Human Prolyl-4-hydroxylase α(I) Transcription Is Mediated by Upstream Stimulatory Factors*

Prolyl-4-hydroxylase α(I) (P4Hα(I)) is the rate-limiting subunit for P4H enzyme activity, which is essential for procollagen hydroxylation and secretion. In the current study, we have characterized the human P4Hα(I) promoter for transcription factors and DNA elements regulating P4Hα(I) expression. Using a progressive deletion cloning approach, we have constructed pGL3-P4Hα(I) recombinant plasmids. We have identified a positive regulatory region at the positions of bp –184 to –97 responsible for ∼80% of the P4Hα(I) promoter efficiency. Three E-boxes were located within this region, and the E-box at position bp –135 explains most of the regulatory capacity. Upstream stimulatory factors (USF1/USF2) were shown to bind on the E-box using chromatin immunoprecipitation assay. Suppression of USF1 and/or USF2 using specific short interference RNA resulted in a significant reduction in P4Hα(I) promoter activity, and overexpressed USF1 or USF2 increased P4Hα(I) promoter activity significantly. Although transforming growth factor β1 increased the USF1/USF2-E-box binding and P4Hα(I) promoter activity, this up-regulatory effect can be largely prevented by USF1/USF2-specific short interference RNA. On the other hand, cigarette smoking extracts, which have been shown to suppress P4Hα(I) expression, inhibited the binding between the USF1/USF2 and E-box, resulting in a reduced P4Hα(I) promoter activity. Furthermore, the E-box on the P4Hα(I) promoter appeared to indiscriminately bind with either USF1 or USF2, with a similar outcome on the promoter efficiency. In conclusion, our study shows that USF1/USF2 plays a critical role in basal P4Hα(I) expression, and both positive (transforming growth factor β1) and negative (cigarette smoking extract) regulators appear to influence the USF-E-box interaction and affect P4Hα(I) expression.

Extracellular matrix (ECM), 2 the structural framework of all tissues, plays a key role in pathogenesis. In arterial wall, ECM holds various types of cells, including endothelial cells, vascular smooth muscle cells (VSMC), and fibroblasts, in their functional places. Dysregulated ECM is frequently associated with atherogenesis, vulnerable plaques, and aneurysm (1)(2)(3)(4)(5). ECM is maintained by a balanced synthesis and degra-dation of mainly collagens and elastins, which are in turn regulated by other molecules, e.g. fibrillin. Collagens form a multigene family with at least 39 members, the genes of which are known to be dispersed throughout at least 15 chromosomes. Collagen molecules consist of three identical or dissimilar polypeptide chains (called ␣ chains) and contain at least one triple helical collagenous domain with repeating (GXY) n sequences, i.e. a glycine residue at every third amino acid, with the frequent presence of proline and 4-hydroxyproline in the X and Y positions, respectively. Most collagen molecules form supramolecular assemblies, and the superfamily can be divided into nine different families based on their polymeric structures or other features, such as collagens that form fibrils (types I-III, V, and XI). In normal arteries, fibrillar collagens I and III are the major constituents of the intima, media, and adventitia, whereas types I, III, IV, and V are in endothelial and VSMC basement membranes.
The biosynthesis of the precursors of fibrillar collagens (called procollagens) involves a number of post-translational modifications. The intracellular modifications require five specific enzymes, three collagen hydroxylases (6,7) and two collagen glycosyltransferases (8,9). Among these hydroxylases, prolyl 4-hydroxylase (P4H), a ␣ 2 ␤ 2 heterotetramer, plays a key role in collagen metabolism. The ␤ subunit is identical to the enzyme protein disulfide isomerase (6,10) and is produced in excess of the ␣ subunit, which contains a major portion of the catalytic sites of the enzyme, and the abundance of the ␣ subunit restricts the enzymatic activity (6). P4H is an intracellular enzyme required for the synthesis and formation of all known types of collagen. It catalyzes the formation of hydroxyproline from proline residues located in repeating Xaa-Pro-Gly triplets in the procollagens during post-translational processing and is the process essential for folding the procollagen polypeptide chains into stable triple helical molecules (6). Inhibition of the P4H enzymes generates unstable collagen that is degraded inside the cell and is not secreted, which is associated with reduced collagen production (5). On the other hand, overexpression of the P4H␣(I) gene is associated with excess collagen production (11). P4H isoenzymes are expressed in most tissues, although there may be tissue specificity with regard to type I or II isoforms for the ␣ subunit (12)(13)(14). Although relatively little data is available for factors affecting P4H expression and activity, P4H may be regulated by nitric oxide, tumor necrosis factor ␣, transforming growth factor ␤ (TGF␤), insulin-like growth factor 1, ␤ fibroblast growth factor, certain cytokines, and cigarette smoking (15)(16)(17)(18)(19)(20)(21)(22). Environmental stimuli, such as hypoxia, can induce P4H␣(I) expression at both transcriptional and post-transcriptional levels, which results in excess accumulation of collagen, a process relevant to fibrosis (23,24).
