Interaction site of GTP binding Gh (transglutaminase II) with phospholipase C.

The GTP binding Gαh (transglutaminase II) mediates the α1B-adrenoreceptor signal to a 69-kDa phospholipase C (PLC). Thus, Gαh possesses both GTPase and transglutaminase activities with a signal transfer role. The recognition sites of this unique GTP binding protein for either the receptor or the effector are completely unknown. A site on human heart Gαh (hhGαh) has been identified that interacts with and stimulates PLC. Expressed mutants of hhGαh with deleted C-terminal regions lost the response to(-)-epinephrine and GTP and failed to coimmunoprecipitate PLC by the specific Gh7α antibody. The interaction regions were further defined by studies with synthetic peptides of hhGαh and a chimera in which residues Val665-Lys672 of hhGαh were substituted with Ile707-Ser714 residues of human coagulation factor XIIIa. Thus, eight amino acid residues near the C terminus of hhGαh are critical for recognition and stimulation of PLC.

The G␣ h protein, transglutaminase II (TGase II), 1 is unique in that the enzyme exhibits two distinct enzyme activities, namely guanosine triphosphatase (GTPase) and TGase, with a signal transfer role (Ref. 1; see also Refs. 2 and 3). The GTPase function of G␣ h differs from other TGases, coagulation factor XIIIa (FXIIIa), keratinocyte, and epidermal transglutaminases (4). G␣ h , which is species specific in molecular mass, directly interacts with ␣ 1 -adrenoreceptor (5, 6) and a 69-kDa PLC in the activation process (7,8). Physiological TGase role of G␣ h remains unclear (4). However, it has been suggested that TGase II is involved in control of cell growth and differentiation (9 -11) and activation of cytosolic phospholipase A 2 (12).
The amino acid sequences of all TGases including G␣ h show high homology in the middle portions of the polypeptides, which include the TGase active site and a calcium binding region (13). However, the N-and C-terminal regions of G␣ h do not share sequence homology among TGases. This divergence is particularly greater at the C-terminal domain of G␣ h , giving rise to the hypothesis that this region may play a significant role in hormone signaling. In this study, evidence for a direct interaction between the region of G␣ h and PLC is demon-strated. This interaction activates PLC.

EXPERIMENTAL PROCEDURES
Isolation and Mutagenesis of hhG␣ h cDNA-Full-length cDNA of human heart G␣ h (hhG␣ h ) was isolated from a human heart cDNA library by polymerase chain reaction (PCR) using two oligonucleotides, CCCGACCATGGCCGACGAGGAGCTGGTCT (5Ј-primer) and TGGGC-CAGGGGCACATTCCATTTC (3Ј-primer), synthesized from the known nucleotide sequences of human endothelial TGase II (14). After the PCR product (ϳ2.1 kilobases) was cloned into the pCR TM II without purification, the insert was digested with KpnI and NotI and cloned into the modified eukaryotic expression vector pMT2Ј to yield pMT2ЈhhG␣ h (1). Four 10-amino acid-truncated mutants of hhG␣ h were generated by introducing a TAA stop codon using the following oligonuceotides: 1) CTTCACAGCCTTCAGCTTGTCGCT, 2) GTTCACCACCAGCTTGTG-GAGGCC, 3) GAGCGGCACGAGGTCCATTCTCAC, and 4) TTCCTC-CCCTGCCTCCACGGGGTC. The hhG␣ h /human FXIIIa chimera was constructed by ligation of PCR products generated from two different sets of primers containing nucleotide sequences of human FXIIIa (5Јprimer, 5Ј-GAATTCGAATTCCCCGACCATGGCCGAGGAGCTGGTC-TTA-3Ј and 3Ј-primer, 5Ј-GCTGGCTATCAGCTTGTGGAGGCCCAT-GTGGAGCGGCACGAGGTC-3Ј; 5Ј-primer, 5Ј-ATGAGCAGTGACT-CCAAGGCTGTGAAGGGCTTCCGGAATGTCATC-3Ј and 3Ј-primer, 5Ј-GCGGCCGCGCGGCCGCTGGGCCAGGGGCACATTCCATTTC-3Ј) (single underline denotes nucleotide sequence from human FXIIIa; double underline denotes 5Ј-primer encoding two EcoRI restriction sites at 5Ј-primer and two NotI restriction sites at 3Ј-primer). Each PCR product was treated with Klenow fragment of Escherichia coli DNA polymerase I and digested with EcoRI or NotI. The blunt ends of two fragments were ligated and cloned into the eukaryotic expression vector pMT2Ј. Orientation of the constructs was confirmed by restriction enzyme mapping and DNA sequencing.
