G βγβγβγβγ mediate differentiation of vascular smooth muscle cells

Abstract Proliferation and subsequent dedifferentiation of vascular smooth muscle (VSM) cells contribute to the pathogenesis of atherosclerosis and postangioplastic restenosis. The dedifferentiation of VSM cells in vivo or in cell culture is characterized by a loss of contractile proteins such as smooth muscle-specific α-actin and myosin heavy chain (SM-MHC). Serum increased the expression of contractile proteins in neonatal rat VSM cells, indicating a redifferentiation process. RNase protection assays defined thrombin as a serum component that increases the abundance of SM-MHC transcripts. Additionally, serum and thrombin transiently elevated cytosolic Ca2+ concentrations, led to a biphasic extracellular signal-regulated kinase (ERK) phosphorylation, up-regulated a transfected SM-MHC promoter construct, and induced expression of the contractile proteins SM-MHC and α-actin. Pertussis toxin, N17-Ras/Raf, and PD98059 prevented both the serum- and thrombin-induced second phase ERK phosphorylation and SM-MHC promoter activation. Constitutively active Gαq, Gαi, Gα12, and Gα13 failed to up-regulate SM-MHC transcription, whereas Gβγ concentration-dependently increased the SM-MHC promoter activity. Furthermore, the Gβγ scavenger β-adrenergic receptor kinase 1 C-terminal peptide abolished the serum-mediated differentiation. We conclude that receptor-mediated differentiation of VSM cells requires Gβγ and an intact Ras/Raf/MEK/ERK signaling.


INTRODUCTION
Fully differentiated, contractile vascular smooth muscle (VSM) cells are major determinants of blood pressure and flow. In chronic vascular diseases such as hypertension and atherosclerosis, VSM cells proliferate and undergo a phenotypic modulation characterized by local matrix degradation and a loss of contractile function (1). In vivo, dedifferentiated VSM cells can gradually revert towards a more contractile phenotype (2). Interest in the underlying mechanisms and participating signal transduction pathways leading to altered phenotypes of VSM cells has led to extensive study of the VSM cell phenotype both in vivo and in vitro (for review see 3).
Differentiated VSM cells are characterized by high expression levels of contractile proteins such as smooth muscle α-actin (SM-α-actin) and smooth muscle myosin heavy chain (SM-MHC) (4).
The expression of SM-MHC isoforms SM-1 and SM-2 is restricted to smooth muscle cells (5,6) and is downregulated in proliferating cells (7). High expression levels of SM-1/2, therefore, are valuable markers for the differentiated phenotype of VSM cells. Similar to pathological proliferation during vascular disease, VSM cells downregulate SM-1/2 expression in primary culture. Although cultured VSM cells initially retain SM-1/2 expression when cultured on laminin or under serum-free conditions, they undergo morphological changes towards a dedifferentiated phenotype within a few days (8). Patterns of gene expression similar to those in cultured VSM cells from neonatal rats have been observed in neointimal cells within injured vessels (9). Neonatal VSM cells can, therefore, provide an in vitro model for studying phenotypic modulation processes in vascular disease. The mechanisms and signaling pathways that induce a phenotypic modulation towards the contractile phenotype of VSM cells are still largely elusive. 4 differentiate in response to mechanical strain (11). These findings were recently corroborated by applying mechanical forces to cultured whole vessels (12). Interestingly, phenotypic modulation of VSM cells depends on the activation of mitogen activated protein kinases (MAPK) in both experimental settings.
In many cellular systems, the receptor-mediated proliferation and differentiation involves the extracellular signal regulated kinase (ERK) subfamily of MAPK (13,14). ERKs are part of a multi-kinase module through which a variety of extracellular stimuli (growth factors, differentiation signals and cellular stress) are transmitted into the cell (15). Receptor tyrosine kinases (RTKs), upon autophosphorylation and activation of adaptor proteins, recruite Ras and subsequently engage the Raf/MEK/ERK cascade. Alternatively, G protein-coupled receptors (GPCRs) have been shown to stimulate ERKs via the G i -, G q -or G 12/13 -subfamilies of heterotrimeric G proteins. In addition, transactivation of receptor tyrosine kinases has been demonstrated to participate in signaling from GPCRs to ERKs (16)(17)(18). 6 plasmid DNA and 10 µl/well Superfect transfection reagent (Quiagen) for 5 h. After 48 h, cell lysates were prepared using the CAT Enzyme Assay System (Promega). CAT activities were normalized to the protein concentration of each sample as measured by the BCA assay.
Transfection of a promoterless CAT construct served as a base-line indicator, allowing all other promoter constructs to be expressed relative to promoterless activity. All CAT activities (means ± SEM) represent at least three independent transfection experiments with each setting tested in triplicate per experiment. Cotransfection of a viral promoter/β-Gal or LacZ construct to control for transfection efficiency was discontinued since variations in transfection efficiency among independent experimental samples are small (≤ 10%). Furthermore, it has been shown that such constructs interfere with SM-specific promoters, presumably due to competition for common transcription factors (21).

