Upstream stimulatory factors regulate aortic preferentially expressed gene-1 expression in vascular smooth muscle cells.

The phenotypic modulation of vascular smooth muscle cells (VSMC) plays a central role in the pathogenesis of arteriosclerosis. Aortic preferentially expressed gene-1 (APEG-1), a VSMC-specific gene, is expressed highly in differentiated but not in dedifferentiated VSMC. Previously, we identified an E-box element in the mouse APEG-1 proximal promoter, which is essential for VSMC reporter activity. In this study, we investigated the role of upstream stimulatory factors (USF) in the regulation of APEG-1 transcription via this E-box element. By electrophoretic mobility shift assays, recombinant USF1 and USF2 homo- and heterodimers bound specifically to the APEG-1 E-box. Nuclear extracts prepared from primary cultures of rat aortic smooth muscle cells exhibited specific USF1 and USF2 binding to the APEG-1 E-box. To investigate the binding properties of USF during VSMC differentiation, nuclear extracts were prepared from the neural crest cell line, MONC-1, which differentiates into VSMC in culture. Maximal USF1 and USF2 protein levels and binding to the APEG-1 E-box occurred 3 h after the differentiation of MONC-1 cells was initiated. Co-transfection experiments demonstrated that dominant negative USF repressed APEG-1 promoter activity, and USF1, but not USF2, transactivated the APEG-1 promoter. Our studies demonstrate that USF factors contribute to the regulation of APEG-1 expression and may influence the differentiation of VSMC.

Vascular smooth muscle cells (VSMC), 1 the major cell type in blood vessel walls, exhibit a spectrum of phenotypes that change in response to environmental cues (1). The conversion of VSMC from a quiescent to a proliferative phenotype contributes to the pathogenesis of arterial restenosis, hypertension, atherosclerosis, and its related complications (2). The developmental origins of VSMC are diverse, including mesodermal precursors and neural crest cells that can differentiate into VSMC (3). However, the molecular mechanisms controlling the differentiation and development of VSMC, including the regulation of VSMC-specific or selective gene expression is only beginning to be elucidated (4). Thus, identifying these mechanisms may enhance our understanding of VSMC phenotypic regulation and provide potential therapeutic or preventive targets for vascular proliferative syndromes, including atherosclerosis.
Using deletion and mutation analysis, we previously identified an E-box motif (CAGCTG) located in exon 1 of the APEG-1 gene (bp ϩ39/ϩ44) as a major positive regulatory element (6). In addition to APEG-1, several SMC-specific marker genes contain E-box elements in their promoters including SM ␣-actin, SM-22␣, and SM-myosin heavy chain (7)(8)(9). Basic helixloop-helix (bHLH) and bHLH leucine zipper transcription factor families bind to E-box motifs (10). Previously, we did not detect E12 and E47 binding to the APEG-1 E-box (6), suggesting that other E-box binding proteins are present in these complexes. Upstream stimulatory factors 1 and 2 (USF1 and -2), originally identified as activators of the adenovirus major late promoter (11), are ubiquitously expressed transcription factors (12). USF1 and USF2 recognize and bind to DNA with an E-box motif as either homodimers or heterodimers. The USF family members regulate the expression of several genes (13,14), including SMC-expressed genes (15), and they are known to bind to E-boxes similar to that found in the APEG-1 promoter (16). In this report, we wanted to further elucidate the nuclear proteins binding to the APEG-1 E-box and to investigate their role in the regulation of APEG-1 expression.
Plasmid Constructs and DNA Probes Used-The luciferase reporter plasmids APEG-1 p(Ϫ355/ϩ76), p(Ϫ122/ϩ76), and p(Ϫ122/ϩ76)E-mut were generated by subcloning these fragments into the pGL3-Basic vector (Promega) as described (6). The 0.9-kb USF1 cDNA fragment was cloned by reverse transcription-PCR from mouse aortic SMC RNA. The 1-kb USF2 cDNA fragment from psvUSF2 was provided by Dr. M. Sawadogo (Houston, TX). The Drosophila expression plasmids, pPAC and phsp82LacZ, were provided by Dr. T. Maniatis (19). The expression vectors of USF1 and USF2 were constructed by subcloning the USF1 and USF2 inserts into pcDNA3.1(Ϫ) (Invitrogen) and pPAC. To generate a dominant negative USF expression construct, we amplified a fragment of the USF1 cDNA (encoding amino acids 213-310) by PCR. This fragment was cloned into the pCMV/myc/nuc vector (Invitrogen) that contains a nuclear localization signal and subsequently into pcDNA3 (Invitrogen). All constructs were confirmed by sequencing.