The promoter of rat P4H␣(I) has been described by Takahashi et al. (24). However, little is known with regard to the functional features of human P4H␣(I) promoter (AL731563). In the present study, we characterized the elements in the promoter of human P4H␣(I). We found that the E-box-like sequence (CACGGG), which has been shown to bind upstream stimulatory factors (USF) (25)(26)(27), in the position of bp Ϫ135 is vital for P4H␣(I) transcription. Upstream stimulatory factor (USF1/USF2) appears to be the transcription factor that controls the basal P4H␣(I) expression. Factors including TGF␤1 and cigarette smoking extract (CSE) up-regulated and down-regulated the USF-Ebox interaction and affected P4H␣(I) expression.

MATERIALS AND METHODS
Plasmid Construction-pGL3-basic (Promega, Madison WI) was used to construct pGL-P4H␣(I) reporter vector to evaluate transcription efficiency. We used a PCR-based method to clone serially deleted P4H␣(I) promoter fragments into the pGL3-basic vector at the KpnI and HindIII sites. Genomic DNA from HeLa cells was used as the template for the cloning. The inserts were confirmed for the correct sequence and orientation by direct sequencing. We named each recombinant pGL3 vector according to the size of the inserted P4H␣(I) fragment ( Table 1). The same antisense oligonucleotide primer that corresponded to the intron 1 region IVS1 ϩ 76 -IVS1 ϩ 96 (5Ј-CCAAAGC-TTACCACCACAGCGGGAAGGAAT-3Ј) was used for the generation of all of the clones. A KpnI linker was incorporated into the 5Ј-end of the sense primers, and a HindIII linker was incorporated into the 5Ј-end of the antisense primers. The digested PCR products were subcloned into pGL3-basic at KpnI and HindIII sites. The mutated construct pGL-145M was obtained by site-directed mutagenesis with a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions using primers specifically introducing a mutation at the E-box site of the P4H␣(I) pGL-145 vector. The primers used for site-directed mutagenesis are as follows (underlines refer to mutated nucleotides): sense 5Ј AGC AGG GAG GCA CTC TCT TGG CGC CTG C-3; antisense 5Ј-GCA GGC GCC AAG AGA GTG CCT CCC TGC T-3Ј. The E-box mutation was confirmed by the direct sequencing. The USF1-pSG5 and USF2-pSG5 expression vectors were generous gifts from the laboratory of Dr. M. Sawadogo at the University of Texas MD Anderson Cancer Center, Houston, TX. Details of the vectors were described previously (28).
