5′ CArG degeneracy in smooth muscle α-actin is required for injury-induced gene suppression in vivo (original) (raw)
Substitution of SMα-actin 5 ′ CArGs with the c-fos SRE consensus CArG had no effect on cell-specific gene expression in vivo. To determine the importance of the reduced SRF binding affinity of CArG-A and CArG-B in controlling cell-specific expression of SM α_-actin_, we generated transgenic animals containing an SRE consensus CArG element substituted for CArG-A (SRE-A), CArG-B (SRE-B), or CArG-A and CArG-B (SRE-AB) in the SM α_-actin_ promoter (Figure 1). Our studies employed a 2,560-bp α_-actin_ construct plus the first intron (2,784 bp) of the promoter, which mimics expression patterns of the endogenous gene in transgenic mice (5). Surprisingly, in contrast to studies in cultured cells in which these substitutions resulted in relaxed cell specificity (7), substitution of the SRE consensus CArG for either one or both SM α_-actin_ 5′ CArGs had no detectable effect on cell-specific gene expression across at least 3 independent transgenic founder lines per construct in adult (Figure 2A), E13.5 (data not shown), or E16.5 (Figure 2B) mice. Results of LacZ staining in whole organs, tissues, and embryos were further verified by histological analyses (Figure 2, C and D). Studies indicated complete retention of cell-specific expression patterns of the SRE-substituted mutant transgenes throughout development and maturation. Taken together, these results contradict our initial hypothesis that the SM α_-actin_ 5′ degenerate CArGs are required for cell-specific gene expression and demonstrate that the conservation of degeneracy of the 2 SM α_-actin_ 5′ CArG elements across multiple species and millions of years of evolution must serve some alternative function.
Schematic diagram of wild-type and SRE-substituted mutant/LacZ promoter constructs used to generate transgenic mice. To determine the importance of the reduced SRF binding affinity of the degenerate CArGs for controlling SM α_-actin_ expression, we generated mutant LacZ promoter constructs in which a c-fos SRE consensus CArG was substituted for the CArG-A and/or CArG-B element(s) within the SM α_-actin_ promoter previously shown to mimic expression of the endogenous gene. The resulting mutations are shown in bold.
SRE-substituted transgenic mice showed no loss of SMC-specific SM α_-actin_ expression during normal development and maturation. (A) Analysis of LacZ expression in whole adult tissues of WT and SRE-AB transgenic mice indicated that the SRE substitutions had no effect on transgene expression in adult tissues. Sm. int., small intestine. (B) Analysis of LacZ expression in E16.5 whole embryos from wild-type and SRE-AB transgenic mice indicated no loss of specificity upon replacement of SM α_-actin_ CArG-A and CArG-B with the c-fos SRE consensus CArG. (C) Histological analyses of adult tissues indicated no loss of SMC-specific SM α_-actin_ expression in SRE-substituted transgenic mice. Tissues were paraffin-embedded, sectioned, and stained with H&E. LacZ expression was specific to SMCs in wild-type aorta and in aortas of SRE-AB transgenic mice across multiple independent founder lines. Magnification, ×40. In SRE-substituted transgenic mice, the same LacZ expression pattern was found in all SMC-containing tissues examined, including esophagus and small intestine (data not shown). (D) Histological examination of LacZ- and eosin-stained aortas of E16.5 embryos from wild-type and SRE-substituted transgenic mice indicated no loss of specificity in the mutants. LacZ expression was restricted to the SMC-containing layers of all tissues examined, including the aorta (magnification, ×40), esophagus, bronchi, and small intestine (data not shown).