Preparation of Nuclear Extracts and in Vitro Translated Products-Nuclear extracts were prepared from cultured cells as described (20). Protein concentrations of nuclear extracts were measured by the Bradford dye-binding method (21) with the Bio-Rad protein assay reagent. In vitro translated products were prepared using the TNT T7 Quick Coupled Transcription/Translation system (Promega).
Electrophoretic Mobility Shift Assays (EMSA)-Two 18-bp oligonucleotide probes, containing the E-box sequence, 5Ј-GGGCCTCAGCTGG-TCAG-3Ј, or mutated E-box, E-mut 5Ј-GGGCCTCAGCacGGTCAG-3Ј, were synthesized according to the mouse APEG-1 exon 1 sequence (6). USF E-box consensus oligonucleotide was synthesized in the context of the APEG-1 promoter, 5Ј-GGGCCTCACGTGGTCAG-3Ј. After annealing, the double-stranded oligonucleotides were end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase (New England Biolabs) and purified with G-50 spin columns (Amersham Pharmacia Biotech). The binding reaction contained 20,000 cpm of DNA probe, 2 g of poly(dI-dC), 25 mM HEPES (pH 7.9), 40 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, and either 10 g of nuclear protein or in vitro translated products. In cold competition assays, a 100-fold molar excess of unlabeled double-stranded oligonucleotides, including APEG-1 Ebox, APEG-1 E-mut, and E-box consensus oligonucleotides, were added to the binding reaction 30 min before the addition of radiolabeled probe. For mobility supershift experiments, 2 g of USF1 (C-20) and USF2 (N-18) rabbit polyclonal antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were added to the binding reactions and incubated at room temperature for 30 min before the addition of radioactive probe. Protein-DNA complex formation was carried out on ice for 30 min and resolved by a 5% nondenaturing polyacrylamide gel electrophoresis using 0.5ϫ Tris borate-EDTA buffer (44.5 mM Tris-base, 44.5 mM boric acid, 1 mM EDTA, pH 8.0) at 4°C. Gels were dried, and bands were detected using autoradiographic film.
Western Blot Analysis-Nuclear proteins (30 g) were separated on 6% SDS-PAGE and transferred to nitrocellulose filters (22). The blot was probed with USF1 or USF2 polyclonal rabbit antibodies. Signal was detected on Western blots using enhanced chemiluminescence (SuperSignal West Pico kit; Pierce).
RNA Extraction and Northern Blot Analysis-Total RNA from cultured cells was extracted using a MINI RNA isolation kit (Qiagen). Total RNA (10 g) was denatured, fractionated on 1.3% formaldehydeagarose gels, and subsequently transferred to NitroPure filters (Osmonics). The filters were then hybridized with random primed, [␣-32 P]dCTP-labeled APEG-1, USF1, USF2, and SM-␣ actin cDNA probes. To correct for the differences in RNA loading, blots were subsequently hybridized to a 32 P-labeled oligonucleotide probe complementary to 28 S rRNA. The blots were exposed to a phosphoscreen and x-ray film. The radioactivity was measured on a PhosphorImager by using ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA).
Transient Transfections and Reporter Activity Assays-APEG-1 p(Ϫ355/ϩ76), p(Ϫ122/ϩ76), and p(Ϫ122/ϩ76)E-mut promoter-reporter plasmids were used for transient transfection assays. For assays in RASMC, 2 ϫ 10 5 cells/well were plated in triplicate in six-well plates, incubated for 24 h. Using FuGENE 6 transfection reagent (Roche Molecular Biochemicals), cells were transfected with 1 g of the APEG-1 promoter-reporter plasmid plus various amounts of dominant negative USF expression construct and empty vector, pcDNA3, to make the total amount of plasmid DNA to be the same. Forty-eight hours later, the cells were harvested for luciferase and ␤-galactosidase activity assays. For D.Mel-2 cells, 2 ϫ 10 5 cells/well were plated in triplicate in 12-well plates, 1 day before transfection. The cells were transfected using CellFECTIN reagent (Invitrogen) and harvested after 24 h. Luciferase activity was measured using Promega's Luciferase Assay System and normalized to ␤-galactosidase activity.