Expression of hhG␣ h Proteins and Preparation of Membranes-Transfection and membrane preparation were performed using the method of Nakaoka et al. (1). COS-1 cells (5 ϫ 10 6 per 100-mm dish) were cotransfected with plasmids containing ␣ 1B -adrenoreceptor cDNA (4 -5 g of cDNA) and hhG␣ h or its mutants (8 -10 g of cDNA) using the DEAE-dextran method. The cells were grown for 48 -72 h after transfection. The membranes prepared from the transfected COS-1 cells were suspended in a buffer (20 mM Hepes, pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.5 mM dithiothreitol, 10% glycerol, and protease inhibitors) and were stored at Ϫ80°C until use.
Analysis of Expressed hhG␣ h Protein and Its Mutants-All expressed proteins were estimated by immunoblotting with guinea pig liver TGase II and G h7␣ antibodies (1,6). Membrane proteins (100 g) were solubilized with 1% sodium cholate and subjected to SDS-polyacrylamide gel electrophoresis (10% gel). The proteins were transferred to Immobilon-P (Millipore) and probed with the antibodies by the methods of Baek et al. (6). Antibody cross-reactivity of proteins was visualized with chemiluminescence reagent (DuPont NEN) using Kodak XAR-5 film. GTP binding ability of the partially purified expressed proteins (50 ng/tube) was measured by GTP-mediated inhibition of the TGase activity in the presence and absence of GTP. Partial purification of the expressed proteins was achieved by Q-Sepharose chromatography. The lysates (5 mg of protein) of transfected COS-1 cells were solubilized with 0.4% sucrose monolaurate (SM) at 4°C for 1 h. The extracts were applied to a Q-Sepharose column (0.4 ml) equilibrated with a buffer containing 20 mM Hepes, pH 7.4, 1 mM EGTA, 0.5 mM dithiothreitol, 70 mM NaCl, and 0.05% SM. The columns were washed with 3-5 ml of 170 mM NaCl in the same buffer. The hhG␣ h and its mutant proteins were eluted with 500 mM NaCl in the same buffer. The yields were analyzed * This work was supported by National Institutes of Health Grant RO1GM45985. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Dept. of Molecular Cardiology (FFB-37), Research Institute, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195. Tel.: 216-444-8860; Fax: 216-444-8372. 1 The abbreviations used are: TGase II, transglutaminase II; FXIIIa, coagulation factor XIIIa; GTP␥S, guanosine 5Ј-O-(3-thiotriphosphate); PCR, polymerase chain reaction; PLC, phospholipase C; SM, sucrose monolaurate; hhG␣ h , human heart G␣ h . by immunoblotting using G h7␣ antibody and Coomassie Blue staining following SDS-polyacrylamide gel electrophoresis. The TGase activity of the purified proteins was determined in the presence of 0.5 mM CaCl 2 and 1 mM dithiothreitol by evaluating Coimmunoprecipitation-For coimmunoadsorption of PLC or the ␣ 1adrenoreceptor with hhG␣ h and its mutant proteins, G h7␣ antibodyprotein A-agarose was prepared according to the method of Schneider et al. (15). The membranes (150 g/sample) in HSD buffer (20 mM Hepes, pH 7.4, 150 mM NaCl, and 0.5 mM dithiothreitol) were solubilized with 0.4% SM in the presence of 50 M GTP␥S and 2 mM MgCl 2 at 4°C for 1 h. The extracts were incubated with the antibody-agarose or preimmune-agarose (30 l suspension/assay) at 4°C for 1 h. The resins were collected by centrifugation at 2000 rpm for 15 min and washed three times with HSD buffer containing 5 M GTP␥S, 1 mM MgCl 2 , and 0.05% SM. The antibody resin-bound PLC activity was measured using 50 M CaCl 2 for 20 min at 30°C in a 100-l final volume. The PLC activity absorbed to preimmune-protein A-agarose was negligible and taken as a nonspecific binding. Inhibition of the coimmunoprecipitation of PLC by peptides was determined using the membrane extracts, which were preincubated with peptides. The samples were incubated with the antibody-agarose (30 l/tube) in the presence of 50 M GTP␥S and 2 mM MgCl 2 at 4°C for 2 h. The resin-bound PLC was determined under the same conditions as described above. The phosphatidylinositol 4,5bisphospate was 100 M (1200 cpm/nmol) in the assay.
