Na/H Exchanger Regulatory Factors Control Parathyroid Hormone Receptor Signaling by Facilitating Differential Activation of Gα Protein Subunits*

The Na/H exchanger regulatory factors, NHERF1 and NHERF2, are adapter proteins involved in targeting and assembly of protein complexes. The parathyroid hormone receptor (PTHR) interacts with both NHERF1 and NHERF2. The NHERF proteins toggle PTHR signaling from predominantly activation of adenylyl cyclase in the absence of NHERF to principally stimulation of phospholipase C when the NHERF proteins are expressed. We hypothesized that this signaling switch occurs at the level of the G protein. We measured G protein activation by [35S]GTPγS binding and Gα subtype-specific immunoprecipitation using three different cellular models of PTHR signaling. These studies revealed that PTHR interactions with NHERF1 enhance receptor-mediated stimulation of Gαq but have no effect on stimulation of Gαi or Gαs. In contrast, PTHR associations with NHERF2 enhance receptor-mediated stimulation of both Gαq and Gαi but decrease stimulation of Gαs. Consistent with these functional data, NHERF2 formed cellular complexes with both Gαq and Gαi, whereas NHERF1 was found to interact only with Gαq. These findings demonstrate that NHERF interactions regulate PTHR signaling at the level of G proteins and that NHERF1 and NHERF2 exhibit isotype-specific effects on G protein activation.

The parathyroid hormone receptor (PTHR) 2 is a Family B G protein-coupled receptor (GPCR) that regulates extracellular mineral ion homeostasis and bone growth and turnover. Interaction with its cognate ligands, PTH or the PTH-related peptide (PTHrP), stimulates adenylyl cyclase and phosphatidylinositol-specific phospholipase C (PLC) (1,2). In some cases, occupancy of the PTHR activates only one signaling pathway. For example, in vascular smooth muscle cells, PTH stimulates adenylyl cyclase but not PLC (3,4), whereas in keratinocytes (5,6), cardiac myocytes (7,8), and lymphocytes (9 -11), the PTHR activates PLC but not adenylyl cyclase. In osteoblasts and kidney tubule cells, PTH activates both adenylyl cyclase and PLC (12)(13)(14). Occupancy of the PTHR activates multiple G␣ proteins, and the physiologic responses to PTH may result from contributions of both ␣ and ␤␥ subunits. However, the particular G protein subunit to which the receptor couples varies in a cell-specific manner. Moreover, PTHR stimulation of PLC may arise through activation of G q (4) or G i/o (15,16).
The Na/H exchanger regulatory factor (NHERF) family consists of four related proteins as follows: NHERF1 and NHERF2 that contain two tandem PSD-95/Discs large/ZO-1 (PDZ) domains and an ezrin-binding domain, and NHERF3 and NHERF4 that possess four PDZ domains but no ezrinbinding domain (17). NHERF1 (also known as ezrin-binding phosphoprotein 50, EBP50) shares 52% amino acid identity with NHERF2, also called NHE3 kinase A regulatory protein (E3KARP) (18). NHERF1 and NHERF2 are implicated in protein targeting and in the assembly of protein complexes. They recruit various GPCRs, ion transporters, and other proteins to the plasma membrane of epithelia and other cells (19 -22).
Despite the similarity between their PDZ domains, NHERF proteins exhibit different affinities for PDZ-binding partners. Some NHERF targets, like Taz (23), the PMCA2b Ca 2ϩ -ATPase (24), and the LPA 5 receptor (25) preferentially bind NHERF2. Furthermore, NHERF2 may display distinct binding specificity and physiologic function that is not shared by NHERF1. NHERF2 but not NHERF1, for instance, specifically interacts with PLC-␤3 and plays a key role in PLC-␤3 activation by the PDZ domain-mediated interaction (26). Ca 2ϩdependent inhibition of NHE3 requires an NHE3-NHERF2-␣actinin-4 complex for oligomerization and endocytosis (27). NHERF2 specifically interacts with the LPA 2 receptor and defines the specificity and efficiency of receptor-mediated PLC-␤3 activation (28).
Mahon et al. (21) reported that NHERF2 inhibited adenylyl cyclase by stimulating inhibitory G␣ i and increased PLC in PS120 cells transfected with the PTHR. In contrast, NHERF1 increased PTH-stimulated cAMP accumulation in ROS 17/2.8 cells (29). Adding to the variability of effects, both NHERF1 and NHERF2 increased PTH-stimulated PLC activity or intracellular calcium in PS120 cells, opossum kidney cells, and ROS 17/2.8 cells (21, 29 -31), although no differences in PTH-stimulated cAMP formation were found in wild-type and NHERF1-null proximal tubule cells (32,33) or in CHO-N10-R3 cells in the presence or absence of NHERF1 (34).