USF1 and USF2 Bind to an E-box Element in Exon 1 of the
Mouse APEG-1 Gene-We previously identified an E-box element within exon 1 of the APEG-1 gene (bp ϩ39/ϩ44), which was essential for APEG-1 promoter activity in VSMC (6). We wanted to test the hypothesis that USF proteins, which are capable of binding E-box motifs (CAGCTG) (15), may be important for the regulation of APEG-1 in VSMC. To determine whether USF1 or USF2 proteins could bind to the APEG-1 E-box, we performed EMSA using in vitro translated products of USF1 and USF2 incubated with radiolabeled APEG-1 E-box, APEG-1 E-mut, or E-box consensus oligonucleotides as probes. USF1 and USF2 homodimers and USF1/USF2 heterodimers bound to APEG-1 E-box and E-box consensus (CACGTG) probes (Fig. 1, left and right panels, respectively). USF proteins did not bind to the mutated APEG-1 E-box oligonucleotide (CAGCAC) (Fig. 1, middle panel). The migration of USF1 homodimer is faster than USF2, and the migration of USF1/USF2 heterodimer is intermediate between these two homodimers.
USF Binding Activity in VSMC-To determine whether USF proteins bind to the APEG-1 E-box in VSMC, we prepared nuclear extracts from RASMC. EMSA using APEG-1 E-box (Fig. 2, left panel) and consensus E-box oligonucleotide probes (Fig. 2, right panel) revealed the presence of strong DNAprotein complexes (a and b) when incubated with nuclear extracts from RASMC. There was a faint supershifted band after USF1-specific antibody was added (q), and two supershifted bands after USF2-specific antibody or a combination of USF1/ USF2 antibodies were added (*). Moreover, only the intensity of band a decreased with use of USF1 and USF2 antibodies. The specificity of the DNA-protein complex was also assessed by adding an unlabeled E-box competitor in the binding reaction. As shown in the right lane of each gel, bands a and b were competed away by the 100-fold molar excess of E-box cold competitor. Our results indicated that USF1 and USF2 specifically bound to the APEG-1 E-box in RASMC (band a).
Early Induction of USF Binding Activity and Protein Expression during MONC-1 Differentiation-The MONC-1 neural crest cell line differentiates into smooth muscle cells in culture (18,23). We used this model to study USF binding and expression during the differentiation of MONC-1 into smooth muscle cells. We harvested total RNA and nuclear extracts from undifferentiated MONC-1 cells and from cells in differentiation medium for 3, 6, 9, 12, 24, 48, 96, and 144 h. There were three major DNA-protein complexes detected in MONC-1 cell nuclear extracts (arrows a, b, and c in Fig. 3). EMSA analysis revealed that DNA-binding complexes reached their maximum 3 h after the initiation of MONC-1 cell differentiation (Fig. 3A). To identify the components of the APEG-1 E-box DNA-protein complexes, we performed supershift assays (Fig. 3B). Only the a complex disappeared when USF1-or USF2-specific antibodies were added (Fig.  3B). In the presence of USF1 antibody, a single supershifted complex (Fig. 3B, q) formed. A faint slower migrating complex appeared at 3 h, but this faint complex was also visible in the absence of USF1 antibody (Fig. 3B, lane 2). The USF1supershifted complex (q) was not as intense as band a, which disappeared. These data suggest that in addition to a partial supershift, the USF1 antibody caused a disruption in complex a. Two more pronounced supershifted complexes formed in the presence of the USF2 (Fig. 3B, *) antibody. Again, the slower migrating supershifted complex corresponded in migration speed with a faint complex in the absence of antibody (Fig. 3B, lane 2). However, the increased intensity in the presence of USF2 antibody suggests that this is a true supershifted complex (Fig. 3B, upper (*) complex). To investigate the specificity of binding, we added unlabeled competitor (APEG-1 E-box, APEG-1 E-mut, and E-box consensus oligonucleotides) to the binding reactions. Unlabeled APEG-1 Ebox oligonucleotide competed away DNA-protein complexes (a, b, and c) (Fig. 3C). The E-box consensus oligonucleotide mainly competed away complex a, with some diminishment in complexes (b) and (c). In contrast, APEG-1 E-mut competed away only the fastest migrating complex (c). These data suggest that complexes a and b are specific and that complex a contains the USF proteins. In contrast, it is likely that complex c is binding to the flanking sequence outside of the APEG-1 E-box site.
We next determined the mRNA and protein levels of USF1 and USF2 during SMC differentiation. By Northern blot analysis, we found there was no significant change in USF1 or USF2 mRNA levels in MONC-1 during differentiation (Fig.  4A). The levels of APEG-1 and SM ␣-actin mRNA increased throughout differentiation. Moreover, binding of USF to the APEG-1 promoter (3-h peak) preceded APEG-1 mRNA induc-tion (as early as 6 h) after the initiation of MONC-1 differentiation. By Western blot analysis, we found that the expressions of USF proteins reached their peaks 3 h after the initiation of MONC-1 cell differentiation (Fig. 4B), which correlated with the peak USF binding activity (Fig. 3A).
