On the basis of our previous studies on RhoA signaling in smooth muscle cells (SMC), we hypothesized that RhoA-mediated nuclear translocalization of the myocardin-related transcription factors (MRTFs) was important for regulating SMC phenotype. MRTF-A protein and MRTF-B message were detected in aortic SMC and in many adult mouse organs that contain a large SMC component. Both MRTFs upregulated SMC-specific promoter activity as well as endogenous SM22α expression in multipotential 10T1/2 cells, although to a lesser extent than myocardin. We used enhanced green fluorescent protein (EGFP) fusion proteins to demonstrate that the myocardin factors have dramatically different localization patterns and that the stimulation of SMC-specific transcription by certain RhoA-dependent agonists was likely mediated by increased nuclear translocation of the MRTFs. Importantly, a dominant-negative form of MRTF-A (ΔB1/B2) that traps endogenous MRTFs in the cytoplasm inhibited the SM α-actin, SM22α, and SM myosin heavy chain promoters in SMC and attenuated the effects of sphingosine 1-phosphate and transforming growth factor (TGF)-β on SMC-specific transcription. Our data confirmed the importance of the NH2-terminal RPEL domains for regulating MRTF localization, but our analysis of MRTF-A/myocardin chimeras and myocardin RPEL2 mutations indicated that the myocardin B1/B2 region can override this signal. Gel shift assays demonstrated that myocardin factor activity correlated well with ternary complex formation at the SM α-actin CArGs and that MRTF-serum response factor interactions were partially dependent on CArG sequence. Taken together, our results indicate that the MRTFs regulate SMC-specific gene expression in at least some SMC subtypes and that regulation of MRTF nuclear localization may be important for the effects of selected agonists on SMC phenotype.
- serum response factor
vascular smooth muscle cell (SMC) differentiation is a very important process during vasculogenesis and angiogenesis, and it is recognized that alterations in SMC phenotype play a role in the progression of several prominent cardiovascular disease states, including atherosclerosis, hypertension, and restenosis (33, 38). The identification of the transcription factors involved in this process has been complicated by the lack of terminal differentiation in this cell type and the fact that SMC derive from multiple locations, including local mesoderm, cardiac neural crest, the proepicardial organ, and possibly circulating stem cells (see Ref. 33 for a review). A completely SMC-specific transcription factor has yet to be described, and evidence suggests that SMC differentiation may be regulated by interactions between multiple transcription factors with overlapping expression patterns (4, 5, 11). We and others (14, 19, 24, 26, 27, 29) have shown that the SMC differentiation marker genes are regulated by serum response factor (SRF), a MADS (MCM1, Agamous, Deficiens, SRF) box transcription factor that binds to conserved CArG boxes found in nearly all of the SMC marker gene promoters. Since SRF is ubiquitously expressed and regulates a variety of other muscle-specific genes as well as early response genes c-fos and early growth response factor-1 (Egr-1) (2, 6, 44, 46, 47), it is clear that additional mechanisms are involved.
An important breakthrough in the study of the molecular mechanisms that regulate SMC differentiation was the discovery of the myocardin family of SRF cofactors by Wang et al. (48). The founding member of this family, myocardin, is selectively expressed in the heart and SMC and very powerfully transactivates SMC differentiation marker gene expression by physically interacting with SRF. Importantly, genetic deletion of myocardin resulted in embryonic lethality at embryonic day 10.5 due to failure of SMC differentiation in the mesodermal cells surrounding the descending aorta (22). In more recent studies, Pipes et al. (34) used a chimeric mouse model to demonstrate that myocardin knockout cells could populate the developing aorta, suggesting that myocardin-independent mechanisms are also likely to be important for SMC differentiation.
