Alpha1-adrenergic receptor signaling via Gh is subtype specific and independent of its transglutaminase activity.

Tissue transglutaminase (TGase II) is a Ca2+- and thiol-dependent enzyme that catalyzes the post-translational modification of proteins via the formation of ε(γ-glutamyl) lysine bonds. We have shown previously that the GTP-binding protein, Gh, is a TGase II that mediates intracellular signaling by the α1B-adrenergic receptor (AR) (Nakaoka, H., Perez, D. M., Baek, K. J., Das, T., Husain, A., Mison, K., Im, M.-J., and Graham, R. M. (1994) Science 264, 1593-1596). Here, we evaluated the ability of Gh as compared with Gq to mediate receptor-stimulated inositol phosphate turnover by the three α1-subtypes (α1A, α1B, and α1D). In addition, we questioned if the transglutaminase function of Gh is involved in its receptor signaling activity. A mutant form of a human TGase II cDNA in which the codon for the active site cysteine (Cys277) was replaced by serine was cloned into the mammalian expression vector pMT2′. Compared with wild-type TGase II, no transglutaminase activity was observed with transient transfection of this Cys→Ser mutant in COS-1 cells. However, like wild-type TGase, the Cys→Ser mutant mediated receptor-stimulated inositol phosphate turnover when cotransfected with an α1B-AR cDNA. Gαq supported α1-AR-mediated inositol phosphate turnover by all three receptor subtypes. By contrast, although both the wild-type and Cys→Ser construct mediated receptor signaling by the α1B AR and α1D AR, the α1A-AR was unable to interact with Gh. However, a Gh-dependent signaling phenotype could be rescued by a chimeric α1A construct in which the third intracellular loop of the α1A-AR was replaced by that of the α1B-AR. Thus, the signaling function of Gh is independent of its transglutaminase activity and is α1-AR subtype specific. This subtype specificity of the interaction between α1 ARs and Gh involves important determinants in their third intracellular loops.

Tissue transglutaminase (TGase II) is a Ca 2؉ -and thioldependent enzyme that catalyzes the post-translational modification of proteins via the formation of ⑀(␥-glutamyl) lysine bonds. We have shown previously that the GTP-binding protein, G h , is a TGase II that mediates intracellular signaling by the ␣ 1B -adrenergic receptor (AR) ( Here, we evaluated the ability of G h as compared with G q to mediate receptor-stimulated inositol phosphate turnover by the three ␣ 1 -subtypes (␣ 1A , ␣ 1B , and ␣ 1D ). In addition, we questioned if the transglutaminase function of G h is involved in its receptor signaling activity. A mutant form of a human TGase II cDNA in which the codon for the active site cysteine (Cys 277 ) was replaced by serine was cloned into the mammalian expression vector pMT2. Compared with wild-type TGase II, no transglutaminase activity was observed with transient transfection of this Cys3 Ser mutant in COS-1 cells. However, like wild-type TGase, the Cys3 Ser mutant mediated receptor-stimulated inositol phosphate turnover when cotransfected with an ␣ 1B -AR cDNA. G ␣q supported ␣ 1 -AR-mediated inositol phosphate turnover by all three receptor subtypes. By contrast, although both the wild-type and Cys3 Ser construct mediated receptor signaling by the ␣ 1B AR and ␣ 1D AR, the ␣ 1A -AR was unable to interact with G h . However, a G h -dependent signaling phenotype could be rescued by a chimeric ␣ 1A construct in which the third intracellular loop of the ␣ 1A -AR was replaced by that of the ␣ 1B -AR. Thus, the signaling function of G h is independent of its transglutaminase activity and is ␣ 1 -AR subtype specific. This subtype specificity of the interaction between ␣ 1 ARs and G h involves important determinants in their third intracellular loops.
