Dopamine Receptor-interacting Protein 78 Acts as a Molecular Chaperone for Gγ Subunits before Assembly with Gβ*

Heterotrimeric G proteins play a central role in intracellular communication mediated by extracellular signals, and both Gα and Gβγ subunits regulate effectors downstream of activated receptors. The particular constituents of the G protein heterotrimer affect both specificity and efficiency of signal transduction. However, little is known about mechanistic aspects of G protein assembly in the cell that would certainly contribute to formation of heterotrimers of specific composition. It was recently shown that phosducin-like protein (PhLP) modulated both Gβγ expression and subsequent signaling by chaperoning nascent Gβ and facilitating heterodimer formation with Gγ subunits (Lukov, G. L., Hu, T., McLaughlin, J. N., Hamm, H. E., and Willardson, B. M. (2005) EMBO J. 24, 1965-1975; Humrich, J., Bermel, C., Bunemann, M., Harmark, L., Frost, R., Quitterer, U., and Lohse, M. J. (2005) J. Biol. Chem. 280, 20042-20050). Here we demonstrate using a variety of techniques that DRiP78, an endoplasmic reticulum resident protein known to regulate the trafficking of several seven transmembrane receptors, interacts specifically with the Gγ subunit but not Gβ or Gα subunits. Furthermore, we demonstrate that DRiP78 and the Gβ subunit can compete for the Gγ subunit. DRiP78 also protects Gγ from degradation until a stable partner such as Gβ is provided. Furthermore, DRiP78 interaction may represent a mechanism for assembly of specific Gβγ heterodimers, as selectivity was observed among Gγ isoforms for interaction with DRiP78 depending on the presence of particular Gβ subunits. Interestingly, we could detect an interaction between DRiP78 and PhLP, suggesting a role of DRiP78 in the assembly of Gβγ by linking Gγ to PhLP·Gβ complexes. Our results, therefore, suggest a role of DRiP78 as a chaperone in the assembly of Gβγ subunits of the G protein.

Seven transmembrane receptors (7TM-Rs) 3 mediate a large variety of physiological events, from simple responses to endocrine mediators to complex behavioral events. Many diverse extracellular signals transduce signals via heterotrimeric G proteins (1). G protein heterotrimers are composed of ␣, ␤, and ␥ subunits. To date, 16 G␣, 5␤, and 11␥ human genes have been identified, and splice variants for many of these gene products exist (2). Interest originally was centered on G␣ since it possesses the switch that activates and deactivates signal transduction through guanylyl nucleotide exchange and hydrolysis, respectively. Furthermore, the interaction of G proteins with receptors and effectors is dependent on the subtype of G␣ in the G protein heterotrimer. Evidence supporting an increasing contribution of G␤␥ subunits to receptor and effector interaction has accumulated (for review, see Ref. 3). G␤ and G␥ subunits tightly associate as a constitutive heterodimer (4). Human G␥ isoforms are well conserved among vertebrates, although they differ considerably from each other, especially in the N-terminal region (5). Reports suggest that G␥ subunits differ in their ability to interact with G␤ to form specific heterodimers, to interact with 7TM-Rs, and to modulate effectors (6 -9). G␥ subunits are post-translationally modified by isoprenylation (10,11), and it has been suggested that G␤␥ assembly into a stable heterodimeric complex precedes cytosolic prenylation of the C-terminal CAAX motif on G␥ (12). It is also generally agreed that both isoprenylation of G␥ and association with G␣ are essential for plasma membrane targeting of the G␤␥ dimer (12)(13)(14).
To date little is known about mechanistic aspects of G␤␥ heterodimer formation, i.e. what facilitates their initial meeting and determines which combinations are formed in vivo. It has been demonstrated that phosducin-like protein (PhLP) interacts with G␤␥ subunits (15,16). More recent work has identified PhLP as a chaperone for the G␤ 1 subunit as interactions can be detected even in the absence of G␥ (17,18). In the latter study reduction in PhLP levels using RNA-mediated interfer-ence resulted in an inhibition of functional G␤␥ expression and signaling due to a reduction in the ability to form heterodimers between nascent G␤ and G␥ subunits (18). Phosphorylation of PhLP is necessary to promote G␥ recruitment, suggesting that the assembly of G␤␥ may be a regulated biosynthetic event (18,19). These results were also confirmed in Dictyostelium discoideum since G␤ and G␥ association was reduced in a phlp1 Ϫ strain (20) as well as in Caenorhabditis elegans where disruption of the G␤ subunit interacting homologue of PhLP (Mau- 8) results in a loss of G protein signaling (21).
While screening for interactions between accessory/scaffold proteins and partners of 7TM-R core complexes (receptor, G protein, or effector), we observed that the dopamine receptor-interacting protein 78 (DRiP78) could interact specifically with G␥ 2 subunits in the absence of G␤ 1 . DRiP78 (or DnaJc14; standing for DnaJ (Hsp40) homolog, subfamily C, member 14) is an ER membrane-bound protein known to regulate the transport to plasma membrane of various 7TM-Rs including D1 dopamine, M2 muscarinic acetylcholine, and AT1 angiotensin II receptors (22,23). DRiP78 recognizes a conserved FXXXFXXXF motif in these particular target proteins, and this motif is found in a number of other 7TM-Rs. In this study we suggest that DRiP78 also acts as a chaperone associated with G␥ to modulate G␤␥ assembly.

EXPERIMENTAL PROCEDURES
Materials-Reagents were obtained from the following sources: Dulbecco's modified Eagle's medium high glucose and Lipofectamine 2000 Transfection reagent from Invitrogen; fetal bovine serum, anti-FLAG M2 monoclonal antibody, and protein A-Sepharose from Sigma-Aldrich; Alexa555 goat antimouse IgG and Alexa647 goat anti-rabbit IgG from Molecular Probes (Eugene, OR). Coelenterazine H and Coelenterazine 400a from Cedarlane (Hornby, ON, Canada). Unless otherwise stated, all chemicals were of reagent grade and were obtained from Sigma.