Dominant Negative USF Represses the Activity of the APEG-1 Promoter in RASMC-RASMC exhibit high levels of endogenous USF binding activity ( Fig. 2 and data not shown); thus, we wanted to inhibit USF binding activity to determine its role in regulating the APEG-1 promoter. To perform these experiments, we designed a dominant negative USF expression construct that encodes a protein lacking the basic DNA-binding domain but retains the dimerization domain (24). In addition, this construct contains a nuclear targeting sequence. We confirmed the nuclear localization of dominant negative USF by immunofluorescent staining of RASMC transfected by this construct (data not shown). Using in vitro translated products, dominant negative USF suppressed the binding of USF1 and USF2 to APEG-1 E-box in a dose-dependent manner (Fig. 5A). Co-transfection of dominant negative USF with the APEG-1 promoter construct in RASMC resulted in a 50% reduction in APEG-1 promoter activity (Fig. 5B). These results confirm that endogenous USF proteins have a role in the regulation of the APEG-1 promoter.
USF1 Participates in Activation of the APEG-1 Promoter-To examine the role of USF1 and USF2 transcription factors in regulating the APEG-1 promoter, we used Drosophila cells (D.Mel-2 cells), which do not contain functional USF (25). Drosophila cells were transfected with APEG-1 promoter p(Ϫ122/ ϩ38) reporter construct, USF1 and USF2 expression constructs, and phsp82LacZ to normalize for transfection efficiency. We have previously shown that APEG-1 promoter construct (Ϫ122/ϩ38) contains nearly full promoter activity (6). The overexpression of USF1 produced a dose-dependent increase in luciferase activity with a maximum of ϳ9-fold (Fig. 6). Interestingly, USF2 homodimers and USF1/USF2 heterodimers did not activate the APEG-1 promoter in Drosophila cells (Fig. 6). These data suggest that USF2 is not an activator of the APEG-1 promoter, and when in a heterodimer with USF1, USF2 represses APEG-1 promoter activity. In addition, the USF1 expression construct did not activate the mutated APEG-1 E-box construct, APEG-1 p(Ϫ122/ϩ76)E-mut, indicating that USF1 transactivation is through this site. DISCUSSION We have demonstrated previously that high levels of APEG-1 gene expression in VSMC require the E-box motif in exon 1 (6). In the present study, we showed that USF transcription factors bind to this E-box and contribute to the regulation of APEG-1 gene expression, in a positive or a negative manner, depending on the USF family member.
USF proteins belong to the group B bHLH transcription factors (26,27). They bind preferentially to the CACGTG E-box (28), which is different from the core CAGCTG E-box motif in the APEG-1 gene. The group A bHLH proteins, such as the ubiquitously expressed E2A proteins E12 and E47, recognize and bind as heterodimers to the CAGCTG type of E-box site (28). We have previously shown that E2A proteins from RASMC nuclear extracts do not bind to the APEG-1 E-box (6). In the present study, recombinant USF (Fig. 1) or USF in RASMC nuclear extracts (Fig. 2) bound to the APEG-1 E-box and regulated APEG-1 promoter activity. These results further support the idea that USF factors can recognize and regulate promoters through this type of E-box site.
E-box elements and their associated bHLH transcription factors, such as MyoD, are known to be important in the regulation of striated muscle-specific gene expression and differ-FIG. 2. USF1 and USF2 proteins in RASMC nuclear extracts bind to the APEG-1 E-box. Nuclear extracts (NE) harvested from RASMC were incubated with rabbit polyclonal antibodies (Ab) against USF1 or USF2 and 32 P-labeled APEG-1 E-box or E-box consensus oligonucleotides. For competition assays, an unlabeled APEG-1 E-box oligonucleotide was incubated with nuclear extracts before the radiolabeled probes were added. Arrows a and b indicate the DNA-protein complexes. USF1 and USF2 supershifted complexes are indicated by dots and asterisks, respectively. entiation of striated muscle cells in a tissue-specific manner (29,30). However, no such tissue-restricted bHLH family member has been identified in mature arteries; nor have they been shown to regulate VSMC-expressed genes. While it is not likely that ubiquitously expressed bHLH family members can solely account for expression of VSMC-restricted genes, Johnson and Owens (15) have previously demonstrated that USF factors contribute to the regulation of SM ␣-actin, a SMC differentiation marker gene. One may hypothesize that a ubiquitously expressed factor, such as USF, heterodimerizes with a smooth muscle-specific factor to promote cell type-restricted gene expression. Alternatively, USF factors may be incorporated into a complex of transcription factors, including co-activators, whose stereospecific assembly allows smooth muscle cell-restricted gene expression. Consistent with this hypothesis, Qyang et al. (12) proposed that the ability of USF to activate a given promoter was dependent on cell-specific co-activators.