Adrenergic receptors (ARs) 1 are members of a receptor su-perfamily that exert their physiological effects through coupling to G-proteins. Strictly conserved among this superfamily is the presence of seven ␣-helical transmembrane-spanning domains connected by hydrophilic loops alternately exposed to the intra-and extracellular environment. The intracellular loops bind and activate the receptor-coupled G-proteins (1,2). ␣ 1 ARs acting via pertussis-and cholera-toxin insensitive Gproteins result in inositol phosphate turnover by activating phosphoinositide-specific phospholipase C (PI-PLC) (3). Three ␣ 1 -AR subtypes (␣ 1A , ␣ 1B , and ␣ 1D ) have been clearly identified based on molecular cloning, and pharmacological and biochemical studies (3). All three subtypes activate PI-PLC (3,4). In cotransfection studies it has been found that G ␣q and various members of this G ␣ subunit family (G ␣11 , G ␣14 , and G ␣16 ) can support ␣ 1 AR-mediated PI-PLC␤1 activation (5). In addition, we have reported previously that activation of a 69-kDa PI-PLC, which is distinct from PI-PLC␤, is mediated by G h , a GTP-binding protein that is not a member of the heterotrimeric G-protein family (6,7). G h is identical to tissue transglutaminase type II (TGase II) (7), a Ca 2ϩ -and thiol-dependent acyl transferase that catalyzes the formation of an amide bond between the ␥-carboxamide groups of peptide-bound glutamine residues and the primary amino groups in various compounds, including the ⑀-amino group of lysines in certain proteins (8). It is expressed ubiquitously in mammalian tissues both in membrane and cytosolic fractions and is a member of a larger family of transglutaminases (9). Despite sequence homology between the members of this family, TGases are encoded by distinct genes and differ in their structure and biological functions (9).
TGase II has been suggested to be involved in a variety of biological processes including tumor growth (10), stimulussecretion coupling (11), programmed cell death (12), receptormediated endocytosis (13), and extracellular matrix organization and cell adhesion (14). However, recent studies question the protein cross-linking role of TGase II and suggest that other functions, in addition to or apart from its transglutaminase activity, may contribute to its cellular function(s). For example, the correlation between TGase II gene expression and the extent of protein cross-linking found in cells is not always apparent (15). Also, the involvement of TGase II/G h in membrane signaling appears to be independent of its transglutaminase activity, since it can be demonstrated at physiological levels of Ca 2ϩ that are well below those required for enzyme activation (7). Moreover, recent studies using permeabilized cells indicate that at the physiological levels of the ATP and GTP nucleotides intracellular transglutaminase activity is virtually zero, even in the presence of a high concentration of Ca 2ϩ (10 M) (16). In fact, only when Ca 2ϩ was 100 M and nucleotide * This study was support by a project grant and Eccles Award from the National Health and Medical Research Council, Australia. 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 1 The abbreviations used are: AR(s), adrenergic receptor(s); G-protein, guanine nucleotide-binding regulatory proteins; EC 50 , concentration of agonist causing 50% of the maximal response; PI-PLC, phospho-levels were low or absent could transglutaminase activity be observed. Along the same lines, there is recent evidence that a role of TGase II in the regulation of cell cycle progression is independent of its transglutaminase activity (15).
Based on these considerations, this study was designed to obtain more definitive evidence that the receptor signaling function of G h is independent of its transglutaminase activity. This was achieved by evaluating the receptor-signaling activity of a mutant form of TGase II that retains its ability to bind and hydrolyze GTP, but lacks transglutaminase activity. In addition, we investigated the ability of G h , as compared with the heterotrimeric G-protein G q , to interact not only with the ␣ 1B -AR but also with the other ␣ 1 -AR subtypes.