Cell Culture and Transfection-HEK293 cells were grown in Dulbecco's modified Eagle's medium high glucose supplemented with 10% fetal bovine serum and transfected using Lipofectamine 2000 as per the manufacturer's instructions. Cells were plated at a density of 3 ϫ 10 5 cells/well in 6-well plates. Unless otherwise indicated, experiments were carried out 24 h after transfection.
Bioluminescence Resonance Energy Transfer (BRET) and Bimolecular Fluorescence Complementation (BiFC) Experiments-HEK293 cells were co-transfected with vectors expressing the GFP-and Rluc-fusion proteins (1 g of each cDNA was transfected into each well of a 6-well plate, and total DNA/dish was kept constant by adding pcDNA vector as required). The exceptions to this rule were BRET saturation experiments where increasing amounts of the GFP-tagged partner was transfected, again with a constant total amount of total DNA transfected. Twenty-four hours after transfection cells were harvested and washed once with phosphate-buffered saline (PBS). The cells were then suspended in PBSϩ (PBS ϩ 0.1% glucose) and distributed into 96-well microplates (white Optiplate; Perkin-Elmer Life and Analytical Sciences). Most experiments were conducted using the BRET 2 technology using coelenterazine 400a at a final concentration of 5 M. Signals were collected on a Packard Fusion instrument (Perkin-Elmer Life and Analytical Sciences) using either 410/80-nm (luciferase) and 515/30-nm (GFP) band pass filters for GFP 10 constructs. BRET 1 was also used for constructs incorporating YFP using coelenterazine H as a substrate. Whether or not BRET occurred was determined by calculating the ratio of the light passed by the 515/30 filter (luciferase) to that passed by the 410/80 filter (GFP) for BRET 2 or by the 450/58 (luciferase) and 535/25-nm band pass filters (YFP) for BRET 1 . This ratio is referred to as the BRET ratio. To avoid possible variations in the BRET signal resulting from fluctuation in the relative expression levels of the energy donor and acceptor, we designed transfection conditions to maintain constant GFP 10 /Rluc expression ratios in each experimental set. BiFC signals were determined by the measurement of the light that passed by the 535/25-nm band pass filters (YFP). BRET background was determined under conditions where resonance energy transfer between Rluc and GFP either could not or did not occur. This was accomplished by expressing Rluc or Rluc-tagged proteins either alone or together with GFP or GFP-tagged proteins, none of which interact physiologically. The background was the same regardless of which of the aforementioned individual proteins or combinations of proteins were expressed, and it has been subtracted to yield net BRET.
Immunoprecipitation-48 h after transfection into 100-mm 2 dishes (for these experiments 4 g of each cDNA was transfected into each dish, and total DNA levels/dish were kept constant by adding pcDNA vector as required) cells were washed with PBS and harvested. Samples were lysed in 0.8 ml of radioimmune precipitation assay buffer (50 mM Tris, pH 7.5, 10 mM MgCl 2 , 150 mM NaCl, 0.5% sodium deoxycholate, 1% Nonidet P-40, 0.1% SDS, 2 g/ml aprotinin, 1 M/ml leupeptin, 10 g/ml soybean trypsin inhibitor, 1 mM phenylmethylsulfonyl fluoride, 100 mM iodoacetamide, and DNase I). The lysate was solubilized by incubation at 4°C for 30 min, precleared with 50 l of protein A-Sepharose beads at 4°C for 1 h, and clarified by centrifugation at 14,000 rpm for 10 min. A, GFP 10 -G␥ 2 (revealed using anti-G␥ 2 -specific antisera) is coimmunoprecipitated (IP) with FLAG-tagged DRiP78, and the interaction was decreased in the presence of G␤ 1 down to the level of background (there is some nonspecific binding to the FLAG matrix). Controls demonstrate that FLAG-tagged G␤ 1 can also immunoprecipitate GFP 10 -G␥ 2 . Results are representative of three independent experiments. IB, indicates immunoblot. B, cells expressing DRiP78-Rluc and different GFP 10 -tagged constructs were harvested, and BRET 2 assays were performed 24 h post-transfection. Data are expressed as BRET 2 ratio means Ϯ S.E. for at least three different experiments. The asterisk indicates p Ͻ 0.05 compared with controls using a two-tailed paired Student's t test. C, untagged DRiP78 competes for the BRET interactions between DRiP78-Rluc and GFP 10 -G␥ 2 , ␤ 2 AR-Rluc and GFP 10 -G␥ 2 , and Rluc-G␤ 1 and GFP 10 -G␥ 2 but not between ␤ 2 AR-Rluc and ␤ 2 AR-GFP 10 . DRiP78-Rluc/CD4-GFP 10 interactions were used as a negative control in all cases and subtracted off as described under "Experimental Procedures" and shown in B. Data are expressed as net BRET 2 ratio with means Ϯ S.E. for at least three different experiments. The double asterisks indicates p Ͻ 0.01 compared with controls using a two-tailed paired Student's t test. D, interactions detected using BRET between G␤ subunits tagged with eGFP and DRiP78-Rluc in the absence or presence of G␥ 2 . Data are expressed as net BRET 1 ratios (subtracting BRET for eGFP-G␤ 1 /CD4-Rluc used as negative controls) with means Ϯ S.E. for at least three different experiments. The asterisk indicates p Ͻ 0.05 compared with negative control using two-tailed paired Student's t test.