The molecular mechanism by which NHERF association with PTHR promotes switching of receptor signaling between adenylyl cyclase and phospholipase C is not known. It has been speculated that the NHERF proteins may promote G q -mediated signaling by tethering G␣ q effectors such as PLC (26,35), PKC (36), and PKD (37) in the vicinity of receptors. However, it is also possible that NHERF-GPCR interactions might directly modulate the G protein-coupling preferences of the receptors. The most direct and unambiguous way to determine the influence of NHERF1/2 on PTHR signaling is to measure effects on G protein activation. We show here that NHERF1 increases PTH-stimulated PTHR coupling to G␣ q but not to G␣ s or G␣ i . In contrast, NHERF2 decreases PTH-induced G␣ s and increases G␣ q and G␣ i activation. These data reveal that NHERF-PTHR interactions can directly influence receptor coupling to G proteins.

EXPERIMENTAL PROCEDURES
HA.11 and His 6 monoclonal antibodies were obtained from Covance (Berkeley, CA). NHERF1 rabbit polyclonal antibody was purchased from Affinity Bioreagents (Golden, CO). NHERF2 rabbit polyclonal antibody was kindly provided by Dr. R. A. Frizzell (University of Pittsburgh). Polyclonal G␣ s antibody was obtained from Millipore (Billerica, MA). G␣ q monoclonal antibody was provided by BD Transduction Laboratories. G␣ i polyclonal antibody was from NewEast Biosciences (Malvern, PA). Ni-NTA-agarose was provided by Qiagen (Valencia, CA). Horseradish peroxidase-conjugated goat antirabbit secondary antibody was from Pierce. Horseradish peroxidase-conjugated sheep anti-mouse antibody was from GE Healthcare. Lipofectamine 2000 and geneticin, protein A-Sepharose 4B conjugate, and rec-protein G-Sepharose 4B conjugate were obtained from Invitrogen. Protease inhibitor mixture Set I and pertussis toxin were from Calbiochem. Human PTH(1-34) was purchased from Bachem (Torrance, CA). [ 35 S]GTP␥S (1,250 Ci/mmol) was purchased from PerkinElmer Life Sciences. FuGENE 6 was purchased from Roche Applied Science. All other reagents were from Sigma.
Membrane Preparation-Plasma membranes were isolated by differential centrifugation at 4°C as described previously (1,39,40). Briefly, after cells achieved confluence, they were rinsed with cold PBS and then incubated with hypotonic buffer (10 mM Hepes, 0.5 mM EDTA, pH 7.4) for 15 min. Swollen cells were harvested, collected by centrifugation (1000 ϫ g for 10 min), and resuspended in ϳ9 volumes of 10 mM Tris, 1 mM EDTA, pH 7.4, with proteinase inhibitor mixture set I. Cells were disrupted with 20 -40 strokes in a "loose" Dounce homogenizer on ice. The lysates were centrifuged at 1000 ϫ g for 10 min to remove unbroken cells, large cell debris, and some nuclei. The supernatant was further centrifuged at 30,000 ϫ g for 20 min. The membrane pellet was resuspended in freezing buffer (10 mM Hepes, 0.1 mM EDTA, pH 7.4) at a protein concentration of 5-10 g/l and rapidly frozen in liquid nitrogen. Membranes were then stored at Ϫ80°C until used.
GTP␥S Binding and Immunoprecipitation of GTP␥S-bound G␣ Subunits-The comparative changes in G␣-[ 35 S]GTP␥S binding immunoprecipitated by specific G␣ subtype-specific antisera were used to delineate PTHR coupling to distinct G␣ protein subunits. [ 35 S]GTP␥S binding to G proteins was performed with a modification of previously described methodologies (1,40). Frozen membrane aliquots (150 g) were incubated with 100 l of assay buffer (10 mM Hepes, 100 mM NaCl, 5 mM MgCl 2 , pH 7.4) containing 5 M GDP, 5 nM [ 35 S]GTP␥S, and 100 nM PTH(1-34) at 30°C for 5 min (unless otherwise stated). Incubations were terminated by the addition of 800 l of ice-cold assay buffer and immediate transfer to an ice bath. Cell membranes were recovered from the reaction mixture by centrifugation at 20,000 ϫ g for 10 min, and the resulting supernatant was removed. Membrane pellets were solubilized, and immunoprecipitation of [ 35 S]GTP␥S bound to G␣ subunits was measured as described below under "Coimmunoprecipitation and Immunoblot Analysis." After Sepharose beads were washed three times, the beads were resuspended with 100 l of 0.5% SDS and incubated at 85-90°C for 2-3 min. The entire contents of each tube were transferred to a vial containing 5 ml of scintillation mixture, and radioactivity was measured by ␤-emission spectrometry. Nonspecific binding was determined in the presence of 100 M GTP␥S.