The results of these studies indicate that the receptor signaling function of G h is independent of its protein cross-linking activity. Moreover, while G q supports PI-PLC activation by all three ␣ 1 -AR subtypes, the interaction of G h with ␣ 1 -AR is subtype specific, since it is observed with the ␣ 1B and ␣ 1D subtypes but not with the ␣ 1A -AR. This subtype specificity involves determinants in the third intracellular loop, since G h -dependent signaling can be restored with a chimeric construct in which the third intracellular loop of the ␣ 1A -AR is replaced with that of the ␣ 1B -AR. Vectors and DNA Constructs-The ␣ 1 -AR and rat TGase II (G h ) constructs used were the hamster ␣ 1B -AR cDNA, the rat ␣ 1A -AR cDNA and the rat ␣ 1D -AR cDNA, and rat TGase II cDNA cloned into the modified eukaryotic expression vector, pMT2Ј, as described previously (17)(18)(19)(20). All of the following cDNAs were also subcloned into pMT2Ј between the EcoRI (5Ј) and NotI (3Ј) sites in its polylinker: G ␣q , originally provided in pCMV (21); human TGase II, kindly provided in pSG5 by Dr V. Gentile (University of Naples); and the mutant form of human TGase II containing a serine codon instead of the codon for the active site cysteine (Cys 277 ), originally provided in the prokaryotic expression vector, pET-8c (22). This mutant is designated here as mTG. Although the studies reported here utilized the wild-type rat TGase II versus the mutant human TGase II, the rat and human cDNAs show 90% amino acid homology, and both support ␣ 1 -AR-mediated PI-PLC activation (data not shown). In all cases, the nucleotide sequences in the cloning site regions were confirmed after subcloning. For transfection, plasmid DNAs were purified by CsCl-gradient centrifugation followed by Biogel A-50m (Bio-Rad) column chromatography, as described (1,23).
Cell Culture and Transfection-COS-1 cells (American Type Culture Collection) were cultured and transiently transfected with the indicated constructs using the DEAE-dextran method, as described previously (1,18). This method provides a reproducible transfection efficiency of 30 -40%, as determined by in situ staining of cells transfected with pSV-LacZ, a plasmid encoding the reporter, ␤-galactosidase, and treatment of the cells with 0.2% 5-bromo-4-chloro-3-indoyl ␤-D-galactoside. Cells were harvested 72-h post-transfection.
Western Blotting-Membranes (20 g of protein) were dissolved in SDS sample buffer, followed by boiling for 5 min and then subjected to SDS-polyacrylamide gel electrophoresis, as described previously (1,7). The resolved proteins were electroblotted onto Immobilon-P membranes and then immunostained with detection using the ECL chemiluminescence system (Amersham Corp.), as described previously (7). G h was detected using a monoclonal antibody to guinea pig liver TGase II (25), and G q was detected using a polyclonal antiserum that was kindly provided by Dr. Michael Crouch (John Curtin School of Medicine, Australian National University). The relative levels of protein expression were determined by densitometric scanning of the bands visualized by immunoblotting.
Phosphatidylinositol Hydrolysis in Intact Cells-PI hydrolysis by intact, transfected COS-1 cells was determined largely as described previously (4,18,19). Briefly, 1 day after transfection, the cells were seeded onto 12-well plates and labeled for 20 -24 h with myo-[ 3 H]inositol at 10 Ci/ml in RPMI 1640 inositol-free medium containing 5% fetal calf serum. After labeling, the cells were washed and incubated in serum-free Dulbecco's modified Eagle's medium for 4 h, followed by a 20-min incubation with 10 mM LiCl plus 0.1 mM ascorbic acid and 5 ϫ 10 Ϫ5 M DL-propranolol. Various drugs were added, as indicated, and the reactions were then stopped by the addition of 20 mM formic acid. Cells were sonicated and the supernatant fractions applied to 1-ml packed AG 1-X8 columns. The columns were washed with 20 ml of 5 mM inositol, and then total IPs were eluted with 2 ml of 1 M ammonium formate, 0.1 M formic acid. Radioactivity in the eluted fractions was determined by ␤-spectrometry as described (4). Differences in the number of cells/dish were normalized, based on the basal accumulation of inositol phosphates or the total counts incorporated into the labeled intact cells, which were roughly proportional to the number of cells/ dish, as also reported by Wong and Ross (2). The accumulation of total (IP 1 and IP 2 plus IP 3 ) [ 3 H]inositol phosphates in different experiments is thus given either as counts/minute compared with basal release or as the increase over basal levels. To avoid interassay differences, all treatments e.g. ␣ 1B alone, ␣ 1B plus TG, and ␣ 1B plus G q were always evaluated together following the transfection of a single batch of cells with the various cDNAs. Results are expressed as the mean Ϯ S.E. (error bars). An analysis of variance and Student's t test were used to determine significant differences (p Ͻ 0.05).