SRF binding activity of the c-fos SRE consensus CArG was greater than SRF binding activity of the SMα-actin 5 ′ CArGs in the context of the SMα-actin promoter. Based on observations that substitution of the c-fos SRE consensus CArG for the SM α_-actin_ 5′ CArGs had no detectable effect on SMC specificity in vivo, we questioned whether SRF binding affinity was altered as significantly as would be predicted based on previous EMSA analyses of a wide range of SRE consensus CArG substitution mutants (16–19). EMSAs previously done in our lab showed that the SRE consensus CArG, in the context of the c-fos promoter, bound SRF with greater activity than either CArG-A or CArG-B in the context of the SM α_-actin_ promoter (7). However, there was no direct evidence that substitution of the SRE consensus CArG box alone for CArG-A and/or CArG-B in the context of the SM α_-actin_ promoter resulted in increased SRF binding activity to the CArG region. In particular, it was not known whether the reduced SRF binding affinity of the SM α_-actin_ CArGs, as compared to the SRE consensus CArG, was independent of the G/C substitutions and determined by other sequences within or flanking the SM α_-actin_ CArGs. Therefore, we performed EMSA experiments using in vitro translated SRF and a 95-bp SM α_-actin_ probe containing both 5′ α_-actin_ degenerate CArGs. Competition experiments were performed with double-stranded, unlabeled wild-type SM α_-actin_, SRE-AB, SRE-A, and SRE-B complexes, all 95 bp in length. Also included in the competition experiments for comparison was a 95-bp, double-stranded, unlabeled c-fos promoter complex, which included the SRE consensus CArG. The SRE-AB, SRE-A, and SRE-B cold competitor complexes competed for SRF binding with the labeled probe more effectively than did the wild-type SM α_-actin_ competitor (Figure 3). Not surprisingly, the c-fos DNA complex was the most effective competitor for SRF binding. Importantly, all SRE-substituted mutant DNA complexes competed more effectively than the wild-type competitor did. Therefore, the SRE-substituted SM α_-actin_ CArGs exhibited the expected increase in SRF binding affinity, and the lack of an effect of the SRE G/C substitution mutations is thus unlikely to be the consequence of unique sequences within or surrounding the SM α_-actin_ CArGs. Our results indicate that increasing the SRF binding affinity of the SM α_-actin_ CArGs had no detectable effect on SMC selectivity of the gene, at least during normal development and maturation.
Oligonucleotides containing the c-fos SRE consensus CArG substitution competed for SRF binding activity more effectively than did wild-type SM α_-actin_ oligonucleotides (containing CArG-A and/or CArG-B) in EMSAs. (A) In vitro translated SRF and a 95-bp radiolabeled probe harboring the CArG-containing region of the SM α_-actin_ promoter were used for EMSAs. Unlabeled 95-bp double-stranded oligonucleotides containing either the SM α_-actin_ 5′ CArGs (WT) or the SRE consensus CArG substituted for CArG-A (SRE-A), CArG-B (SRE-B), or CArG-A and CArG-B (SRE-AB) in the context of the SM α_-actin_ promoter were used as cold competitors at approximately 50-, 100-, and 200-fold excess over labeled probe. The SRE CArG in the context of the c-fos promoter (fos) was used as a cold competitor at approximately 50-fold excess over labeled probe. The SRF band was supershifted (SRF SS) by the addition of 2 μg of anti-SRF rabbit polyclonal antibody. Unprog. lysate, unprogrammed control lysate. (B) Densitometry was performed on the SRF bands (see Figure 5A) and results were plotted relative to maximal SRF binding to the radiolabeled probe in the absence of cold competitor. Results are representative of 3 independent experiments. Statistical analyses were performed using 1-way ANOVA. We found statistically significant differences between percentage of SRF binding in the presence of WT cold competitor and percentage of SRF binding in the presence of SRE-AB, SRE-A, or SRE-B cold competitor under all but 1 condition (×50 WT versus ×50 SRE-AB) across multiple experiments (data not shown).
SRE-AB promoter transgene showed an attenuated response compared to the wild-type promoter transgene following injury in mice. The G/C substitutions are completely conserved across all species in which the promoter has been cloned, and they clearly influence SRF binding. Therefore, we hypothesized that the degenerate CArG elements may have evolved to provide a means to differentially regulate CArG-dependent SMC genes compared to CArG-dependent growth response genes under conditions in which SRF is elevated. We further postulated that such a condition might occur following vascular injury, which is known to simultaneously induce CArG/SRF-dependent growth response genes such as c-fos (20) and downregulate CArG/SRF-dependent SMC gene expression (21).