Coimmunoprecipitation and Immunoblot Analysis-Interaction of G proteins with NHERF1 or NHERF2 was analyzed as described previously (41). In brief, 6-well plates of CHO-R3 cells were transiently transfected with His-NHERF1, His-NHERF2, or empty vector. Forty eight hours later, the cells were lysed with 1% Lubrol, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 7.4, and 150 mM NaCl. Solubilized materials were incubated overnight at 4°C with Ni-NTA-agarose or incubated with anti-G␣ s polyclonal antibody, G␣ q monoclonal antibodies, or G␣ i polyclonal antibody for 1 h at 4°C and then protein A or protein G-Sepharose 4B conjugate was added to each sample and incubated overnight at 4°C. Total lysates and immunoprecipitated protein, eluted by the addition of SDS sample buffer, were analyzed by SDS-polyacrylamide gels and transferred to Immobilon-P membranes (Millipore) using the semi-dry method (Bio-Rad). Membranes were blocked overnight at 4°C with 5% nonfat dried milk in Trisbuffered saline plus Tween 20 and incubated with different antibodies (anti-His antibody (1:1000); anti-G␣ s (1:500); anti-G␣ q (1:250); anti-G␣ i (1:500); anti-HA (1:1000); anti-NHERF1 (1:1000); anti-NHERF2 (1:4000); or anti-actin (1:2000)) for 2 h at room temperature. The membranes were then washed and incubated with goat anti-rabbit IgG conjugated to horseradish peroxidase or anti-sheep mouse IgG conjugated to horseradish peroxidase (1:5000) for 1 h at room temperature. Protein bands were revealed with a luminol-based enhanced chemiluminescence substrate.
Overlay Assay-His/S-tagged PDZ1 and PDZ2 domain proteins (1 g/lane) were spotted on nylon membrane as described previously (42). The membranes were blocked with blot buffer for 30 min at room temperature and then overlaid with 100 nM GST-tagged C-terminal 22-amino acid fragments of wild-type PTHR (PTHR-ctETVM) or its mutant form PTHR-ctETVA in blot buffer for 1 h at room temperature. The blots were washed and incubated with horseradish peroxidase-conjugated anti-GST antibody (1:3000). Interactions of the GST fusion proteins with the PDZ domains were visualized by chemiluminescence.
Rescue forms of human NHERF1 and NHERF2 (resNHERF1 and resNHERF2, respectively) resistant to their respective shRNA were generated by introducing three silent mutations in the NHERF1 and NHERF2 sequences using primers listed in Table 1. The mutations were constructed by PCR using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The fidelity of plasmids was confirmed by sequencing (ABI PRISM 377, Applied Biosystems, Foster City, CA). Rescue shRNA scrambled constructs containing the 29-mer scrambled cassette were purchased from OriGene.
FRET-G␣ i activation was measured in real time in live cells by fluorescence resonance energy transfer (FRET) as described previously (43,44). Briefly, HEK-293 cells plated on MatTek (Ashland, MA) dishes were maintained in Hepes buffer with 0.1% (w/v) bovine serum albumin at 22°C. Cells were observed using a 40 ϫ 1.30 NA oil immersion objective on a Nikon A1s confocal microscope attached to a Ti-E inverted base. Subunit rearrangement of G␣ i and G␤␥ was measured by FRET between YFP-tagged G␣ i and cerulean-tagged G␤ 1 ␥ 2 bimolecular fluorescence complementation. FRET signal was measured as the normalized FRET ratio (nFRET) of the YFP and cyan fluorescent protein emission (F YFP /F CFP ) (45).