Ligand Binding-The ligand binding characteristics of the expressed receptors were determined in a series of radioligand binding studies performed exactly as described previously (18,19), except using [ 3 H]prazosin, an ␣ 1 -specific antagonist, as the radioligand. Binding data were analyzed using the iterative non-linear, curve-filling program LIGAND.
Construction of the ␣ 1A (i 3 -B)-AR chimera-To construct an ␣ 1A -AR chimera in which the native third intracellular loop (i 3 ) was replaced with i 3 from the ␣ 1B -AR, designated ␣ 1A (i 3 -B)-AR, a Phoenix mutagenesis strategy (26) was employed using the hapaxoterministic restriction endonuclease, BstXI (Fig. 1). This approach was necessary because of the lack of convenient restriction sites to allow construction of the chimera by a conventional mutagenesis strategy. Briefly, the plasmid containing the cDNA for the hamster ␣ 1B was used as a template to amplify a DNA fragment by PCR, which corresponded to the region encoding its i 3 as well as nine residues at the C-terminal end of the fifth transmembrane segment that are identical in both the ␣ 1A and ␣ 1B ARs. The 5Ј(sense) (5Ј-CCATCATTCTGGCCATCATTCTGGTCATGTACTG-CCGGGTCTAC-3Ј, 44 mer) and 3Ј (antisense) (5Ј-CCAGCGTCTTGGC-CAGCGTCTTGGCTGCTTTCTTTTCCCTGGAGAAC-3Ј, 47 mer) primers used for PCR amplification each contained two BstXI recognition sites (underlined) such that with subsequent digestion of the resulting PCR fragment with BstXI, the liberated 5Ј and 3Ј termini would be complementary to the unique 5Ј and 3Ј termini generated by digestion of pMT2Јr␣ 1A -AR at the BstXI sites (BstXIa and BstXIb, Fig. 1) flanking the i 3 region of the ␣ 1A -AR cDNA. pMT2Јr␣ 1A -AR was then digested with BstXI to yield four fragments each containing unique 5Ј and 3Ј termini, which cannot self-hybridize. The ␣ 1B -AR cDNA-amplified PCR fragment was digested with BstXI and isolated after gel purification. The four BstXI-generated ␣ 1A -AR cDNA fragments were ligated with T 4 DNA ligase in the presence of a 10-fold molar excess of the digested and gel-purified ␣ 1B -AR cDNA PCR fragment (i 3 -B) and used to transform Escherichia coli (DH5␣). Resulting ampicillin-resistant colonies (due to the expression of an ampicillin resistance gene in pMT2Ј) containing i 3 -B incorporated into the ␣ 1A -AR cDNA-backbone were identified by colony hybridization using i 3 -B as a probe. A positive colony was then used to isolate plasmid containing the chimeric construct, pMT2Ј ␣ 1A (i 3 -B)-AR, and the sequence of the chimeric construct was confirmed.