RNase protection assay
RNA isolation, generation of DNA templates and hybridization conditions have been described previously (11). In brief, 10 µg of total RNA were hybridized with a radiolabled probe covering the alternatively spliced C-terminal exons of SM-1 and SM-2 variants of rat SM-MHC. After overnight incubation at 42°C, non-hybridized fragments were digested with a diluted RNase A/T1 mixture. The remaining protected fragments (380 nt for SM-2 and 261 nt for SM-1, respectively) were separated by denaturing (8% urea) polyacrylamide gel electrophoresis and exposed to Amersham Hyperfilm at -80 °C for 2-24 h. Bands were excised and counted in a liquid scintillation counter. Equal loading was controlled by hybridization of a second aliquot with a rat glutaraldehyde-phosphate-dehydrogenase (GAPDH)-radiolabled probe. 8 recorded with a 12-bit CCD-camera. After correction for background signals, intracellular Ca 2+ concentrations were calculated as described (22). R max , R min and F380 min / F380 max were determined in fura 2-loaded cells equilibrated for 3 h in HBS supplemented with 1 µM ionomycin and either

Regulation of SM-MHC expression by serum components
The presence of serum is essential to grow VSM cells in primary cell culture. In addition to its mitogenic properties, we observed that fetal calf serum enhanced the expression of contractile proteins in neonatal rat VSM cells. Dual staining of proliferating cell nuclear antigen (PCNA) and smooth muscle-specific α-actin or SM-MHC revealed that after short-term (24 h) exposure to serum, PCNA expression was positive, whereas SM-α-actin was poorly detectable (Fig. 1A). VSM cells as compared to serum-starved controls (QM). The relatively low expression levels of SM-1/SM-2 in serum-starved VSM cells allowed us to study the effect of single compounds on SM-1/SM-2 expression. Both, 1 µM angiotensin II and 10 ng/ml PDGF-BB did not alter the SM-1/SM-2 expression significantly, whereas treatment with 10 ng/ml TGF-β resulted in a further reduction of SM-1/SM-2 steady-state expression (Fig. 2). In contrast, 1 U/ml thrombin increased SM-1/SM-2 expression by the 10 ± 0.9-fold. The substantial increase in SM-1/SM-2 steady-state expression most likely represents an upregulation of the transcriptional activity, although changes in RNA stability and/or turnover cannot be ruled out.
Since the inhomogeneous response to vasoactive peptides may rely on the presence or absence of the corresponding receptors in cultured VSM cell preparations, functional coupling of receptors in

Serum, thrombin and TRAP induce biphasic ERK phosphorylation
Since both proliferative and differentiating signals can be transmitted via ERK1/2, depending on the cellular context and the transient or sustained character of their activation, we studied the kinetics of ERK phosphorylation in VSM cells. A rapid and transient phosphorylation of ERK1/2 was elicited by serum, PDGF-BB, EGF, thrombin, TRAP, LPA, and -to a lower extent -by angiotensin II. The early ERK1/2 phosphorylation was maximal after 3-5 min for all agonists.
Only serum, thrombin, TRAP and LPA elicited a delayed second phase ERK phosphorylation (Fig. 4A). The second phase ERK phosphorylation appeared approximately 45 min after agonist application and continuously rose for another 2 h. In contrast, a delayed ERK phosphorylation was absent in response to PDGF-BB, EGF, or angiotensin II. The weak and monophasic ERK1/2 activation by angiotensin II may rely on low AT 1 receptor expression in our VSM cell preparation