Two myocardin-related transcription factors, MRTF-A/megakaryoblastic leukemia 1 (MKL-1) and MRTF-B/MKL-2, have been identified that have similar transcriptional activity to myocardin. These factors are thought to be expressed more widely, and their role in regulating cell type-specific gene regulation is much less clear. MKL-1 message is expressed strongly in a variety of human tissues that have a large SMC component, including the aorta and bladder, and we and others have detected MRTF-A/MKL-1 protein in multiple SMC lines, including primary rat aortic SMC, A7R5, A10, and PAC-1 as well as mouse embryonic stem cells and multipotential 10T1/2 cells (9, 23). Overexpression of MRTF-A/MKL-1 strongly upregulated several muscle-specific promoters in a variety of cell types and, importantly, was sufficient to activate endogenous expression of SMC differentiation marker genes in embryonic stem cells. MRTF-B was originally shown to have little effect on CArG-dependent transcription (49). However, its human homologue, MKL-2, strongly activated the atrial natriuretic factor, SM22α, and SM α-actin promoters in HeLa cells (39). Since these proteins share >80% homology, the reason for this discrepancy is currently unknown. Two separate groups have recently shown that genetic disruption of MRTF-B led to a lethal defect in pharyngeal arch remodeling and that this phenotype was accompanied by a failure of SMC differentiation of the cardiac neural crest cells that populate the cardiac outflow tract (18, 32). Interestingly, both Li et al. (21) and Sun et al. (45) have shown that MRTF-A knockout females have a nursing defect that is accompanied by a loss of SMC differentiation marker gene expression that normally occurs in the myoepithelial layer of the mammary gland during lactation.
Identification of the mechanisms that regulate the myocardin transcription factors will be very important for our understanding of their role in cell type-specific gene regulation. Miralles et al. (30) were the first to demonstrate that the activity of MRTF-A was regulated by RhoA-dependent signaling. These authors demonstrated in NIH 3T3 cells that MRTF-A resided nearly exclusively in the cytoplasm in serum-starved cells, that MRTF-A translocated to the nucleus upon serum stimulation, and that this process was regulated by RhoA-dependent changes in actin polymerization. The RPEL domains in the NH2 terminus of MRTF-A were shown to bind to G-actin (35), and this interaction was shown to be important for retaining MRTF-A in the cytoplasm. A recent study indicated that MRTF-B localization may be regulated by the same mechanism (17). Interestingly, myocardin is constitutively nuclear, and it has been suggested that lack of conservation in the second RPEL domain of myocardin may inhibit cytoplasmic retention (30). Several groups have reported somewhat contrasting results on MRTF-A localization (9, 39). For example, Du et al. (9) reported that in primary rat aortic SMC, MRTF-A/MKL-1 was constitutively nuclear in serum-starved cells or in the presence of dominant-negative RhoA. In addition, Selvaraj and Prywes (39) did not observe nuclear translocation of MKL-1 upon serum stimulation of NIH 3T3 and HeLa cells (39). The reason for these discrepancies is unknown.
We have previously shown that RhoA is an important determinant of SMC differentiation marker gene expression and that MRTF-A was required for the upregulation of SMC-specific transcription observed upon treatment of aortic SMC with the strong RhoA agonist sphingosine 1-phosphate (S1P) (23, 25). The goals of the present study were to determine the contributions of the MRTFs to SMC-specific gene regulation, to test whether regulation of MRTF nuclear localization is an important signaling mechanism for controlling SMC-specific transcription, and to further characterize the differences between the three myocardin transcription factors.
MATERIALS AND METHODS
Plasmids and proteins.
Myocardin, MRTF-A and MRTF-B plasmids were a generous gift of D.-Z. Wang (University of North Carolina, Chapel Hill, NC). MRTF deletions lacking the NH2-terminal RPEL domains were created by PCR. The dominant-negative ΔB1/B2 MRTF-A was a generous gift of R. Treisman (Cancer UK, London) and has been described previously (30). MRTF-A/myocardin chimeras that fused NH2-terminal fragments of MRTF-A to COOH-terminal fragments of myocardin were made by PCR. An exogenous XhoI restriction site (that codes LE) was inserted at the MRTF-A/myocardin junction to facilitate cloning, and junction sites were placed in regions of low homology (more details on the MRTF-A/myocardin chimeras can be found in Fig. 3 and are available on request). The myocardin double mutation S72P, S76E that restored RPEL2 was made using the QuikChange method (Stratagene). All myocardin factors were subcloned into a flag-tagged pcDNA3.1 and/or an enhanced green fluorescent protein (EGFP) expression vector.
Cell culture, transient transfections, and reporter assays.