Cell Culture, Transfection, and Luciferase Assays-HeLa cells purchased from the American Type Culture Collection (Masassas, VA) were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum at 37°C with 5% CO 2. Human aortic VSMCs were purchased from Cambrex BioScience, Inc. (Walkersville, MD). VSMCs were cultured in smooth muscle basal medium containing 10% fetal bovine serum (Cambrex BioScience, Inc.), and the cells in passages 6 -8 were used for the experiments. Plasmid DNA was prepared using a plasmid extraction kit (Qiagen Inc., Valencia, CA). One microgram of pGL-P4H␣(I) hybrid DNA and/or 1 g of siRNA or USF-pSG5 were transfected into the cells cultured in a 6-well plate with Lipofectamine 2000 (Invitrogen), and 0.25 g of pSV␤gal (Promega), which expresses ␤-galactosidase, was co-transfected to allow normalization of the transfection efficiency. After 4 h, the cells were treated with or without 10 ng/ml TGF␤1 (Leinco Technologies, Inc., St. Louis, MO) for 24 h in the same medium but containing 0.2% fetal bovine serum. We also treated the transfected cells with CSE as described previously (29). Briefly, the smoke of a research cigarette (2R4F from Tobacco Health Research, University of Kentucky; nicotine, 0.85 mg/cigarette; tar, 9.70 mg/cigarette) was bubbled into a flask containing 20 ml of prewarmed F12K complete medium. The ignited cigarette was completely consumed in puffs by 10-ml syringe suction at one end of the tube within 5 min. The aqueous smoke extract was filtered through 0.2-m syringe filter, and the pH of the CSE was adjusted to 7.4. The concentration of CSE was calculated in arbitrary units as cigarette equivalents/ml of the medium. CSE was used to treat the cells within 30 min of extraction. After 24 h, the cells were harvested and lysed in reporter lysis buffer (Promega), and cell lysates were assayed for luciferase and ␤-galactosidase activities. Luciferase activity was measured with an FB12 luminometer (Berthold Detection Systems, Pforzheim, Germany) using luciferin (Promega) as the substrate. ␤-Galactosidase was assayed in an assay buffer (Promega), and the activity was measured at 420 nm in a spectrophotometer. The luciferase activity, which reflects the promoter activity, was normalized to the ␤-galactosidase to account for the variation in transfection efficiency.
For siRNA transfection, 100 nM siRNA was transfected into the cells cultured in a 6-well plate with Lipofectamine 2000. The siRNA duplex (Integrated DNA Technologies, Houston, TX) sequences for USF1 were: sense, 5Ј-GAC CCA ACC AGU GUG GCU ATT-3Ј and antisense, 5Ј-UAG CCA CAC UGG UUG GGU CTT-3Ј. The siRNA duplex sequences for USF2 were: sense, 5Ј-CCU CCA CUU GGA AAC GGU ATT-3Ј, and antisense, 5Ј-UAC CGU UUC CAA GUG GAG GTT-3Ј. As the negative control siRNA, we used scrambled siRNA designed and tested by Ambion, Inc. (catalog number 4611G, Austin, Texas), which has no significant sequence similarities to mouse, rat, or human gene sequences.
Chromatin Immunoprecipitation (ChIP) Assay-The ChIP assay was performed using the histone H3 ChIP assay kit according to the manufacturer's protocol (Upstate Biotechnology, Charlottesville, VA). In brief, ϳ3 ϫ 10 7 cells were used per ChIP assay. The cells were cross-

TABLE 1 Primers for the generation of serially deleted P4H␣(I) promoter fragments
The same antisense oligonucleotide primer was used for the generation of all of the clones shown. A KpnI linker (CTGGTACC), shown in bold, was incorporated into the 5Ј-end of the sense primers, and a HindIII linker (CCAAAGCTT) was incorporated into the 5Ј-end of the antisense primers. The IVS1 indicates the location within the first intron counting from the end of exon 1 (Fig. 1A). The insertion sites of the cloned DNA fragments are at the multiple cloning region of the ϩ5 (KpnI) and ϩ53 (HindIII) upper stream of the luciferase gene.

