Targeted Inactivation of Gh/Tissue Transglutaminase II*

The novel G-protein, Gh/tissue transglutaminase (TGase II), has both guanosine triphosphatase and Ca2+-activated transglutaminase activity and has been implicated in a number of processes including signal transduction, apoptosis, bone ossification, wound healing, and cell adhesion and spreading. To determine the role of Gh in vivo, the Cre/loxPsite-specific recombinase system was used to develop a mouse line in which its expression was ubiquitously inactivated. Despite the absence of Gh expression and a lack of intracellular TGase activity that was not compensated by other TGases, theTgm2 −/− mice were viable, phenotypically normal, and were born with the expected Mendelian frequency. Absence of Gh coupling to α1-adrenergic receptor signaling in Tgm2 −/− mice was demonstrated by the lack of agonist-stimulated [α-32P]GTP photolabeling of a 74-kDa protein in liver membranes. Annexin-V positivity observed with dexamethasone-induced apoptosis was not different inTgm2 −/− thymocytes compared withTgm2 +/+ thymocytes. However, with this treatment there was a highly significant decrease in the viability (propidium iodide negativity) of Tgm2 −/−thymocytes. Primary fibroblasts isolated fromTgm2 −/− mice also showed decreased adherence with culture. These results indicate that Gh may be importantly involved in stabilizing apoptotic cells before clearance, and in responses such as wound healing that require fibroblast adhesion mediated by extracellular matrix cross-linking.

Transglutaminases (TGases) 1 are a family of thiol-and Ca 2ϩdependent acyl transferases that catalyze the formation of an amide bond between the ␥-carboxamide groups of peptidebound glutamine residues and the primary amino groups in various compounds, including the ⑀-amino group of lysine in certain proteins (1). Seven distinct transglutaminases have been described (reviews in Refs. 2-4 and 5). In addition to G h , also known as tissue TGase (TGase II, 74 -80 kDa), these include, the enzymatically inactive band 4.2 (72-77 kDa), involved in the cytoskeletal network; plasma factor XIIIA (fXIIIA, 75 kDa) involved in catalyzing formation of the fibrin clot at sites of blood coagulation; keratinocyte TGase (TGase I, 90 kDa), which plays a major role in terminal differentiation of epithelia, and in the formation of the cornified cell envelope of the epidermis; epidermal TGase (TGase III, 77 kDa), involved in differentiating epidermal and hair follicle cells; prostate TGase (TGase IV, 65-77 kDa), which, in rodents results in the formation of the copulatory plug through cross-linking of proteins in the seminal vesicle secretion (1); and TGase X (TGase 5, 80 kDa), a novel TGase gene isolated from human keratinocytes. Two new TGases (VI and VII) have recently been identified. 2 G h /TGase II has G-protein signaling and TGase protein cross-linking activities. It is ubiquitously expressed in mammalian tissues (6) and is found both extracellularly at the cell surface in association with the extracellular matrix (7) and intracellularly, where it is both membrane-associated and cytosolic. G h has been implicated in a variety of cellular processes including signal transduction (8), cell adhesion, and spreading (9), wound healing, apoptosis, and bone ossification (10).
In the extracellular matrix, G h cross-links and stabilizes a number of substrates such as laminin-nidogen, fibronectin, fibrinogen, collagen, osteonectin, osteopontin, and the cell adhesion molecule C-CAM (16 -22). G h overexpression in fibroblasts enhances cell attachment (9,21) and in endothelial cells, reduced expression, achieved by the use of antisense techniques, results in decreased cell adhesion and spreading (23). G h has recently been shown to mediate cell adhesion in fibroblasts by acting as a ␤ 1 and ␤ 3 but not ␤ 2 integrin-associated coreceptor for fibronectin; an action that is independent of its TGase activity (24). A rat punch biopsy wound healing model, followed over 6 days, demonstrated increased G h expression and activity at sites of neovascularization and invasion of the fibrin matrix and then in the granulation tissue matrix during healing (25). An intracellular role for G h is suggested by the finding that its activity is down-regulated in the myocardium of humans with cardiac failure (26). Also, transgenic overexpression * This work was supported in part by Grant 980199 from the National Health and Medical Research Council. 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.