Receptor Binding-Receptor binding was performed as described previously (34)  Adenylyl Cyclase-Adenylyl cyclase activity was determined by assay of cAMP accumulation as described previously (34). Briefly, HEK-293R cells transfected with scrambled shRNA, NHERF1 shRNA, or NHERF2 shRNA in 24-well plates were labeled with 0.5 Ci of [ 3 H]adenine for 2 h. The cells were then treated with vehicle or 100 nM PTH  in the presence of phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (1 mM) for 15 min. The reaction was terminated by addition of 1 M TCA. cAMP was isolated by the two-column method. Intracellular where F is the measured fluorescence intensity; F max is the fluorescence measured after addition of 10 M ionomycin; F min is the fluorescence measured after addition of 10 mM EGTA, and K d is the dissociation constant of the dye-Ca 2ϩ complex (520 nM) (46).
Statistics-Data are presented as the mean Ϯ S.E., where n indicates the number of independent experiments. Multiple Three silent mutations (underlined) were introduced in the shRNAs targeting NHERF1 or NHERF2.

Gene sequences
comparisons were evaluated by analysis of variance with posttest repeated measures analyzed by the Duncan procedure using Prism (GraphPad Software, Inc., San Diego). Differences of p Ͻ 0.05 were assumed to be significant.

RESULTS
Specificity of PTH-induced G␣ Subunit Activation-We first determined the constitutive expression of G␣ protein subunits in membranes of PS120-R cells. Fig. 1A shows that PS120-R cells express G␣ s , G␣ q , and G␣ i proteins in cell membranes. We then measured PTH-stimulated [ 35 S]GTP␥S binding to G␣ s , G␣ q , and G␣ i . PTH increased [ 35 S]GTP␥S binding to G␣ s protein by 3.6-fold (Fig. 1B) NHERF2 Inhibits PTH-stimulated G␣ s and Increases G␣ q and G␣ i Activation-Mahon et al. (21) reported that NHERF2 switched PTHR signaling from adenylyl cyclase to PLC in PS120-R cells stably transfected with NHERF2 (PS120-R-N2). Pertussis toxin pretreatment of PS120-R-N2 cells markedly inhibited PTH activation of PLC and enhanced activation of adenylyl cyclase, implying that PTH stimulates G i/o proteins when the PTHR is bound to NHERF2. We tested the effects of NHERF2 on resting and PTH-stimulated G␣ s , G␣ q , and G␣ i exchange. PS120-R cells express small amounts of NHERF1 (47) but not NHERF2 (Fig. 3A), whereas PS120-R-N2 cells express NHERF2 mostly at cell membranes (Fig. 3A). NHERF2 did not affect basal G␣ s , G␣ q , or G␣ i activity but significantly blunted PTH-stimulated G␣ s activation (Fig. 3B). Conversely, NHERF2 augmented G␣ q and G␣ i activation (Fig. 3B). These effects occurred without a detectable change in cell membrane PTH receptor abundance (Fig. 3C), suggesting that the action of NHERF2 on G␣ GTP exchange is not due to altered abundance of the PTHR or of ligand binding to the PTHR.  NHERF1 and NHERF2 Differentially Regulate PTH-stimulated G␣ Activation-The ability of PTH to increase the coupling of the receptor to G␣ s and G␣ q correlates with ligandinduced, receptor-dependent sensitivity of adenylyl cyclase and PLC signaling (1). Based on the described differences of NHERF1 and NHERF2 regulation of PTHR signaling (21,29,34), we hypothesized that NHERF1 and NHERF2 might differentially regulate PTH-stimulated G␣ protein activation. To compare the effects of NHERF1 and NHERF2 on G s activation, we used CHO cells stably transfected with PTHR (CHO-R), which lack detectable expression of NHERF1 (34) or NHERF2 (data not shown) but express similar levels of G␣ s , G␣ q , and G␣ i proteins (40). We transiently transfected CHO-R cells with His-NHERF1 or His-NHERF2, resulting in similar levels of cell membrane expression (Fig. 4A). In the absence of NHERF1 or NHERF2, PTH activation of G␣ s was greater than that of G␣ q , but no effects on G␣ i activation were detected (Fig. 4B). Neither NHERF1 nor NHERF2 affected basal G␣ activity (data not shown). In the presence of NHERF1, PTH significantly enhanced G␣ q activity without an effect on G␣ s or G␣ i (Fig. 4B). Thus, NHERF1 selectively promotes receptor coupling to G␣ q . In contrast, NHERF2 significantly inhibited PTH-stimulated G␣ s but enhanced PTH-induced activation of G␣ q and G␣ i (Fig.  4B). NHERF2 therefore influences receptor coupling to all three G ␣ proteins, promoting opposite effects on G␣ s and G␣ i , but like NHERF1 increasing PTH-dependent G␣ q activity. NHERF1 and NHERF2 did not affect receptor number as evidenced by comparable PTH binding to the PTHR (Fig. 4C), consistent with previous reports on CHO and ROS 17/2.8 cells (29,34).