The precleared lysate was incubated with primary antibody (M2 mouse anti-FLAG or anti-HA monoclonal antibodies 1:400 dilution) for 30 min, then 50 l of protein A-Sepharose beads were added, and the mixture was incubated for 1 h. After extensive washing with radioimmune precipitation assay buffer, the immunoprecipitated proteins were eluted from beads with 50 l of SDS sample buffer and resolved by SDS-PAGE, and Western blots were performed as described (29). In some cases, to resolve low molecular weight proteins such as the G␥ subunit, we modified our SDS-PAGE system by using a modified Tricine gel system with a spacer layer between the stacking gel and the separating gel (30). Immunoblots were probed with either a monoclonal anti-GFP antibody (BD Bioscience, 1:1000), rabbit polyclonal anti-G␥ 2 antibodies (Santa Cruz, 1:1000), an anti-PhLP antibody (Santa Cruz, 1:1000) or with a monoclonal anti-FLAG M2 (Sigma, 1:1000), monoclonal anti-HA (Covance, 1:1000 dilution), or rabbit polyclonal anti-DRiP78 (1:1000 dilution; a generous gift from Dr, Norbert Tautz). For the experiments using small interfering RNA directed against DRiP78, we used polyclonal anti-FLAG antibodies (Sigma, 1:400 dilution) or monoclonal anti-HA antibodies (Covance, 1:400 dilution) for immunoprecipitation. Horseradish peroxidase-conjugated secondary antibodies were also from Sigma (anti-mouse or anti-rabbit, 1:20,000).
Confocal Microscopy-Twentyfour hours post-transfection using the same conditions described for BRET experiments, HEK293 cells were harvested and seeded on laminin-coated coverslips for 4 h at 37°C. The cells were then fixed for 20 min in PBS, pH 7.4, containing 3% (w/v) paraformaldehyde, then washed 3 times in PBS and incubated for 1 h at room temperature in PBS containing 2% normal goat serum plus 0.2% (v/v) Triton X-100. Excess serum was removed, and the cells were incubated overnight at 4°C with primary antibody diluted in PBS containing 1% normal goat serum and 0.04% (v/v) Triton X-100 (1:100 dilution of rabbit anti-calnexin (StressGen) or 1:200 dilution of mouse M2 monoclonal anti-FLAG (Sigma)). The coverslips were then washed with PBS, drained, and incubated for 1 h at room temperature with the appropriate secondary antibody. The coverslips were washed with PBS, drained, and mounted onto glass slides using show the results of co-immunoprecipitation of HA-tagged G␥ 2 with FLAG-tagged G␤ 1 in vector-or shRNA-co-transfected cells, respectively. The last lane represents a control experiment demonstrating the HA antibodies can also immunoprecipitate G␥ 2 and co-immunoprecipitate FLAG-tagged G␤ 1 . Data are representative of three independent experiments in this case. C, reduction of DRiP78 levels in HEK293 cells results in degradation of G␥ 2 but not G␤ 1 . 24 h post-transfection cells were treated with cycloheximide (100 g/ml) for 3 h, and then cell lysates were processed for Western blots. Vector (lanes 1, 3, 5, and 7)-and shRNA-transfected (lanes 2, 4, 6, and 8) are shown for FLAG-tagged G␤ 1 (lanes 1-4) and HA-tagged G␥ 2 (lanes 5-8). Data are shown from two independent experiments in this case.

DRiP78 Is a Chaperone in G␤␥ Dimer Assembly
a drop of 0.4% 1,4-diazabicyclo(2.2.2)octane (DABCO)/glycerol medium. Coverslips were fixed to the slides with nail polish. Confocal microscopy was performed with an unmodified Zeiss LSM-510 system (with photo-multiplier tube detection) using a 63/1.4 oil Plan-Apochromat objective. Some experiments were also performed using a Leica SP2 system. GFP (green) was excited at 488 nm with an argon laser, emitting fluorescence at 510 nm. Secondary antibodies were goat anti-mouse Alexa-Fluor 555-conjugated or goat anti-rabbit Alexa-Fluor 647-conjugated IgG. Alexa-Fluor 555 was excited with a HeNe1 laser at 543 nm, emitted fluorescence at 570 nm, and is represented in red. Alexa-Fluor 647 was excited with a HeNe2 laser at 633 nm, emitted fluorescence at 668 nm, and is represented in blue. For deconvolution, the three-dimensional image data were transferred to an AZ-10 work station (Azunis Technologies), and images were processed with Huygens 2 software with a theoretical pointspread function running on LINUX RedHat. Deconvoluted Z-stacks were transferred back into LSM-510 software for three-dimensional reconstruction. Control experiments omitting primary antibodies revealed absent or very low level background staining. Experiments with primary antibodies on non-transfected cells indicated no nonspecific staining.

RESULTS
Interaction of DRiP78 with 7TM-R Core Complexes-To initially characterize possible interactions with 7TM-R signaling complexes with DRiP78, we immunoprecipitated FLAG-tagged DRiP78. Co-immunoprecipitation of GFP 10 -G␥ 2 with DRiP78 was seen, suggesting that FLAG-tagged DRiP78 associates with G␥ 2 as part of a complex (Fig.  1A, first lane). In the presence of coexpressed G␤ 1 , however, levels of co-immunoprecipitated G␥ 2 were significantly reduced (Fig. 1A, second lane), suggesting that DRiP78 may be an alternative partner for G␥ 2 . Controls also demonstrate that FLAG-tagged G␤ 1 can immunoprecipitate GFP 10 -G␥ 2 (Fig. 1A, last lane), demonstrating the efficacy of our immunoprecipitation protocol.
