C-terminal mutation of G protein beta subunit affects differentially extracellular signal-regulated kinase and c-Jun N-terminal kinase pathways in human embryonal kidney 293 cells.

G protein β and γ subunits (Gβ and Gγ) form a complex that is involved in various signaling pathways. We reported that the C-terminal 10 amino acids of Gβ are required for association with Gγ (Yamauchi, J., Kaziro, Y., and Itoh, H. (1995) Biochem. Biophys. Res. Commun., 214, 694-700). To evaluate further the significance of the C-terminal region of Gβ in the formation of a Gβγ complex and its signal transduction, we constructed several C-terminal mutants and expressed them in human embryonal kidney 293 cells. The mutant lacking the C-terminal 2 amino acids (ΔC2) failed to associate with Gγ, whereas deletion of the C-terminal amino acid (ΔC1), replacement of Trp at −2 position by Ala (W339A), and addition of six histidines ((His)6) at the C terminus did not affect the association with Gγ. We also studied the effect of these mutations on the activation of mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) and c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK). Co-expression of the ΔC2 or (His)6 mutant with Gγ did not activate MAPK/ERK at all, whereas the ΔC1 or W339A mutant showed the MAPK/ERK activation. The JNK/SAPK activity was stimulated by the W339A, ΔC2, or (His)6 mutant, but not by the ΔC1 mutant. These results suggest that the C-terminal region of Gβ participates differentially in the signaling for MAPK/ERK and JNK/SAPK activations in mammalian cells.

Heterotrimeric guanine nucleotide-binding regulatory proteins (G proteins) 1 mediate signals from a variety of cell surface receptors to effector molecules (1)(2)(3)(4)(5). G proteins are composed of ␣, ␤, and ␥ subunits. Binding of ligands to G protein-coupled receptors stimulates the dissociation of G␣ and G␤␥, which regulate, independently or cooperatively, a variety of effector molecules.
Normally, G␤ and G␥ associate tightly and function as a complex. G␤ contains seven WD repeating units, each of which consists of approximately 40 amino acids and ends mainly with Trp-Asp (WD) (6). Recently, Wall et al. (7) and Lambright et al. (8) reported the x-ray crystallographic structure of heterotrimer of G␣ i1 ␤ 1 ␥ 2 and G␣ i/t ␤ 1 ␥ 1 , respectively, and the structure of G␤ 1 ␥ 1 was analyzed by Sondek et al. (9). These reports have revealed that the structure of G␤␥ in the trimeric complex is not very much different from that in the dimeric complex. The WD repeat provides a rigid scaffold of ␤-propeller structure, which is composed of seven ␤-propeller blades containing four antiparallel ␤-sheets. The N terminus of G␥ forms an ␣-helical coiled coil structure with the N terminus of G␤, and the remainder of G␥ stretches along the ␤-propeller blades, forming multiple interaction sites with G␤. On the other hand, the N-terminal ␣-helix of G␣ binds with a ␤-propeller blade of G␤, and a region designated switch II of G␣ fits into the top of the ␤-propeller.
In the present study, we demonstrated that a few amino acids at the C terminus of G␤ are involved in the complex formation with G␥. Furthermore, we found that mutations in C-terminal amino acids influence differentially the stimulation of MAPK/ERK and JNK/SAPK activities.

MATERIALS AND METHODS
Antibodies-Rabbit polyclonal anti-G␣ o antibody was produced against amino acids 94 -108 of G␣ o and purified by a peptide affinity column. Rabbit polyclonal antibody (06-238) that recognizes a peptide spanning amino acids 127-139 identical in G␤ 1 and G␤ 2 was purchased from Upstate Biotechnology, Inc. Mouse monoclonal antibodies (M2 and 12CA5) against FLAG epitope (8 amino acids, EYKEEEEK) and HA epitope (9 amino acids, YPYDVPDYA) were from Eastman Kodak Co. and Boehringer Mannheim, respectively. Rabbit anti-mouse Ig antibody (55480) was from Cappel.
Cell Culture and Transfection-Human embryonal kidney (HEK) 293 cells were maintained in Dulbecco's modified Eagle's medium (Sigma) containing 1 g/ml kanamycin (Sigma) with 10% heat-inactivated fetal bovine serum (Life Technologies, Inc.). The cells were cultured at 37°C in humidified atmosphere containing 10% CO 2 . Plasmid DNAs were transfected into HEK 293 cells by the calcium phosphate precipitation technique. The final amount of transfected DNA per 60-mm dish was adjusted to 30 g by adding empty vector pCMV5. The medium was replaced 24 h after transfection, and the cells were harvested at 48 h posttransfection.