For the coimmunoprecipitation of the ␣ 1 -adrenoreceptor with hhG␣ h and its mutants, the membranes that coexpressed hhG␣ h or its mutants with ␣ 1B -adrenoreceptor or ␣ 1B -receptor alone were incubated with 5 ϫ 10 Ϫ6 M (Ϫ)-epinephrine at 4°C for 3 h. The membranes were then solubilized with 0.2% SM in HSD buffer in the presence of 5 ϫ 10 Ϫ6 M (Ϫ)-epinephrine at 4°C for 1 h. The recovery of proteins in the extracts usually reached ϳ40% for the receptor and 40 -50% for hhG␣ h and its mutants (see also Ref. 5). Since the hhG␣ h and its mutants were overexpressed 3-5-fold as compared to the ␣ 1 -adrenoreceptor, the membrane extracts (100 fmol of the ␣ 1 -receptor) were incubated with 100 l of G h7␣ antibody-agarose or nonimmune sera-agarose in 300 l (final volume) with gentle rotation at 4°C for 2 h. After centrifugation at 2000 rpm, the ␣ 1B -receptor density (50 l of supernatant/tube) was measured using 3 nM [ 3 H]prazosin (final) in the presence or absence of 10 Ϫ4 M phentolamine in 200 l (final volume) after removing excess (Ϫ)-epinephrine through a dried 3-ml Sephadex G-25 column. The preimmune sera-agarose-treated samples were used as controls to calculate the amounts of the receptor immunoprecipitated for each sample.
Other Assays-PLC activity was evaluated by the method of Im et al. (8). The ␣ 1 -adrenoreceptor density was determined using the radiolabeled ␣ 1 -specific antagonist [ 3 H]prazosin (5). The ␣ 1B -adrenoreceptoractivated PLC stimulation was determined after normalizing receptor number. The protein concentration was determined by the method of Bradford (16).

RESULTS AND DISCUSSION
The isolated full-length hhG␣ h cDNA was an exact match in the nucleotide and deduced amino acid sequences with the human endothelial TGase II (14). To identify interaction sites of hhG␣ h with the ␣ 1 -adrenoreceptor and PLC, systematic 10amino acid-deleted mutants of hhG␣ h cDNA(s) were generated from the C-terminal end (Fig. 1). The full-length hhG␣ h cDNA and truncated hhG␣ h cDNA(s) were cotransfected into COS-1 cells with ␣ 1B -adrenoreceptor cDNA. The expressed proteins were recognized by the G h7␣ antibody as well as guinea pig TGase II antibody and were of the expected sizes with ϳ80 kDa for full-length hhG␣ h and a decrease in size as the length of nucleotide deletion increased (Fig. 2A). The ␣ 1B -receptor was also expressed, resulting in 2-3 pmol/mg protein of [ 3 H]prazosin binding. The expressed hhG␣ h proteins exhibited both GTP binding and TGase activities ( Fig. 2B and inset). The Ca 2ϩstimulated TGase activities of expressed hhG␣ h and its truncated mutants were completely inhibited with Ͼ100 M GTP, and the inhibitory potency (IC 50 ) of GTP was in the range of 20 -50 M for all hhG␣ h proteins, also confirming that GTP is a negative regulator for the TGase of G␣ h (1, 3, 4).