Human P4H␣(I) Promoter
linked with 1% formaldehyde at 37°C for 10 min and rinsed twice with ice-cold phosphate-buffered saline. The cells were harvested by a brief centrifugation. Cell pellets were resuspended in SDS lysis buffer (50 mmol/liter Tris-HCl, pH 8.1, 10 mmol/liter EDTA, 1% SDS, and protease inhibitors). Sonication was performed on ice using a Sonifier II 450 (Branson, Danbury, CT) with a 3-mm tip set at duty cycle 20 and an output level 2 to achieve chromatin fragments ranging between 200 and 1000 bp in size, followed by centrifugation at 15,000 ϫ g for 10 min at 4°C. Supernatants were collected and diluted 10-fold in ChIP dilution buffer (a 20-l aliquot was removed to serve as an input sample) followed with preimmunoprecipitation clearing with 80 l of a mixture of salmon sperm DNA/protein A/protein G at 4°C with rotation for 30 min. Immunoprecipitation was carried out with 2 g of anti-USF1 or -USF2 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C overnight with rotation. After immunoprecipitation, 60 l of a mixture of salmon sperm DNA/protein A/protein G was added and incubated at 4°C with rotation for 30 min and followed by brief centrifugation. The precipitates were washed twice (5 min each at 4°C) with low salt buffer, once with high salt buffer, and once with LiCl buffer. Then the precipitates were washed again with TE buffer (  ufacturer's protocol. The RNA solution was treated with 1 unit of RNase-free DNase I (1 unit/l; Promega) in a final volume of 20 l at 37°C for 15 min to remove trace amounts of genomic DNA contamination. The DNase I was inactivated by incubation at 75°C for 10 min. One microgram of total RNA was used for reverse transcription. The mRNAs were reverse-transcribed into cDNAs with an iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) containing a mixture of oligo(dT) and random primers. Real time PCR was performed using an iCycler iQ real time PCR detection system with a SYBR Green I Supermix kit (Bio-Rad) with the program running for 40 cycles at 95°C for 20 s and 60°C for 1 min. The PCR efficiency for the primers was examined by serially diluting the cDNA templates. The melting curve analysis was performed over the range 55-95°C by monitoring SYBR Green fluorescence at increasing temperatures (0.5°C increment changes at 10-s intervals). PCR-specific products were determined as a clear single peak, at the melting curves, of Ͼ80°C. The specificity of primers was also confirmed as a single band with the correctly amplified fragment size through agarose gel electrophoresis of the real time RT-PCR products. Real time PCR was duplicated for each cDNA sample. Each gene mRNA level was acquired from the value of the threshold cycle (Ct) of the real time PCR as related to that of ␤-actin through the formula 2 ⌬Ct (⌬Ct ϭ ␤-actin Ct Ϫ gene of interest Ct).

Identification of a Positive Regulatory Element in the P4H␣(I)
5Ј-Flanking Sequence-To clarify the regions responsible for the basal transcriptional regulation of the P4H␣(I) gene, we prepared several reporter constructs by cloning the 5Ј-flanking sequence of the P4H␣(I) gene (Fig. 1A) into the upstream of the luciferase reporter. We then transfected HeLa cells with these constructs containing the progressively deleted 5Ј-flanking regions of the P4H␣(I) gene (Fig. 1B) and measured the luciferase activity of the resulting cell extracts. The deletion of the 5Ј-flanking sequence from bp Ϫ580 to Ϫ184 did not affect the transcription efficiency (Fig. 1B). However, a deletion up to bp Ϫ97 resulted in an ϳ80% decrease in the transcription activity. These results suggest that the basal transcription of the P4H␣(I) gene was positively regulated by a region located between bp Ϫ184 and Ϫ97. One or more specific factors interacting with a cis-element(s) located between bp Ϫ184 and Ϫ97 may regulate the P4H␣(I) promoter efficiency.
Effects of Mutation in the Ϫ135 E-box-Through a bioinformatics search, we have located three consensus E-box-like elements (CACGGG) located between bp Ϫ184 and Ϫ97. They are at bp Ϫ169, Ϫ158, and Ϫ135 (Fig. 1A). To confirm the importance of the E-boxes in P4H␣(I) gene transcription, we generated a wild-type pGL-145 (Fig. 2) and a mutated reporter gene construct pGL-145M in which the Ϫ135 E-box-like sequence (CACGGG) was mutated to CACTCT. We then transfected HeLa cells with the pGL-184, pGL-145, and pGL-145M vectors (Fig. 2). The deletion of the 5Ј-flanking sequence from Ϫ184 to Ϫ145, which removed the Ϫ169 and Ϫ158 E-boxes, did not affect the transcription efficiency. The mutation of the Ϫ135 E-box (pGL-145M), however, resulted in an ϳ60% reduction in the activity. These results suggest that the basal transcription of the P4H␣(I) gene was positively regulated by the Ϫ135 E-box.
Involvement of USF1/USF2 in the P4H␣(I) Promoter Activity-The USF is one of the main families of transcription factors binding to the E-box (30,31). To investigate the involvement of USFs in transactivating the P4H␣(I) promoter, we first examined the effects of silencing USF1 and USF2 by siRNAs on the activity of the P4H␣(I) promoter. Transfection of USF1 siRNA and/or USF2 siRNA into HeLa cells showed a significantly decreased pGL-145 promoter activity (Fig. 3A). Using HeLa cells, we further examined whether USF1/USF2 silencing could also suppress P4H␣(I) expression. Indeed, the mRNA levels of the P4H␣(I) gene measured by the quantitative real time RT-PCR were significantly reduced when USF1/USF2 were silenced (Fig. 3B). In the meantime, negative control siRNA had no effect on P4H␣(I) expression. Our experiments further demonstrated that the degree of inhibition by the USF1 or USF2 siRNA appeared to be similar, and treatment with both siRNAs did not induce any additive effect comparable to the reduction by a single siRNA. These results suggest that the transcription of the P4H␣(I) gene was positively regulated by the USF1/USF2, either one of which would be efficient for P4H␣(I) expression.