Cell Lysis and Immunoprecipitation-To analyze the association of G␤ mutants with FLAG-G␥, cells were transfected with 10 g of G␤ and 10 g of G␥ DNAs. The transfected cells were washed twice with phosphate-buffered saline and suspended in lysis buffer A (20 mM HEPES-NaOH (pH 7.5), 5 mM MgCl 2 , 150 mM NaCl, 1 mM phenylmethanesulfonyl fluoride, 1 g/ml leupeptin, 1 mM EDTA, and 1% Lubrol-PX) in a total volume of 600 l and incubated for 10 min on ice. To analyze the interaction of G␤␥ complex with G␣ o , 10 g of G␤, 10 g of G␥, and 10 g of G␣ o DNAs were transfected. The transfected cells were lysed in lysis buffer A containing 100 M GDP. The cell lysates were centrifuged at 14,000 rpm for 10 min at 4°C in a microcentrifuge. The supernatants were incubated at 4°C for 1 h with mouse anti-FLAG antibody (1 g) and mixed gently at 4°C for 1 h with protein A-Sepharose CL-4B (Pharmacia Biotech Inc.), which was preabsorbed with rabbit anti-mouse Ig antibody (1 g). The immune complexes were precipitated by centrifugation and washed four times with lysis buffer A.
SDS-Polyacrylamide Gel Electrophoresis and Immunoblotting-Samples were boiled in Laemmli sample buffer (50 mM Tris-HCl (pH 6.8), 2% SDS, 30 mM dithiothreitol, and 10% glycerol). The boiled samples were separated by SDS-polyacrylamide gel electrophoresis, and the proteins were transferred to BA81 nitrocellulose membranes (Schleicher & Schuell). After blocking the membranes, the separated proteins were immunoblotted with rabbit anti-common G␤ or G␣ o antibody. The bound antibodies were visualized by the enhanced chemiluminescence detection system using anti-Ig antibody conjugated with horseradish peroxidase as a secondary antibody (Amersham Life Science Inc.).
Guanine Nucleotide Exchange of Immunoprecipitated G␣␤␥-The cells were transfected with cDNAs of G␣ o , G␤ mutants, and FLAG-G␥ and lysed in buffer A. The cell lysates were used for immunoprecipitation with mouse anti-FLAG antibody as described above, and the complexes were incubated in buffer A containing 500 M GDP or GTP␥S at 30°C for 2 h. The reaction was stopped by chilling on ice. The complexes were washed with buffer A, boiled in Laemmli sample buffer, and subjected to immunoblot analysis using rabbit anti-G␣ o or anti-common G␤ antibody.
Kinase Assays-The MEK kinase activity of Raf was measured as described previously (25). The cells were transfected with 10 g of pLNCX-RafFH6, 10 g of G␤ DNA, and 10 g of G␥ DNA. The trans-fected cells were starved in the serum-free medium containing 1 mg/ml bovine serum albumin for 24 h. The cells were lysed in 600 l of lysis buffer B (20 mM HEPES-NaOH (pH 7.5), 3 mM MgCl 2 , 100 mM NaCl, 2 g/ml aprotinin, 10 g/ml leupeptin, 1 mM EGTA, 1 mM Na 3 VO 4 , 10 mM NaF, 20 mM ␤-glycerophosphate, and 0.5% Nonidet P-40). After centrifugation at 14,000 rpm for 10 min in a microcentrifuge, the epitopetagged c-Raf-1 was immunoprecipitated from aliquots of the supernatants with mouse anti-FLAG antibody (1 g) and protein A-Sepharose CL-4B preabsorbed with rabbit anti-mouse Ig antibody (1 g). The immunoprecipitates were washed twice with lysis buffer B and twice with reaction buffer A (20 mM HEPES-NaOH (pH 7.5), 5 mM MgCl 2 , 0.5 mM MnCl 2 , 0.2 g/ml aprotinin, 0.1 g/ml leupeptin, and 0.1 mM EGTA). The precipitates were incubated at 30°C for 20 min in reaction buffer A with 3.3 g of recombinant MEK, 6.6 g of recombinant kinase-negative MAPK, 200 M ATP, and 5 Ci of [␥-32 P]ATP. Recombinant (His) 6 MEK and kinase-negative GST-fused MAPK were produced in E. coli and purified as described before (34,35). The reaction was stopped by adding 4 ϫ Laemmli sample buffer and heating at 95°C for 5 min. The boiled samples were separated by SDS-polyacrylamide gel electrophoresis, and the incorporation of radioactive phosphate into the MAPK was measured by an imaging analyzer (Fuji BAS 2000).

