Glucose up-regulates thrombospondin 1 gene transcription and transforming growth factor-beta activity through antagonism of cGMP-dependent protein kinase repression via upstream stimulatory factor 2.

Thrombospondin 1 (TSP1) transcription is stimulated by glucose, resulting in increased TGF-beta activation and matrix protein synthesis. We previously showed that inducible expression of the catalytic domain of cGMP-dependent protein kinase (PKG) inhibits glucose-regulated TSP1 transcription and transforming growth factor (TGF)-beta activity in stably transfected rat mesangial cells (RMCs(tr/cd)). However, the molecular mechanisms by which PKG represses glucose-regulated TSP1 transcription are unknown. Using a luciferase-promoter deletion assay, we now identify a single region of the human TSP1 promoter (-1172 to -878, relative to the transcription start site) that is responsive to glucose. Further characterization of this region identified an 18-bp sequence that specifically binds nuclear proteins from mesangial cells. Moreover, binding is significantly enhanced by high glucose treatment and is reduced by increased PKG activity. Gel mobility shift and supershift assays show that the nuclear proteins binding to the 18-bp sequence are USF1 and -2. USF1 and USF2 bound to the endogenous TSP1 promoter using a chromatin immunoprecipitation assay. Glucose stimulates nuclear USF2 protein accumulation through protein kinase C, p38 MAPK, and extracellular signal-regulated kinase pathways. Increased PKG activity down-regulates USF2 protein levels and its DNA binding activity under high glucose conditions, resulting in inhibition of glucose-induced TSP1 transcription and TGF-beta activity. Overexpression of USF2 reversed the inhibitory effect of PKG on glucose-induced TSP1 gene transcription and TGF-beta activity. Taken together these data present the first evidence that USF2 mediates glucose-induced TSP1 expression and TSP1-dependent TGF-beta bioactivity in mesangial cells, suggesting that USF2 is an important transcriptional regulator of diabetic complications.

It is established that elevated blood glucose levels are a significant risk factor for the development of microvascular complications of diabetes, including diabetic nephropathy (1). Hyperglycemia stimulates an increase in TGF-␤ 1 activity (2)(3)(4), which has been shown to be a major mediator of the fibrotic changes in the pathogenesis of diabetic nephropathy (5,6).
TGF-␤ is synthesized and secreted as latent complex (Latent TGF-␤). It must be converted to the active state before binding to its receptors and eliciting cellular functions. Latent TGF-␤ can be activated by a number of factors (7)(8)(9). The matricellular protein, thrombospondin1 (TSP1), is a major physiological regulator of TGF-␤ activation (10 -12). TSP1, a 420-kDa homotrimer with individual subunits of ϳ145 kDa, is a multifunctional protein whose expression is controlled by many factors (13)(14)(15)(16). Our earlier work showed that glucose-stimulated increases in TSP1 expression are responsible for the activation of TGF-␤ in mesangial cells when exposed to high glucose concentrations (30 mM), which contributes to the accumulation of extracellular matrix proteins (17). Furthermore, high glucose mediates increases in TSP1 expression and TSP1-dependent TGF-␤ bioactivity through down-modulation of nitric oxide (NO)/cGMP-dependent protein kinase (PKG) signaling (18,19).
PKG is a serine/threonine kinase consisting of an aminoterminal regulatory and a COOH-terminal catalytic domain within one polypeptide chain. Binding of cGMP by the regulatory domain leads to activation of the catalytic domain (20). Although PKG is best known as a regulator of vascular smooth muscle cell (VSMC) contractility, VSMC phenotype, cardiac contractility, intracellular calcium signaling, platelet aggregation, and cytoskeletal reorganization (21)(22)(23)(24), PKG also regulates expression of multiple genes such as c-fos, mitogen-activated protein kinase phosphatase I, gonadotropin-releasing hormones, soluble guanylate cyclase, osteopontin, and TSP1 (18, 19, 22, 24 -30). Activation of the cAMP response, serum response, activator protein-1 elements, and transcriptional regulator TF II-I are reported to be involved in the PKG-regulated gene transcription (28,(31)(32)(33).