These experiments do not address whether DRiP78 and G␥ 2 interact more directly or as part of a larger signaling complex. BRET can be used to directly monitor protein/ protein interactions in living cells (31). We measured BRET using DRiP78 tagged with Renilla luciferase in combination with different potential partner proteins tagged with GFP. These fusion proteins have all been previously demonstrated to be functional (24,32) with the exception of DRiP78-Rluc. However, as we demonstrate below, DRiP78-Rluc and FLAG-tagged DRiP78 have similar functional effects. The ␤ 2 AR contains a theoretical DRiP78 binding motif; thus, we evaluated the potential interaction between the ␤ 2 AR and DRiP78. There was a small but significant interaction between DRiP78-Rluc and ␤ 2 AR-GFP 10 as compared with the negative controls (Fig. 1B). Interestingly, a large BRET signal was detected between G␥ 2 and DRiP78 in accord with our immunoprecipitation data. However, no signal was detected between G␤ 1 and DRiP78 (whether or not other signaling partners such as G␥ 2 or the receptor were present) or between G␣ i or G␣ s and DRiP78. For meaningful information to be taken from resonance transfer experiments, it is essential to devise negative controls using proteins that express at similar levels to those being tested and which are localized to the same subcellular compartment. A CD4-GFP 10 fusion protein was targeted to membranes and expressed at similar levels as our experimental constructs. Background BRET signals were, thus, measured using DRiP78-Rluc and CD4-GFP 10 , yielding a basal BRET ratio FIGURE 3. DRiP78 and G␥ 2 interactions occur in the ER. A, colocalization of DRiP78 and G␥ 2 is observed at the ER. DRiP78 is an ER-membrane-bound protein that colocalizes with calnexin, an ER marker (bottom panel). When GFP 10 -G␥ 2 is co-expressed with DRiP78, colocalization of the three proteins occurs in ER (calnexin)-labeled compartments (top panel). In each of these panels, the right image represents a merged image of the two or three labels shown in the rest of each panel. Data are representative of three independent experiments. B, interference with anterograde ER-Golgi trafficking does not effect the interaction between DRiP78-RLuc and GFP 10 -G␥ 2 . We assessed the interaction using BRET, after blocking anterograde transport at different steps using wild type (as control) and two dominant negative Sar1 GTPases (Sar1 H79G and Sar T39N). Data were expressed as net BRET with the mean Ϯ S.E. of at least three independent experiments. MAY 4, 2007 • VOLUME 282 • NUMBER 18 of 0.075 Ϯ 0.005, a ratio similar to that measured when either DRiP78-Rluc or ␤ 2 AR-Rluc expressed alone (0.067 Ϯ 0.006 and 0.0725 Ϯ 0.004, respectively).

DRiP78 Is a Chaperone in G␤␥ Dimer Assembly
One measure of the specificity of interactions measured using BRET is that the BRET signal should be reduced by coexpression of an untagged version of one of the putative interacting partners. Co-expression of FLAG-tagged DRiP78 reduced BRET signals between GFP 10 -G␥ 2 and DRiP78-Rluc, between Rluc-G␥ 2 and GFP 10 -G␤ 1 , and between GFP 10 -G␥ 2 and the ␤ 2 AR-Rluc but had no effect on the BRET signal between ␤ 2 AR-GFP 10 and ␤ 2 AR-Rluc (Fig. 1C). The fact that the interaction between receptor equivalents in the ␤ 2 AR homodimer was not affected by co-expression of FLAG-tagged DRiP78 suggests that DRiP78 does not play a role in the assembly of the ␤ 2 AR dimer and confirms the specificity of the DRiP78 interaction with G␥ 2 . Expression levels of either the GFP 10 -or Renilla luciferase-tagged partners were not altered in these experiments, suggesting that the reduction in BRET signals were in fact due to competition by the untagged DRiP78 (supplemental Figs. 1 and 2).
To confirm that DRiP78 did not interact with other G␤ subunits, we co-expressed eGFP-tagged versions of G␤ 1-5 with DRiP78-Rluc with and without G␥ 2 . Compared with our positive control for this experiment (i.e. Kir3.1-Rluc/eGFP-G␤ 1 in the presence or the absence of coexpressed G␥ 2 (32)), very little BRET was detected over the background of our negative control (CD4-Rluc/eGFP-G␤ 1 , Fig. 1D). Again, the lack of BRET signals could not be attributed to significant differences in expression of the different eGFP-tagged G␤ subunits (supplemental Fig. 3). These data are consistent with a role for DRiP78 as a G␥-interacting protein.
The fact that no BRET signals were detected between G␤ subunits even in the presence of G␥ suggested that G␤␥ does not interact with DRiP78, an observation that is supported by the experiments described below.
We have recently demonstrated that the receptor and G␤␥ initially interact in the ER (33), and thus, the effect of DRiP78 on this interaction makes sense if DRiP78 modulates the formation of G␤␥. To begin to characterize potential roles for DRiP78 in assembly or function of G␤␥, we first attempted to modulate endogenous levels of DRiP78 using small interfering RNA. Using a vector that also expresses eGFP, we were able to show that the shRNA construct was expressed ( Fig. 2A, left panel) and that it results in a reduced expression of DRiP78 ( Fig. 2A, right panel).