¶ 29) transfected with full-length G h cDNA showed an increase in both spontaneous and induced apoptosis. Furthermore, inhibition of neuroblastoma and human promonocytic cell G h expression results in decreased susceptibility to retinoic acid-induced apoptosis (30). In vivo, the expression of G h coincides with apoptosis during formation of the interdigital web (31) and is observed in hypertrophic chondrocytes during endochondral ossification (18,31,32), in myoblasts during differentiation of skeletal muscle (31), and during embryo implantation and postpartum involution of the uterine epithelium (33).
To evaluate the in vivo role of G h /TGase II, we report here the development of a Tgm2-loxP knockin mouse, which allowed inactivation of both Tgm2 alleles after cross-breeding with animals expressing Cre-recombinase.

Generation of Floxed and Knockout Tgm2 Mice-Clones encoding
Tgm2 were isolated from an 129SVJ mouse genomic DNA library (Stratagene) using rat G h cDNA as a probe (34). A binary approach based on the Cre/loxP site-specific recombination system of bacteriophage P1 (35) was used to develop mouse lines in which Tgm2 can either be ubiquitously inactivated or selectively inactivated in specific tissues. A gene-targeting construct in which loxP sites were inserted in the same orientation into introns 5 and 8, which flank exons 6 -8 (encoding the TGase catalytic core domain of G h ) was generated. For positive selection, a hygromycin resistance gene under the control of the phosphoglycerate kinase (PGK) promoter was inserted in antisense orientation immediately 3Ј to the loxP site in intron 5. The PGK/ hygromycin cassette was also flanked by frt sites to allow Flprecombinase-mediated excision, should the presence of the cassette interfere with normal mRNA splicing (Fig. 1A). The targeting construct was transfected into W9.5 embryonic stem cells by the Genetically Modified Mouse Laboratory (GMML), Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia, and homologous recombination events were confirmed by Southern blot analysis using a 5Ј probe (probe 1) external to the targeting construct and a 3Ј probe (probe 2) internal to the targeting construct ( Fig. 1). Chimeric mice were generated and back-crossed with C57BL/6 mice to obtain heterozygous floxed (flanked by loxP sites) Tgm2 t/ϩ mice. Knockout animals (Tgm2 Ϫ/Ϫ ) were generated by crossing heterozygous or homozygous floxed mice with transgenic mice expressing Cre recombinase in the germline under the control of the human cytomegalovirus minimal promoter (36). Mice were initially genotyped by Southern blot analysis of BamHI-digested tail genomic DNA. PCR analysis of tail DNA using 3 primers (P1: forward primer, 5Ј-CATGAATCAGGATGCATCTG-3Ј; P2: forward primer, 5Ј-TAGGGATACAAGAAGCATTG-3Ј; P3: reverse primer, 5Ј-GACAAAGGAGCAAGTGTTAC-3Ј) was performed to genotype the animals in the successive generations.
Western Blot Analysis-Recombinant rat G h was expressed and purified as described (37). Liver and heart tissues were placed in a hypertonic buffer (10 mM Tris-HCl, pH 7.5, 1.4 mM EGTA, 12.5 mM MgCl 2 ) that included a protease inhibitor mixture (Roche Molecular Biochemicals). Tissues were minced and homogenized in an ice bath with a mechanical homogenizer, filtered through a 70 m Nylon cell strainer, and further homogenized in a Dounce homogenizer. Intact cells and organelles were pelleted (600 ϫ g, 10 min, 4°C). Membrane and cytosol BamHI; H, HindII; C, ClaI; loxP and frt, sequences recognized by Cre and Flp recombinases, respectively. The location of probes and primers used for genotyping is indicated. B, genotyping by Southern blot analysis of BamHI-digested genomic tail DNA using probe 1. C, genotyping by PCR screening of genomic tail DNA using primers P1, P2, and P3. D, Northern blot analysis of 20 g of heart total RNA using Tgm2 3Ј-untranslated region probe. E, RT-PCR of 0.5-1 g of heart total RNA using exons 3 and 11 forward and reverse primers, respectively. F, Western blot analysis of liver membranes (100 g) using an anti-bovine TGII polyclonal antibody (G h7␣ ). G, left panel; Western blot analysis of heart cytosol (100 g) using an anti-human TGII monoclonal antibody (CUB 7402) and right panel, heart cytosol (30 g) using an anti-guinea pig TG II polyclonal antibody (06 -471). P (0.1 g), purified recombinant rat G h ; ϩ/ϩ, wild type; ϩ/Ϫ, heterozygous knockout; Ϫ/Ϫ, homozygous knockout; t/ϩ, heterozygous floxed; t/t, homozygous floxed; t/Ϫ, heterozygous floxed, and heterozygous knockout.