Previous studies from our lab have shown that SM marker genes undergo significant transcriptional repression 7 days after vascular injury. In particular, vascular injury studies were done in transgenic mice harboring an SM α_-actin_/LacZ, SM myosin-HC/LacZ, or SM22/LacZ transgene. Seven days after vascular injury, all transgenic mice examined showed significantly reduced expression of the LacZ transgene in the media and developing intima compared to contralateral control vessels. Expression was increased 14 days after injury but was still well below expression levels in the uninjured control artery (21). To determine if the G/C substitutions within the SM α_-actin_ 5′ CArGs played a role in controlling gene expression in response to vascular injury, we performed wire injury studies on at least 5 animals from each of 2 independent SRE-AB founder lines and 2 wild-type SM α_-actin_ founder lines. It should be noted that since we were assessing a transgene and not the endogenous gene, this approach allowed us to determine pathways that control gene expression within the injured vessel without altering the injury response per se. That is, both our wild-type and SRE-AB transgenic mice contained their normal endogenous SM α_-actin_ genes such that there were no differences in function or in the nature of the injury responses between the 2 groups of mice. As such, one can study fundamental mechanisms that control transcription of SM α_-actin_ in the context of the normal vascular injury response. Notably, substitution of the SRE consensus CArG for both CArG-A and CArG-B resulted in a significant attenuation in the transgene response to injury. Unlike the wild-type SM α_-actin_ gene, which was dramatically downregulated in response to injury, the SRE-AB transgene was expressed in the media and the developing intima of the injured carotid artery at levels equivalent to those seen in the uninjured contralateral control vessel (Figure 4). These results provide evidence that the G/C substitutions contained within the SM α_-actin_ 5′ CArG elements and the resulting reduction in SRF binding affinity are critical for controlling gene expression in response to pathophysiological stimuli such as vascular injury.
Analysis of LacZ expression in mouse carotid arteries indicated that the SRE-AB transgene is differentially expressed compared with the wild-type SM α_-actin_ transgene 7 days following wire injury. Results are representative of at least 5 animals from each of 2 WT and 2 SRE-AB independent transgenic founder lines. Small arrowheads indicate the internal elastic lamina, and large arrowheads indicate the external elastic lamina. Magnification is indicated in each panel.
Quantitative real-time RT-PCR showed increased levels of SRF expression following vascular injury. To begin to understand possible mechanisms by which substitution of the SM α_-actin_ 5′ CArGs with the c-fos SRE consensus CArG resulted in an attenuated response to vascular injury, we analyzed real-time RT-PCR data from injured rat carotid artery samples to determine the expression pattern of SRF at various time points following injury. SRF levels increased within 1 hour following injury and peaked at 24 hours following injury (Figure 5). At 24 hours, SRF expression levels were approximately 4 times that of the uninjured control and decreased back to control levels by 7 days after injury. These results indicate that, even under circumstances of increased SRF expression, the SM α_-actin_ degenerate CArGs provide a regulatory mechanism that abrogates CArG activation of SM α_-actin_ in response to injury. Taken together, our data suggest that SRF can distinguish between muscle-specific and growth factor–specific CArG boxes in vascular injury. Moreover, our results suggest that there may be mechanisms that normally regulate SRF binding to degenerate CArG elements that must be absent or significantly attenuated under conditions of vascular injury. However, it remains to be determined whether this distinction was a result of the CArG sequence alone or whether other factors were involved in helping SRF to distinguish muscle-specific from non–muscle-specific CArG elements.
Temporal expression analysis by real-time RT-PCR of endogenous SRF in balloon-injured carotid arteries showed increased SRF expression following injury. SRF expression was normalized to 18S rRNA expression in the injured and uninjured contralateral control vessel. Each time point represents the mean ± SE of the injured (SRF:18S) vessel normalized to the uninjured (SRF:18S) vessel (n = 4 animals per time point).