In the presence of the DRiP78 shRNA, the ability of FLAGtagged G␤ 1 and HA-tagged G␥ 2 to co-immunoprecipitate was reduced compared with vector-transfected cells (Fig. 2B). Under these conditions, expression of both G␤ 1 and G␥ 2 was detected, and the effect of altering DRiP78 levels was determined on G␤␥ subunits that did not possess the larger tags used for BRET experiments. To determine whether DRiP78 played a role in G␥ 2 stability, we treated vector-or shRNA-transfected (along with either HA-tagged G␥ 2 or FLAG-tagged G␤ 1 alone) cells for 3 h with 100 g/ml cycloheximide. Our results clearly show that under conditions of reduced DRiP78 expression, G␥ 2 levels are also markedly reduced (compare lanes 5 with 6 and 7 with 8 in Fig. 2C), whereas G␤ 1 levels are not altered (compare lanes 1 with 2 and 3 with 4 in Fig. 2C). These data indicate that FIGURE 4. Effect of DRiP78 on G␤␥ association. A, specificity of DRiP78/G␥ 2 interaction confirmed by a BRET saturation experiment. BRET saturation experiments using a constant background of DRiP78-Rluc with increasing amounts of GFP 10 -G␥ 2 alone (pcDNA3.1 control) or with co-transfected G␤ 1 . Co-expression of G␤ 1 reduces BRET between DRiP78 and G␥ 2 . Specificity is also indicated in that no saturable interactions were detected between CD4-Rluc and GFP 10 -G␥ 2 expressed at similar levels. BRET 50 values were 0.09 and 0.11 for DRiP78-Rluc/GFP 10 -G␥ 2 with or without G␤ 1 , respectively. Data are expressed as the mean Ϯ S.E. of at least three different experiments and were fit using nonlinear regression or linear regression in the case of the negative control. B, cells expressing DRiP78-Rluc, YFP 1-158 -G␤ 1 (YN-G␤ 1 ) and YFP 159 -238 -G␥ 2 (YC-G␥ 2 ) constructs were harvested, and bimolecular fluorescence complementation assays were performed 24 h post-transfection. Data are expressed as RFU means Ϯ S.E. of at least three different experiments. The asterisk indicates p Ͻ 0.05 compared with controls (pcDNA3.1 versus DRiP78) using two-tailed paired Student's t test.
in the absence of stoichiometric amounts of G␤, DRiP78 protects G␥ 2 from degradation if de novo synthesis is blocked and also reduces the interaction of nascent G␤ and G␥, which is indicative of a putative role for DRiP78 as a specific G␥ chaperone.
The role of receptor/DRiP78 interactions is likely related to maturation of receptor-based signaling complexes as has been shown for the D1 dopamine receptor (22). To determine the functional role of the DRiP78/G␥ interaction, we first determined whether DRiP78 and G␥ 2 were colocalized in the ER. DRiP78, as mentioned above, is an ER membranebound protein that colocalizes with calnexin, a classical ER marker (Fig. 3A, bottom panel). When GFP 10 -G␥ 2 was coexpressed with DRiP78, colocalization of the two proteins occurs in the ER, i.e. in calnexin-labeled compartments (Fig. 3A,  top panel). We have recently demonstrated that blocking anterograde receptor transport from the ER with dominant negative Sar1 GTPases has no effect on receptor/G␤␥ interactions (33). We, therefore, tested whether the DRiP78/G␥ interaction was sensitive to dominant negative Sar1. When compared with a control experiment using the wild type Sar1, we could detect no effects of the dominant negative Sar1 GTPases H79G or T39N on the DRiP78/G␥ 2 interaction (Fig. 3B). These results also suggested that the interaction between the two proteins occurs in the ER.
Specificity of the DRiP78/G␥ 2 interaction was also verified using BRET saturation assays (32,34). Experiments performed in cells expressing increasing levels of GFP-G␥ 2 and fixed amounts of DRiP78-RLuc showed that BRET also increased hyperbolically (Fig. 4A), The fact that the BRET between DRiP78-Rluc and GFP 10 -G␥ 2 reached a plateau value as the amounts of expressed GFP 10 -G␥ 2 increased indicates a specific rather than nonspecific protein-protein interaction. The interaction between DRiP78 and G␥ 2 was also competed with G␤ 1 , and this resulted in both a lower maximal BRET signal as well as a slightly reduced BRET 50 affinity parameter (0.58 -0.74). To more carefully assess the effect of DRiP78 on the G␤␥ subunit association, we used a more direct protein complementation assay for G␤␥ interaction based on the reconstitution of a functional YFP carried by either G␤ or G␥. BiFC (35,36) was assessed between YFP 1-158 -G␤ 1 and YFP 159 -238 -G␥ 2 when co-expressed with DRiP78-Rluc. If G␤ 1 and G␥ 2 associate, YFP fluorescence associated with these subunits will be reconstituted. In the absence of association, as when each subunit is expressed alone, no fluorescence was detected as demonstrated in Fig. 4B (columns 3 and 4). When tagged G␤ 1 ␥ 2 were expressed together without DRiP78, a significant fluorescent signal was measured. However, when DRiP78 was co-expressed with tagged G␤ 1 ␥ 2 , the level of fluorescence detected drops by almost 50%. Taken together, these data suggest that G␤ 1 and DRiP78 are mutually exclusive alternative partners for the G␥ 2 subunit. We next wished to examine a possible role for DRiP78 as a chaperone for G␥, analogous to the role of PhLP for G␤.
