mutant PLC-␤3 and NHERF2 were reacted with anti-FLAG antibody, and the precipitate was then probed with anti-NHERF2 antibody. Mutation of Thr to Ala at the Ϫ2position or Leu to Ala at the last position resulted in complete loss of NHERF2 binding to PLC-␤3. In contrast, mutation of the other residues to Ala had no effect (Fig. 4). The results indicate that these Thr and Leu residues of PLC-␤3 participate in the interaction with NHERF2.
The Second PDZ Domain of NHERF2 Is Required for Interaction with PLC-␤3-In order to determine which regions of NHERF2 take part in the interaction, we expressed and purified wild type as well as various fragments of NHERF2 as His 6 -tagged recombinant proteins as described previously (28). COS7 cell extracts were incubated with the recombinant FIG. 2. PLC-␤3 specifically interacts with NHERF2. The extracts of COS7 cells expressing FLAG-tagged PLC-␤3 with His 6 -NHERF2 or HA-NHERF1 were immunoreacted with anti-FLAG antibody. The precipitates were analyzed with antibodies to the epitopes tagged to each protein; anti-His antibody for His 6 -NHERF2, anti-HA antibody for HA-NHERF1. IP, immunoprecipitation.

TABLE I NHERF2 interacts specifically with the PDZ domain binding motif of PLC-␤3
The C-terminal heptapeptide of PLC-␤3 was tested for its ability to bind NHERF2 in the yeast two-hybrid system revealed by the induction of the yeast reporter genes HIS3 and ␤-galactosidase. HIS activity (measured as percentage of colonies growing on histidine-lacking medium) was as follows: ϩϩϩ (Ͼ60%), ϩϩ (30 -60%), ϩ (Ͻ30%), Ϫ (no significant growth). ␤-gal (time needed for yeast colonies to turn blue in X-gal filter lift assays at room temperature) was as follows: ϩϩϩ (Ͻ20 min), ϩϩ (20 -60 min), Ϫ (no detection). pBHA clone pGAD10 clones HIS3 ␤-Galactosidase 1. Interaction of NHERF2 with PLC-␤3. A, interaction of GST-NHERF2 with PLC-␤3. The extract of COS7 cells endogenously expressing PLC-␤3 was incubated with GST-NHERF2 immobilized on glutathione beads. The precipitates were detected with anti-PLC-␤3 antibody. B, PLC-␤3 interacts with NHERF2 through its C-terminal PDZ binding motif. The extracts of COS7 cells expressing FLAG-tagged wild type or the C-terminal deleted form of PLC-␤3 were incubated with GST-NHERF2, and the precipitated proteins were detected with anti-FLAG antibody.
NHERF2 constructs immobilized on Ni 2ϩ -nitrilotriacetic acid beads. The precipitates were then probed with anti-PLC-␤3 antibody. As shown in Fig. 5A, entire NHERF2 and fragments containing the PDZ2 domain were associated with PLC-␤3. The data are consistent with the results obtained in the yeast two-hybrid system. Although the C-terminal region of NHERF2 by itself did not effectuate the direct binding to PLC-␤3, the PDZ2 domain containing this region associated more tightly with PLC-␤3 than PDZ2 alone (Fig. 5A).
To confirm the above results, we undertook further immunoprecipitation experiments. The ERM binding domain-deleted form and the PDZ2-deleted form of NHERF2 were generated as described under "Experimental Procedures." The cell extracts containing FLAG-PLC-␤3 and one of each recombinant NHERF2 were immunoreacted with anti-FLAG antibodies. Subsequently, the presence of NHERF2 bound to PLC-␤3 was analyzed by immunoblot assay with anti-NHERF2 antibody. As shown in Fig. 5B, both wild type and the ERM binding domain-deleted form of NHERF2 associated with PLC-␤3, but the PDZ2-deleted form did not, suggesting that PDZ2 is essential for the interaction.
NHERF2 Increases Phosphoinositides Hydrolysis Induced by Carbachol-COS7 cells transfected with the genes for muscarinic receptor type 1 and various recombinant forms of NHERF2 were labeled with [ 3 H]inositol as described under "Experimental Procedures." After 12 h, the cells were treated with carbachol, and inositol phosphate release into the cytosol was measured. As shown in Fig. 6A, the accumulation of inositol phosphates upon carbachol treatment increased 3-4-fold in the wild-type NHERF2-transfected cells relative to vectortransfected cells. On the other hand, the PDZ2-deleted NHERF2 had no effect.
