Transforming Growth Factor β Regulates the Expression of the M2 Muscarinic Receptor in Atrial Myocytes via an Effect on RhoA and p190RhoGAP*

Transforming growth factor β (TGFβ) signaling is involved in the development and regulation of multiple organ systems and cellular signaling pathways. We recently demonstrated that TGFβ regulates the response of atrial myocytes to parasympathetic stimulation. Here, TGFβ1 is shown to inhibit expression of the M2 muscarinic receptor (M2), which plays a critical role in the parasympathetic response of the heart. This effect is mimicked by overexpression of a dominant negative mutant of RhoA and by the RhoA kinase inhibitor Y27632, whereas adenoviral expression of a dominant activating-RhoA reverses TGFβ inhibition of M2 expression. TGFβ1 also mediates a decrease in GTP-bound RhoA and a reciprocal increase in the expression of the RhoA GTPase-activating protein, p190RhoGAP, whereas total RhoA is unchanged. Inhibition of M2 promoter activity by TGFβ1 is mimicked by overexpression of p190RhoGAP, whereas a dominant negative mutant of p190RhoGAP reverses this effect of TGFβ1. In contrast to atrial myocytes, in mink lung epithelial cells, in which TGFβ signaling through activation of RhoA has been previously identified, TGFβ1 stimulated an increase in GTP-bound RhoA in association with a reciprocal decrease in the expression of p190RhoGAP. Both effects demonstrated a similar dose dependence on TGFβ1. Thus TGFβ regulation of M2 muscarinic receptor expression is dependent on RhoA, and TGFβ regulation of p190RhoGAP expression may be a cell type-specific mechanism for TGFβ signaling through RhoA.

Transforming growth factor ␤ (TGF␤) signaling is involved in the development and regulation of multiple organ systems and cellular signaling pathways. We recently demonstrated that TGF␤ regulates the response of atrial myocytes to parasympathetic stimulation. Here, TGF␤ 1 is shown to inhibit expression of the M 2 muscarinic receptor (M 2 ), which plays a critical role in the parasympathetic response of the heart. This effect is mimicked by overexpression of a dominant negative mutant of RhoA and by the RhoA kinase inhibitor Y27632, whereas adenoviral expression of a dominant activating-RhoA reverses TGF␤ inhibition of M 2 expression. TGF␤ 1 also mediates a decrease in GTP-bound RhoA and a reciprocal increase in the expression of the RhoA GTPase-activating protein, p190RhoGAP, whereas total RhoA is unchanged. Inhibition of M 2 promoter activity by TGF␤ 1 is mimicked by overexpression of p190RhoGAP, whereas a dominant negative mutant of p190RhoGAP reverses this effect of TGF␤ 1 . In contrast to atrial myocytes, in mink lung epithelial cells, in which TGF␤ signaling through activation of RhoA has been previously identified, TGF␤ 1 stimulated an increase in GTP-bound RhoA in association with a reciprocal decrease in the expression of p190RhoGAP. Both effects demonstrated a similar dose dependence on TGF␤ 1 . Thus TGF␤ regulation of M 2 muscarinic receptor expression is dependent on RhoA, and TGF␤ regulation of p190RhoGAP expression may be a cell type-specific mechanism for TGF␤ signaling through RhoA.
Control of the autonomic response of the heart is essential for the regulation of cardiac automaticity, heart rate, and contractile force (1). Autonomic regulation of the heart is deter-mined by the balance between the response to stimuli from the sympathetic and parasympathetic nervous systems. Much effort has focused on the regulation of sympathetic responsiveness and its role in cardiovascular diseases such as heart failure, cardiac hypertrophy, and the genesis of arrhythmias (2). Much less is known concerning the regulation of parasympathetic responsiveness despite its clear role in cardiovascular disease. Parasympathetic stimulation of the heart has been shown to play a role in the protection of the heart from the development of arrhythmias (3,4), whereas parasympathetic dysfunction is associated with sudden death due to arrhythmias in patients with diabetic autonomic neuropathy (5). Parasympathetic stimulation of the heart decreases contractile rate via the binding of acetylcholine to atrial M 2 muscarinic receptors (M 2 ), 5 which results in the dissociation of the heterotrimeric G-protein G i2 into ␣ i2 and ␤␥ subunits (6). The latter activates the G-protein-dependent inward rectifying K ϩ channel, (GIRK1) 2 / (GIRK4) 2 , resulting in hyperpolarization of the atrial myocyte membrane and a decrease in the rate of contraction (6). The relationship between the control of expression of these proteins and the response of the heart to parasympathetic stimulation has been demonstrated in the porcine heart and in atrial myocytes (7,8). Previously, we and others have demonstrated that TGF␤ regulates the response of atrial myocytes to parasympathetic stimulation via an effect on the expression of G␣ i2 and M 2 (9 -11). Here we demonstrate that TGF␤ regulates M 2 expression by a novel mechanism involving alterations in RhoA activity.