Characterization of the DRiP78 Interaction with G␥ 2 -The basic paradigm for a molecular chaperone is that it will recognize and selectively bind unfolded or misfolded but not correctly processed proteins to assist in the formation of stable and correctly matured final products (37). After our initial result using the shRNA construct, we tried to characterize the role of the DRiP78 interaction with G␥ 2 . G␤ subunits, when expressed alone, are relatively less stable than when in complex with G␥ (38), although we did not see this when we overexpressed G␤ 1 perhaps due to complementation by endogenous G␥ subunits (Fig. 2C). G␥ subunits, at least in reticulocyte systems, were somewhat more stable in the absence of G␤ (Refs. 38 -40; for review, also see Ref. 41)). However, in cells expressing the DRiP78 shRNA construct, G␥ 2 was clearly less stable. We, therefore, measured the stability of G␤ 1 and G␥ 2 , each tagged with GFP either expressed alone, with their complementary partner, or with DRiP78 in living cells. Cells were again incubated initially with 100 g/ml cycloheximide for 1 h to block de FIGURE 5. Effect of DRiP78 on G␤ and G␥ subunit stability. A, GFP-G␥ 2 was expressed in combination with empty vector (pcDNA3.1), G␤ 1 alone, DRiP78 alone, and DRiP78 and G␤ 1 together to verify the effect of these proteins of GFP-G␥ 2 stability. B, GFP 10 -G␤ 1 was expressed in combination with empty vector (pcDNA3.1), G␥ 2 alone, DRiP78 alone, and DRiP78 and G␥ 2 together to verify the effect of these proteins of GFP 10 -G␤ 1 stability. In both A and B cells were transfected and harvested 24 h post-transfection and treated with cycloheximide (100 g/ml) for the times indicated followed by measures of GFP fluorescence. Data are expressed as RFU means Ϯ S.E. of at least three different experiments. MAY 4, 2007 • VOLUME 282 • NUMBER 18 novo protein synthesis, and then measurements of expressed GFP fluorescence were made at different time points in the continued presence of cycloheximide. When GFP 10 -G␥ 2 was expressed alone, GFP fluorescence decreased by ϳ30% in 3 h (Fig. 5A). Conversely, co-expression of G␤ 1 or DRiP78 alone or in combination completely preserved the stability of GFP 10 -G␥ 2 . In contrast, when GFP 10 -G␤ 1 stability was measured, only G␥ 2 or the combination of G␥ 2 and DRiP78, but not DRiP78 alone, was able to maintain GFP 10 -G␤ 1 stability (Fig. 5B). The stability of G␤ 1 alone or G␤ 1 co-expressed with DRiP78 was diminished by ϳ20% over the course of the experiment. These results again highlight the selective effect of DRiP78 on G␥ 2 over G␤ 1 and suggest a role for DRiP78 in maintaining G␥ 2 in a conformation suitable for proper folding and assembly with G␤.

DRiP78 Is a Chaperone in G␤␥ Dimer Assembly
To achieve signaling specificity, multiple combinations of individual G protein subunits must be associated together as unique heterotrimers. Are there unique molecular chaperones which facilitate assembly of particular G␤␥ pairs or G␣␤␥ heterotrimers? To begin to address these questions, we determined whether different G␤␥ combinations were sensitive to DRiP78 using BiFC. As seen in Fig. 6A, functional YFP molecules could be reconstituted with several G␤␥ combinations (also see Ref. 36). G␤ 1 containing combinations of G␤␥ were all sensitive to either DRiP78-FLAG or DRiP78-Rluc. However, combinations of G␤␥ bearing G␤ 2 or G␤ 4 were insensitive to DRiP78. This data were presented as net BiFC with the relevant YN-G␤ x construct expressed with YC vector subtracted as a negative control. Still, some of the BiFC combinations produce reasonably low fluorescence, and this could be a confounding factor. However, a similar but consistent pattern is seen in our data and that of the Berlot group with respect to fluorescent intensity of G␤␥ pairs using BiFC (36).
Individual G␥ subunits were also able to compete for the interaction between DRiP78-Rluc and GFP 10 -G␥ 2 as measured using BRET. We used untagged G␥ 1 , G␥ 2 , G␥ 3 , G␥ 11 , and a mistargeted G␥ 2mito , which leads to mitochondrial localization of G␤␥ (12) to verify the specificity of interaction of DRiP78 with different G␥ subunits. Our results show that the interaction between DRiP78 and G␥ 2 can be competed to varying extents by all the G␥ isoforms we used (Fig. 6B). However, at equivalent amounts of transfected cDNA for tagged and untagged G␥ 2 , the interaction between GFP 10 -G␥ 2 and DRiP78-RLuc was not totally abolished, and a more stringent competition assay was not performed. The most potent competitors of the DRiP78Rluc/GFP 10 -G␥ 2 interaction were G␥ 2 , G␥ 3 , and G␥ 2mito . G␥ 2 and G␥ 3 are very similar at the sequence level, which may explain why they compete the DRiP78-Rluc/GFP 10 -G␥ 2 to a similar extent. The experiments with G␥ 2mito confirm again that the interaction between G␥ and DRiP78 occurs before targeting (or mistargeting) of the complex out of the ER. The less homologous the isoform was to G␥ 2 , the less potent it was in competing the DRiP78Rluc/ GFP 10 -G␥ 2 interaction. G␥ 1 , with the lowest homology of all isoforms tested, reduced the interaction to the least extent, whereas G␥ 7 and G␥ 11 showed intermediate effects. These results demonstrate a certain specificity of DRiP78 for certain isoforms of G␥ (G␥ 2 and G␥ 3 in particular), suggesting that other chaperones might be more specific for the other G␥ isoforms.
Interactions between G␤-and G␥-specific Chaperones-The fact that G␤ 1 -containing combinations of G␤␥ were more sensitive to DRiP78 led us to investigate whether or not DRiP78 might play a role in specific formation of particular G␤␥ complexes. We, thus, looked for interactions between DRiP78 and the G␤-specific chaperone, PhLP. We could detect interactions FIGURE 6. Specificity of interaction of DRiP78 among G␥ isoforms and G␤␥ combinations. A, cells expressing CD8 (as a control), DRiP78-Rluc, or DRiP78-FLAG were co-expressed with YFP 1-158 -G␤ 1,2,4 (YN-G␤ x ) and YFP 159 -238 -G␥ 1,2,3,7, or 11 (YC-G␥ x ) constructs and harvested, and bimolecular fluorescence complementation assays were performed 24 h post-transfection. Data are expressed as net RFU (i.e. with negative control values for YN-G␤ x and the C-terminal half of YFP subtracted) with means Ϯ S.E. of at least three different experiments. Note that in our hands G␤ 4 ␥ 1 and G␤ 4 ␥ 7 combinations do not form as measured using BiFC. B, the interaction between DRiP78-Rluc and GFP 10 -G␥ 2 was competed with different "cold" G␥ subunits. Cells were harvested, and BRET 2 assays were performed 24 h post-transfection. Data are expressed as net BRET 2 ratios with means Ϯ S.E. of at least three different experiments. The asterisk indicates p Ͻ 0.05, and the double asterisks indicate p Ͻ 0.01 compared with controls using a two-tailed paired Student's t test.