SMC from rat thoracic aorta were isolated, cultured, and transfected as previously described (24, 41). In short, cells were maintained in 24-well plates in 10% serum and were transfected 24 h after a plating at 70–80% confluency using the transfection reagent Superfect (Qiagen), as per protocol. The SM22α promoter (from −450 to +88), SM α-actin promoters (from −2,560 to +2,784), SM myosin heavy chain promoter (from −4,200 to +11,600), and c-fos promoter (from −356 to +109) used in this study have been previously described (20, 26, 28). In some experiments, myocardin, MRTF-A, MRTF-B, or variants thereof were cotransfected along with the promoter-luciferase constructs.
Before the agonist treatments, SMC were placed in serum-free media for 24 h, whereas 10T1/2 cells were placed in 0.2% charcoal-treated serum (to remove serum lipids). Cells were treated with S1P (1 μM), 10% serum, transforming growth factor (TGF)-β (1 ng/ml), and PDGF-BB (20 ng/ml), and luciferase assays were performed after 24 h. In some experiments, the Rho-kinase inhibitor Y-27632 (10 μM) was added 15 min before the addition of agonist. The S1P used in these experiments was obtained from Matreya and was maintained in 4 mg/ml fatty acid free BSA, which was used as a vehicle control. Relative promoter activities are expressed as the means ± SE computed from a set of at least three separate transfection experiments. We did not cotransfect a viral promoter/LacZ construct as a control for transfection efficiency since we have previously shown that such constructs exhibit unknown and variable squelching effects on the SM-specific promoters, presumably due to competition for common transcription factors (41). Moreover, we have found that inclusion of such controls is unnecessary in that variations in transfection efficiency between independent experimental samples are routinely very small (<10%) (41).
Analysis of MRTF expression.
Adult C57/Black6 mice were killed, and blood was removed by perfusing phosphate-buffered saline through the vasculature via a puncture of the left ventricle. Tissues were excised and homogenized in RIPA buffer plus inhibitors by sonication. Lysates were clarified by centrifugation, and protein concentrations were determined by using bicinchoninic acid protein assay (Pierce). One-hundred fifty micrograms of total protein from each tissue lysate were run on an 8% SDS polyacrylamide gel and subsequently transferred to nitrocellulose. MRTF-A was detected by using MRTF-A antiserum generously provided by R. Treisman (Cancer Research UK, London, UK). In vitro translated MRTF-A was prepared by using TNT T7 Coupled Reticulocyte Lysate System (Promega) and was run along side the lysates as a positive control. Semiquantitative PCR was used to measure MRTF-B expression. In brief, RNA was prepared from cell and tissues as above using TRIzol reagent (Invitrogen) and quantified by Ribogreen Assay (Molecular Probes). cDNA was generated with the iScript cDNA synthesis kit (Bio-Rad) using 1 μg of RNA per manufacturer's protocol. The following exon-spanning primers were used for amplification reactions: MRTF-B, 5′-atgaggaagccatcaagcag-3′ and 5′-atctgctgactgtgcaca-3′; GAPDH, 5′-atgggtgtgaaccacgaagaa−3′ and 5′-ggcatggactgtggtcatga-3′.
Visualization of myocardin factor localization.
The myocardin factors were subcloned into the EGFP-C3 vector (Clontech). SMC and 10T1/2 cells were transfected with fusion protein plasmids as described above, maintained in 10% serum overnight, and then held in serum-free media for 24 h. After addition of agonist, myocardin factor localization was monitored in real time on an inverted fluorescent microscope, and images were taken at 5-min intervals for 80 min using a Spot digital camera. Micrographs were converted into time-lapse movies using Isee imaging software and QuickTime. To quantify localization in the entire cell population, cells were fixed after 80 min of treatment in 4% paraformaldehyde, and localization was scored into three separate categories: nuclear, diffuse, and cytoplasmic.
Gel shift analyses.
SRF and flag-tagged myocardin factors were translated in vitro using the Promega T7 TnT kit. Binding reactions contained 1 μl of SRF, 2 μl of myocardin factor, 20,000 counts/min of a 32P-labeled oligonucleotide probe containing CArGA, CArGB, or the intronic CArG from the rat SM α-actin promoter, and 0.20 μg dIdC in binding buffer (10 mM Tris, pH 7.5, 50 mM NaCl, 100 mM KCl, 1 mM DTT, 1 mM EDTA, and 5% glycerol). Reactions were incubated for 30 min before loading on nondenaturing 4% polyacrylamide gel that was prerun at 170 V for 1 h. Electrophoresis was performed at 170 V in 0.25× TBE (45 mM Tris-borate and 1 mM EDTA). Gels were dried and exposed to film for 24–72 h at −80°C. For supershift studies, 1 μl of M2 Flag antibody (Sigma) was added after 20 min of incubation.