The PTHR, through its C-terminal ETVM PDZ recognition sequence, interacts with NHERF1 by binding to PDZ1 and

. NHERF2 inhibits PTH-stimulated G␣ s binding and increases both G␣ q and G␣ i binding in PS120-R cell membranes.
A, equal amounts of membrane (Mem) and cytosolic (Cyt) proteins (20 g) from PS120-R cells or PS120-R-N2 cells were immunoblotted with NHERF2 antibody. B, PTH-stimulated [ 35 S]GTP␥s binding to G␣ s , G␣ q , or G␣ i protein was measured in PS120-R or PS120-R-N2 cell membranes. Data are summarized as the mean Ϯ S.E. of four independent experiments. **, p Ͻ 0.01; ***, p Ͻ 0.001, compared with PS120-R plus PTH group. C, cell surface binding of 125 I-PTH(1-34) in PS120-R cells or PS120-R-N2 cells was measured as described under "Experimental Procedures." Data are summarized as the mean Ϯ S.E. of triplicate determinations. PDZ2 (34,48). PTHR interactions with NHERF2 PDZ domains have not been described. Here, we simultaneously compared the interactions of GST-tagged C-terminal 22 amino acid peptide fragments of the wild-type PTHR (PTHR-ctETVM) and a mutant form PTHR-ctETVA, which cannot bind NHERF1, with PDZ1 and PDZ2 domains of NHERF1 and NHERF2 (Fig.  5A). The results show that the PTHR preferentially interacts with the PDZ1 domain of NHERF1 and PDZ2 of NHERF2. These associations were abolished with the PTHR harboring the mutated PDZ interaction motif.
Additional examination of the NHERF-mediated switch of G protein activation was undertaken in HEK-293 cells, which constitutively express NHERF1 and NHERF2. In HEK-293 cells, transfected with wild-type PTHR-ETVM (Fig. 5B), PTH significantly activated G␣ s , G␣ q , and G␣ i (Fig. 5C), consistent with a previous report that PTH promoted activation of G␣ s , G␣ q/11, and G␣ i in HEK-293 cells (1). The PTHR-ETVA, which does not bind NHERF (49), showed decreased PTH-stimulated G␣ q activation and increased PTH-stimulated activation of G␣ s (Fig. 5C), although G␣ i activation was absent. To delineate the individual effects of NHERF1 and NHERF2 on G protein activation, endogenous NHERF1 or NHERF2 expression was silenced by RNA interference. A scrambled shRNA was used as a control. NHERF1 or NHERF2 shRNA reduced endogenous NHERF1 or NHERF2 levels by 78 and 82%, respectively, compared with a scrambled control (Fig.  5D). NHERF1 shRNA did not interfere with NHERF2 expression, and conversely, NHERF2 shRNA did not affect NHERF1 expression, demonstrating the specificity of the knockdown of endogenous NHERF1 and NHERF2 by their respective shRNAs. Neither shRNA affected basal [ 35 S]GTP␥S binding to G␣ subunits (data not shown). Knockdown of NHERF1 expression selectively inhibited PTH-stimulated activation of G␣ q (Fig. 5E). Silencing NHERF2 expression, in contrast, significantly increased PTH-stimulated G␣ s and inhibited G␣ i activation. To rule out off-target effects of shRNA, we generated NHERF1 and NHERF2 rescue constructs (resNHERF1 and resNHERF2) harboring silent mutations to their respective shRNA and then conducted rescue experiments in HEK-293R cells. Expression of resNHERF1 blocked shNHERF1 inhibition of PTH-stimulated activation of G␣ q (Fig. 5F). resNHERF2 abolished shNHERF2 increases of PTH-stimulated G␣ s and inhibition of G␣ i activation.
Further and independent characterization of the dynamic interactions of NHERF1 and NHERF2 with G␣ i was conducted using real time FRET in living cells. PTH activated G␣ i in HEK-293 cells transfected with PTHR-ETVM compared with the cells transfected with PTHR-ETVA (Fig. 5G). NHERF1 shRNA did not affect PTH-induced G␣ i , whereas the FRET signal was abolished with shNHERF2. These data further confirm that NHERF2 specifically increases PTH-induced G␣ i activation. Taken together, these results provide a mirror image of the effects of individual NHERF1 and NHERF2 actions on CHO-R cells and show that NHERF1 augments receptor-mediated stimulation of G␣ q but has no effect on stimulation of G␣ i or G␣ s , whereas NHERF2 enhances receptor-mediated stimulation of both G␣ q and G␣ i but decreases stimulation of G␣ s .