DRiP78 Is a Chaperone in G␤␥ Dimer Assembly
between the two proteins using BRET, which were significantly higher than two independent negative controls, i.e. DRiP78-Rluc and soluble YFP and PhLP-YFP and CD4-Rluc (Fig. 7A). Co-expression of G␤ 1 , G␥ 2 , or G␤ 1 ␥ 2 did not affect the interaction between PhLP and DRiP78, suggesting that they may interact constitutively and independently of their respective cargoes. This interaction was also shown to be saturable, attesting to its specificity, whereas the negative control was effectively linear (subtracted to show net BRET) and was not competed by G␤␥ co-expression in either BRET or co-immunoprecipitation experiments (Fig. 7, B and C). Results suggest that DRiP78 might recognize specific combinations of G␤/PhLP. Furthermore, we show that PhLP can reduce the BRET between DRiP78 and G␥ 2 , similar to the effect of G␤ 1 (Fig. 8A, also see Fig. 3), hinting that it may be involved in recruiting the nascent G␥ subunit-DRiP78 complex to G␤. Interestingly again, there may be a specificity to the chaperone effects of PhLP; expression of either PhLP, DRiP78, or the two together had no effect on the formation of G␤ 3 ␥ 4 or G␤ 4 ␥ 4 as measured by BiFC (Fig.  8B). DRiP78 inhibited the formation of G␤ 1 ␥ 2 , and PhLP favored formation of this complex. Co-expression of both chaperones had an intermediate effect. In the case of G␤ 2 ␥ 2 , DRiP78 had no effect on formation of this complex. However, there was a slight increase in G␤ 2 ␥ 2 in the presence of PhLP. This led us to investigate further the interaction of other G␤ subunits with PhLP. As shown by a number of other studies (17,18,20), we detected a saturable BRET interaction between Rluc-G␤ 1 and PhLP-YFP (Fig. 8C). This was competed by untagged G␤ 2 but not by untagged G␤ 3 , corroborating the notion that PhLP is specific for G␤ 1 and G␤ 2 . The effect of PhLP on the different G␤␥ subunit complexes was not affected by co-expression of stoichiometric amounts of G␣ s (Fig. 8D). However, G␣ s did reduce the effect of DRiP78 on G␤␥ complex, suggesting that heterotrimeric G protein assembly may have distinct requirements for chaperones or they may act cooperatively with each other. Taken together, our results indicate that proteins which interact with individual G protein subunits during biosynthesis may play a role in formation of specific signaling complexes.

DISCUSSION
Although much is known about the physiological roles that G␤␥ subunits play in modulating effectors, less is known about how the heterotrimeric G protein is assembled. It has become clear in recent years that signaling specificity may be in part manifested by the formation of protein complexes composed of unique combinations of receptor, G proteins, effectors, and regulatory molecules (for review, see Ref. 42). We have recently shown that neither dominant negative Sar 1 nor Rab 1 affected G␤␥ formation per se, confirming that they initially interact before the leave the ER (33). We further demonstrated in this study that the initial interaction of G␤␥ with the receptor also occurs in the ER. The G␤␥ complex requires heterotrimer formation together with isoprenylation of the G␥ for plasma membrane targeting (12). G␤␥ remains ER-localized when expressed without G␣. Co-expression of G␣ leads to strong plasma membrane localization of the heterotrimer (12, 14,43,44). G␣ subunits seem to interact with nascent receptor-G␤␥ complexes after they leave the ER on their way to the Golgi apparatus (33). However, it has become clear that assembly of the G␤␥ complex may also be an event regulated by other proteins as well.
Recently, it was suggested that PhLP-1 in concert with the cytosolic chaperonin complex acts as a molecular chaperone for the G␤␥ assembly by specifically interacting with the G␤ subunit (17,18,20). Furthermore, it has been shown that G␥ subunits may facilitate release of G␤ from PhLP-chaperonin complex. Here, we demonstrate using a variety of techniques that DRiP78 assumes an analogous role by associating with G␥ and, thus, may also regulate the assembly of the G␤␥ complex.
DRiP78 is an ER membrane-bound protein known to regulate the transport to plasma membrane via a FXXXFXXXF motif of various seven transmembrane receptors such as dopamine D1, acetylcholine M2 muscarinic, and angiotensin II AT1 receptors (20,22,23). Overexpression or sequestration of DRiP78 leads to retention of D1 dopamine receptors in the ER, reduced ligand binding, and slower kinetics of receptor glycosylation (Ref. 22; for review, see Ref. 45). To date DRiP78 is known for its role in the regulation of the critical exit step of 7TM-Rs from the ER. We also demonstrate that DRiP78 can interact with the ␤ 2 AR, which contains the FXXXFXXXF motif in its proximal C-terminal tail, suggesting a similar role for DRiP78 in ␤ 2 AR maturation. It is possible that DRiP78 acts as a chaperone for the receptor/G␤␥ assembly event as well since this also occurs in the ER (33). G␥ subunits and DRiP78 initially colocalize in the endoplasmic reticulum, presumably facing the cytosolic compartment where they can still interact with G␤. This interaction is not blocked by dominant negative Sar 1 constructs which inhibit anterograde transport to the Golgi. DRiP78 competes with G␤ for the interaction with G␥ and may facilitate its release from the chaperone as well. The fact that shRNA knockdown of endogenous DRiP78 in HEK293 cells also reduce formation of G␤␥ and results in a more rapid loss of G␥ 2 if de novo synthesis is blocked suggests that DRiP78 regulates the stability of G␥ in the absence of its heterotrimeric partners. DRiP78 can, therefore, be considered as a chaperone for G␤␥ assembly by protecting G␥ from degradation until it correctly associates with G␤.