RESULTS
As shown in Fig. 2, membranes prepared from COS-1 cells transiently transfected with the ␣ 1B -AR cDNA alone showed a low level of endogenously expressed TGase II that was evident as Ca 2ϩ -sensitive TGase activity. As reported previously (7), this Ca 2ϩ -stimulated TGase activity was sensitive to inhibition by monodansylcadaverine or GTP␥S. The intrinsic COS-1 cell TGase activity was unaltered by cotransfection of ␣ 1B plus mTG, or ␣ 1B plus G q , although in the cells transfected with ␣ 1B ϩ mTG the membrane expression of the mutant TGase II was readily apparent (Fig. 2). This confirms that although, as previously shown, the TGase II mutant still binds and hydrolyzes GTP (22), it lacks TGase activity because of replacement of its active site cysteine with a serine residue. COS-1 cells cotransfected with ␣ 1B plus the wild-type TGase II cDNA showed markedly increased membrane expression of TGase II that was evident by immunoblot analysis. Although basal TGase activity (i.e. TGase activity in the absence of Ca 2ϩ ) was unaltered in these membranes, TGase activity was markedly increased with Ca 2ϩ activation, and this increase could be inhibited by both monodansylcadaverine and GTP␥S (Fig. 2).
In intact COS-1 cells transiently transfected with ␣ 1B -AR cDNA, stimulation with the adrenergic agonist, (Ϫ)epinephrine (in the presence of the ␤-AR antagonist, DL-propranolol), caused a significant increase in IP accumulation (Fig. 3). This response was due to expression of ␣ 1 -ARs and interaction with endogenously expressed G-proteins and PLC, since it could be inhibited by the ␣-antagonist, phentolamine. Native COS-1 cells do not express ␣ 1 ARs and stimulation with ␣ 1 agonists does not increase IP accumulation (data not shown). Both the wild-type TGase II and the mutant TGase II, as well as G q , significantly enhanced (Ϫ)epinephrine-stimulated IP accumulation when cotransfected with the ␣ 1B -AR cDNA, and these (Ϫ)epinephrine-mediated responses could be inhibited with phentolamine. These findings indicate that the mutant TGase II, although lacking TGase activity, can still support ␣ 1B -AR-mediated PLC activation. This was also evident from the time-and dose-dependent activation of PLC observed in COS-1 cells cotransfected with ␣ 1B ϩ mTG versus ␣ 1B alone (Fig. 4, A and B). As shown in Fig. 4A, the time-dependent increase in IP accumulation with ␣ 1B ϩ mTG, and also ␣ 1B ϩ G q , was significantly greater than with ␣ 1B alone. The EC 50 for (Ϫ)epinephrinestimulated IP accumulation decreased from 0.1 Ϯ 0.02 M in cells transfected with ␣ 1B alone to 0.03 Ϯ 0.01 M (p Ͻ 0.05) in ␣ 1B ϩ mTG cells, and in these latter cells maximal IP accumulation was significantly greater (Fig. 4B). Since ␣ 1B -AR density was similar in these studies (11.9 Ϯ 1.2 or 8.7 Ϯ 0.7 pmol/mg membrane protein in cells transfected with ␣ 1B or ␣ 1B ϩ mTG, respectively), these findings indicate that the mutant TG was expressed more abundantly and interacted more efficiently with the ␣ 1B -AR than endogenously expressed G-proteins mediating PLC activation. In cells transfected with ␣ 1B ϩ G q , the potency of (Ϫ)epinephrine (EC 50 0.4 Ϯ 0.2 M) for IP accumulation was similar to that observed with ␣ 1B alone. However, consistent with the overexpression of G q , and despite similar levels of receptor expression in the cells cotransfected with ␣ 1B ϩ G q (10.5 Ϯ 1.3 pmol/mg membrane protein), maximal IP accumulation was significantly greater in the ␣ 1B ϩ G q cells than in the cells transfected with ␣ 1B alone.