Northern Blot Hybridization and Reverse Transcription (RT)-PCR-Total RNA was extracted from heart and liver tissue using the Totally RNA kit (Ambion) according to the manufacturer's instructions. A 1-kb fragment encoding exon 13 and the 3Ј-untranslated region of Tgm2 was used as a probe in Northern blot analyses. Superscript One-Step RT-PCR system (Life Technologies) was used to perform RT-PCR with the following primers: exon 3 forward primer, 5Ј-GCTTCATCTACCAAGGC-3Ј and exon 11 reverse primer, 5Ј-GCTGGTTCGATGAGAAGGC-3Ј.
Transglutaminase Assay-Fibroblast and thymocyte cell lysates were prepared as described (38). Hearts and livers were harvested as follows. Mice were anesthetized with a mixture of xylazine (20 mg/kg) and ketamine (100 mg/kg) and anticoagulated with a single bolus of heparin (5000 units/kg) administered intravenously into the right internal jugular vein. After 5 min, the left internal jugular vein was cannulated, and the animals were perfused with heparinized saline at a rate of 4 ml/min following transection of the right common carotid artery, until all blood was cleared. TGase activity (37) of cytosol and membrane preparations from heart and liver, as well as cell lysates from fibroblasts and thymocytes, was assayed 40 min after addition of 0 mM (basal), 300 M (80% maximal), or 2 mM (maximal) CaCl 2 . The specificity of the TGase assay was confirmed by addition of the competitive substrate inhibitor, monodansylcadaverine. GTP␥S inhibition of TGase activity is inversely proportional to Ca 2ϩ activation and is greatest under conditions where TGase activity is minimal (37). GTP␥S inhibition of TGase activity was therefore evaluated at ϳ80% of maximal Ca 2ϩ -activated TGase activity.
Cardiac Hemodynamic Assessment-Age-matched Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ mice were anesthetized with xylazine (20 mg/kg) and ketamine (100 mg/kg) given intraperitoneally, connected to a rodent ventilator after endotracheal intubation and placed on a thermostatically controlled heating pad. The right carotid artery was cannulated with a 1.4 F pressure transducer (Millar Instruments, Houston, TX), which was advanced into the ascending aorta and then left ventricular cavity. Pressure measurements were recorded in both the left ventricle and the ascending aorta at a sampling frequency of 2000 Hz with a Biopac MP-100 data acquisition system (Biopac Systems Inc., Santa Barbara, CA). Data was subsequently analyzed to determine aortic and left ventricular pressures and heart rate; maximum rates of pressure development (dP/dt max ) and relaxation (dP/dt min ) were calculated from the first derivative of the left ventricular pressure.
Statistical Analyses-All comparisons were performed using the unpaired Student's t test with p Ͻ 0.05 considered significant.

RESULTS AND DISCUSSION
Targeted Disruption of Tgm2-The Cre/loxP site-specific recombination system of bacteriophage P1 was used to develop mouse lines in which Tgm2 can either be ubiquitously inacti- vated or selectively inactivated only in specific tissues. Heterozygous (t/ϩ) and homozygous (t/t) "Tgm2-floxed (flanked by loxP sites)" mouse lines were generated using a gene-targeting vector in which loxP sites, for Cre-mediated excision, were inserted to flank exons 6 -8, which encode the TGase catalytic core domain of G h . mRNA splicing after Cre-mediated excision, results in a frameshift that introduces a number of downstream stop codons, thereby ensuring disruption of G h . Tgm2floxed mice were crossed with mice expressing Cre ubiquitously under the control of the human cytomegalovirus (CMV) minimal promoter to generate Tgm2 knockouts (heterozygous, ϩ/Ϫ; homozygous, Ϫ/Ϫ). Genotyping of progeny by Southern blot analysis of BamHI-digested genomic tail DNA using probe 1 allowed Tgm2-floxed (t, 10 kb), wild type (ϩ, 8 kb) and Credeleted (Ϫ, 4 kb) alleles to be identified (Fig. 1B). Successive generations were genotyped using primers, P1, P2, and P3. P2 and P3 amplifies the wild type (100 bp) and/or floxed (140 bp) alleles, and P1 and P3 amplifies a 180-bp product after Cremediated deletion (Fig. 1C). Northern blots of total RNA isolated from ϩ/ϩ, ϩ/Ϫ, Ϫ/Ϫ, t/Ϫ, and t/t mice were analyzed using either full-length rat G h cDNA (data not shown) or a 1-kb probe encoding exon 13 and the 3Ј-untranslated region of Tgm2 (Fig. 1D). This demonstrated Tgm2 transcripts of both the appropriate size and abundance in Tgm2-floxed mice, indicating normal Tgm2 transcription/splicing despite intron manipulation and the absence of Tgm2 transcripts (full-length or truncated) in the knockout mice. These results were confirmed by RT-PCR (Fig. 1E) using primers directed to exons 3 and 11.