All animal usage procedures described in these studies conformed to National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee at University of North Carolina-Chapel Hill.
MRTF-A and MRTF-B were expressed in aortic SMC and many SM-containing tissues.
Previous studies (49) have demonstrated that the MRTFs are expressed more widely than myocardin. However, MRTF-A message expression is extremely high in the human aorta and fairly high in bladder, stomach, intestine, and in many SMC or SMC-like cell lines including primary rat aortic SMC, A10, PAC-1, and 10T1/2 (9), suggesting that it may have an important role in SMC. To get a better idea of MRTF-A protein expression in adult mouse SMC-containing tissues, we performed Western analysis. Results shown in Fig. 1 demonstrate that MRTF-A levels were high in aorta, bladder, lung, and uterus and in rat aortic and A7r5 SMC cultures. The size of immunoreactive bands varied slightly between tissues, indicating that MRTF-A may be posttranslationally modified or processed in a cell type-specific manner. Because of the lack of a suitable MRTF-B antibody, we used semiquantitative RT-PCR to measure MRTF-B expression in adult mouse SMC-containing tissues. Figure 1B demonstrates that MRTF-B message was more evenly distributed but was relatively high in aorta, lung, stomach, liver, and rat aortic SMC cultures.
The MRTFs regulated SMC differentiation marker gene expression.
Myocardin has been shown to regulate SMC differentiation (10, 22, 52), but the role played by the MRTFs in this process is less clear. To directly compare activation of SMC-specific gene expression by all three myocardin factors, we expressed them in multipotential 10T1/2 cells. This cell line is a very useful and consistent model for studying the regulation of SMC-specific transcription because many endogenous SMC-specific differentiation marker genes, including SM α-actin, SM22α, and calponin, can be induced by agonists such as TGF-β or S1P (15, 23). As shown in Fig. 2, all three of the myocardin factors transactivated the SM α-actin and SM22α promoters and stimulated the endogenous expression of SM22α. However, transactivation by myocardin was significantly greater than that induced by MRTF-B and in most instances was greater than that induced by MRTF-A, even though these transcription factors were expressed at similar levels. Our data also support previous studies that demonstrated that the NH2-terminal actin-binding region of both MRTF-A and MRTF-B had inhibitory effects on their activities.
MRTF activity was regulated by nuclear localization.
MRTF-A nuclear localization has been shown to be regulated by RhoA signaling (23, 30), and we hypothesize that changes in MRTF nuclear localization may help explain the effects of certain environmental cues on SMC-specific transcription. To begin to test this, we constructed EGFP fusion proteins so that we could monitor myocardin factor localization in real time. Fusion protein expression was confirmed by Western blot using an anti-EGFP antibody, and results from cotransfection experiments demonstrated that these fusion proteins significantly activated the SM α-actin promoter, suggesting that the EGFP moiety did not dramatically disrupt myocardin protein function (data not shown). Initially, 10T1/2 cells were serum starved for 24 h, fixed, and scored for localization into three separate categories: nuclear, diffuse, and cytoplasmic. Myocardin localization was constitutively nuclear even after prolonged serum deprivation, whereas MRTF-B localization was nearly completely cytoplasmic under the same conditions (Fig. 3). Interestingly, MRTF-A could localize to the nucleus or cytoplasm but was often found in a more diffuse pattern. In addition, neighboring cells with dramatically different MRTF-A localization patterns were frequently observed. Since it has been shown that RhoA activity is downregulated in confluent SMC and other cell types (31, 54), we also tested whether cell density could affect MRTF-A localization. As shown in Fig. 3B, the percentage of cells containing MRTF-A exclusively in the nucleus was significantly inhibited in confluent cells versus cells that were 50–70% confluent. Because these localization patterns correlated fairly well with the relative activities of the myocardin factors (Fig. 2), we tested whether serum starvation had differential effects on myocardin and MRTF-A activity. Results shown in Fig. 3C demonstrate that transactivation by myocardin was unaffected by serum withdrawal. In contrast, MRTF-A activity was reduced by ∼70%, a result in excellent agreement with previous studies (30).