To demonstrate that the (Ϫ)epinephrine-stimulated increase in IP accumulation was due to activation of a PI-PLC, we evaluated the effects of the aminosteroid PI-PLC inhibitor  1A (i 3 -B). The plasmid pMT2Ј containing the ␣ 1A -AR cDNA has four recognition sites for the hapaxoterministic restriction endonuclease BstXI. Two of these sites (BstXIa and BstXIb) flank the i 3 -encoding region of the ␣ 1A -AR. A DNA fragment encoding the ␣ 1B -AR i 3 -region was generated by PCR amplification of the ␣ 1B -AR cDNA using sense (5Ј) and antisense (3Ј) primers that contained BstXI sites at their 5Ј ends. The ␣ 1B -AR residues encoded by the PCR product are shown and encompass nine residues at the C terminus of the fifth transmembrane, which are identical in the ␣ 1A -AR, followed by the i 3 residues (boxed). The corresponding ␣ 1A -AR residues are also shown. Ligation of the BstXI-digested ␣ 1A -AR cDNA with a 10-fold molar excess of the BstXI-digested and purified PCR product yielded a chimeric ␣ 1A construct that now contained the ␣ 1B -AR i 3 -encoding region.
U73122 and its inactive analogue U73343. As shown in Fig. 5, U73122 inhibited both the activation of PI-PLC mediated by interaction of the ␣ 1B -AR with endogenous G-proteins and the activation of PI-PLC mediated by the interaction of the ␣ 1B -AR with either the mutant TGase II or G q . By contrast, U73343 had no effect.
We next evaluated the ability of the ␣ 1D and ␣ 1A ARs to mediate PLC-activation via an interaction with either the wildtype or mutant TGase II, or with G q . Although, as shown previously, the ␣ 1D -AR receptor was expressed at much lower levels than the ␣ 1B -AR, the increase in IP accumulation in COS-1 cells cotransfected with the G-proteins was significantly greater than in cells transfected with the ␣ 1D -AR cDNA alone (Fig. 6). In contrast, although able to mediate PLC activation through an interaction with G q , the ␣ 1A -AR receptor did not interact with either TGase II or the mutant TGase II (Fig. 7A). This was also evident from dose-response studies, which despite similar levels of receptor expression (4.4 Ϯ 0.2, 4.3 Ϯ 0.2, and 4.1 Ϯ 0.1 pmol/mg membrane protein for ␣ 1A , ␣ 1A ϩ TG, or ␣ 1A ϩ mTG, respectively) showed no change in (Ϫ)epinephrine potency (EC 50 0.8 Ϯ 0.4, 1.0 Ϯ 0.2, or 1.3 Ϯ 0.7 M for ␣ 1A , ␣ A ϩ TG, or ␣ 1A ϩ mTG) or in maximal IP accumulation in cells co-expressing the ␣ 1A -AR and either TGase II or the mutant TGase II, as compared with the ␣ 1A -AR alone (Fig. 8A). The EC 50 for (Ϫ)epinephrine-stimulated IP accumulation was also not significantly different in cells cotransfected with the ␣ 1A -AR plus G q (0.5 Ϯ 0.2 M). However, consistent with overexpression of G q in these cells, maximal IP accumulation was significantly increased (Fig. 8A). In these studies, the levels of ␣ 1A -AR expression were similar for ␣ 1A plus G q versus ␣ 1A alone (4.4 Ϯ 0.2 and 4.0 Ϯ 0.2 pmol/mg protein, respectively). An inability of the ␣ 1A -AR to interact with TGase II or the mutant TGase II was also evident when receptor expression was altered by transfecting COS-1 cells with increased amounts of ␣ 1A -AR cDNA (Fig. 8B). In these studies, receptor expression ranged from 2 to 15 pmol/mg membrane protein. Failure of the ␣-AR to mediate PLC activation through an interaction with G h was not due to differential expression of TG or mTG, since the levels of TG or mTG expression were not significantly different with cotransfection of cDNAs for any of the ␣ 1 -AR subtypes (data not shown).