In our previous studies we generated stably transfected rat mesangial cells with tetracycline-regulated expression of the catalytic domain of PKG-I to directly regulate the activity of PKG independent of cGMP levels (19). Using these stably transfected rat mesangial cells, we showed that expression of the catalytic domain of PKG significantly repressed glucoseinduced but not basal TSP1 gene transcription and TSP1-dependent TGF-␤ activation (19). However, the molecular mechanisms by which glucose up-regulates and PKG represses glucose induction of TSP1 transcription are unknown.
In previous work, using the human TSP1 promoter we showed that high glucose concentrations (30 mM) up-regulate the activity of a 2.033-kilobase region of the human TSP1 promoter and that increased PKG activity inhibits high glucose-stimulated reporter activity (19). These data suggest that this 2.033-kilobase promoter region contains important regulatory elements required for PKG-mediated repression of glucose-induced TSP1 gene expression. In the present study we investigated the molecular mechanisms by which PKG mediates repression of glucose-induced TSP1 gene transcription in mesangial cells by using a promoter-deletion approach. We now report identification of a region Ϫ932 to Ϫ915 relative to the transcription initiation site in the TSP1 promoter (18-bp region) that is a binding site for upstream stimulatory factors (USFs), which are critical for glucose-induced TSP1 transcription. Furthermore, we now show that PKG represses glucose stimulation of TSP1 transcription and TGF-␤ activity through down-regulation of USF2 protein and DNA binding, implicating USF2 is a significant transcriptional regulator of diabetic complications.

EXPERIMENTAL PROCEDURES
Chemicals, Reagents, and Antibodies-RPMI 1640 medium with Lglutamine without glucose was purchased from Invitrogen. Insulintransferrin-sodium selenite liquid media supplement, minimal essential medium nonessential amino acid solution, and sodium pyruvate solution were purchased from Sigma. Synthetic oligonucleotides were purchased from Qiagen (Valencia, CA). Wild type and mutant USF consensus oligonucleotides and antibodies used for supershift assay, including USF-1 (sc-229, sc8983), USF-2 (sc-862), c-Myc (sc-674), Max (sc-197), were purchased from Santa Cruz Biotechnology, Inc. (Santa, Cruz, CA). Monoclonal antibody 133, raised against human platelet TSP1 stripped of TGF-␤ activity, was purified by our laboratory in a joint effort with the University of Alabama at Birmingham Hybridoma Core Facility (34). Luciferase assay reagent, passive lysis buffer, and dual-luciferase reporter assay system were purchased from Promega (Madison, WI). Human recombinant TGF-␤1 was purchased from R&D Systems, Inc. (Minneapolis, MN)). Neomycin sulfate (G418) was obtained from ICN Biomedicals Inc. (Aurora, Ohio). MAPK kinase (MEK) inhibitor PD 98059, p38 MAPK inhibitor SB 202190, c-Jun aminoterminal kinase inhibitor SP600125, and PKC inhibitor bisindolylmaleimide I were purchased from Calbiochem. Dominant negative plasmids and control plasmids for MEK and p38 MAPK were obtained from Dr. Lin Mei from the University of Alabama at Birmingham. Chromatin immunoprecipitation assay kit was obtained from Upstate Biotechnology (Lake Placid, NY).
Cell Culture-Primary rat mesangial cells (RMCs) (passage 3) were a generous gift from Dr. Anne Woods, University of Alabama at Birmingham. As described previously (19), RMCs were stably transfected with pcDNA6/TR and pcDNA4/TO/myc-His C/CD (a construct of the catalytic domain of PKG-I␣). These stable transfectants were labeled as RMCs(tr/cd) and cultured in RPMI 1640 medium supplemented with 20% heat-activated fetal bovine serum, 5 mM D-glucose, 2 mM L-glutamine, 1% (v/v) nonessential amino acids, 2 mM sodium pyruvate, 10 g/ml transferrin, 5 ng/ml sodium selenite, 0.6 IU, 1.5 g/ml blasticidin, and 250 g/ml zeocin. Serum-free RPMI 1640 media with 5 g/ml transferrin, 5 ng/ml sodium selenite, and 5 mM D-glucose were used to quiescent RMCs(tr/cd). 1 g/ml tetracycline was used to induce the expression of the catalytic domain of PKG to increase PKG activity.