MRTF localization was regulated by specific agonists.
We also monitored the localization of MRTF-A and MRTF-B in real time following treatment with several different agonists that have been shown to have variable effects on SMC differentiation marker gene expression, including 10% serum, S1P, TGF-β, and PDGF-BB. In cells that contained MRTF-A in the cytoplasm, stimulation of cells with 10% serum or S1P (both strong activators of RhoA signaling) caused MRTF-A to translocate to the nucleus within a time span of ∼45–60 min (see Fig. 4A and Quicktime movies in the online version of this article, which contains supplemental data). To better quantify the effects of these agonists, we fixed cells after 80 min and scored for localization. Results shown in Fig. 4B demonstrate that serum or S1P treatment significantly increased the number of cells that exhibited nuclear localization of the MRTFs and significantly decreased the number of cells exhibiting cytoplasmic or diffuse localization. Pretreatment of cells with the Rho-kinase inhibitor Y-27632 completely inhibited the effects of S1P on nuclear translocalization of MRTF-A. Somewhat surprisingly, Y-27632 did not seem to cause MRTF-A that was already nuclear (∼20% of cells) to translocate to the cytoplasm during the time frame of these studies, suggesting that nuclear retention and/or export is regulated by a separate mechanism. Treatment of cells with TGF-β slightly increased MRTF-A nuclear localization but had no effect on MRTF-B. PDGF-BB had no effect on localization of either of the MRTFs. Parallel transfection experiments demonstrated that the effects of S1P, serum, and PDGF-BB on MRTF localization correlated well with their abilities to stimulate SM α-actin promoter activity in this model system (Fig. 4C). In contrast, although TGF-β had only minor effects on MRTF-A nuclear translocation at 80 min, it strongly stimulated SM α-actin promoter activity. Analysis of MRTF localization in TGF-β-treated cells at 3, 6, and 24 h demonstrated no further changes in MRTF localization.
As shown in Fig. 5, MRTF nuclear localization in SMC was slightly different. Under serum-starved conditions, a significantly greater proportion of SMC contained MRTFs in the nucleus or in the diffuse pattern instead of exclusively in the cytoplasm. Also, S1P- and serum-induced nuclear translocation occurred much more quickly (∼10 min vs. ∼60 min) (Fig. 5A and supplemental data).
MRTF nuclear translocation was required for SMC-specific transcription.
Results presented so far indicate that MRTF-A and MRTF-B are expressed in aortic SMC, can activate SMC-specific transcription when overexpressed, and show increased nuclear localization upon stimulation with S1P or serum. To test whether nuclear localization is important for SMC differentiation marker gene expression in SMC, we used a dominant-negative generated by Miralles et al. (30) that contains deletions to two NH2-terminal basic domains that were shown to be required for nuclear localization of MRTF-A (ΔB1/B2). These authors showed that ΔB1/B2 MRTF-A trapped endogenous MRTF-A in the cytoplasm through dimerization, thus inhibiting its activity as a transcription factor (30). Indeed, ΔB1/B2 MRTF-A acted as a dominant-negative in our model system, attenuating, in a dose-dependent manner, the increase in SM α-actin activity mediated by overexpression of wild-type MRTF-A (Fig. 6A). Importantly, ΔB1/B2 MRTF-A inhibited the activity of the SM22α, SM α-actin, and SM myosin heavy chain promoters in SMC, indicating that MRTF activity is important for regulating SMC differentiation marker gene expression (Fig. 6B). ΔB1/B2 MRTF-A also inhibited the upregulation of SM22α and SM α-actin promoter activity in 10T1/2 cells treated with S1P or TGF-β (Fig. 6, C and D). Interestingly, ΔB1/B2 significantly augmented activation of the c-fos promoter by S1P and TGF-β in this model (Fig. 6E). Taken together, these data indicate that nuclear translocation of the MRTFs regulates SMC-specific transcription and may serve as an important link between the extrinsic cues that regulate SMC function and the transcriptional machinery that ultimately determines SMC phenotype.
Myocardin factor nuclear localization was regulated by NH2 terminus-dependent and NH2 terminus-independent mechanisms.