NHERF2 but Not NHERF1 Binds G␣ i and Regulates Second
Messenger Signaling Pathways-NHERF1 interacts directly with G␣ q (35). We analyzed NHERF1 and NHERF2 binding to G␣ s , G␣ q , and G␣ i . CHO-R cells were transiently transfected with His-NHERF1 or His-NHERF2. Neither NHERF1 nor NHERF2 interacted with G␣ s (data not shown). However, both NHERF1 and NHERF2 bound G␣ q , and the interactions were enhanced in the presence of PTH (Fig. 6A). These data are consistent with the report that NHERF1 interacts with G␣ q and to a greater extent with G␣ q -R183C, . a constitutively active G␣ q mutant (35). We next investigated the interaction of NHERF1 and NHERF2 with G␣ i . NHERF2 but not NHERF1 coimmunoprecipitated with G␣ i by using Ni-NTA-agarose followed by immunodetection with a G␣ i antibody (Fig. 6B). Likewise, the interaction could be detected by using a G␣ i antibody for immunoprecipitation and immunodetection of NHERF2 with a His antibody. The association was also enhanced in the presence of PTH (Fig. 6B). Importantly, NHERF2 interacted with endogenous G␣ i in native CHO cells not expressing the PTHR (Fig. 6C), indicating that the interaction between NHERF2 and G␣ i is receptor-independent.
Since NHERF1 and NHERF2 differentially regulate PTHstimulated G␣ i activation, their effects on PTH-induced second messenger signaling pathways in the presence or absence of pertussis toxin should differ. Therefore, we examined the effect of NHERF1 or NHERF2 on PTH-stimulated cAMP formation and [Ca 2ϩ ] i , as an index of PLC activity, in HEK-293R cells. Pertussis toxin (100 ng/ml) pretreatment for 16 h markedly increased PTH activation of adenylyl cyclase (Fig. 7A), without affecting the magnitude of PTH-induced [Ca 2ϩ ] i (Fig. 7B, top) , consistent with a previous report regarding pertussis toxin effects on PTH-stimulated cAMP formation and [Ca 2ϩ ] i (15). Silencing NHERF1 did not affect PTH-stimulated cAMP formation or maximal [Ca 2ϩ ] i . Pertussis toxin increased PTHstimulated cAMP accumulation and decreased [Ca 2ϩ ] i in the presence of shNHERF1 (Fig. 7, A and B, middle). As expected, pertussis toxin had no further action on PTH-stimulated cAMP production or [Ca 2ϩ ] i after knockdown of NHERF2 (Fig. 7, A  and B, bottom). Taken together, these results show that NHERF1 and NHERF2 differentially regulate PTH-stimulated G protein activation and second messenger signaling pathways.

DISCUSSION
PTH activates multiple second messenger signaling pathways that are reportedly coupled by distinct G proteins to the PTHR. Indirect approaches suggest that the PTHR is capable of coupling to G␣ s and to multiple G␣ q family members (2) and G␤␥ subunits (50). However, to the best of our knowledge, only a single report analyzed PTHR activation of G proteins (1). Schwindinger et al. (1) measured PTHR coupling to G␣ proteins in HEK-293 cell lines heterologously expressing the PTHR at low (C20; 40,000 receptors/cell) or high (C21; 400,000 receptors/cell) and in ROS 17/2.8 osteosarcoma cells, which constitutively express 72,000 receptors/cell (51). The ability of PTH(1-34) to activate G␣ s and G␣ q , measured by [␣-32 P]GTP-␥-azidoanilide binding followed by immunospecific G␣ subunit detection, correlated with the magnitude of ligand-induced receptor-dependent sensitivity of adenylyl cyclase and PLC signaling pathways. PTH promoted activation of G␣ s , G␣ q/11 , and G␣ i , albeit to a lesser extent, in C21 but not in C20 cells. In ROS 17/2.8 cells, PTH activated G␣ s to a greater degree than in C21 cells but had no effect on G␣ q/11 or G␣ i . NHERF proteins had not been discovered that at the time those studies were performed. Our data now permit some comparison of the results and show that the expression of NHERF1 and NHERF2 are, at least in part, responsible for the heterogeneity of the cellular responses to PTH. NHERF1 permits selective activation of G␣ q , whereas NHERF2 attenuates PTHstimulated G␣ s , while enhancing PTH-induced activation of G␣ q and G␣ i . These findings explain why PTH stimulated G␣ q and G␣ i in C21 cells, which express both NHERF1 and NHERF2, but not in ROS 17/2.8 cells, which express neither. Consistent with this interpretation, PTH does not stimulate [Ca 2ϩ ] i in native ROS 17/2.8 cells (29). Upon transfection of NHERF1, however, PTH significantly increased [Ca 2ϩ ] i . Notably, relatively high PTHR density is required for efficient activation of PLC, in contrast to receptor activation of adenylyl cyclase that occurs at physiologic receptor concentrations (52). This may contribute to the limited PLC activation in ROS 17/2.8 cells.