Mink lung epithelial cells (MLECs-clone 32) stably transfected with the TGF-␤ response element of the human plasminogen activator inhibitor-1 gene promoter fused to firefly luciferase reporter gene were a generous gift from Dr. D. B. Rifkin (New York University Medical Center). Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% calf serum, L-glutamine, and 200 g/ml G418.
TSP1 Promoter-Reporter Constructs-A luciferase reporter plasmid containing the Ϫ2033 to ϩ750 region of the human TSP1 gene promoter was generously provided by Dr. Paul Bornstein (University of Washington) and used as a template. Eight deletion mutants of TSP1 (Ϫ1172), TSP1 (Ϫ1112), TSP1 (Ϫ1052), TSP1 (Ϫ992), TSP1 (Ϫ932), TSP1 (Ϫ878), TSP1 (Ϫ548), and TSP1 (Ϫ330) were prepared using forward primers derived from different 5Ј positions (the numbers in parentheses contain KpnI site) of the human TSP1 promoter and the same reverse primer (5Ј-gctagctgtagcaggaagcacaagag-3Ј, containing NheI site). The PCR was performed using TSP (Ϫ2033/ϩ750) as a template and high fidelity Pfx polymerase from Invitrogen. PCR products were digested with KpnI and NheI and subcloned into a luciferase plasmid PGL 3-Enhancer vector (Promega), and their sequences were confirmed by sequencing.
Transfection and Luciferase Assay-RMCs(tr/cd) were seeded into 6-well plates at a density of 3ϫ10 4 cells/ml for 1 day and made quiescent in serum-free RPMI 1640 media with or without tetracycline (1 g/ml) for 2 days. Then cells were transiently transfected using Effectene transfection reagent (Qiagen) with (1 g) TSP1 promoter luciferase reporter plasmid. For co-transfection experiments, cells were transiently transfected with 1 g of TSP1 promoter luciferase reporter plasmid as well as different amounts of USF2 expression vector or empty vector (pSG5) (generous gifts from Dr. Michele Sawadogo, the University of Texas, MD Anderson Cancer Center). pRL-SV40 (0.02 g) (Promega) was used as an internal control. Transfected cells were treated with normal (5 mM) or high (30 mM) glucose for 1 day, and the luciferase activities were assayed using the dual-luciferase assay kit (Promega) according to the manufacturer's directions.
Nuclear Extract Preparation-RMCs(tr/cd) were cultured in tissue culture dishes with growth media (RPMI 1640 media containing 20% fetal bovine serum). After reaching 80 -85% confluence, cells were made quiescent by changing into serum-free RPMI 1640 media for 2 days in the presence or absence of 1 g/ml tetracycline (to induce the expression of catalytic domain of PKG resulting in increased PKG activity) (19) and then cultured in serum-free RPMI 1640 media with 5 or 30 mM glucose. After 24 h, cells were washed with cold phosphate-buffered saline (pH 7.4), scraped off the dishes, and spun down. Then nuclear extracts were prepared as described previously (35). Briefly, cell pellets were resuspended in 500 l of hypotonic lysis buffer (10 mM Hepes-KOH (pH 7.9), 10 mM KCl, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol, 0.5 mM phenymethylsulfonyl fluoride) and lysed for 15 min on ice. Then 32 l of 10% Nonidet P-40 was added to the suspension and incubated on ice for 10 min. Nuclei were spun down and re-suspended in 100 l of nuclear extraction buffer (20 mM Hepes-KOH (pH 7.9), 420 mM NaCl, 1.2 mM MgCl 2 , 0.2 mM EDTA, 25% glycerol, 0.5 mM dithiothreitol, 0.5 mM phenymethylsulfonyl fluoride, 5 g/ml pepstatin, and 5 g/ml leupeptin). The nuclei were incubated with the extraction buffer on ice for 20 min and spun down. The supernatant (nuclear extract) was stored at Ϫ80°C. Protein concentration of nuclear extracts was determined using the Bio-Rad protein assay kit.