Results shown in Fig. 2 demonstrate that the NH2 terminal actin-binding domains of MRTF-A and MRTF-B inhibit their activity. Studies have shown that the NH2-terminal MRTF RPEL motifs bind nonpolymerized G-actin to trap the MRTFs in the cytoplasm and that RhoA-dependent actin polymerization reduces the G-actin pool to release this inhibitory mechanism (35). Interestingly, although the removal of the NH2-terminal RPEL domains increased MRTF activity, the NH2-terminally truncated MRTFs (especially ΔN MRTF-B) were still less active than was myocardin, suggesting that additional regulatory mechanisms may affect their nuclear localization or activity. Further supporting this idea was our finding that ΔN MRTF-B did not localize exclusively to the nucleus under serum-free conditions and was found in a diffuse pattern in ∼50% of the cells examined.
To further identify regions in the myocardin factors that govern their nuclear localization, we made a series of chimeric molecules that replaced NH2-terminal portions of myocardin with those of MRTF-A. The myocardin used for these experiments was the 935-amino acid form that contains the entire NH2 terminus. Although the myocardin NH2 terminal region is highly homologous to that of the MRTFs, this form of myocardin is still constitutively nuclear. The chimeric constructs that were generated are depicted in Fig. 7A along with the percentage of cells that showed nuclear localization for each chimera. Interestingly, MRTF-A sequences up to the second RPEL domain did not significantly affect localization of myocardin, whereas replacement with an MRTF-A sequence that also included the second basic (B2) and third RPEL domains decreased nuclear localization to 57%. Another large decrease in nuclear localization (to 33%) was seen upon inclusion of an MRTF-A sequence that contained the region containing basic domain 1 (B1). Further inclusion of COOH-terminal MRTF-A sequences gradually reduced chimera localization to that of MRTF-A (20%).
It has been suggested that variations in myocardin that disrupt the actin-binding RPEL motif in RPEL domain 2 may be responsible for the constitutive nuclear localization of full length myocardin. To directly test this hypothesis, we mutated the divergent sequences in RPEL2 back to a consensus RPEL domain and monitored nuclear localization and transcriptional activity of the mutated protein (see Fig. 7B). This myocardin RPEL2 mutant (MC-R2) localized to the nucleus in 100% of the cells examined and had transcriptional activity that was identical to that of wild-type myocardin (data not shown).
Differential SRF binding also regulates myocardin factor activity.
Our results suggest that differential nuclear localization explains, at least in part, the differences in myocardin factor activity observed in these studies. However, once in the nucleus, myocardin factor activity may also be regulated by differential binding to SRF. To analyze myocardin factor binding to SRF, we performed electromobility shift assays. We also wanted to test whether variations in SRF binding contributed to interactions with specific myocardin factors. Thus, we used the three SM α-actin CArG elements (A, B, and intronic) that have dramatically different abilities to bind SRF due to G/C substitutions in their A/T rich regions. It is important to note that we used the ΔN versions of the MRTFs in this assay because actin binding to the full length MRTFs interferes with ternary complex formation in gel shift assays (30, 39) (and data not shown). Results shown in Fig. 8 demonstrate that SRF binds to all three CArGs with varying affinity (intronic>B>>A). All three myocardin factors formed higher-order complexes on the intronic CArG and could be supershifted with anti-flag antibody (Fig. 8, lanes 13–15). However, even though these SRF cofactors were present at equal amounts (Fig. 8, inset), myocardin binding to SRF was stronger than ΔN-MRTF-A and much stronger than ΔN-MRTF-B. There were also differences in complex formation that were CArG specific. For example, the intensity of the MRTF-A-containing ternary complex (relative to SRF binding alone) was much less for the weaker CArGs than for the strong intronic CArG, a difference that was not apparent with myocardin (compare lane 7 to 11 and lane 6 to 10 in Fig. 8).