RESULTS
In a previous study, we found that the C-terminal region of G␤ is involved in the complex formation with G␥ (33). To identify which amino acid residue(s) within the last 10 amino acids in the C-terminal region is required for the association with G␥, we constructed several C-terminal deletion mutants that lack the last six (⌬C6), two (⌬C2), and one (⌬C1) amino acid(s) (Fig. 1). Since G␥ tagged with FLAG epitope at the N terminus can be co-immunoprecipitated with G␤, FLAG-G␥ was utilized to analyze the complex formation with the Cterminal mutants of G␤ (33). The G␤ mutants and FLAG-G␥ were expressed at a similar level in HEK 293 cells (Fig. 2B). Fig. 2A shows that the ⌬C6 and ⌬C2 mutants failed to associate with FLAG-G␥, whereas the ⌬C1 mutant could form a G␤␥ complex. Next, we constructed a mutant (W339A) in which the second amino acid (Trp) from the C terminus was replaced by Ala (Fig. 1). The W339A mutant could form a complex with FLAG-G␥ ( Fig. 2A), suggesting that the presence of an amino acid residue at the Ϫ2 position of G␤ appeared to be important for the complex formation with G␥. The addition of six histidines ((His) 6 ) to the C terminus of G␤ was not inhibitory to the interaction of G␤ with G␥ ( Fig. 2A).
The results of x-ray crystal structure analysis suggest that the C terminus of G␤ may be located near the N terminus of G␣ (7,8). We analyzed the effect of the C-terminal mutations of G␤ on the interaction with G␣. Addition of the FLAG sequence to the N terminus of G␥ has no effect on either the association with G␤ or the formation of a G␣␤␥ complex (33). In the presence of FLAG-G␥, the ⌬C1, W339A, and (His) 6 mutants retained their full ability to interact with G␣ o (Fig. 3). Furthermore, we examined the effect of the C-terminal mutations on the GTP-dependent dissociation of G␣ from G␤␥. The cells were transfected with cDNAs of G␣ o , C-terminal mutants of G␤, and FLAG-G␥. The cells were lysed, and the ternary complexes were immunoprecipitated with anti-FLAG antibody. The immune complexes were incubated with a buffer containing GDP or GTP␥S (Fig. 4). The ternary complexes formed with G␤ mutants of ⌬C1, W339A, and (His) 6 showed the GTP␥Sdependent dissociation similar to the one formed with wild type G␤.
G␤␥ has been shown to stimulate the MAPK/ERK signaling pathway (22)(23)(24)(25). To explore the effect of the C-terminal mutations of G␤ on MAPK/ERK activation, cDNAs of G␤ mutants, wild type G␥, and HA-ERK2 were co-transfected into HEK 293 cells. HA-ERK2 was immunoprecipitated with anti-HA antibody, and the kinase activity was assessed using myelin basic protein as a substrate. It has been shown that MAPK/ERK activation by G i -coupled receptors is mediated by G␤␥, whereas the activation by G q/11 coupled receptors is mediated by G q/11 (22)(23)(24)(25). Carbachol stimulated the HA-ERK2 activity 2-and 5-fold in HEK 293 cells transfected with m2-muscarinic and m1-muscarinic acetylcholine receptors, respectively (data not shown). Crespo et al. (22) and Faure et al. (23) demonstrated that the ERK2 activity is stimulated 4-and 2-fold, respectively, by overexpression of G␤␥ in COS cells. In HEK 293 cells, overexpression of G␤␥ induces a weak phosphorylation of endogenous ERK2 (25). As shown in Fig. 5A, co-expression of wild type G␤ and G␥ activated HA-ERK2 about 1.5-fold, whereas the ⌬C1 and W339A mutants had a relatively weak effect on ERK2 activation, and the ⌬C2, ⌬C6, and (His) 6 mutants could not activate ERK2 at all. Furthermore, we examined the effect of the C-terminal mutations of G␤ on MEK kinase activity of c-Raf-1. Cells were co-transfected with cDNAs of various G␤ mutants, wild type G␥, and RafFH6, which is c-Raf-1 tagged with FLAG epitope and six histidines. The RafFH6 was immunoprecipitated with anti-FLAG antibody, and its MEK kinase activity was assayed using recombinant MEK and recombinant kinase-negative MAPK. The MEK kinase activity of RafFH6 was stimulated more than 1.5-fold by wild type G␤ and G␥. On the other hand, the activation of RafFH6 by the ⌬C1 and W339A mutants was less potent than that by the wild type G␤, and the ⌬C2, ⌬C6, and (His) 6 mutants failed to activate RafFH6 (Fig. 5B). Effects of the C-terminal mutations on the stimulation of c-Raf-1 activity were comparable to those on the stimulation of ERK2 activity. Transfection of G␤ or G␥ alone fails to activate ERK2 (see Fig. 8A, and Refs. 22, 23, and 25) and c-Raf-1 (see Fig. 8B). It is noteworthy that the (His) 6 mutant could stimulate neither ERK2 nor c-Raf-1.