All mutants, as well as hhG␣ h , elevated the basal PLC activity as compared to that of the ␣ 1B -receptor alone (Fig. 3A), indicating induction of precoupled protein complexes resulting from overexpression of the proteins (1, 18). The (Ϫ)-epinephrine-mediated activation of PLC was increased approximately 2-fold with hhG␣ h and the ⌬K676 mutant, whereas the ⌬L656 and ⌬E646 mutants lost the agonist-mediated PLC stimulation, exhibiting a level similar to that of the ␣ 1B -receptor alone. The ⌬N666 mutant stimulated PLC upon activation of the ␣ 1 -receptor, but to a lesser extent than wild type. These data suggested that a region comprising 20 amino acids between His 657 and Lys 677 is critical for coupling to the ␣ 1B -receptor or PLC.
The possibility that the receptor and/or PLC recognition sites were deleted was examined by coimmunoprecipitation. The mutants, ⌬K676, ⌬N666, and ⌬L656, coexpressed with the ␣ 1B -adrenoreceptor, coimmunoprecipitated Ͼ90% ␣ 1 -receptor as wild type did, indicating that the receptor interaction site on these mutants was intact (Fig. 3B). However, the mutant, ⌬E646, consistently coimmunoprecipitated less receptor (ϳ80%) than other mutants but more than the receptor alone (ϳ35%). Although less coimmunoprecipitation of the receptor with this mutant suggested that this region on hhG␣ h might contain the receptor interaction site, this point should be further investigated. Coimmunoprecipitation of the receptor with membrane extract from the expressed ␣ 1B -receptor alone was probably due to complex formation between the internal G␣ h and the receptor.
The loss of PLC interaction site was then assessed by coimmunoprecipitation (Fig. 3C). The results revealed that the ⌬K676 mutant coimmunoprecipitated PLC as effectively as the wild type, whereas the ⌬L656 and ⌬E646 mutants failed to coimmunoprecipitate PLC, showing a similar level to that of the ␣ 1B -receptor alone. The ⌬L666 mutant again showed lower coimmunoprecipitation of PLC than hhG␣ h but higher than the⌬L656 and ⌬E646 mutants. The loss of the PLC interaction site was further confirmed by determining PLC stimulation in response to GTP (Fig. 4, A and B). As expected, the basal levels of PLC in membranes expressing hhG␣ h and mutants were increased 3ϳ6-fold compared to the ␣ 1B -receptor alone. Within these increases, the PLC basal activity gradually decreased as the deletion size increased (Fig. 4A). In the presence of GTP, the expressed hhG␣ h and the mutant ⌬K676 increased PLC stimulation 2-fold (Fig. 4B). Increases in deletion size also resulted in a gradual decrease of GTP-mediated PLC stimulation. Mutants ⌬L656 and ⌬E646 lost ability to stimulate PLC in response to GTP. These results were consistent with the finding from the coimmunoprecipitation studies and strongly suggested that a region between His 657 and Lys 677 on hhG␣ h contained a PLC interaction site.
To further define this putative PLC interaction site, four overlapping peptides corresponding to the deleted regions of hhG␣ h were synthesized and tested for their ability to inhibit coimmunoprecipitation of PLC (Fig. 5A). Peptide 4 (Leu 661 -Lys 672 ), among the four peptides, was able to inhibit coimmunoprecipitation of PLC (Fig. 5B). Coimmunoprecipitation of PLC was inhibited in a concentration-dependent manner, and at 100 -200 M of the peptide, the inhibition reached ϳ80%, suggesting that other interaction site(s) probably exist (Fig.  5C). The competition potency (IC 50 ) of peptide 4 for the interaction between hhG␣ h and PLC was ϳ20 M.