Western blots of liver (Fig. 1F) or heart (data not shown) membrane developed with a polyclonal anti-bovine TGase II antibody G h7␣ (42) showed a 74-kDa band in Tgm2 ϩ/ϩ mice that was of lesser intensity in Tgm2 ϩ/Ϫ mice and absent in Tgm2 Ϫ/Ϫ mice. Westerns blots of heart (Fig. 1G) and liver (data not shown) cytosols or membranes (data not shown) were developed using commercial monoclonal anti-human TGase II (CUB7402) or polyclonal anti-guinea pig TGase II (Upstate Biotechnology, 06 -471) antibodies, respectively. The monoclonal antibody (Fig. 1G, left panel) recognized a 74-kDa band corresponding to G h in Tgm2 ϩ/ϩ mice, which was less intense in Tgm2 ϩ/Ϫ and absent in Tgm2 Ϫ/Ϫ mice (Fig. 1F). The polyclonal anti-guinea pig TGase II antibody (Fig. 1G, right panel) recognized a single band of 74 kDa in Tgm2 ϩ/ϩ mice that was progressively weaker, but nonetheless still present, in Tgm2 ϩ/Ϫ and Tgm2 Ϫ/Ϫ mice, indicating that this commercial polyclonal anti-guinea pig TGase II antibody (06 -471) recognizes two or more proteins of ϳ74 kDa, only one of which is G h . It has been suggested (25) that the three additional bands (ϳ60, 50, and 20 kDa) recognized by the monoclonal antibody are proteolytically degraded G h products. However, the absence of the full-length G h band in Tgm2 Ϫ/Ϫ animals developed with the G h7␣ polyclonal antibody or the monoclonal antibody (Fig. 1F), and the absence of smaller molecular size bands in blots developed with both polyclonal antibodies (Fig. 1, G and  F, left panel), make this unlikely. These findings indicate a lack of specificity of the commercial antibodies and question previously reported data generated using these antibodies.
TGase Inactivation-The absence of G h cross-linking activity in Tgm2 Ϫ/Ϫ mice was demonstrated by TGase activity assays. To minimize the contribution to TGase activity by fXIIIA and other TGases in blood, animals were anticoagulated and perfused with heparinized saline before tissue collection. TGase activity of liver and heart cytosol (Fig. 2, A and B) and membrane preparations (data not shown) was markedly decreased in Tgm2 Ϫ/Ϫ as compared with Tgm2 ϩ/ϩ mice. There was no significant difference in activity between the Tgm2 ϩ/ϩ and Tgm2 ϩ/Ϫ animals. The competitive substrate inhibitor, monodansylcadavarine, although used at a concentration (40 M) that was submaximal, reduced the Ca 2ϩ -stimulated TGase activity in all samples from 100 to 23-40% (data not shown), confirming the specificity of the assay.
The small amount of TGase activity observed in the Tgm2 Ϫ/Ϫ heart and liver preparations may reflect residual blood (and therefore fXIIIA) contamination of the samples or compensation by other intracellular TGases. To address the issue of compensation, TGase activity was evaluated in the presence of GTP␥S, which inhibits the TGase activities of G h and TGase III (43), but not that of other TGases. As shown in Fig. 2, A and B, the TGase activity of both Tgm2 ϩ/ϩ and Tgm2 ϩ/Ϫ samples in the presence of GTP␥S was equivalent to that of Tgm2 Ϫ/Ϫ . Furthermore, the TGase activity of Tgm2 Ϫ/Ϫ samples was not inhibited by GTP␥S, indicating the TGase activity observed in the Tgm2 Ϫ/Ϫ mice is not contributed by a GTP-sensitive TGase and is most likely contributed by an extracellular TGase. This was confirmed by performing TGase assays on primary fibroblast cultures established from Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ heart and lung tissue. No residual TGase activity was observed in cell lysates from knockout primary cultured fibroblasts, whereas robust TGase activity was evident in Tgm2 ϩ/ϩ cells (Fig. 2C). This indicates that the small amount of activity observed in the Tgm2 Ϫ/Ϫ hearts and livers can be attributed to a small amount of blood contamination, and therefore residual activity of fXIIIA or other blood-borne TGases.