To study the role of USF in the regulation of APEG-1 during the differentiation of VSMC, we used an in vitro MONC-1 differentiation system (18). We found that in MONC-1 cells, USF1 and USF2 protein levels and their APEG-1 E-box binding activity peaked very early after differentiation was initiated (Fig. 3). Moreover, expression and binding of USF factors to the APEG-1 E-box preceded expression of the APEG-1 gene (Fig. 4). Therefore, the increased amounts and binding activity of USF proteins may contribute to APEG-1 promoter regulation during VSMC differentiation. These data were supported by functional studies showing that overexpression of a dominant negative USF (that prevents wild-type USF protein binding) decreased APEG-1 promoter activity (Fig. 5).
APEG-1, like SM ␣-actin that is regulated by USF factors, is rapidly down-regulated in injured blood vessels or in dedifferentiated VSMC in culture (5). We now show that USF1 is able to transactivate the APEG-1 promoter; however, USF2 is un-  complexes (a, b, and c), detected in these nuclear extracts. B, for supershift reactions, nuclear extracts were incubated with rabbit polyclonal antibodies directed toward USF1 and USF2. USF1-and USF2-supershifted complexes were indicated by dots and asterisks, respectively. C, in competition assays, the unlabeled APEG-1 E-box, APEG-1 E-mut, and E-box consensus oligonucleotide probes were used as competitors. Hours, the duration of time after differentiation was initiated.
FIG. 4. The mRNA and protein levels of USF1 and USF2 in the differentiating MONC-1 cells. A, total RNA extracted from MONC-1 cells at different time points of differentiation was analyzed by Northern blot analysis using radiolabeled probes for USF1, USF2, APEG-1, and SM ␣-actin. RNA loading was verified by hybridization to a radiolabeled oligonucleotide complementary to the 28 S ribosomal RNA. B, for Western blotting, the nuclear extracts prepared from MONC-1 cells at different time points of differentiation were resolved, transferred, and hybridized with antibodies against USF1 or USF2. able to increase APEG-1 promoter activity (Fig. 6). Even more importantly, co-transfection of an identical amount of USF2 expression plasmid was able to suppress USF1 transactivation of the APEG-1 promoter. These data suggest that USF2 may be acting as a repressor, as described previously in the regulation of plasminogen activator inhibitor-1 promoter (31). Taken together, our data suggest that the relative amounts of USF1 and USF2 and the way in which they interact as heterodimers may determine the level of APEG-1 expression. This effect of USF2 on APEG-1 is different from that described for the SM ␣-actin promoter, in which both USF1 and USF2 caused promoter transactivation (15). Nevertheless, both studies demonstrate that the USF factors modulate the expression of genes, APEG-1 and SM ␣-actin, that are acutely regulated by the differentiation status of VSMC. We propose that USF factors, by altering the expression of key genes, may play an important role in the differentiation capability of VSMC.

FIG. 5. Dominant negative USF suppresses the binding of USF to the APEG-1 E-box and APEG-1 promoter activity in RASMC.
A, EMSA were carried out using various amounts of in vitro-translated products of USF1, USF2, and USF dominant negative mutant, which lacks the DNA-binding domain. B, RASMC were transiently transfected with APEG-1 promoter-reporter constructs, p(Ϫ355/ϩ76), and USF dominant negative mutants or the corresponding empty vector to make the final amount of DNA per well identical. Luciferase activity was determined 48 h after transfection, and the data were normalized to ␤-galactosidase activity. Data represent the average of three experiments, and error bars represent S.E. Normalized luciferase activity is presented as a percentage of APEG-1 promoter activity without dominant negative USF.
FIG. 6. E-box-dependent transactivation of the APEG-1 promoter by USF1 in Drosophila cells. Drosophila D.Mel-2 cells were co-transfected with APEG-1 promoter-reporter construct, p(Ϫ122/ϩ76) or p(Ϫ122/ϩ76)E-mut, and the indicated amounts of pPAC-USF1 and pPAC-USF2 expression plasmids. To make the total amount of plasmid DNA the same, the empty vector pPAC was added to produce a final concentration of 3 g/well. Cells were harvested after 24 h. Data represent the average of three experiments, and error bars represent S.E. Normalized luciferase activity is presented as -fold induction of APEG-1 promoter activity.