We then overexpressed USF1 or USF2 in HeLa cells and human aortic VSMCs by co-transfecting pGL-145 vectors with USF1-or USF2-pSG5 vectors. As shown in Fig. 3C, USF1 and USF2 significantly up-regulated P4H␣(I) promoter activities as measured by luciferase activity in VSMCs. The same effect was also observed in HeLa cells (data not shown). In addition, USF1 appeared to induce a higher P4H␣(I) promoter activity than did USF2 (Fig. 3C). Such differential effect was not elucidated by the siRNA-mediated USF1 and/or USF2 suppression.
Direct Binding of USF1/USF2 to the P4H␣(I) Promoter-To further assess the involvement of USF1/USF2 in P4H␣(I) transcription, we examined whether USF transcription factors could bind to the E-box in the P4H␣(I) promoter in vivo. The binding of E-box-binding proteins USF1/USF2 to the P4H␣(I) promoter was illustrated by the ChIP assay (Fig. 3D). We then examined the effect of USF1/USF2 silencing on the binding activity of the USF1/USF2 using the ChIP assay. We found a decreased binding activity of both USF1 and USF2 to the E-box (Fig.  3D). Meanwhile, the negative control siRNA had a minimal effect on the binding.
Effects of TGF␤1 on P4H␣(I) Promoter Activity-P4H␣(I) has a central role in the biosynthesis of collagens. TGF␤1 can induce up-regulated collagen synthesis. To evaluate whether and how the P4H␣(I) promoter would respond to TGF␤1 stimulation, pGL-145 luciferase plasmids were transfected into HeLa cells. The normalized luciferase activity was calculated after TGF␤1 treatment. We found a 2-fold increase in the luciferase activity after TGF␤1 treatment (Fig. 4A). This effect was blocked by USF1/USF1 silencing using siRNAs (Fig. 4A), whereas negative siRNA had no effect. We further demonstrated the effect of TGF␤1 treatment on the P4H␣(I) mRNA levels in HeLa cells (Fig. 4B). We found a 1.8-fold increase in P4H␣(I) mRNA levels, and this effect was blocked by USF1/USF2 silencing using siRNA (Fig. 4B). The same effect was also shown in VSMCs (data not shown). ChIP assay FIGURE 2. Mutating the E-box reduces P4H␣(I) promoter activity. Plasmids pGL-184, pGL-145, and the mutated construct pGL-145M were transfected into HeLa cells. The relative luciferase activities were assessed. Results are expressed as the ratio of luciferase activity in each extract relative to that in the extract from HeLa cells transfected with pGL3-basic. The data were normalized to the activity of the co-transfected pSV␤gal plasmid and presented as the means Ϯ S.D. of three independent experiments.
showed TGF␤1 increased in the band intensity for the binding between the USF1/USF2 and the E-box in the P4H␣(I) promoter (Fig. 4C). Furthermore, negative control siRNA had no effect on the binding activity. Taken together, these results suggest that the transcription of the P4H␣(I) gene was positively regulated by TGF␤1 through the USF1/ USF2 and E-box interaction.
Effects of CSE on P4H␣(I) Promoter Activity-Previously, we have shown (15) that CSE could down-regulate P4H␣(I) expression in endothelial cells and VSMCs. To evaluate the responsiveness of the P4H␣(I) promoter to CSE treatment, pGL-145 luciferase plasmids were transfected into HeLa cells, and the normalized luciferase activity was calculated after CSE treatment. We found an ϳ50% reduction in the luciferase activity by the CSE treatment (Fig. 5A). We further showed that CSE also decreased the P4H␣(I) mRNA levels in HeLa cells (Fig. 5B), which was consistent with our previous report (15) on endothelial cells and VSMCs. ChIP assay demonstrated that CSE decreased the band intensity for the binding between the USF1/USF2 transcription factor and the E-box in the P4H␣(I) promoter (Fig. 5C). These results suggest that the transcription of the P4H␣(I) gene was negatively regulated by CSE through USF1/USF2-E-box interaction.