Activation of Ras/Raf/MEK/ERK is a prerequisite for SM-MHC promoter activation by serum and thrombin
Enhanced SM-1/SM-2 expression in response to thrombin (Fig. 2) is indicative of a G protein- Additionally, the MEK inhibitor PD98059 largely reduced the thrombin-stimulated SM-MHC promoter activity (Fig. 5C). Since PD98059 was dissolved in DMSO, the effect of the solvent on the SM-MHC promoter activity was assessed in parallel. DMSO (up to 0.5%) further increased the thrombin-stimulated CAT activity almost 1.8-fold, an effect that was also blocked by line with its described IC 50 to inhibit ERK1/2 phosphorylation (23). Correspondingly, in VSM cells the serum-or thrombin-mediated ERK1/2 phosphorylation was largely reduced by 5-20 µM

Differentiation of VSM cells requires pertussis-toxin-sensitive G proteins
The transient elevation of [Ca 2+ ] i and biphasic ERK1/2 phosphorylation induced by thrombin could be mimicked with the tethered ligand of the PAR-1 receptor, TRAP. PAR-1 receptors couple to the G i , G q , and G 12/13 -subfamilies of heterotrimeric G proteins (24). The putative involvement of G 12/13 in the regulation of the SM-MHC promoter was tested by overexpressing constitutively active (GTPase-deficient) mutants of Gα 12 (pCIS/Gα 12 Q229L) and Gα 13 (pCIS/Gα 13 Q226L). Both constructs and their combination failed to significantly induce SM-MHC promoter activity over a wide range of transfected cDNA concentrations (data not shown).
The biological activity of these constructs has been previously demonstrated by their ability to induce contraction of VSM cells (25). G q and G i proteins couple to phospholipases C β to release [Ca 2+ ] i from inositol-1,4,5-trisphosphate-sensitive stores. The possible role of the G i -class of heterotrimeric G proteins was assessed by pretreating cells with 200 ng/ml pertussis toxin (PTX) for at least 18 h. Inactivation of G i proteins by this protocol was demonstrated by a more than 80% reduction of [Ca 2+ ] i signals in response to LPA (Fig. 6A). PTX reduced the peak [Ca 2+ ] i after thrombin stimulation by about 40% (Fig. 6A). The partial block of thrombin-induced [Ca 2+ ] i transients in PTX-pretreated VSM cells reflects coupling to both G i and G q/11 . To further test whether the G i -subfamily also participates in the prolonged ERK1/2 activation, serum-starved VSM cells were pretreated with PTX. Subsequent addition of serum, thrombin, and LPA left early ERK1/2 phosphorylation almost unaltered, whereas the second phase of ERK1/2 activation was completely abrogated in PTX-pretreated cells (Fig. 6B).
Since sustained ERK activation may be required for the regulation of transcriptional activity, we  (Fig. 7C).
Finally, the G i protein-dependent redifferentiation in response to thrombin and serum was confirmed by analyzing the expression of contractile proteins in untransfected cells. In whole cell lysates from VSM cells stimulated with thrombin or serum and normalized for protein content, an increased expression of SM-α-actin and of SM-MHC was detected (Fig. 8). When incubated in the continuous presence of PTX, thrombin and serum failed to increase the expression of both contractile proteins (Fig. 8). Hence, these data demonstrate, that Gβγ released from G i proteins link proximal signaling to the Ras/Raf/MEK/ERK-cascade to mediate the in vitro redifferentiation of vascular smooth muscle cells shown in Fig. 1.