Since the N-terminal region of c-Raf-1 associates physically with the C-terminal region of G␤ (37), we tested whether the C-terminal mutants of G␤ are able to interact with c-Raf-1. As shown in Fig. 6, the (His) 6 mutant could not bind with c-Raf-1, but other mutants could bind. It is likely that six histidine residues added at the C terminus of G␤ may sterically inhibit the association of G␤ with c-Raf-1. It is suggested that the association of G␤ with c-Raf-1 by itself is not sufficient for activation of c-Raf-1, although the association may be required for its activation.
It has recently been reported that the signaling from G protein-coupled receptors to JNK/SAPK involves G␤␥ and that overexpression of G␤␥ enhances JNK/SAPK activity (26). We examined the effect of the G␤ C-terminal mutants on JNK/ SAPK stimulation. The cells were co-transfected with cDNAs of each G␤ mutant and wild type G␥ together with HA-JNK1. After lysis, HA-JNK1 was immunoprecipitated, and its kinase activity was assayed using GST-c-Jun as a specific substrate. Wild type G␤ and G␥ increased the activity of JNK1 approximately 5-fold (Fig. 7A). In contrast to the results of MAPK/ERK and c-Raf-1 activations (Fig. 5), the ⌬C1 mutant had little ability to stimulate JNK1, whereas the ⌬C2, ⌬C6, and (His) 6 mutants showed moderate stimulatory effect. The stimulatory effect of the W339A mutant on JNK1 was almost indistinguish-able from that of the wild type G␤ (Fig. 7A). Although the ⌬C2 and ⌬C6 mutants had little ability to associate with G␥, they activated JNK1 significantly. In order to examine whether G␤ may activate JNK1 in the absence of interaction with G␥ in the cells, we utilized a N-terminal deletion mutant of G␤. Since the N-terminal region of G␤ is essential to form an ␣-helical coiled coil structure with the N terminus of G␥, the deletion of this region of G␤ completely prevents the dimer formation with G␥ (9,33,38). As shown in Fig. 7B, the ⌬N38 mutant, which lacks N-terminal 38 amino acids, stimulated the activity of JNK1 approximately 3-fold. Furthermore, we transfected with G␤ or G␥ alone and measured the JNK1 activity. Fig. 8C shows that the transfection of G␤ alone caused the activation of JNK1. These results suggested that the overexpression of G␤ alone can stimulate the JNK1 activity in the cells. DISCUSSION The mutant (⌬C1) lacking the C-terminal amino acid of G␤ retained the full ability to associate with G␥ and to form a ternary complex with G␣ (Figs. 2-4). Although the C-terminal Asn-340 of G␤ 1 participates in the specific interaction with Asn-62 of G␥ 1 (corresponding to Asn-59 of G␥ 2 ) (9), the deletion of Asn-340 of G␤ 1 did not affect the association with G␥ 2 . The interaction does not seem to be necessary for the complex formation. Removal of the two amino acid residues from the G␤ C terminus dramatically abolished the binding of G␤ with G␥ (Fig. 2). Since G␤ associates with G␥ at multiple sites (9), it was unexpected that the truncation of the last two amino acid residues resulted in the loss of the ability to form a G␤␥ complex. We first thought that the Trp-339 may be important for the complex formation, but the replacement of Trp-339 by Ala (W339A) did not show any effect on the association of G␤ with G␥ (Fig. 2). It appears that the presence of an amino acid residue at the Ϫ2 position of G␤ is important for the G␤␥ complex formation.