TGF␤ signals by binding to the type II TGF␤ receptor, resulting in the phosphorylation of the type I receptor, activin receptor-like kinase (ALK) 5. Activated ALK5 then phosphorylates intracellular signaling molecules including members of the Smad family of transcription factors, Smad2/3, which bind to the common Smad4 to form a complex that is translocated to the nucleus. However, TGF␤ is also known to signal through small GTP-binding proteins such as Ras and RhoA (12). TGF␤ regulation of Ras has been shown to result in pleiotrophic effects that involve either the stimulation or the inhibition of Ras activity that is specific either for cell type or for the stage of embryonic development (9,(13)(14)(15). The Rho family of small GTP-binding proteins is involved in many cellular processes including cell adhesion, migration, and transformation (16). Bhowmick et al. (17) demonstrated that TGF␤-dependent epithelial-to-mesenchymal transdifferentiation (EMT) in a nontransformed mouse mammary epithelial cell line was dependent on an increase in the level of GTP-bound RhoA. Active RhoA was also found to be increased downstream of TGF␤ in mink lung epithelial cells (Mv1Lu) (17). TGF␤ signaling via the inhibition of RhoA activity has not been demonstrated.
RhoA activation is regulated by cycling between the inactive GDP-bound form of RhoA to the active GTP-bound membrane-associated form. RhoA activity can be regulated by at least three different classes of molecules. Rho-GDP dissociation inhibitors (RhoGDIs) regulate RhoA activity by extracting the inactive GDP-bound form of RhoA from the membrane and forming inactive cytosolic complexes (18). The second class of molecules, RhoGEFs (guanine nucleotide exchange factors), stimulates the dissociation of Rho-GDP and enhance GTP binding, thereby increasing RhoA activity (19,20). A third class of molecules, RhoGAPs (GTPase-activating proteins (GAPs)), bind to Rho-GTP and stimulate RhoGTPase activity, which decreases the level of GTP-bound RhoA, thereby decreasing RhoA signaling. Wolf et al. (21) and Arthur and Burridge (22) recently demonstrated that overexpression of p190RhoGAP inhibits Rho-dependent stimulation of cell signaling. Whether TGF␤ regulates RhoA activity through any of these molecules is unknown.
Here we present data that TGF␤ 1 mediates a RhoA-dependent decrease in M 2 muscarinic receptor protein, mRNA, and promoter activity in cultured embryonic chick atrial myocytes. This effect is mediated via a decrease in GTP-bound RhoA and is associated with the stimulation of p190RhoGAP expression by TGF␤ 1 . The importance of this mechanism for the regulation of the levels of RhoA activity in systems in which TGF␤ signaling is mediated via RhoA is further supported by the finding that TGF␤ 1 increases GTP-bound RhoA and decreases p190RhoGAP expression in TGF␤ 1 -stimulated mink lung epithelial cells.

EXPERIMENTAL PROCEDURES
Plasmids-The chicken M 2 muscarinic receptor promoter-luciferase reporter was a gift of Dr. Neil Nathanson (23), pCDNA3-Myc-N19RhoA and pRK5 Myc-L63RhoA were gifts of Dr. Alan Hall, and WT p190RhoGAP and p190RhoGAP R1283A were gifts of Dr. Ian Macara. Y27632 was a gift from Dr. Y. Takae, Yoshitomi Pharmaceutical Industries, Osaka, Japan. The DN-Smad4 was a kind gift of Dr. Akiko Hata (24).