Functional coupling of G h to ␣ 1 -adrenergic receptor (␣ 1 -AR) signaling was investigated in the Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ mice by [␣-32 P]GTP photolabeling of purified liver membranes in the absence or presence of the ␣ 1 -AR agonist, (Ϫ)epinephrine, or in the presence of (Ϫ)epinephrine plus the antagonist, phentolamine (Fig. 3). In Tgm2 ϩ/ϩ liver membranes, (Ϫ)epinephrine treatment resulted in a significant enhancement (2-fold, p Ͻ 0.01) in [␣-32 P]GTP labeling of the 74-kDa G h ; a response that was completely inhibited by pretreatment with the ␣ 1 -antagonist, phentolamine (Fig. 3). In Tgm2 Ϫ/Ϫ liver membranes, however, only labeling of a ϳ40 kDa G-protein, probably G␣ i , was observed. (Fig. 3). These findings indicate both that G h is functionally coupled to the ␣ 1 -AR, and that other GTP-binding TGases are unable to substitute for G h in ␣ 1 -AR-mediated signaling.
expected Mendelian frequency. The Tgm2 Ϫ/Ϫ animals have normal separation of their digits and open eyelids, indicating that developmental apoptosis is not impaired. In addition, homozygous knockouts breed normally and have no problems with parturition. These findings indicate that G h is not critically involved in reproduction or in maturational apoptosis.
In the thymuses of young mice, immature thymocytes that are not selected to differentiate into T-cells, are cleared by apoptotic cell death. The percentage and absolute numbers of the different thymocyte populations in 4 -6-weekold Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ animals, identified by expression of CD4 and CD8 receptors, was determined by flow cytometry. There was no difference between the Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ mice with respect to the percentage (Ϯ 1 S.E.) of CD4 Ϫ CD8 ϩ (Tgm2 ϩ/ϩ 10.9 Ϯ 0.5%, Tgm2 Ϫ/Ϫ 10.5 Ϯ 0.6%), or CD4 ϩ CD8 ϩ (Tgm2 ϩ/ϩ 81.9 Ϯ 1.0%, Tgm2 Ϫ/Ϫ 81.9 Ϯ 0.6%) thymocytes. This indicates that normal thymocyte apoptotic turnover of immature T-cells is not affected by the lack of G h . Because dexamethasone induces apoptosis and increases TGase activity in the thymus in vivo (41), the effect of intraperitoneal dexamethasone administration was assessed. In Tgm2 ϩ/ϩ thymocytes, TGase activity was marginally increased at 8 h and markedly increased at 24 h after dexamethasone treatment, and the latter was significantly greater than the 24 h value in Tgm2 Ϫ/Ϫ cells (p Ͻ 0.05; Fig. 4A). In contrast to the effects of dexamethasone in Tgm2 ϩ/ϩ cells, in Tgm2 Ϫ/Ϫ thymocytes, TGase activity at 24 h was slightly but not significantly increased over that observed at 8 h (Fig. 4A). Moreover, whereas GTP␥S inhibited (by 62%, p Ͻ 0.001) the 24 h increase in TGase activity of Tgm2 ϩ/ϩ thymocytes, TGase activtiy in Tgm2 Ϫ/Ϫ cells was GTP␥S-insensitive (not shown). This lack of increase in thymic TGase activity in the Tgm2 Ϫ/Ϫ animals was associated with smaller thymuses and less TUNEL positivity (Fig. 4, B and C). To ascertain if this was because of decreased cell death or an increased rate of clearance of dead cells from the thymus by macrophages (44), flow cytometry analysis was performed on isolated thymocytes. Cells were cultured for 8 or 24 h in the absence or presence of 1 M dexamethasone, although only the cells at 8 h were further evaluated because of the marked loss of Tgm2 Ϫ/Ϫ cells with 24 h of dexamethasone treatment. Early apoptosis results in membrane exposure of phosphatidylserine that is recognized by the phospholipidbinding protein, Annexin-V. As seen in Fig. 4D, although Annexin-V positivity of both Tgm2 ϩ/ϩ and Tgm2 Ϫ/Ϫ thymocytes increased with dexamethasone treatment (p Ͻ 0.0001), there was no significant difference between the Tgm2 Ϫ/Ϫ and Tgm2 ϩ/ϩ cells. However, propidium iodide staining of dead cells (Fig. 4E), an index of late-stage cell death, showed a small but highly statistically significant decrease (p Ͻ 0.0001) in the  viability of Tgm2 Ϫ/Ϫ thymocytes. This indicates that Tgm2 Ϫ/Ϫ thymocytes are more susceptible to dexamethasone-induced cell death, and that the decreased TUNEL positivity of Tgm2 Ϫ/Ϫ thymuses is the result of increased clearance of dead cells. This, in turn, likely contributes to the decreased size of dexamethasone-treated Tgm2 Ϫ/Ϫ thymuses compared with those from Tgm2 ϩ/ϩ animals. Thus, G h -dependent crosslinking is likely importantly involved in stabilizing apoptotic cells before clearance, as suggested by Piredda et al. (29).