The findings that a region of 12 amino acids between Leu 661 and Lys 672 in hhG␣ h contains a PLC interaction site were FIG. 3. Coupling ability of expressed hhG␣ h and its mutants. A, epinephrine-stimulated inositol 1,4,5-triphosphate (IP 3 ) accumulation in membranes from COS-1 cells coexpressed with ␣ 1B -receptor and hhG␣ h or its mutants. The ␣ 1 -receptor-mediated PLC stimulation was determined after normalizing receptor number (100 fmol/tube) in a 100-l final volume. Receptor number was normalized, since the receptor number is the determinant in signal manipulation, not G-protein number (17), and the expression level of hhG␣ h and its mutants was also 3-5-fold higher than the receptor level. The results are the mean Ϯ S.E. of three independent experiments performed in duplicate. ␣ 1B -AR, ␣ 1B -adrenoreceptor; Ep, (Ϫ)-epinephrine; Ph, phentolamine. B, remaining ␣ 1B -adrenoreceptor in the supernatants after coimmunoprecipitation with hhG␣ h and its mutants using G h7␣ antibody-protein A-agarose. The ␣ 1 -adrenoreceptor absorbed to preimmune protein A-agarose was less than 5% as compared to the resin-untreated samples. The preimmune-resin-treated samples were taken as 100% for each sample. C, immunoadsorption of a complex of PLC with hhG␣ h and its mutants by G h7␣ antibody-protein A-agarose. The PLC activity absorbed to preimmune protein A-agarose was negligible and taken as nonspecific binding. The data shown are the mean Ϯ S.E. of three independent experiments in triplicate. refined by a chimera, hhG␣ h /FXIIIa, in which eight amino acid residues Val 665 -Lys 672 of hhG␣ h were substituted with the corresponding region (Ile 707 -Ser 714 ) of human factor XIIIa (see Fig. 1) (13). This region of FXIIIa was chosen because FXIIIa does not interact with or stimulate PLC or bind GTP (4), and this region of FXIIIa is distinct among TGases (13). The chimera was expressed ϳ7-fold higher than endogenous G␣ h and to the similar level of hhG␣ h (Fig. 6A). The chimera protein exhibited GTP binding and TGase activity at the same levels as the wild type (Fig. 6B). In addition, using the partially purified chimera when the GTP-mediated inhibition of TGase activity was titrated, the inhibition was similar to ⌬K676 and wild type, indicating that substitution of this region did not change GTP binding affinity (data not shown). The chimera also failed to stimulate PLC in response to GTP (Fig. 6C) and upon activation of the ␣ 1B -receptor (data not shown). The G h7␣ antibody did not coimmunoprecipitate PLC but effectively coimmunoprecipitated the receptor (data not shown). These findings clearly demonstrate that the C-terminal region of hhG␣ h from Val 665 to Lys 672 is a critical site for interaction and stimulation of PLC.
The substituted region of the chimera has a significant change in properties of the amino acids (Fig. 1). Thus, four charged amino acids (Asn 667 , Glu 669 , Asp 671 , and Lys 672 ) were substituted for serine, except Asp 671 . Hydrophobic amino acids (Val 665 , Val 666 , and Phe 668 ) were also changed to smaller (V666A and F668M) or larger (V665I) amino acids. Although it has been suggested that replacement of a bulky side chain of hydrophobic amino acids can result in loss of activity due to unfavorable van der Waals interactions (19), overall hydrophobicity of the substituted amino acids, however, remains similar. Therefore, it is unlikely that a hydrophobic interaction is responsible for coupling of hhG␣ h to PLC. Three charged amino acids in this region, i.e. a hydrophilic interaction, probably play a critical role in the contact of hhG␣ h with PLC.
In heterotrimeric G-protein-mediated signaling systems, the near C-terminal domain of the ␣-subunit appears to contain an effector contact region (adenylyl cyclase with G␣ s (20 -22) and cGMP-phosphodiesterase with G␣ t (23)). Despite extensive primary structural differences between G␣ h and ␣-subunits of the heterotrimeric G-proteins, our data indicate that G␣ h seems to share this common structural feature in signaling. Our data also suggest that the carboxyl domain of G␣ h with its primary structure distinct from other transglutaminases is likely to be involved in signaling functions, including receptor and GTP binding sites.