Western Blot-For detection of USF1 and USF2, equal amounts of protein in the nuclear extract from RMCs(tr/cd) were subjected to 10% SDS-PAGE gel under reducing conditions. After electrophoretic transfer to nitrocellulose membranes and blocking, the membranes were incubated with polyclonal anti-USF1 and anti USF2 antibodies (1:500 dilution) for 1 h at room temperature. After intensive washing, secondary antibody (1:10,000 dilution) was used for the detection of immunoreactive bands with the enhanced chemiluminescence detection system (Pierce). Equal amounts of protein in the conditioned media from RMCs(tr/cd) were subjected to 8% SDS-PAGE gel under reducing conditions and transferred to nitrocellulose membranes to detect TSP1 levels by using monoclonal anti TSP1 antibody (antibody 133) as described previously (18). Equal loading and transfer of protein samples were assayed by staining the membranes with Ponceau S.
Electrophoretic Mobility Shift Assay (EMSA) and Supershift Assay-Oligonucleotides (oligos) used for EMSA were end-labeled with 32 P by T4 polynucleotide kinase (Invitrogen). 2-5 ϫ 10 4 cpm probe was incubated with 3 g of nuclear extracts in a 20-l volume of binding reaction buffer (10 mM Tris-Cl (pH 7.5), 100 mM NaCl, 10% glycerol, 50 ng/ml poly(dI/dC)) on ice for 20 min. In competition experiments, a 50ϫ and/or 100ϫ excess amount of unlabeled competitor was mixed with the labeled probe before being added to the binding mixture. The binding reaction was then allowed to proceed for 20 min on ice. In supershift experiments, 3 g of control IgG or 3 g of specific antibodies were incubated with 3 g of nuclear extracts in a 20-l volume of binding reaction for 20 min on ice followed by incubation with the labeled probe on ice for an additional 30 min. All binding mixtures were separated by 4 -20% gradient Tris-buffered TBE gels (Bio-Rad). The gels were dried and exposed to film.
Chromatin Immunoprecipitation Assay-To detect the in vivo asso-ciation of nuclear proteins with the TSP1 promoter, the chromatin immunoprecipitation assay was conducted as described by Upstate Biotechnology with some modifications. In brief, primary mesangial cells (p3-6) or stably transfected mesangial cells (RMCs(tr/cd)) were cultured in a 10-cm culture dish. After being made quiescent in serumfree RPMI 1640 media for 2 days, cells were treated with normal glucose (5 mM) or high glucose (30 mM) for 24 h. After treatment, protein-DNA complexes were fixed by 1% formaldehyde in phosphatebuffered saline. The fixed cells were washed and lysed in SDS lysis buffer with protease inhibitors and sonicated on ice. After centrifugation at 500 ϫ g for 1 min, one portion of the precleared supernatant was used as DNA input control, and the remaining supernatant was subdivided into aliquots and then incubated overnight at 4°C with nonimmune rabbit immunoglobulin G (IgG; Santa Cruz) or IgG antibodies to USF1 or USF2. The immunoprecipitated complexes of antibody-protein-DNA were collected using a protein A slurry, washed successively with low salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl (pH 8.1), 150 mM NaCl), high salt buffer (same as the low-salt buffer but with 500 mM NaCl), LiCl buffer (0.25 M LiCl, 1% Nonidet P-40, 1% deoxycholate, 1 mM EDTA, 10 mM Tris-HCl (pH 8.1), and Tris-EDTA (pH 8.0), and then eluted with elution buffer (1% SDS, 100 mM NaHCO 3 ). The cross-linking of protein-DNA complexes was reversed by incubation with 5 M NaCl at 65°C for 4 h, and DNA was digested with 10 mg of proteinase K (Sigma)/ml for 1 h at 45°C. The DNA was then extracted with phenol-chloroform, and the purified DNA pellet was resuspended in H 2 O and subjected to PCR amplification with the forward primer, 5Ј-GTCTGCTCTGTAAATAGCTG-3Ј, and the reverse primer, 5Ј-GTGCTTCTTAACGTGACTCC-3Ј, which were specifically designed from the TSP1 promoter. The 80-bp PCR products were resolved by 3.5% agarose-ethidium bromide gel electrophoresis, visualized by UV.