Cell Culture-Embryonic chick atrial myocyte cultures were prepared by a modification of the method of DeHaan (25) as described (26). Since we had previously demonstrated that culture of atrial myocytes in 6% lipoprotein-depleted serum, LPDS, stimulated the expression of the M 2 muscarinic receptor, studies of the effects of TGF␤ were carried out in atrial myocytes cultured in LPDS or in 0.6% fetal bovine serum to maximize the signal. LPDS was prepared as described previously (26). Cell culture media and supplies were from Invitrogen. Mv1Lu were obtained from Dr. Harold Moses and cultured as described (17). Note that the effects of TGF␤ 1 on M 2 promoter activity in atrial myocytes were similar whether cells were incubated in LPDS or 0.6% fetal bovine serum (see Figs. 2 and 4).
RNase Protection Analysis-An M 2 muscarinic receptor RNase protection probe was generated from an XbaI fragment derived from chicken M 2 cDNA (nucleotides 97-1139) subcloned into pBluescript and linearized with BamHI (27). Using T7 RNA polymerase (Roche Applied Science) in the presence of [ 32 P]UTP (800 Ci/mmol, PerkinElmer Life Sciences), this template gave a 449-nucleotide antisense riboprobe. The glyceraldehyde phosphate dehydrogenase RNase protection probe, used as a normalizing control to ensure equal sample loading, was generated from a cDNA template that was linearized with HindIII. Using T3 RNA polymerase, this template gave a 250nucleotide antisense riboprobe. Probes were purified by PAGE on a 6% gel, and the major band corresponding to the predicted molecular weight for the riboprobe was excised and eluted overnight. Total RNA was isolated from primary cultures of embryonic chick atrial cells 14 days in ovo plated at 4 ϫ 10 5 cells/cm 2 using guanidinium CsCl 2 centrifugation as described (28). RNase protection was carried out as described previously (26). Riboprobes were hybridized to 15 g of total RNA prepared from cells treated with either vehicle (4 mM HCl and 0.5 mg/ml bovine serum albumin) or TGF␤ 1 (R&D systems Inc. Minneapolis, MN) as indicated. The samples were treated with RNase and analyzed by PAGE on 6% gels containing 8 M urea followed by autoradiography. Autoradiographic exposure was 8 h for the M 2 muscarinic receptor and 2 h for glyceraldehyde phosphate dehydrogenase. The relative intensity of the bands was determined by densitometry scanning using NIH Image Pro.
Western Blotting Analysis-Embryonic chick atrial cells from hearts of embryos 14 days in ovo were grown for 3 days in LPDS at 4 ϫ 10 5 cells/cm 2 and then incubated with either vehicle or TGF␤ 1 as indicated. On the fourth day, whole cell lysates were analyzed by SDS-PAGE using a 12% gel. Equal amounts of protein were loaded as determined by a DC protein assay (Bio-Rad). Equal loading was determined by Coomassie Blue staining. Western blot analysis was carried out as described (26). Immunodetection of the M 2 receptor was performed using a rabbit polyclonal antibody from Research & Diagnostic Antibodies, Benicia, CA. Polyclonal goat anti-RhoGDI, antibody number A-20 (Santa Cruz Biotechnology, Santa Cruz, CA), and monoclonal mouse anti-p190RhoGAP antibodies (BD Transduction Laboratories) were also used for immunodetection as indicated. The relative intensity of the bands was determined by densitometry scanning using NIH Image Pro.