DISCUSSION
Because collagen is the major structural protein in the ECM, which provides the framework for all tissues including arteries, there is little doubt that factors regulating collagen synthesis, process, secretion, and degradation will play a significant role in health and disease. It is therefore important to understand the molecular mechanisms regulating P4H␣(I) expression, a rate-limiting subunit for the P4H enzyme, which is critical for procollagen modification and secretion (32,33). Our study unequivocally shows that the E-box-like sequence located in the region of bp Ϫ184 to Ϫ97 upstream of the P4H␣(I) promoter is the positive regulator and is responsible for nearly 80% of the basal promoter activity. As far as we are aware, this is the first characterization of the human P4H␣(I) promoter (12,34).
Although there is a significant sequence homology between human and rodent P4H␣(I), the regulatory elements appear to be different between them. Takahashi and colleagues showed that the positive regulatory region is located at positions bp Ϫ246 to Ϫ165 in the rat P4H␣(I) promoter (24). We showed that this positive regulator is located at bp Ϫ184 to Ϫ97, particularly the E-box at bp Ϫ135 in the human P4H␣(I) promoter. Although their study did not show experimental evidence for any cis-acting element in P4H␣(I) regulation, our experiments have shown convincing evidence for the involvement of the USF family and E-box interaction in regulating P4H␣(I) expression. This conclusion is supported by the siRNA-specific suppression of the USF1/USF2, USF1/ USF2 overexpression and the E-box mutagenesis. It should be noted, however, our study only demonstrates the regulatory elements for basal expression. DNA elements and corresponding transcription factors regulating the response to stimulation are likely to be different from those governing the basal expression. Results are expressed as the ratio relative to HeLa cells without siRNA treatment. C, human aortic smooth muscle cells were co-transfected with 1 g of pGL-145 hybrids and 1 g of empty expression vector (empty pSG5) or expression constructs for USF1 or USF2 using Lipofectamine 2000. 0.25 g of pSV␤gal was co-transfected to allow normalization of transfection efficiency. After 28 h, the cells were harvested for luciferase assay. Results in each extract are expressed as the ratio relative to that in the extract from human aortic smooth muscle cells co-transfected with pGL-145 and empty pSG5. D, ChIP assays were performed with anti-USF1 or anti-USF2 antibody for the HeLa cells treated with (ϩ) or without (Ϫ) 100 nM USF1 siRNA or USF2 siRNA for 24 h. Primers for the P4H␣(I) 5Ј-flanking region from Ϫ145 to ϩ38 were used to amplify the DNA isolated from the ChIP assay. All data are presented as the means Ϯ S.D. of three independent experiments. IP, immunoprecipitation.
USF1 and USF2 are ubiquitous proteins characterized by highly conserved C-terminal basic helix-loop-helix and leucine zipper domains responsible for dimerization and DNA binding activities (28,35,36). The E-box, the consensus sequence of which is CANNTG, is known as the binding site for USF and regulates the transcription of several genes (37,38). They appear to be involved in the regulation of developmental and tissue-specific or gene-specific expression of a number of genes (39,40), including those for surfactant protein A (41,42), steroidogenic factor 1 (43), and carboxyl ester lipase (44). Our study adds a collagen regulator P4H␣(I) as another USF target in regulating ECM metabolism in addition to its suppressing effect on cathepsin B expression (45). Together with a long list of genes including those involved in glucose metabolism, cholesterol metabolism, cell proliferation, detoxification, and hypoxia that can be potentially regulated by the USF (46 -54), it appears clear that USF may have a role in vascular disease more than we have appreciated so far.
The major USF species present in most tissues and cell lines is the heterodimer between USF1 and USF2, which interact with E-box (36). USF1 homodimers are less abundant, and USF2 homodimers are scarce (55,56), which may have different effects (57,58). In our study, however, we have shown that USF1 and USF2 appear to be interchangeable in P4H␣(I) promoter activity. Suppressing either one will result in a similar degree of inhibition in P4H␣(I) promoter activity; suppressing both does not induce an additive suppressing effect, which is consistent with the recent reports on TGF␤1 (59,60).