To demonstrate the generality of the effects of NHERF1 and NHERF2, we studied the PTH-dependent activation of heterotrimeric G proteins in several cell lines that have been used as models to analyze the heterogeneity of the responses to PTH. We chose PS120 fibroblasts, CHO-derived cell lines, and HEK-293 cells, all of which were transfected with either the WT-PTHR or a mutated form with an impaired PDZbinding motif. Our results are summarized in Table 2. Several conclusions can be drawn from these results. For instance, the data clearly indicate that coupling of the PTHR to G␣ i absolutely requires the expression of NHERF2. Furthermore, in the absence of NHERF1 and NHERF2, the receptor couples primarily to G␣ s . Finally, G␣ q coupling is significantly increased by either NHERF1 or NHERF2.
These results are consistent with previous reports examining the downstream effects of PTH in various cell lines and now provide a mechanism to reconcile apparent discrepancies between different cells or cell lines. Furthermore, our findings provide a solid experimental framework to explain the complex spectrum of cellular responses to PTH. For example, in PS120 cells stably transfected with the PTHR, the expression of NHERF2 decreased cAMP responses and increased PLC-dependent effects (21). Our data show that this effect is due to increased G␣ i and decreased G␣ s activation in the presence of NHERF2. Likewise, HEK-293R robustly activate cAMP and Ca 2ϩ -dependent responses. The present findings support the conclusion that this is due to NHERF1/NHERF2-dependent modulation of the pattern of G protein activation downstream of the PTHR in these cells.  . NHERF2 but not NHERF1 binds G␣ i . A, PTH increased the interaction of G␣ q protein with NHERF1 or NHERF2. CHO-R cells were transfected with His-NHERF1 or His-NHERF2. His-tagged proteins were precipitated with Ni-NTA-agarose. The precipitated protein was then immunoblotted (IB) with G␣ q antibody. B, PTH increased the interaction of G␣ i proteins with NHERF2. CHO-R cells were transfected with His-NHERF1 or His-NHERF2. His-tagged proteins were precipitated with Ni-NTA-agarose or endogenous G␣ i protein was precipitated with G␣ i antibody. The precipitated proteins were immunoblotted with either G␣ i or His antibody. C, interaction of G␣ q or G␣ i subunits with NHERF1 or NHERF2 in CHO cells was measured and described as above. Data are representative of three independent experiments. IP, immunoprecipitated.
Despite the similarity of their PDZ domains and C-terminal ezrin-binding domain, NHERF1 and NHERF2 exhibit different affinities for PDZ-binding partners and GPCR signaling (20,25,53,54). These results show that the PTHR binds preferentially to PDZ1 of NHERF1 and PDZ2 of NHERF2. Because binding to both NHERF proteins is mediated by the same C-terminal sequence, the PTH1R cannot simultaneously interact with NHERF1 and NHERF2. Thus, the relative levels of expression of NHERF1 and NHERF2 may influence the differential coupling of the PTHR to different downstream signaling pathways. Importantly, although it is clear that the expression of NHERF1 and NHERF2 influences PTHR signaling, some variations among cell lines have been reported, indicating the influence of additional factors in PTHR signaling. For instance, NHERF1 increases PTH-stimulated cAMP accumulation in ROS 17/2.8 cells (29) but decreases cAMP responses in OKH cells (30). Furthermore, PTH-stimulated cAMP production in wild-type and NHERF1-null proximal tubule cells was comparable (32,33), as it is in CHO-N10-R3 cells in the presence or absence of NHERF1 (34). Our results indicate that NHERF1 has no effects on PTH-induced G␣ s or G␣ i activation; therefore, the different effects of NHERF1 on the production of cAMP by these cell lines are probably due to other factors that remain unidentified.