Luciferase Assay-Embryonic chick atrial cells 14 days in ovo were plated on 6-well dishes at 4 ϫ 10 5 cells/cm 2 and grown in medium supplemented with 6% LPDS or 0.6% fetal bovine serum. On the second culture day, 0.5 g of a 789-bp fragment of the chicken M 2 promoter ligated to the 5Ј end of a firefly luciferase reporter gene in a pGL3 expression vector (M 2 -Luc) was transiently transfected into atrial myocytes with 0.2 g of pCMV-␤-galactosidase (Clontech) using FuGENE 6 transfection reagent (Roche Applied Science) as recommended by the manufacturer. Total DNA was maintained at 2.1 g/well by the addition of pBlueScript DNA. Sixteen hours prior to harvesting, cells were incubated with TGF␤ 1 as indicated. On the fourth day, cells were washed in phosphate-buffered saline and solubilized in lysis buffer, 425 l/plate (24 mM glycyl-glycine, 15 mM MgSO 4 , 4 mM EGTA, 1% Triton X-100, and 1 mM dithiothreitol). The cell extract was sonicated three times for 10 s and centrifuged at 13,000 ϫ g for 3 min at 4°C, and the supernatant assayed for luciferase and ␤-galactosidase activity as described (29). In some experiments, cells were co-transfected with pCDNA3-Myc-N19RhoA (dominant negative, DN) or pRK5-Myc-L63RhoA (dominant activating, DA). In some experiments using 12-well dishes, cells were transfected with 150 ng of the M 2 promoter-luciferase reporter construct using the Lipofectamine transfection reagent (Invitrogen). Where indicated, cells were co-transfected with either WT-p190RhoGAP or p190RhoGAP R1283A mutant or DN-Smad4 mutant constructs at the concentrations indicated. Total DNA was maintained at 1 g/well using pBlueScript according to the manufacturer's recommendation. Twenty-four hours following transfection, cells were incubated with the indicated concentrations of TGF␤ 1 in medium supplemented with 0.6% fetal bovine serum for an additional 24 h and harvested, and luciferase activity was determined as indicated above.
Measurement of GTP-bound RhoA-GTP-bound RhoA was determined as described (30). Chicken atrial myocytes 14 days in ovo were cultured at 4 ϫ 10 5 cells/cm 2 for 3 days in medium supplemented with either 6% LPDS plus vehicle or TGF␤ 1 as indicated. Atrial myocytes were harvested on the fourth day, and whole cell lysates were precipitated with Rhotekin (Upstate Biotechnology, Lake Placid, NY), a mouse glutathione S-transferase fusion protein with the Rho binding domain specific for the active GTP-bound RhoA (31). Rho binding domain precipitates were analyzed by SDS-PAGE using a 15% gel, and immunodetection of RhoA was carried out using a purified polyclonal rabbit anti-RhoA, antibody number 119 (Santa Cruz Biotechnology, Santa Cruz, CA) (32,33). For determination of total RhoA, 50 g of whole cell lysate from each sample that had not been precipitated by Rhotekin was subjected to SDS-PAGE and Western blot analysis. TGF␤ 1 inhibition of GTP-bound RhoA was calculated as -fold change normalized to total RhoA for each sample. Equal loading was determined by Coomassie Blue staining. For studies of GTP-bound RhoA in Mv1Lu, cells were washed and harvested, and GTP-bound RhoA and total RhoA were determined as described for atrial myocytes.
Adenoviral Infection of Atrial Myocytes with Ad-DA-RhoA-Recombinant adenovirus encoding a DA-L63RhoA under the control of a tetracycline-controlled transactivator has been described previously (34). Atrial myocytes were cultured to near confluence. Medium was removed, and cells were co-infected for 2 h with a virus encoding the transactivator and a virus expressing either GFP or a DA-RhoA under the positive control of the tetracycline-controlled transactivator at the indicated multiplicities of infection. Virus was removed, and cells were incubated for 24 h, after which they were treated with either TGF␤ or vehicle in medium containing 0.6% fetal bovine serum. Cells were harvested for Western blot analysis as described above.
Statistical Analysis-All data are plotted as the mean Ϯ S.E. Statistical significance was determined by a paired t test analysis. A value of p Ͻ 0.05 was considered significant.

Effect of TGF␤ 1 on M 2 Muscarinic Receptor Expression-To
determine the mechanism by which TGF␤ 1 inhibited the expression of the M 2 muscarinic receptor in atrial myocytes, we chose growth conditions that would maximize M 2 muscarinic receptor expression. We had previously demonstrated that growth of embryonic chick atrial myocytes in LPDS markedly increased both the expression of the M 2 muscarinic receptor and the physiologic response of atrial myocytes to the muscarinic agonist carbamylcholine (32). For this reason, embryonic chick atrial cells 14 days in ovo were cultured in LPDS and subsequently incubated with either vehicle or 5 ng/ml TGF␤ 1 for 16 h. Western blot analysis of cell lysates demonstrated that TGF␤ 1 markedly decreased the level of total M 2 receptor protein (Fig. 1A) by a mean of 46 Ϯ 6.3% (n ϭ 3, p Ͻ 0.01, Fig. 1B) when compared with control.