Site-directed Mutagenesis-Point mutations were introduced into the USF binding site (Ϫ924 to Ϫ919) in the TSP1 promoter reporter construct (Ϫ2033/ϩ750) using a QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA). The primers used to mutate the USF binding site are 5Ј-gactctggagccagagagttctgcaaattctcc-3Ј and 5Јggagaatttgcagaactctctggctccagagtc-3Ј (purchased from Qiagen). PCR was performed in a 50-l volume with Pfu polymerase, 10 ng of DNA template TSP1 (Ϫ2033/ϩ750), and 125 ng of each of the primers using the following conditions: 95°C for 30 s for 1 cycle and 95°C for 30 s, 55°C for 1 min, and 68°C for 12 min for 18 cycles. The PCR products were treated with DpnI (10 units) for 60 min at 37°C. XL1-blue supercompetent cells were transformed with DpnI-treated PCR mixtures as described in the instruction manual and plated on ampicillin plates. Plasmids were prepared from individual colonies and sequenced to confirm the correctness of introduced mutations.
TGF-␤ Assay-Total and active TGF-␤ levels in the condition media were assayed using the plasminogen activator inhibitor-1/luciferase assay as described previously (18). Briefly, mink lung epithelial cells were plated into 24-well tissue culture plates at 1.6 ϫ 10 5 cells/ml and incubated for 3-4 h for optimal attachment. After aspiration of the growth medium from the attached cells, 500 l of condition media were added. To measure total TGF-␤, conditioned media samples were heatactivated for 3 min at 100°C. Samples and TGF-␤ standards were incubated overnight at 37°C. After incubation, cells were lysed with lysis buffer at room temperature for 20 min. Lysates were analyzed for luciferase activity using a microplate Luminometer (Berthold Detection Systems) after the injection of 100 l of substrate solution and recorded as relative light units. The mean values of triplicate samples were converted into concentrations of TGF-␤ (picomolar) using a standard curve obtained with human recombinant TGF-␤1.
Statistical Analysis-Data are expressed as the mean Ϯ S.D. Statistical evaluation of the data was performed using the paired Student's t test, considering a p value of Ͻ0.05 as significant.

Glucose-mediated Increases in TSP1 Promoter Activity in RMCs(tr/cd) Are Supported by TSP1 Promoter Region Ϫ1172 to
Ϫ878 -We previously showed that high concentrations of glucose mediate increases in TSP1 expression and TSP1-dependent TGF-␤ bioactivity through down-modulation of NO-dependent PKG signaling (18). Moreover, we generated stable transfectants expressing the catalytic domain of PKG in mesangial cells using an inducible expression system (tetracycline-induced gene expression (tet-on)). With the expression of the catalytic domain of PKG to increase PKG activity, glucoseinduced TSP1 expression was inhibited at the transcriptional level (19). To identify the promoter elements that regulate TSP1 gene transcription in mesangial cells in response to high glucose concentrations and increased PKG activity, a series of TSP1 promoter-luciferase reporter constructs was generated (Fig. 1A) and transiently transfected into RMCs(tr/cd). The promoter activity was measured by assaying the luciferase activity as described under "Experimental Procedures" and was normalized to Renilla luciferase activity. As shown in Fig.  1B, left panel, the longest construct, TSP (Ϫ2033), gave rise to a 2.3-fold increase in TSP1 promoter activity in response to high concentrations of glucose (30 mM). Deletion of a 295-bp region (Ϫ1172 to Ϫ878) totally abolished high glucose responsiveness, suggesting that this 295-bp region is important for high glucose-induced TSP1 transcription. Increased PKG activity inhibited 30 mM glucose-induced TSP1 promoter activity to basal levels (5 mM glucose) (Fig. 1B, right panel). These data suggest that the region between Ϫ1172 and Ϫ878 in the TSP1 promoter is involved in PKG-mediated repression of TSP1 transcription under high glucose conditions.