In addition, we established HeLa cell lines expressing wild type or deletion mutants of NHERF2. When the cells were fractionated employing a method previously used by Yao et al. (31), both PLC-␤3 and wild type NHERF2 were co-localized in the Triton X-100-soluble membrane fraction. The localization of PLC-␤3 was not influenced by overexpression of NHERF2, or vice versa. The subcellular localization of the PDZ2-deleted mutant was the same as that of the wild type. The ERM binding domain-deleted form was predominantly localized in the cytosolic fraction, although a small amount of the ERM binding domain-deleted form was detected in the membrane fraction (data not shown). The data suggest that the ERM binding domain may influence the localization of NHERF2 in cells.
The accumulation of inositol phosphates after carbachol treatment in cells expressing wild type or the ERM binding domain-deleted form of NHERF2 was similar to that in the control cells. However, carbachol treatment of cells expressing the PDZ2-deleted form of NHERF2 only barely induced accumulation of inositol phosphates. This suggests that the muscarinic receptor-mediated PLC activity may be regulated by the PDZ2-mediated interaction of NHERF2 with PLC-␤3 (Fig. 6B). DISCUSSION Using the yeast two-hybrid system, we found a molecule that interacted with PLC-␤3. Since the C-terminal four amino acids of target proteins are sufficient to serve as the binding site for PDZ domain-containing proteins (33, 34), we used a C-terminal heptapeptide of PLC-␤3 as bait and identified NHERF2 as a PLC-␤3-interacting protein. NHERF2, also called E3KARP,

FIG. 3. Specific interaction between NHERF2 and PLC-␤3 in vivo.
A, immunoprecipitation (IP) using anti-NHERF2 antibody. The extracts of COS7 cells expressing NHERF2 together with one of the PLC-␤ isotypes (isotopes 1-4) were allowed to immunoreact with anti-NHERF2 antibody. The precipitates were analyzed with antibodies against each PLC-␤ isotype. B, immunoprecipitation using the anti-FLAG antibody. To confirm the result in A, the extract of COS7 cells expressing NHERF2 together with one of FLAG-tagged PLC-␤ isotypes were immunoprecipitated with anti-FLAG antibody. The precipitates separated by SDS-polyacrylamide gel electrophoresis were probed with anti-FLAG antibody or anti-NHERF2 antibody.
FIG. 4. The Thr and Leu residues at the C terminus of PLC-␤3 are essential for the interaction with NHERF2. Each residue of the PDZ binding motif in PLC-␤3 was mutated, in turn, to Ala using polymerase chain reaction with mutagenic primers. The extract of COS7 cells expressing NHERF2 with one of the point-mutated forms of PLC-␤3 was allowed to immunoreact with anti-FLAG antibody. The precipitated proteins, separated by SDS-polyacrylamide gel electrophoresis, were probed with anti-NHERF2 antibody. IP, immunoprecipitation. was initially cloned by two-hybrid screening using the C-terminal tail of the Na ϩ /H ϩ exchanger type 3 (NHE3) (35). Because this protein was required for the cAMP-dependent inhibition of NHE3, it was thought to be a regulator of NHE3 (28). However, further studies revealed that NHERF2 does not act by itself but that it takes part in the regulation of NHE3 by anchoring other regulators. Recent studies have shown that NHERF can target PKA to NHE3 by associating with ezrin, a member of the ERM family, via the ERM-binding domain in its C-terminal region (36). By analogy, we suggest that NHERF2 may act as a scaffold in PLC-␤-mediated signal transduction.
To date, two isotypes of NHERF, NHERF1 (EBP50) and NHERF2 (E3KARP) have been cloned (29,35). These proteins share 52% amino acid sequence identity. They also have two in tandem PDZ domains and an ERM domain in common. In addition, it has been suggested that they play common roles in linking the cytoskeletal proteins to the plasma membrane. The ERM-binding domain interacts with ERM family proteins such as ezrin, radixin, and moesin, which associate with the cytoskeleton (36,37). Moreover, the ERM-binding domain interaction with activated ezrin suggests that the molecular interaction may be influenced by intracellular signaling events (38,39). NHERF2 also associates with NHE3 via the second PDZ domain and other C-terminal region (28). They interact with the four C-terminal residues of cytoplasmic domains of the ␤ 2 -adrenergic receptor and the cystic fibrosis transmembrane conductance regulator through their first PDZ domain (29,37).