Since the third intracellular loop (i 3 ) of G-protein-coupled receptors appears to be importantly involved in the activation of their coupled G-protein, we questioned whether these loops, which differ markedly in their primary structure among the various ␣ 1 -ARs, could account for the observed subtype difference in PLC activation via G h (TGase II). To this end, we constructed a chimeric ␣ 1A -AR in which its i 3 was replaced by that of the ␣ 1B -AR. This ␣ 1A (i 3 -B)-AR chimera was expressed at similar levels to the wild-type ␣ 1A -AR, and its binding of the radioligand [ 3 H]prazosin and of (Ϫ)epinephrine was unaltered (data not shown). In addition, its ability to mediate (Ϫ)epinephrine-stimulated IP accumulation via G q was similar to that of the wild-type ␣ 1A -AR (Fig. 7B). However, by contrast with the wild-type ␣ 1A -AR, the chimera supported (Ϫ)epinephrine stimulated PLC activation via both TGase II and the mutant TGase II (Fig. 7B). DISCUSSION The results of this study provide strong support for our initial contention that the receptor-signaling function of TGase II/G h is independent of its transglutaminase activity. Thus, receptor-mediated IP turnover was unaltered when tested with a mutant form of TGase II that lacks protein cross-linking activity. Together with the finding that intracellular transglutaminase activity is absent in the presence of physiological cytosolic concentrations of nucleotides and Ca 2ϩ (16), this suggests that the major biological role of TGase II may be as a FIG. 2. Transglutaminase activity in membranes from COS-1 cells transfected with the hamster ␣ 1B -AR cDNA (␣ 1B ) or ␣ 1B plus cDNAs for rat TGase II (TG), the mutant human TGase II (mTG), or G ␣q (G q ). Cells were transfected with ␣ 1B (2 g) alone or with the various G-proteins (10 g each). Upper panel, membranes prepared from the transfected cells were subjected to immunoblotting after SDSpolyacrylamide gel electrophoresis resolution, as described under "Experimental Procedures" using a monoclonal antiserum to TGase II (25) (bands 1 to 3 from left to right) or an antiserum to G ␣q (band 4). In cells transfected with ␣ 1B ϩ TG or ␣ 1B ϩ mTG a 74 -80-kDa band is apparent (upper arrow). The slight difference in size between the rat TG (␣ 1B ϩ TG band) and human TG (␣ 1B ϩ mTG band) is due to a previously noted species difference (7). In cells transfected with ␣ 1B ϩ G q , a 42-kDa band is apparent (lower arrow). mediator of membrane signaling rather than as an enzyme involved in the cross-linking of intracellular proteins. Nevertheless, it is possible that the transglutaminase activity of TGase II is relevant to processes such as programmed cell death, in which nucleotide generation is likely to be impaired and intracellular Ca 2ϩ levels are increased. Whether the multifunctional nature of TGase II/G h also has a role in the physiological functioning of cells remains unclear. Since GTP binding by TGase II is inhibited by Ca 2ϩ (9), it is possible that the rise in intracellular Ca 2ϩ associated with receptor stimulation modulates receptor input by limiting TGase-mediated effector activation. Such a mechanism would be analogous to, but distinct from, the limitation of receptor input that results from the activation of the GTPase activity of G q or transducin following interaction with their effectors, PLC␤ or cGMP-dependent phosphodiasterase, respectively (27,28).