Localization of a 55-bp TSP1 Promoter Region (Ϫ932 to Ϫ878) That Binds glucose-induced Nuclear Proteins and Mediates the Repression of TSP1 Transcription by Increased PKG under High Glucose Conditions-To characterize the cis-acting elements in the 295-bp region (Ϫ1172 to Ϫ878), first we determined whether the 295-bp region binds any nuclear proteins in response to high glucose. Five overlapping oligonucleotides (Oligo I, II, III, IV, and V) spanning the entire 295-bp TSP1 promoter region were synthesized ( Fig. 2A), labeled, and used as probes to perform EMSAs with nuclear extracts from normal or high glucose-treated stably transfected mesangial cells (RMCs(tr/cd)) in the absence or presence of tetracycline (Fig.  2B). Only oligo V gave rise to a major band in the EMSA (lane 1). This band was significantly enhanced by high glucose treatment (lane 2). Increased PKG activity diminished this protein-DNA complex (lane 4), suggesting that oligo V contains cis-acting elements involved in PKG-mediated repression of glucose-induced TSP1 transcription. The specificity of this protein-DNA complex was confirmed in the competition assays using cold oligo V as the competitor in the EMSA to abrogate the formation of this complex (Fig. 2C). To rule out the possibility that the stable transfection of mesangial cells alters cell regulation, primary cultured mesangial cells were also used in the above EMSA assay to confirm the association of oligo V with glucose-induced nuclear proteins. Early passage primary mesangial cells (p3-p6) were treated with normal or high glucose media for 24 h, and then nuclear proteins were extracted, and EMSA were performed. As shown in Fig. 2D, oligo V also specifically gave rise a major band in EMSA, which was significantly enhanced by high glucose (30 mM) treatment of primary mesangial cells.
To determine whether the 55-bp TSP1 promoter region (Ϫ932 to Ϫ878) mediates the repression of TSP1 transcription by increased PKG activity under high glucose conditions, four additional deletion mutants of TSP1 promoter: TSP (Ϫ1112), TSP (Ϫ1052), TSP (Ϫ992), and TSP (Ϫ932) were generated (Fig. 3A). These constructs were transiently transfected into mesangial cells, and promoter activity was evaluated by quantifying the luciferase activity (Fig. 3B). Consistent with the EMSA results (Fig. 2B), transfection studies with these mutants indicate that the 55-bp region from Ϫ932 to Ϫ878 is necessary for glucose-induced TSP1 transcription and that it is also important for PKG repression of TSP1 transcription.
Identification of an 18-bp Sequence of the TSP1 Promoter That Specifically Binds Glucose-induced USF1 and USF2-To identify the nuclear proteins that bind to oligo V (55-bp region of TSP1 promoter Ϫ932 to Ϫ878), we performed a computer analysis using the GCG program, which revealed that this 55 bp contains putative binding sequences for several transcription factors, including NF-1, c-Jun, USF, heat shock transcription factor, C/EBP, and E12. To specifically identify which transcription factors might be important for glucose regulation of TSP1 transcription, we performed competition assays to identify the specific nuclear protein binding sequence within this 55-bp region. Four oligonucleotides (A, B, C, and D) derived from the 55-bp oligo V were synthesized and used as competitors in the EMSA. As shown in Fig. 4B, only oligo A competed efficiently for the nuclear protein binding, suggesting that an 18-bp sequence within the 55-bp region specifically binds the nuclear proteins induced by high glucose concentrations.
The 18-bp sequence contains a CAGATG motif at Ϫ924 to Ϫ919, which resembles the CANNTG motif of classical E-box cognates and binds both Myc family members (36) and USF proteins (37,38). This suggests that the nuclear proteins binding to the 18-bp sequence might be members of the Myc family or USF proteins. Thus, we immunologically probed the protein-DNA complex with antibodies against c-Myc, c-Max, USF-1, and USF-2 in gel supershift assays as described under "Experimental Procedures." For these experiments, gel shifts were carried out with radiolabeled duplex oligo V (55 bp). As shown in Fig. 5A, only anti-USF1 and -USF2 antibodies supershifted the band (lanes 5, 6, 11, and 12), indicating that the nuclear proteins binding to the 18-bp sequence are USF1 and USF2, possibly as heterodimers. Competition studies showed that unlabeled USF oligonucleotide efficiently competes for USF binding, whereas mutant USF oligonucleotide failed to compete (Fig. 5B). These data strongly suggest that the 18-bp sequence in the TSP1 promoter has a USF binding site.