A number of different G␥ subunits were able to compete for the G␥ 2 /DRiP78 interaction when expressed alone. However, when we expressed different G␤ subunits (which do not interact with DRiP78 per se), not all of the combinations were sensitive to DRiP78. This may suggest that either DRiP78 is involved in the formation (or perhaps prevention of formation) of specific G␤␥ combinations. Alternatively, it may suggest that there are other DRiP78 homologues involved in the formation of other combinations. DRiP78 is a member of a large family of J-domain proteins modulating protein-protein interactions (for review, see Ref. 46). Members of the J domain family of molecular chaperones contain a protein-protein interaction domain of ϳ70 amino acids, the so-called J domain, as their only common characteristic. The 702-amino acid HDJ3 protein is highly homologous to rat DRiP78, suggesting that it or other J domain proteins may subserve overlapping but non-identical roles in the assembly of specific G proteins (47). There are no obvious FXXXFXXXF motifs in the G␥ subunits, suggesting that the interaction may be mediated through another domain on DRiP78. This would make intuitive sense if the DRiP78 or other DnaJ proteins act to assemble multiple components of 7TM-R signaling complexes.
Another novel finding of our study is that PhLP and DRiP78 interact physically, and both exhibit specificity toward different combinations of G␤␥. PhLP predominantly affects the formation of G␤␥ combinations containing G␤ 1 or G␤ 2 and has little effect on combinations containing G␤ 3 or G␤ 4 . DRiP78 can interact with several G␥ subunits by virtue of the fact that several G␥ subunits can compete for G␥ 2 interaction with DRiP78. Although it is clear that DRiP78 does not interact directly with any of the G␤ subunits, its effects are more pronounced when G␤ 1 is involved. This may suggest that the DRiP78⅐PhLP complex is also important in selective formation of different G␤␥ pairs. Moreover, expression of the G␣ subunit may also alter this pattern even though we demonstrate that DRiP78 does not interact with G␣. Proteins that may be considered chaperones for G␣ subunits include the J domain containing cysteine string protein (CSP; Ref. 48), GIV, Daple, and FLJ000354 (49). These proteins may be specifically involved in assembling G protein heterotrimers and may also represent potential interacting proteins for PhLP and/or DRiP78. An earlier report noted that G␤␥ subunits also interacted with the cytosolic chaperone Hsp90 (50), suggesting that other chaperones may be involved in the formation of G␤␥ as well. Other factors that may govern the FIGURE 8. Effects of DRiP78 and PhLP on G␤␥ maturation. A, representative BRET saturation experiments using a constant background of DRiP78-Rluc with increasing amounts of GFP 10 -G␥ 2 in the presence of G␤ 1 , PhLP, or both. BRET 50 values were 0.65, 0.22, 0.51, and 0.99 for DRiP78-Rluc/GFP 10 -G␥ 2 with and without PhLP1, PhLP ϩ G␤ 1 , and G␤ 1 , respectively. Data are expressed as the mean Ϯ S.E. of at least three different experiments and were fit using nonlinear regression. B, cells expressing DRiP78-Rluc, PhLP1, or both with YFP 1-158 -G␤ 1 (YN-G␤ 1 ) and YFP 159 -238 -G␥ 2 (YC-G␥ 2 ) were harvested, and bimolecular fluorescence complementation assays were performed 24 h post-transfection. Data are expressed as RFU means Ϯ S.E. of at least three different experiments. The asterisk indicates p Ͻ 0.05 compared with controls (pcDNA3.1 versus experimental condition using two-tailed paired Student's t test. C, representative BRET saturation experiments using a constant background of Rluc-G␤ 1 with increasing amounts of PhLP-YFP in the presence of untagged G␤ 2 or G␤ 3 . BRET 50 values were 0.82, 0.22, and 0.78 for Rluc-G␤ 1 /PhLP-YFP with pcDNA3, untagged G␤ 2 , or untagged G␤ 3 , respectively. Data are expressed as the mean Ϯ S.E. of at least three different experiments and were fit using nonlinear regression. D, cells expressing DRiP78-Rluc, PhLP1, or both with G␣ s , YFP 1-158 -G␤ 1 (YN-G␤ 1 ) and YFP 159 -238 -G␥ 2 (YC-G␥ 2 ) were harvested, and bimolecular fluorescence complementation assays were performed 24 h post-transfection. Data are expressed as RFU means Ϯ S.E. of at least three different experiments. The asterisk indicates p Ͻ 0.05 compared with controls (pcDNA3.1 versus experimental condition using two-tailed paired Student's t test). selectivity and/or kinetics of specific signaling complex formation include modulation of the chaperone complexes by cellular signaling pathways. For example, PhLP phosphorylation by casein kinase 2 modulates the release of the PhLP⅐G␤ subcomplex from the cytosolic chaperonin complex (19). Also, we have recently demonstrated that G␤␥ subunits can interact with their effector molecules such as Kir3.1 potassium channels before being trafficked to the cell surface and that these complexes become receptor-sensitive in internal compartments as well (32). The fact that certain chaperones such as CSP or DRiP78 can also interact with effector molecules such as voltage-gated calcium channels (51)(52)(53) or 7TM-Rs such as such as dopamine D1, acetylcholine M2 muscarinic, and angiotensin II AT1 receptors (20,22,23) and the ␤ 2 AR (this study) hints that these chaperones may also be involved in the formation or trafficking of larger or perhaps even specific complexes of receptor, G protein heterotrimer, and effector. Elaborating the mechanistic aspects of signaling complex formation before trafficking to their ultimate destinations would certainly enhance our understanding of specificity and rapidity inherent in cellular signaling.