The effect of TGF␤ 1 on M 2 receptor mRNA expression and promoter activity was examined to determine whether the regulation of M 2 muscarinic receptor by TGF␤ 1 occurred at the level of transcription. RNase protection analysis of cells incubated with TGF␤ 1 demonstrated that TGF␤ 1 significantly decreased the level of M 2 mRNA expression (Fig. 1, C and D) by a mean of 56 Ϯ 8.3% (n ϭ 3, p Ͻ 0.05) when compared with control. Finally, atrial myocytes were transfected with a construct expressing the chick M 2 muscarinic receptor promoter ligated to a luciferase reporter (23) and incubated with either vehicle or TGF␤ 1 . TGF␤ 1 decreased M 2 promoter activity by 56 Ϯ 4% (n ϭ 5, p Ͻ 0.001) when compared with control, Fig. 2A.
The Smad family of transcription factors is an important mediator of many TGF␤ signaling events. To determine whether TGF␤ 1 inhibition of M 2 promoter activity was Smaddependent, atrial myocytes were co-transfected with the M 2 promoter luciferase reporter and increasing concentrations of a construct expressing a truncated Smad4 containing only the DNA binding domain, which behaves as a dominant negative mutant (24). Expression of the DN-Smad4 completely reversed the inhibition of M 2 promoter activity by TGF␤ 1 in a dose-dependent manner (Fig. 2A).
Role of RhoA in TGF␤ 1 Inhibition of M 2 Receptor Expression-In addition to Smads, RhoA has been implicated in several TGF␤ signaling pathways. To determine whether RhoA played a role in TGF␤ 1 -mediated inhibition of M 2 promoter activity, atrial myocytes were co-transfected with the M 2 promoter luciferase reporter and a construct expressing a dominant negative mutant of RhoA, N19-RhoA (35). N19-RhoA expression inhibited M 2 promoter activity 74 Ϯ 4% (n ϭ 3, p Ͻ 0.004) when compared with control (Fig. 2B), demonstrating that N19-RhoA mimicked the effect of TGF␤ 1 on M 2 promoter activity. Y27632, a specific inhibitor of Rho kinase, also inhibited M 2 promoter activity (data not shown). Furthermore, Western blot analysis of the expression of the M 2 receptor in cells incubated with Y27632 demonstrated that Y27632 also mimicked the effect of TGF␤ 1 by inhibiting M 2 receptor expression (Fig. 2C). To further determine the role of Rho in the regulation of M 2 receptor expression, atrial myocytes were co-transfected with the M 2 promoter luciferase reporter and a construct expressing a dominant activating mutant of RhoA, L63-RhoA (35). DA-RhoA expression not only reversed the effect of TGF␤ on M 2 promoter activity but also stimulated activity 1.35 Ϯ 0.02 (n ϭ 4, p Ͻ 0.01) -fold above control (Fig. 3A). Co-infection of atrial myocytes with adenoviruses expressing the tetracycline transactivator, and GFP had no effect on TGF␤ inhibition of M 2 muscarinic receptor expression as determined by Western blot analysis (Fig. 3B). However, co-infection with the transactivator plus an adenovirus expressing the DA-L63RhoA completely reversed the effect of TGF␤ on M 2 expression (Fig. 3B). These data support the conclusion that TGF␤ inhibition of M 2 receptor expression is dependent on RhoA.