On the other hand, recent reports have shown that NHERF1 is phosphorylated by G protein-coupled receptor kinase 6A under basal conditions, while NHERF2 is not phosphorylated at all (36,40). Our results show that NHERF2 interacts with PLC-␤3; NHERF1 does not. These observation raise the possibility that each member of the NHERF family may be involved in the separate cellular events, although they sometimes have target molecules in common.
With respect to the PLC-␤ isotypes, immunoprecipitates from extracts of cells co-expressing each PLC-␤ isotype and NHERF2 showed that NHERF2 specifically interacted with PLC-␤3 but only barely with other types. Hall et al. suggested that the C-terminal motif (DTPL-COOH) of PLC-␤1 might be a target site of NHERF (33). In the yeast two-hybrid system, NHERF2 also associated with the motif of PLC-␤1 (data not shown). However, NHERF2 failed to interact with PLC-␤1 in in vivo binding assays, indicating that the specificity of an interaction through the PDZ domain may be determined by other regions as well as the binding motif itself. These data also suggest that the PLC-␤ isotypes act in different signaling pathways, although they share the common ability to hydrolyze phosphatidylinositol 4,5-bisphosphate.
Studies of the molecular interactions of proteins with PDZ domains have shown that the PDZ domains interact with specific carboxyl termini carrying the consensus sequence of X(S/ T)X(V/L/I). The PDZ domain of cortactin-binding protein 1 and PICK1 interact with the C terminus of the somatostatin receptor (QTSI-COOH) and protein kinase C-␣ (QSAV-COOH), re- spectively (16,41). In some cases, other regions besides the PDZ domains are required for PDZ domain-mediated interaction. For example, the second PDZ domain and the carboxyl terminus of NHERF2 form a unit when they interact with NHE3 (28). In this study, mutations of the Thr or Leu residues in the PLC-␤3 tail to Ala caused the PLC-␤3 or NHERF2 to disappear from immune complexes with anti-NHERF2 antibodies or anti-FLAG antibodies. The pull-down assay using fragments of NHERF2 showed that it was the second PDZ domain that associated with PLC-␤3. The C-terminal region of NHERF2 enhanced this association, although it was not directly involved in the interaction. In addition, the mutant in which the second PDZ domain was deleted seemed not to interact with PLC-␤3 based on the immunoprecipitation results. The data, thus, imply that the second PDZ domain and Cterminal region may cooperatively interact with PLC-␤3.
Since PDZ-containing proteins function as scaffolding proteins in many cases, we thought that NHERF2 may have an effect on the enzymatic activity of PLC-␤3. However, we found that purified NHERF2 was unable to change PLC-␤3 activity in vitro (data not shown). When muscarinic receptor type 1 and NHERF2 were co-transfected into COS7 cells, which express PLC-␤3 rather than the other isotypes, expression of NHERF2 led to the enhancement of the PLC activity in response to carbachol treatment. However, the PDZ2-deleted form of NHERF2, which does not interact with PLC-␤3, barely influenced the activation of PLC-␤. In HeLa cells constitutively expressing recombinant NHERF2, expression of the PDZ2-deleted NHERF2 mutant caused blockage of the PLC-␤ activation by carbachol. Thus, by inhibiting the interaction between endogenous NHERF2 and PLC-␤3, the mutant could block the signal that activates PLC-␤3. Accumulation of inositol phosphates upon carbachol treatment in HeLa cells must be mainly due to the activity of PLC-␤3, since large amounts of PLC-␤3 were detected in the immunoassay, while other isotypes were hardly detectable. These results, therefore, suggest that NHERF2 may regulate PLC activity by anchoring PLC-␤3.
In summary, the C-terminal sequence in PLC-␤3 binds to the second PDZ domain of NHERF2. This conclusion is in good agreement with recent demonstrations of PDZ domains recognizing the C terminus of their target proteins. Furthermore, other regions of PLC-␤3 and the C-terminal region of NHERF2 may also be involved in the interaction. The isotype specificity of the interaction between PLC-␤3 and NHERF2 shows that PDZ domain-containing proteins allow signaling molecules to transduce a specific intracellular signaling cascade. Expression of NHERF2 enhances the PLC-␤3 activation by muscarinic receptors, although NHERF2 does not have a direct effect on the PLC-␤3 activity. This result suggests the possibility that NHERF2 may facilitate the regulation of the PLC-␤3 activity by forming complexes with certain signaling molecules. Further investigations are required to determine how NHERF2 organizes the molecules participating in signal transduction and the cross-talk between signaling pathways.