Having demonstrated the binding of transcription factors USF1 or USF2 to the TSP1 promoter in vitro, we further analyzed the in vivo binding of USF to TSP1 promoter by the chromatin immunoprecipitation assay. Primary rat mesangial cells or the stable mesangial cell line (RMC(tr/cd)) was cultured and treated with normal (5 mM) or high glucose (30 mM) for 24 h. After treatment, formaldehyde was added to cross-link the DNA-protein complexes in vivo. After sonication, immunoprecipitation, and reversal of cross-linking, the DNA was purified and used as a template for PCR amplification using primers that encompass the region from Ϫ958 to Ϫ878 containing the putative USF binding site (Fig. 6A). In primary cultured mesangial cells, high glucose treatment caused a marked enrichment of USF1-or USF2-associated TSP1 promoter DNA (Fig. 6B). In stably transfected mesangial cells, glucose also induced increased association of USF1 or USF2 with the TSP1 promoter (Fig. 6C). Moreover, increased PKG activity reduced glucose-induced USF association with the TSP1 promoter but had no effect on USF association with the TSP1 promoter under normal glucose conditions (Fig. 6C). As a negative control, chromatin immunoprecipitation was performed with nonimmune IgG. These data show that the association of USF1 or USF2 with the TSP1 promoter in mesangial cells is enhanced by high glucose treatment, and increased PKG activity diminishes this protein-DNA association. The 18-bp USF Binding Sequence Is Functionally Involved in Glucose-induced TSP1 Transcription-To determine whether the 18-bp USF binding sequence is functionally involved in PKG-mediated repression of glucose-induced TSP1 transcription, we introduced point mutations in the 18-bp USF binding site. In the CAGATG motif, TG was converted to GA (Fig. 7A). Competition assays showed that mutant oligo V failed to com-

FIG. 1. Glucose-mediated increase in TSP1 promoter activity in RMCs(tr/cd) is supported by TSP1 promoter region ؊1172 to ؊878.
A, construction of five deletion mutants of TSP1 promoter: TSP (Ϫ2033), TSP (Ϫ1172), TSP (Ϫ878), TSP (Ϫ548), and TSP (Ϫ330). B, RMCs(tr/cd) were cultured in 6-well plates and made quiescent in serum-free media for 2 days in the presence or absence of tetracycline. Then cells were transiently cotransfected with a series of TSP1 promoter-luciferase reporter constructs as shown in A and internal-control plasmid pRL-SV40. Transfected cells were treated with normal (5 mM) or high glucose (30 mM) media for 1 day. The promoter activity was quantified by assaying luciferase activity as described under "Experimental Procedures" and was normalized to Renilla luciferase activity of internal control. The experiments were repeated three times, and the representative result is shown. Data are represented as the mean of three replicates Ϯ S.D. RLU, relative light units. pete for USF binding (Fig. 7A), indicating that CAGATG in the 18-bp sequence is a USF binding site. The same mutation was introduced into the TSP-luciferase reporter construct (Ϫ2033) to generate a reporter deficient in the USF binding site (mTSP1 (Ϫ2033)). mTSP1 (Ϫ2033) was transiently transfected into RMCs(tr/cd). As expected, high glucose (30 mM) significantly induced TSP1 (Ϫ2033) activity, which was inhibited by increased PKG activity (Fig. 7B). In contrast, high glucose concentrations failed to induce the reporter activity in the promoter construct with the mutated USF binding site. PKG activity did not affect activity of the mutant promoter.
USF2 Protein Levels Are Regulated in Mesangial Cells by Glucose and PKG-We showed that increased PKG activity significantly diminished glucose-induced USF binding (Fig.  2B). To identify whether these changes in USF binding activity reflect changes in the nuclear accumulation of USF1 and/or USF2 protein levels, we performed immunoblot analyses of extracts prepared from RMCs(tr/cd) treated with either normal or high glucose concentrations in the absence or presence of tetracycline as described under "Experimental Procedures." As

FIG. 2. A 55-bp TSP1 promoter region (؊932 to ؊878) binds glucose-induced nuclear proteins.