Genetic studies of the pheromone response pathway in Saccharomyces cerevisiae suggested that two regions of the Ste4 protein (S. cerevisiae G␤) may be involved in the effector activation (39). The first region is localized in the ␣-helical structure of the N terminus and far from the binding sites with G␣ (7,8,39). If this region is involved in the binding with an effector, it is unclear how the activation of effector by G␤␥ is inhibited by G␣. The second region is found in the third WD repeat (39). We made a substitution mutant using Gly at Val-135 that was well conserved between mammalian G␤ and yeast Ste4. However, the V135G mutant of G␤ retained the ability to stimulate the MAPK/ERK and JNK/SAPK activities to an extent similar to that of the wild type G␤ in HEK 293 cells (data not shown). It is possible that the V135G mutation may affect other signaling pathways in mammalian cells or that other mutations in the second region of G␤ may affect ERK/MAPK and JNK/SAPK activities.
We found differential effects of the C-terminal mutations of G␤ on the MAPK/ERK and JNK/SAPK pathways in mammalian cells. The ⌬C1 mutant of G␤, together with G␥, could stimulate the activity of c-Raf-1 and MAPK/ERK but showed only a slight activation of JNK/SAPK. On the other hand, the (His) 6 mutant of G␤ failed to stimulate c-Raf-1 and MAPK/ERK activities in the presence of wild type G␥, and the mutant retained the ability to activate JNK/SAPK (Figs. 5 and 7). The C-terminal amino acid residue of G␤ is localized on the outside of a G␤␥ complex (9). It is possible that the C terminus of G␤ is involved in binding with effector molecule(s). We speculate that the differential effects of these mutations on the activation of MAPK/ERK and JNK/SAPK might be due to the difference of their direct effector(s). The putative effector molecule(s) of G␤␥ in the MAPK/ERK and JNK/SAPK cascades are as yet unidentified, although phosphatidylinositol 3-kinase has been reported to be involved in the pathway from G␤␥ to MAPK/ERK (40). Phosphatidylinositol 3-kinase ␥ (13) may be a candidate for its effector. It has been reported that calcium ion is important in G i -coupled receptor-mediated stimulation of JNK/SAPK (41). The direct regulation of phospholipase C␤ isozymes by G␤␥ (11, 12) may be involved in the signaling pathway.
Recently, Coria et al. (42) reported that the C-terminal region of yeast Ste4 is essential in triggering the yeast phero- mone response cascade. Ste4 has four extra amino acid residues at the C terminus compared with mammalian G␤. The extra amino acid residues may be important for the function of Ste4.
It has been shown that overexpression of Ste4 alone causes an increased response to the pheromone in yeast (43)(44)(45). We found that the ⌬C2 and ⌬C6 mutants, which were unable to associate with FLAG-G␥, could stimulate JNK/SAPK but not MAPK/ERK activity (Figs. 5 and 7). Therefore, we examined whether G␤ alone can stimulate the activity of JNK/SAPK in HEK 293 cells. It was observed that the transfection of G␤ alone could stimulate JNK/SAPK to the same extent as cotransfection of G␤ and G␥ (Fig. 8C). As shown in Fig. 7B, G␥-incompetent G␤ mutant (⌬N38), which lacks N-terminal ␣-helical structure for coiled coil interaction with G␥, has the ability to activate JNK/SAPK in the cells. Taken together, it is suggested that G␤ plays an essential role in the JNK/SAPK pathway in HEK 293 cells. Coso et al. (26) have reported that overexpression of G␤ alone does not stimulate the JNK/SAPK activity in COS cells. The discrepancy may be due to the difference of cell type.
In the course of this study, two groups have reported the regions of G␤ involved in the interaction with effector. Yan and Gautam (46), using a yeast two-hybrid system, showed that the N-terminal 100-amino acid fragment of G␤ associated with adenylyl cyclase type II and G protein-coupled inward rectifier potassium channel 1. Zhang et al. (47) demonstrated that the G␤ mutation in the C-terminal region prevented the stimulation of phospholipase C␤2 in COS cells. We obtained C-terminal mutants of G␤ that retained the ability to interact with G␥ and G␣ yet exhibited dramatic decreases in the MAPK/ERK or JNK/SAPK activation in mammalian cells. Further biological studies using these mutants should throw more light on the role of G␤ in the distinct cellular signaling pathways.