TGF␤ 1 Inhibits RhoA Activity by Decreasing GTP-bound RhoA-To determine whether TGF␤ 1 regulated RhoA activity in atrial myocytes, cells were cultured in medium supplemented with LPDS plus vehicle or LPDS plus increasing concentrations of TGF␤ 1 , and the level of GTP-bound RhoA determined. Whole cell lysates were precipitated with Rhotekin. Western blot analysis of Rhotekin precipitates using an anti-RhoA antibody revealed a marked dose-dependent decrease in GTP-bound RhoA, whereas total RhoA in unprecipitated aliquots of these extracts was unchanged (Fig. 4A). The mean of three experiments similar to that in Fig. 4A demonstrated that when compared with control, TGF␤ 1 decreased GTP-bound RhoA by 42 Ϯ 5% (n ϭ 4, p Ͻ 0.01) at a concentration of 1 ng/ml and 64 Ϯ 7% (n ϭ 4, p Ͻ 0.01) at 5 ng/ml (Fig. 4B). TGF␤ 1 incubation resulted in a small decrease in membrane-bound RhoA, which was not statistically significant (data not shown).
p190RhoGAP Expression Is Stimulated by TGF␤ 1 -A decrease in GTP-bound RhoA in response to TGF␤ 1 might be explained in part by an increase in RhoGDI or RhoGAP. Western blot analysis of whole cell lysates of atrial myocytes cultured in LPDS and incubated with either vehicle or TGF␤ 1 demonstrated that TGF␤ 1 significantly increased the expression of p190RhoGAP by 2.18 Ϯ 0.12-fold (n ϭ 5, p Ͻ 0.01) with no effect on RhoGDI (Fig. 5, A and B).
p190RhoGAP Inhibits M 2 Promoter Activity-We predicted that if M 2 promoter activity is inhibited by a decrease in GTPbound RhoA, then overexpression of p190RhoGAP should mimic the effect of TGF␤ 1 on M 2 promoter activity. Atrial myocytes were transfected with the M 2 receptor luciferase reporter and incubated with either vehicle or TGF␤ 1 or co-transfected either with p190RhoGAP or p190RhoGAP R1283A , a dominant negative mutant of p190RhoGAP that binds to GTP-bound Rho but lacks GAP activity (22, 36 -39). As shown previously, TGF␤ 1 decreased M 2 promoter activity by 46 Ϯ 5% (n ϭ 3, p Ͻ 0.01), whereas p190RhoGAP decreased promoter activity by 36 Ϯ 2% (n ϭ 3, p Ͻ 0.01), Fig. 5C. Co-transfection of atrial myocytes with the M 2 receptor luciferase reporter and p190RhoGAP R1283A had no effect on the basal M 2 receptor promoter activity, Fig. 5C.
If TGF␤ inhibition of M 2 promoter activity is mediated via increased expression of p190RhoGAP, then we predicted that overexpression of DN-p190RhoGAP should prevent TGF␤-mediated inhibition of M 2 receptor expression. Atrial myocytes cotransfected with the M 2 receptor luciferase reporter and increasing concentrations of a construct expressing DN-p190RhoGAP were incubated with TGF␤ 1 . Although TGF␤ 1 inhibited M 2 receptor activity 56 Ϯ 3% (n ϭ 4) in the absence of DN-p190RhoGAP , expression of DN-p190RhoGAP completely reversed the effect of TGF␤ 1 on M 2 receptor promoter activity in a dose-dependent manner (Fig. 5D). These data demonstrate that p190RhoGAP is required for TGF 1 -mediated inhibition of M 2 receptor expression.

TGF␤ 1 Decreases p190RhoGAP Expression and Increases
RhoA Activity in Mv1Lu-To determine whether the regulation of p190RhoGAP expression plays a role in TGF␤-regulated RhoA activity in other cell types, we examined p190RhoGAP expression in Mv1Lu cells in which TGF␤ had previously been shown to stimulate RhoA-dependent EMT. Mv1Lu cells were incubated with increasing concentrations of TGF␤ 1 for 16 h. Western blot analysis of whole cell extracts demonstrated that TGF␤ 1 decreased p190RhoGAP expression in a dose-dependent manner (Fig. 6A), opposite to that shown in atrial myocytes (Fig. 5, A and B). Immunoprecipitation of whole cell extracts used for the determination of p190RhoGAP expression with Rhotekin followed by Western blot analysis with anti-RhoA antibody demonstrated that TGF␤ 1 increased GTP-bound RhoA with a dose dependence similar to that for the inhibition of p190RhoGAP expression, (Fig. 6A). Analysis of three experiments similar to that in Fig. 6A demonstrated that TGF␤ 1 decreased p190RhoGAP by 29 Ϯ 9.8% (n ϭ 3, p Ͻ 0.01) and 54 Ϯ 9.7% (n ϭ 3, p Ͻ 0.01) and increased GTP-bound RhoA by 2.5 Ϯ 0.4-fold (n ϭ 3, p Ͻ 0.05) and 3.17 Ϯ 0.5-fold (n ϭ 3, p Ͻ 0.05) at 0.5 ng/ml and 2 ng/ml TGF␤ 1 , respectively (Fig. 6B). Western blot analysis of replicate membranes demonstrated that TGF␤ 1 had no effect on either total RhoA or RhoGDI (Fig.  6, A and B).