Extensive evidence indicates that myocardin is an important regulator of SMC differentiation. Although the role played by the MRTFs is less clear, several lines of evidence from the present study support their involvement in regulating SMC phenotype. First, both MRTFs are expressed strongly in isolated SMC and in a variety of adult mouse tissues that contain a significant SMC component. Although these studies obviously lack sufficient resolution to determine whether the SMC within these organs express MRTF-A or MRTF-B, our results support a more in-depth examination of MRTF expression by in situ or immunohistochemistry. Second, both MRTFs upregulated SMC-specific promoter activity as well as endogenous SMC differentiation marker gene expression in 10T1/2 cells. 10T1/2 cells do not express myocardin (52), indicating that either of the MRTFs was sufficient for this response. Third, a dominant-negative form of MRTF-A (ΔB1/B2) significantly attenuated SMC-specific transcription in SMC. We and others have previously used dominant-negative myocardin family variants that lack the transactivation domain to inhibit SMC-specific promoter activity. Importantly, however, these variants inhibit all three of the myocardin factors, making it difficult to interpret experiments in SMC that express multiple members of this family (9). In the present study, we used ΔB1/B2 MRTF-A to trap the endogenous MRTFs in the cytoplasm without affecting myocardin, which is constitutively nuclear (48). Taken together, these results indicate that the MRTFs regulate SMC-specific transcription in SMC, perhaps in concert with myocardin. The observations that myocardin and MRTF-A can associate directly through conserved leucine zipper motifs and that these motifs are required for full activity of these transcription factors support this idea (9, 51).
The phenotypes of the myocardin family member knockouts indicate that each member has essential nonredundant functions in the regulation of SMC differentiation marker gene expression (18, 21, 22, 32, 45). The early lethality associated with the myocardin and MRTF-B knockouts as well as the potential redundancy between these very similar transcription factors has made it difficult to determine whether the MRTFs are important for regulating SMC differentiation in the SMC subtypes where they are expressed. It is also unclear whether MRTFs are required for other aspects of SMC function that do not directly involve specification, such as the changes in gene expression that are known to occur during environmental stresses such as hypertension and atherosclerosis. On the basis of the mammary myoepithelial defect observed in MRTF-A knockout mice, the MRTFs may be responsible for the upregulation of SMC marker gene expression that is observed in many SMC-like cells (i.e., myofibroblasts or mesangial cells) following injury (8, 16).
Given the importance of the myocardin family for regulating SMC-specific transcription, the identification of the signaling mechanisms that regulate the expression and/or activities of these transcription factors will be very important to our understanding of the regulation of SMC phenotype. In our studies, agonist-induced upregulation of SMC-specific transcription did not correlate with increased expression of any of the myocardin factors [data not shown (23)]. Instead, our data indicated that RhoA-dependent regulation of MRTF nuclear localization was important and involved cytoplasmic retention of the MRTFs by a mechanism involving G-actin binding to the MRTF NH2 terminus. Results obtained with the MRTF-A/myocardin chimeras and the myocardin RPEL2 domain mutant suggest that other regions are also important, including basic region 1 (B1) and perhaps B2. These regions have already been implicated in nuclear import, and it is likely that a delicate balance exists between these nuclear localization and cytoplasmic retention signals.
Since a number of extrinsic cues that regulate SMC differentiation also regulate RhoA activity, these studies indicate that regulation of MRTF nuclear localization may be an important mechanism by which environmental factors regulate SMC phenotype. The observation that ΔB1/B2 MRTF-A inhibited the induction of the SM22α and SM α-actin promoters by S1P and serum supports a role for MRTF nuclear translocation in this response, as does the direct correlation between the effects of S1P, serum, and PDGF-BB on promoter activity and nuclear translocation of the MRTFs. Interestingly, ΔB1/B2 MRTF-A actually increased c-fos promoter activity in S1P- and TGF-β-treated cells, providing additional evidence that the myocardin factors differentially regulate SRF-dependent growth and SRF-dependent differentiation. Wang et al. (50) have shown that myocardin and the ternary complex factors compete for SRF binding, and the positive effects of ΔB1/B2 MRTF-A on the c-fos promoter would be consistent with this model.