The effects of NHERF2 on adenylyl cyclase activation are somewhat less diverse. NHERF2 expression markedly inhibited adenylyl cyclase in PS120 cells transfected with the PTHR, a result that is consistent with the differential activation of G␣ s and G␣ i induced by NHERF2 expression (21). Both NHERF1 and NHERF2 increase PTH-stimulated PLC activity and intercellular calcium in PS120 cells, opossum kidney cells, or ROS 17/2.8 cells (21, 29 -31). These findings are consistent with the model presented here. Both NHERF1 and NHERF2 increase the activation of PLC␤ and the generation of intracellular Ca 2ϩ transients. However, despite the similarities between NHERF1 and NHERF2, the coupling of the PTHR to calcium signaling is mediated by distinct mechanisms. In HEK-293R cells (which express both NHERF1 and NHERF2), the PTHR stimulates Ca 2ϩ release by a mechanism that is insensitive to pertussis toxin. Thus, in the presence of both NHERF1 and NHERF2, a G␣ q -driven mechanism predominates. Knockdown of NHERF1 has no effect on the magnitude or duration of the Ca 2ϩ transients but reveals significant sensitivity to pertussis toxin. This suggests that G␣ i activation is important for Ca 2ϩ release when only NHERF2 is present. Finally, knockdown of NHERF2 significantly reduces the magnitude of the Ca 2ϩ release response, which in this case remains insensitive to pertussis toxin. A comparison of the magnitude of these responses suggests that NHERF2 is more efficient in the coupling of Ca 2ϩ responses, which are mediated by the engagement of both G q and G i , whereas NHERF1 only supports G qmediated responses.
Because NHERF2 was reported to inhibit adenylyl cyclase by stimulating G␣ i proteins in PS120 cells stably transfected with the PTHR or stably transfected with both the receptor and NHERF2 (21), we employed these cell models to investigate NHERF2 coupling of the PTHR to different G␣ protein subunits. The results show that NHERF2 significantly inhib- ited PTH-stimulated G␣ s activation but activated both G␣ q and G i , consistent with the reported NHERF2 regulation of PTHR second messenger signaling pathways (21). We further detailed the effect of NHERF1 on PTHR activation of G␣ subunits by using CHO cells stably transfected with PTHR (CHO-R cells), which do not constitutively express detectable NHERF1 or NHERF2. Consistent with the results with PS120-R-N2 cells, transient expression of NHERF1 did not affect PTH-induced activation of G␣ s or G␣ i . However, expression of NHERF2 significantly inhibited PTH-induced G␣ s but increased coupling of the PTHR to G␣ q and G␣ i proteins. Nonetheless, some inconsistencies regarding NHERF1 regulation of PTH-stimulated adenylyl cyclase activity have been reported (29,30,(32)(33)(34). These differences may be ascribed to NHERF1 regulation of PTHR trafficking, cell specificity, or as yet unrecognized factors.
The precise mechanistic basis of the differential regulation of the G protein coupling of the PTHR by NHERF1 and NHERF2 remains to be explained. One plausible explanation is suggested by the data shown in Fig. 6. Both NHERF1 and NHERF2 coimmunoprecipitated with G␣ q , but only NHERF2 was able to pull down G␣ i . The specific regions of NHERF1 and NHERF2 that mediate these putative interactions remain unknown. The next question is as follows. How do the interactions with NHERF1 and NHERF2 with G proteins influence PTHR signaling? The simplest model involves the pre-formation of a multimeric complex involving the PTHR, NHERF1/NHERF2, and a heterotrimeric G protein. We propose that the formation of this complex, by bringing the receptor and the G protein into close proximity, facilitates G protein activation. Thus, NHERF1, which binds G␣ q , will promote G q activation, and NHERF2, which binds G␣ q and G␣ i , will enhance the activation of these two G proteins. The differential effects of NHERF1 and NHERF2 on G s activation may be due to the different topology of the interactions of the two PDZ proteins with the PTHR. It is conceivable that NHERF1 binding to PTHR, which is mediated by the N-terminal PDZ1 domain, does not interfere with G s binding. Conversely, NHERF2-PTHR interactions, which involve the central PDZ2 of NHERF2, may induce conformational changes that result in reduced affinity for G␣ s .
In summary, NHERF1 interacts with G␣ q and increases PTH-stimulated PTHR coupling to G␣ q but has no effect on G␣ s or G␣ i . In contrast, NHERF2 interacts with both G␣ q and G␣ i . Therefore, NHERF1 and NHERF2 control PTHR signaling by differential coupling of G␣ proteins to the PTHR. These novel findings explain the differences between NHERF1 and NHERF2 on downstream regulation of PTHR signaling.