DISCUSSION
We had previously demonstrated that TGF␤ treatment of atrial myocytes decreased their response to parasympathetic stimulation (9,15). Data presented here demonstrate that this decrease in parasympathetic response is due at least in part to the inhibition of M 2 muscarinic receptor expression by TGF␤ 1 and that this effect is dependent on RhoA. This conclusion is supported by the finding that TGF␤ 1 inhibition of M 2 muscarinic receptor expression and M 2 promoter activity correlates with a decrease in GTP-bound RhoA. Furthermore, TGF␤ 1 inhibition of M 2 promoter activity is mimicked by co-expression of a dominant negative RhoA mutant, whereas co-expres-  The mechanism by which TGF␤ regulates RhoA-dependent signaling is not known. RhoGDI has been shown to regulate RhoA activity by binding GDP-bound RhoA and transporting it from the membrane to the cytoplasm (18). However, we found no change in the levels of RhoGDI in response to TGF␤. Recently, RhoA has been reported to be targeted for degradation by TGF␤ via ubiquitination involving Par6 and Smurf1 (40). However, we found no change in total RhoA after TGF␤ treatment, eliminating this possible mechanism. It has been well established that RhoGAP decreases RhoA activity by binding to GTP-bound RhoA at the membrane and stimulating RhoA GTPase activity, thus decreasing the level of GTP-bound RhoA (18,19). The finding that the TGF␤ 1 -mediated decrease in GTP-bound RhoA is associated with a reciprocal increase in the level of expression of p190RhoGAP suggested that TGF␤ 1 might regulate RhoA activity in atrial myocytes via an effect on p190RhoGAP expression. The finding that overexpression of wild type p190RhoGAP mimics the effect of TGF␤ 1 on M 2 promoter activity, whereas overexpres-sion of a dominant negative p190RhoGAP reverses the effect of TGF␤ 1 on M 2 promoter activity, further supports this conclusion.
Several studies have demonstrated that increased expression of p190RhoGAP or increased p190RhoGAP activity is associated with decreased RhoA activity. Arthur and Burridge (22) demonstrated that overexpression of p190RhoGAP decreased the ratio of GTP/GDP-bound RhoA and stimulated cell spreading and cell migration by promoting membrane protrusion. Wolf et al. (21) demonstrated that transfection of oligodendroglioma cells with p190RhoGAP resulted in a decrease in GTPbound RhoA and a block in proliferation similar to that seen in cells treated with the Rho kinase inhibitor Y27632. Although our study is the first demonstration that cytokine signaling regulates the expression of p190RhoGAP, the conclusion that p190RhoGAP expression might be regulatable and that changes in p190RhoGAP activity might regulate RhoA activity is further supported by the finding that ouabain treatment of Madin-Darby canine kidney cells resulted in a decrease in GTPbound RhoA accompanied by cell detachment and an increase in the level of p190RhoGAP expression (41). Epidermal growth factor stimulation of c-Src-dependent phosphorylation of p190RhoGAP increases the binding of a constitutively activated RhoA to p190RhoGAP, resulting in a decrease in RhoA signaling (38). Finally, integrin signaling and cadherin expres- C, cells were transfected with the M 2 promoter luciferase reporter alone or the M 2 promoter-reporter plus 100 ng of a construct expressing either WT-p190RhoGAP or a p190RhoGAP R1283A GAP-deficient mutant and incubated for 24 h followed by the addition of 5 ng/ml TGF␤ 1 or vehicle and incubation continued for 24 h. Cells were harvested, and promoter activity was determined as in Fig. 2A (n ϭ 5, **, p Ͻ 0.01). D, experiment was carried out as in panel C except that cells were co-transfected with 0, 100, and 150 ng of the p190RhoGAP R1283A construct, respectively, and incubated with either vehicle or 5 ng/ml TGF␤ 1 as indicated (n ϭ 4). **, p Ͻ 0.01 when compared with TGF␤ 1 without p190RhoGAP R1283A expression. sion have been shown to stimulate the phosphorylation of p190RhoGAP and decrease the ratio of GTP/GDP-bound RhoA (42).