TGF-β was the strongest activator of SMC-specific transcription in these studies and has been shown to activate RhoA in some cell culture model systems (1, 7). The inhibitory effect of ΔB1/B2 MRTF-A on this response indicated that a basal level MRTF activity is required for TGF-β-induced upregulation of SMC-specific promoter activity. It is also possible that our inability to detect partial changes in MRTF localization following TGF-β treatment could explain these results. However, the observations that TGF-β had only minor effects on MRTF localization and that Y-27632 had little to no effect on TGF-β-induced SM α-actin promoter activity suggest that non-RhoA-dependent mechanisms are more important. The effects of TGF-β on SMC-specific transcription are thought to be mediated by activation of the SMADs (SMADs 2 and 3 in particular) (36, 40, 42). Our data would indicate that, whereas both pathways are essential for SMC differentiation marker gene expression, they probably act, at least to some extent, in parallel. Interestingly, Chen et al. (7) demonstrated in Monc-1 cells that Y-27632 inhibited TGF-β-induced, SMC-specific gene expression and that dominant-negative RhoA inhibited SMAD-2 and SMAD-3 nuclear localization. This same study showed that inhibition of RhoA in 10T1/2 cells (by dominant-negative RhoA and C3 exotoxin) yielded more modest results, suggesting that there are probably cell type-specific differences in the importance of RhoA on TGF-β signaling. Two recent studies have shown that myocardin interacts directly with SMAD-3 and SMAD-1 to regulate transcription in smooth and cardiac muscle, respectively (3, 37). Therefore, it will be very important to further identify the mechanisms that integrate the RhoA and TGF-β signaling pathways.
Because many environmental cues are known to affect RhoA activity (see Ref. 12 for a review), the regulation of MRTF localization is probably complicated by many cell type-specific and microenvironmental differences. Our observations that basal MRTF nuclear localization was higher in SMC than in 10T1/2 cells and that nuclear translocation occurred more quickly in SMC in response to treatment support this idea. Whereas our data suggest that cell confluency (a signal that inhibits RhoA signaling) attenuated MRTF nuclear localization, others have shown that application of cell tension (a signal that stimulates RhoA) promoted nuclear localization of Drosophila MRTF (13, 43). The later findings suggest that not only is the MRTF-SRF interaction highly conserved throughout evolution, but the regulation of MRTF translocation by RhoA may be as well. It is also possible that the contrasting results on MRTF localization that have been reported (9, 39) may reflect inherent differences in RhoA signaling between cell types.
Our gel shift assays demonstrated that the transcriptional activities of the myocardin factors are also regulated by their ability to interact with SRF. Even after removal of the inhibitory NH2-terminal region, MRTF-SRF complexes were still much weaker than those formed with myocardin. As with nuclear localization, these differences correlated well with the relative activities of the myocardin factors, suggesting that this parameter is also an important determinant of myocardin factor activity. Another interesting finding of our studies was that, relative to SRF binding, MRTF-A formed a more robust ternary complex at the intronic CArG than at CArGs A or B, which was not the case for myocardin. Little is known about the mechanisms by which the myocardin factors discriminate between SRF bound to the large number of CArG-dependent gene promoters, but these data suggest that variations in CArG sequence may be an important determinant. MRTF-A has been shown to preferentially interact with SRF at the vinculin and SRF promoters but not with SRF at the c-fos or Egr-1 promoters (30). Whereas this difference may be due to known competition for SRF binding between Elk-1 and the myocardin factors, slight differences in SRF conformation (or perhaps DNA bending) due to variations in CArG sequence may also play a role. In support of this, Zaromytidou et al. (53) have recently shown that MRTF-A coimmunoprecipitated with SRF more strongly in the presence of a consensus CArG oligonucleotide than in the presence of a CArG that contained substitutions to the AT-rich region. It is clear that a better understanding of interactions between the different myocardin family members and specific SRF-CArG complexes will be important for delineating the role of the myocardin family in both cell type-specific and gene-specific transcriptional regulation.
In summary, results from the present study indicate that the MRTFs play an important role in the regulation of SMC differentiation marker gene expression in at least some SMC subtypes. Unlike myocardin, MRTF nuclear localization, and hence MRTF transcriptional activity, is regulated by agonists that activate RhoA, and we feel that this signaling pathway may be an important mechanism by which a variety of environmental cues regulate SMC-specific transcription. Since SMC are known to maintain a significant level of plasticity even in adult animals, perhaps signaling mechanisms such as these have evolved as a mechanism by which SMC can quickly and reversibly alter their phenotype in response to environmental cues.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-070953 (to C. P. Mack), HL-081844 and HL-071054 (to J. M. Taylor), and American Heart Association Grants 0130193N (to C. P. Mack) and 0355776U (to J. M. Taylor).
We thank Richard Treisman (Cancer Research UK, London, UK) and Da-Zhi Wang (Univ. of North Carolina, Chapel Hill, NC) for reagents.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society