Previous studies of Swiss 3T3 fibroblasts and endothelial cells indicate that G h has an extracellular role in cell attachment (21,23) and cell spreading (23). In agreement with these studies, it was more difficult to establish primary fibroblast cultures from Tgm2 Ϫ/Ϫ mice than from Tgm2 ϩ/ϩ animals. Thus, despite plating fibroblasts at equal density, fewer Tgm2 Ϫ/Ϫ fibroblasts were adherent after 2 h (20 Ϯ 3% versus 79 Ϯ 8% for Tgm2 ϩ/ϩ , n ϭ 10 fields Ϯ 1 S.E.). This was not because of increased death of non-adherent Tgm2 Ϫ/Ϫ fibroblasts, as confirmed by trypan blue exclusion. Similar results were obtained with cells grown on fibronectin-coated plates (data not shown). Thus, G h may be importantly involved in various physiological and pathological responses, such as wound healing and scar formation, which are mediated by the interaction of fibroblasts with the extracellular matrix.
Cardiac function in 8 -10-week-old Tgm2 Ϫ/Ϫ and Tgm2 ϩ/ϩ animals was evaluated by micromanometry (Table I). Systolic and diastolic blood pressures were measured in the ascending aorta, and maximum rates of pressure development (dP/dt max ) and of relaxation (dP/dt min ) were calculated from the left ventricular pressure. The Tgm2 Ϫ/Ϫ mice showed no statistically significant differences compared with Tgm2 ϩ/ϩ mice, for any of the measured parameters. The role of G h in the maintenance of normal cardiovascular function is unclear. Cardiac failure is associated with both a down-regulation and uncoupling of ␤-AR and hence a relative up-regulation of ␣ 1 -ARs (26). However, down-regulation of G h has also been demonstrated in the setting of cardiac dysfunction (26). ␣ 1 -ARs are thought to contribute little to normal cardiac inotropy and may thus act as a reserve mechanism (for review see Ref. 45). ␣ 1B -ARs, however, do have a significant role in maintaining normal vascular tone, and hence in blood pressure homeostasis (46). The lack of change in either blood pressure or parameters of left ventricular function in this study suggests either that G h in the vasculature contributes minimally to ␣ 1B -AR-mediated vasoconstrictor responses, or that in the absence of G h , vascular tone can be maintained by compensatory signaling pathways. Alternatively, impaired inotropic drive in the hearts of the knockout animals may be counterbalanced by a reduced afterload that results from the impaired ability of the ␣ 1B -AR to mediate vasoconstriction in the absence of G h . Interestingly, cardiacspecific overexpression of G h results in mild hypertrophy and ventricular fibrosis, as well as impaired cardiac function (27), a phenotype that is consistent with that obtained by overexpression of either the wild-type ␣ 1B -AR (47) or a constitutively active ␣ 1B -AR mutant (48). Thus, elucidating the potential in vivo role of G h in cardiovascular homeostasis may require additional evaluations in animals in which G h is selectively inactivated only in the heart or only in the vasculature, or may only be revealed by subjecting the knockout animals to a pathophysiological stress, such as thoracic aortic constriction. Such studies are currently in progress.