Comparison of the effects of TGF␤ on GTP-bound RhoA and the expression of p190RhoGAP in atrial myocytes and Mv1Lu cells suggests the conclusion that the regulation of p190RhoGAP expression might regulate RhoA activity and is consistent with the hypothesis that differences in the regulation of p190RhoGAP expression might account for these pleiotrophic effects of TGF␤ signaling on RhoA activity. TGF␤ has been shown to induce epithelial to mesenchymal transdifferentiation in Mv1Lu via stimulation of RhoA activity (17). TGF␤ stimulation of RhoA activity has been shown to be associated with an early phase peaking at 15 min and a late phase peaking at 12-16 h associated with the formation of stress fibers and actin filaments characteristic of epithelial to mesenchymal transdifferentiation (43). We demonstrated that the increase in GTPbound RhoA in Mv1Lu was associated with a reciprocal decrease in the expression of p190RhoGAP and a late phase. Both the increase in GTP-bound RhoA and the decrease in p190RhoGAP expression demonstrated a similar dose dependence on TGF␤ 1 . Thus in two different cell systems, TGF␤ 1 signaling was mediated via opposing effects on RhoA activity. Taken together with the studies cited above demonstrating that changes in the expression and activity of p190RhoGAP regulated RhoA activity, the demonstration of a reciprocal relationship between the expression of p190RhoGAP and RhoA activity in atrial myocytes and Mv1Lu cells suggests the conclusion that TGF␤ regulation of RhoA via the control of p190RhoGAP expression might constitute a cell type-specific mechanism for RhoA-dependent TGF␤ signaling (Fig. 7).
The role of small GTP-binding proteins in the pleiotrophic effects of TGF␤ is also supported by studies that demonstrate that TGF␤ regulation of the activity of Ras is specific either for cell type or for the stage of embryonic development. Thus TGF␤ increased the level of GTP-bound Ras in untransformed epithelial cells, and dominant negative mutant Ras inhibited TGF␤ stimulation of Smad1 phosphorylation (13). However, TGF␤ suppressed the transformed phenotype in Ras transformed hepatocytes (14). Our data demonstrate that TGF␤ 1 exerts similar cell type-specific pleiotrophic effects on RhoA activity.
RhoA activity is also regulated by RhoA GEFs, which regulate the rate of release of GDP from GDP-bound RhoA (19,20). Studies presented here do not rule out the possibility that changes in RhoA activity in atrial myocytes and Mv1Lu are in part due to TGF␤ regulation of RhoA GEFs.
Our data demonstrating that TGF␤ attenuates the parasympathetic response in atrial myocytes in association with decreased expression of the M 2 muscarinic receptor and G␣ i2 (9) and the role of RhoA and p190RhoGAP in regulating TGF␤ signaling may have important implications in our understanding of cardiovascular disease. Parasympathetic dysfunction in diabetic autonomic neuropathy has been related to the increased incidence of life-threatening arrhythmias and sudden death in the diabetic population (44 -46). Diabetes has been associated with increased TGF␤ signaling in peripheral tissues (47,48). Based on our data demonstrating a decreased parasympathetic response in atrial myocytes treated with TGF␤ (15), parasympathetic dysfunction in diabetes might be secondary to increased TGF␤ signaling. In addition, TGF␤ signaling has been implicated in ventricular remodeling following myocardial infarction, in heart failure, and as a complication of hypertension (49 -51). Increased TGF␤ signaling has also been implicated in the increased interstitial fibrosis associated with the cardiomyopathy and renal dysfunction in diabetes mellitus (44,47). Therefore, understanding the role of specific signaling molecules downstream of TGF␤ may provide novel therapeutic opportunities for the treatment of cardiovascular disease.