As demonstrated here the ␣ 1D -AR also interacts with both TGase II and mutant TGase II in activating the endogenous PI-PLC in COS-1 cells. However, we were not able to demonstrate ligand-induced inositol phosphate release when cells were cotransfected with TG or mTG and the ␣ 1A -AR. Subtype differences in the interaction of ␣ 1 -ARs with various members of the G ␣q family have been demonstrated by Wu et al. (5). In those studies it was shown that while the ␣ 1B -AR interacted with G ␣q , G ␣11 , G ␣14 , and G ␣16 , the ␣ 1A -AR did not interact with G ␣16 , and the ␣ 1D -AR did not interact with G ␣14 or G ␣16 . Consistent with these studies, we also found that G ␣q could mediate ligand-induced inositol phosphate release by all three ␣ 1 -AR receptor subtypes. For the ␣ 1B and ␣ 1D ARs, the EC 50 values for (Ϫ)epinephrine-induced IP accumulation were not markedly different whether these receptors were cotransfected with TGase II/G h or G ␣q . Since it has been demonstrated pre-  Fig. 3. The receptor densities were 0.54 ϩ 0.1, 0.49 ϩ 0.14, 0.36 ϩ 0.13, and 0.58 ϩ 0.19 pmol/mg for ␣ 1D , ␣ 1D ϩ TG, ␣ 1D ϩ mTG, and ␣ 1D ϩ G q , respectively. Results are from four independent experiments performed in duplicate. Asterisks indicate significant differences (*, p Ͻ 0.05; **, p Ͻ 0.01) versus the response in cells transfected with ␣ 1D alone. viously that differences in receptor-G-protein interactions can be discriminated despite overexpression following transient transfection in COS-1 cells (29), it is likely that the ␣ 1B -or ␣ 1D -AR recognizes both TGase II/G h and G ␣q with similar efficacy (whether the slightly lower EC 50 for (Ϫ)epinephrine-stimulated IP accumulation observed in the cells transfected with ␣ 1B ϩ mTG (0.03 Ϯ 0.01 M) compared with the cells transfected with ␣ 1B ϩ Gq (0.4 Ϯ 0.2 M) is physiologically significant remains unclear). Further, since TGase II/G h is expressed ubiquitously in mammalian tissues, and G ␣q is also expressed in a wide variety of cell types, except for some hemopoietic cells (5), it is likely that ␣ 1B -and ␣ 1D -AR signaling involve combined input from both TGase II/G h and members of the G ␣q family. Indeed, such dual coupling to distinct G-proteins may explain why transfection of a dominantly active G ␣q mutant failed to induce transformation in Rat-1 and 3T3 fibroblast cells, whereas transformation of these cells was readily achieved when transfected with a constitutively active ␣ 1B -AR (5).
Since G-protein activation by the G-protein-coupled receptor superfamily involves important determinants in the third intracellular loop, we questioned whether such determinants could also be involved in the subtype-selective interaction of ␣ 1 -ARs with TGase II/G h . As demonstrated here, this indeed appears to be the case, since exchange of the third intracellular loop of the ␣ 1B -AR onto the backbone of the ␣ 1A -AR receptor allowed rescue of a TGase II/G h -dependent signaling phenotype.
The 69-kDa PI-PLC activated by TGase II/G h is not a member of the PLC␤ family (7,30). Recently, based on in vitro studies, we demonstrated that this 69-kDa PI-PLC, in contrast to PLC␤, is resistant to the aminosteroid PI-PLC inhibitor, U73122 (31). Moreover, in intact heart tissue, which expresses TGase II/G h abundantly, ␣ 1 -AR activation of inositol phosphate release is resistant to U73122. With culture of cardiac myocytes, expression of TGase/G h falls markedly (32), whereas expression of G q remains unaltered, and ␣ 1 -AR activation of inositol phosphate release is now U73122-sensitive. This suggests that in vivo myocardial ␣ 1 -ARs activate the U73122resistant 69-kDa PI-PLC via interaction with TGase II/G h , but with culture there is a switch to a U73122-sensitive PLC that is activated by G q . The finding, therefore, that ␣ 1B -AR-mediated IP accumulation is U73122-sensitive (Fig. 5) suggests that either COS-1 cells do not express the U73122-insensitive 69-kDa IP-PLC, or with overexpression of TGase II/G h it promiscuously interacts with an endogenous U73122-sensitive PI-PLC. Resolution of this issue will require further studies including the full identification and characterization of the 69-kDa PI-PLC.  Fig. 3 after stimulation for 30 min with the doses of (Ϫ)epinephrine indicated. B, COS-1 cells were transfected with the amounts of ␣ 1A indicated or with ␣ 1A plus either TG, mTG, or G q , and IP accumulation was determined after stimulation with (Ϫ)epinephrine for 30 min, as detailed in Fig. 3. Results are from four or five independent experiments performed in duplicate. Asterisks indicate significant differences (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001) versus ␣ 1A ϩ G q .