Our experiments have further shown that the positive effect of TGF␤1 on collagen production is at least partly mediated by the up-regulated P4H␣(I) expression, which is regulated through the USF-E-box interaction at the bp Ϫ135 of the P4H␣(I) promoter. Likewise, the same promoter mechanism also appears to be negatively regulated by cigarette smoking, which has been shown to down-regulate P4H␣(I) expression and collagen production (29). Clearly both TGF␤1 and cigarette smoking would regulate ECM metabolism through more than one target. Our study shows the USF-E-box interaction on the P4H␣(I) promoter is one of the targets. The same mechanism may apply to other target genes of TGF␤1 and cigarette smoking if they have the same cis-acting element. However, we should be aware that the promoter elements responsible for basal expression may differ from those FIGURE 4. Effects of TGF␤1 on the P4H␣(I) promoter activity. A, one microgram of pGL-145 hybrids was transfected with (ϩ) or without (Ϫ) 100 nM USF1 siRNA and/or USF2 siRNA into HeLa cells with Lipofectamine 2000. We used scrambled siRNA designed and tested by Ambion, which has no significant sequence similarities to mouse, rat, or human gene sequences, as the negative control siRNA. 0.25 g of pSV␤gal was co-transfected to allow normalization of transfection efficiency. After 4 h, the cells were treated with or without 10 ng/ml TGF␤1 in the same medium but containing 0.2% fetal bovine serum for 24 h. Results are expressed as the ratio of luciferase activity in each cell extract relative to that in the extract from cells transfected with pGL-145 but without TGF␤1 and siRNA treatment. B, HeLa cells were transfected with or without 100 nM USF1 siRNA and/or USF2 siRNA and then treated with or without 10 ng/ml TGF␤1 for 24 h. After 24 h, the P4H␣(I) mRNA level was quantitated by real time RT-PCR and normalized to the housekeeping gene ␤-actin. Results are expressed as the ratio relative to HeLa cells without TGF␤1 and siRNA treatment. C, ChIP assays were performed with anti-USF1 or anti-USF2 antibody for HeLa cells transfected with or without 100 nM USF1 siRNA or USF2 siRNA and then treated with or without 10 ng/ml TGF␤1 for 24 h. Primers for the P4H␣(I) 5Ј-flanking region from bp Ϫ145 to ϩ38 were used to amplify the DNA isolated from the ChIP assay. All data are presented as the means Ϯ S.D. of three independent experiments. IP, immunoprecipitation. Results are expressed as the ratio of luciferase activity in each cell extract relative to that in the extract from cells transfected with pGL-145 but without CSE treatment. B, HeLa cells were treated with or without CSE (0.01 unit) for 24 h. After 24 h, the P4H␣(I) mRNA level was quantitated by real time RT-PCR and normalized to the housekeeping gene ␤-actin. Results are expressed as the ratio relative to HeLa cells without CSE treatment. C, ChIP assays were performed with anti-USF1 or anti-USF2 antibody for HeLa cells treated with or without CSE (0.01 unit) for 24 h. Primers for the P4H␣(I) 5Ј-flanking region from bp Ϫ145 to ϩ38 were used to amplify the DNA isolated from the ChIP assay. All data are presented as the means Ϯ S.D. of three independent experiments. IP, immunoprecipitation. responding to stimulation. Although our experiments show that TGF␤1 may regulate USF-E-box interaction, studies of others show that USF is also a positive transcription regulator for the TGF␤1 gene via the glucose response element (59,60). These findings suggest that, in addition to directly regulating P4H␣(I) expression, USF proteins may regulate collagen or ECM metabolism through TGF␤1 pathways. These effects could be cell type-specific. Together with their effects on growth factors, e.g. insulin-like growth factor-binding proteins (25), it appears that USF may have a significant role in ECM metabolism.
In summary, our study has characterized that USF plays an important role in P4H␣(I) expression. The USF-E-box interaction appears to be the site wherein the positive (TGF␤1) and negative (cigarette smoking) regulators of collagen production and P4H␣(I) expression will act. Further studies on how TGF␤1 or cigarette smoking regulates USF-E-box interaction will be of significant value in understanding vascular diseases, such as vulnerable atherosclerotic plaque or aortic aneurysm that involves dysregulated ECM metabolism.