DISCUSSION
In this study we describe a receptor-mediated signaling pathway leading to differentiation of vascular smooth muscle cells. The thrombin-induced SM-MHC expression is transmitted via the Ras/Raf signaling cascade and leads to a biphasic temporal pattern of ERK1/2 phosphorylation.  (Reusch, unpublished data), activation of ERKs appears to be both sufficient and necessary for receptormediated differentiation of VSM cells. The ERK pathway regulates two mutually opposing processes, cellular proliferation and differentiation, depending on the duration of activation and the cellular context (15,34,35). Our data indicate that a sustained rather than a short-lived ERK phosphorylation is a requirement for the differentiation of VSM cells. Several other cellular models including megakaryocytes (35), thymocytes (37) and PC-12 cells (38) support the idea that the short versus long-term ERK phosphorylation determines the proliferative or differentiating outcome, respectively.
Multiple upstream signaling pathways link receptor activation to phosphorylation of ERK1/2. In VSM cells, the biphasic kinetic pattern of ERK-phosphorylation in response to serum, thrombin, or LPA suggests that at least two independent pathways control the early and delayed phases of ERK phosphorylation. Considering that strong Ca 2+ -signals result from thrombin stimulation of VSM cells, a Ca 2+ -and PKC-dependent formation of Ras/Raf-1 complexes (39, 40) may engage ERKs. Alternatively, Ca 2+ /calmodulin-dependent activation of Pyk2 (41) and subsequent Srcactivation may target Ras either including (42)(43)(44) or bypassing transactivated receptor tyrosine kinases (45, 46). Since Ca 2+ -ionophores evoked large [Ca 2+ ] i signals, but failed to induce a longlived ERK phosphorylation in VSM cells, an isolate Ca 2+ -elevation was not sufficient to mimick the effects of serum components. In PTX-pretreated VSM-cells stimulated with serum, thrombin, or LPA, the remaining activation of G q/11 and G 12/13 induced an early ERK activation, but failed to generate a sustained phospho-ERK signal. Since PTX pretreatment also abolished the contractile protein expression in response to serum and thrombin, we focussed on signaling pathways that are initiated by either Gα i or Gβγ subunits released from activated G i proteins. The inhibition of adenylyl cyclases by Gα i lowers cAMP concentrations and subsequently protein kinase A activity.
Signaling via the Ras/Raf/MEK/ERK cascade is counterregulated by protein kinase A-dependent phosphorylation and inactivation of Raf-1 (47,48). Indeed, forskolin-treatment further reduced the basal ERK-phosphorylation in serum-starved VSM cells (Reusch, unpublished data). Thus, a disinhibition of Ras-Raf-signaling by further reducing the cAMP concentrations in quiescent cells might result in an increased activity of the c-Raf kinase as has been shown for another cell system (49). However, expression of constitutively active Gα i (Q205L) failed to increase the SM-MHC promoter activity. Most strikingly, coexpression of Gβγ subunits mimicked while the Gβγscavenger βARK1ct attenuated the effects of thrombin or LPA. Within the multiple Gβγ effector systems, Gβγ-sensitive PLC-isoforms, phosphoinositide-3-kinases, or further unknown effectors bear the potential to feed into the Ras/Raf/MEK/ERK cascade. Although the molecular mechanisms are currently poorly defined, a growing body of evidence points to a role of Gβγ in initiating the assembly of a multi-protein-complex including β-arrestin and c-Src in clathrincoated pits (50,51). Within these microdomains, a ligand-independent transactivation of receptor tyrosine kinases such as the EGF-receptor may link Gβγ signaling to Ras. Other concepts favour a direct association of Gβγ with Raf-1 (52), or the activation of a Ras-GEF other than Sos1 (52).
Our preliminary data demonstrating that the tyrphostin AG 1478 prevents the thrombin-induced ERK phosphorylation in VSM cells, point to a crucial role of an EGF-receptor transactivation.
Receptors for endogenous vasoconstrictors such as endothelin-1, angiotensin II, and vasopressin or serum components like thrombin or LPA activate the G i -, G q -, and G 12/13 -classes of heterotrimeric G proteins. The Gα q -dependent pathway, via phospholipase C β-catalyzed formation of inositol-1,4,5-trisphosphate, increases [Ca 2+ ] i to activate the calmodulin-dependent myosin light chain kinase. In parallel, activated G 12/13 , via RhoA and Rho-kinase, inhibits a myosin phosphatase (25). Both pathways synergistically control the contraction of VSM cells by increasing the myosin light chain (MLC 20 ) phosphorylation. The additional coupling to G i proteins does not contribute to the acute regulation of contraction (25). It is, therefore, tempting to speculate that Gβγ released from G i proteins function to maintain the contractile phenotype by enhancing the expression of contractile proteins.
In summary, we have defined the G i -component of multiply coupling receptors as a pivotal step in the receptor-mediated expression of contractile proteins. Our data clearly indicate that Gβγsubunits induce a sustained ERK phosphorylation which is critical for the differentiation of VSM cells. In the past, substantial data have been accumulated regarding the serum-factor-dependent promoter regulation of contractile proteins. The addition of our data, demonstrating how receptormediated differentiation signals may be transmitted to the nucleus, may converge to a more 20 clearly defined step-by-step model describing the regulation of contractile protein expression in vascular smooth muscle cells.

FOOTNOTES:
The