Myocardin increased SRF association with the CArG-containing region of the SMα-actin promoter but not of the c-fos promoter in intact chromatin and was downregulated following vascular injury. To assess whether myocardin plays a role in regulating the ability of SRF to distinguish between SM-specific and growth response CArG elements, we used chromatin immunoprecipitation (ChIP) assays to determine the effects of overexpression of myocardin on SRF association with the CArG-containing regions of the SM α_-actin_ and c-fos promoters. Quantitative real-time PCR analysis of precipitated DNA showed that SRF association was markedly enriched at the SM α_-actin_ CArG-containing region in myocardin-overexpressing cells compared to control cells (Figure 6A). However, there was virtually no difference in SRF binding at the c-fos CArG-containing region between cells overexpressing myocardin and control cells. These results suggest that myocardin may contribute to regulation of SM α_-actin_ cell-specific gene expression by selectively enhancing SRF binding to the SM α_-actin_ degenerate CArGs under normal circumstances. In addition, results suggest that loss of myocardin expression would preferentially decrease SM α_-actin_ expression compared to c-fos expression.
Myocardin can differentially regulate SRF binding to degenerate versus consensus CArGs and is decreased following vascular injury. (A) Results of ChIP assays in cultured rat aortic SMCs indicate that SRF binding is enhanced at the CArG-containing region of the SM α_-actin_ promoter but not the c-fos promoter in response to myocardin overexpression. Quantitative PCR was used to detect CArG-containing regions of the SM α_-actin_ and c-fos promoters in chromatin fragments immunoprecipitated with an SRF antibody. Data represent the mean ± SE of the fold increase in SRF association in cells overexpressing myocardin versus control cells in 3 independent experiments. A fold increase value of 1 indicates no change in SRF association in cells overexpressing myocardin versus control cells. *P < 0.05 compared with SRF association at the c-fos CArG region under the same conditions. (B) Temporal expression analysis by real-time RT-PCR of endogenous myocardin in balloon-injured rat carotid arteries showed decreased myocardin expression following injury. Myocardin expression was normalized to 18S rRNA expression in the injured and uninjured contralateral control vessel. Each time point represents the mean ± SE of the injured vessel (myocardin:18S) normalized to that of the uninjured (myocardin:18S) vessel (n = 4 animals per time point). *P < 0.05 compared with myocardin expression prior to injury. (C) Effect of substitution of CArG-A and CArG-B with the c-fos SRE CArG on myocardin responsiveness of SM α_-actin_ promoter activity. SM α_-actin_/luciferase and SRE-AB/luciferase promoter constructs were cotransfected with myocardin into rat aortic SMCs and assayed for luciferase activity. The activity was normalized for protein content. Normalized promoter activities of SM α_-actin_/luciferase and SRE-AB/luciferase in the absence of myocardin were set to 1. Fold induction over basal promoter activity in response to myocardin was calculated. Values represent the mean ± SE of 3 independent experiments. *P < 0.05 compared with WT fold induction under the same conditions.
To determine whether reduced expression of myocardin may contribute to injury-induced suppression of SM α_-actin_ expression, we performed quantitative real-time RT-PCR on rat carotid artery samples taken at various time points following vascular injury. Expression of myocardin mRNA was significantly decreased within 3 days following injury compared to the uninjured control, but it returned to control levels by 7 days after injury (Figure 6B). We have previously shown partial reinduction of SM α_-actin_ gene expression in injured arteries between 1 week and 3 weeks after injury (21). Although there are no high-quality myocardin antibodies available to confirm these results at the protein level, our results suggest that reduced expression of myocardin may contribute to injury-induced suppression of SM α_-actin_, assuming injury does not result in selective stabilization of myocardin mRNA. Moreover, based on our ChIP assays (Figure 6A), our results also suggest that the reduction in myocardin may selectively decrease SRF binding to the SM α_-actin_ degenerate CArGs, but that reduced myocardin has no effect on SRF association with the c-fos CArG. As such, our data support a fundamental mechanism whereby expression of CArG-dependent SMC genes may be selectively repressed in response to vascular injury, even in the presence of increased SRF levels. If this is the case, substitution of the c-fos SRE consensus CArG for the SM α_-actin_ 5′ CArGs would be predicted to decrease myocardin-induced transactivation of SM α_-actin_. Consistent with our hypothesis, cotransfection studies in rat aortic SMCs showed that myocardin-induced transactivation of an SRE-substituted SM α_-actin_ promoter construct was significantly reduced compared to that of a wild-type promoter construct (Figure 6C). Taken together, these results suggest that myocardin can selectively regulate SRF binding to the SM α_-actin_ degenerate CArGs, and loss of myocardin may contribute to repression of SM α_-actin_ in response to vascular injury.