A, the diagram of 5 overlapping oligos (I, II, III, IV, V) spanning the 295-bp TSP1 promoter region (Ϫ1172 to Ϫ878) was shown. B, oligos I-V shown in A were labeled and used as probes to perform EMSA with nuclear extract (N.E.) from RMCs(tr/cd) treated with normal (NG) or high glucose (HG) media for 1 day in the presence or absence of tetracycline (Tet). Only oligo V (TSP1 promoter region Ϫ932 to Ϫ878) assembles a major protein-DNA complex in EMSA. The experiments were repeated three times, and the representative result is shown. C, a competition assay was performed using labeled oligo V as probe, and nuclear extracts (N.E.) from RMCs(tr/cd) treated with high glucose media in the absence or presence of tetracycline. Excess cold oligo V was added as competitor. As the control, no competitor was added in lane 1 and lane 3. D, oligo V was labeled and used as the probe to perform the EMSA with nuclear extract from primary mesangial cells treated with 30 mM glucose for 24 h. Excess unlabeled oligo V was used as the competitor. The experiments were repeated four times, and the representative result is shown. shown in Fig. 8, treatment of RMCs(tr/cd) with high glucose up-regulated USF2 (panel B) but not USF1 (panel A) protein accumulation. Moreover, increased PKG activity reduced glucose-induced USF2 to basal levels (5 mM glucose treatment) (panel B in Fig. 8), suggesting that regulation of TSP1 promoter USF binding activity in mesangial cells by high glucose and PKG reflects in part the regulation of USF2 protein accumulation.
In mesangial cells, high glucose environments activate mitogen-activated protein kinase (MAPK) pathways including extracellular signal-regulated kinase, p38 MAPK and c-Jun amino-terminal kinase (39 -42). To determine whether glucose signaling through MAPK pathways is involved in USF2 expression, RMCs(tr/cd) were incubated with the MEK inhibitor PD98059, the p38 MAPK inhibitor SB202190, dominant nega-tive MEK, and p38 MAPK plasmids or the c-Jun amino-terminal kinase inhibitor SP600125 in the presence of 5 or 30 mM glucose media for 24 h, and then nuclear USF2 protein levels were assayed by immunoblotting. As shown in Fig. 8C, without tetracycline induction, both PD98059 (10 M) and SB202190 (10 M) compounds inhibited 30 mM glucose-mediated USF2 up-regulation. Neither inhibitor showed a significant effect on basal USF2 levels (5 mM glucose treatment). However, with tetracycline induction of PKG activity, neither PD98059 nor SB202190 compounds affected USF2 protein levels under normal or high glucose conditions. Similar results were obtained by using dominant negative MEK and p38 MAPK plasmids to transfect mesangial cells (Fig. 8, D and E). The c-Jun aminoterminal kinase inhibitor had no effect on USF2 protein levels FIG. 5. Identification of USF1 and USF2 binding to oligo V (TSP1 promoter region ؊932 to ؊878). A, supershift assays. RMCs(tr/cd) were cultured and made quiescent for 2 days in the absence or presence of tetracycline (Tet). Then cells were treated with high glucose (HG) media for 1 day, and nuclear proteins were extracted as described under "Experimental Procedures." Oligo V was labeled and used as the probe. Supershift assays were performed in the presence of the indicated antibodies and control IgG (3 g each). Anti-USF1 and USF2 antibodies disrupt and supershift the DNA-protein complex. The blots shown are the representative of three separate experiments. N.E., nuclear extracts. B, competition assays. EMSA was performed using the labeled oligo V and nuclear extracts from RMCs(tr/cd) treated with high glucose media for 1 day in the presence or absence of tetracycline. Excess cold USF consensus oligonucleotide or USF mutant oligonucleotide (mUSF) was added as a competitor. Only USF consensus oligonucleotide competes for the binding of oligo V to glucose-induced nuclear proteins. These experiments were repeated four times, and the representative result is shown. under normal or high glucose conditions in the absence or presence of tetracycline (data not shown). These data suggest that extracellular signal-regulated kinase and p38 MAPK, but not the c-Jun amino-terminal kinase pathways are involved in glucose-mediated up-regulation of nuclear USF2 levels. However, these pathways do not appear to be important for PKG mediated down-regulation of nuclear USF2 levels.
Glucose-induced activation of the MAPK pathway can be mediated by increased PKC activity in mesangial cells (43,44). Moreover, glucose-induced PKC activation has been shown to up-regulate TSP1 expression (45). To address the role of PKC in the regulation of high glucose-induced USF2 expression, the PKC inhibitor (bisindolylmaleimide I (Bis)) was used. In preliminary dose-response experiments, 100 nM Bis was found to be an optimal concentration (data not shown). As shown in Fig.  8F, without tetracycline induction Bis (100 nM) compound inhibited 30 mM glucose-mediated USF2 up-regulation. However, with tetracycline induction of PKG activity, Bis compound did not affect USF2 protein levels under normal or high glucose conditions. This result suggests that PKC activity is involved in glucose-induced USF2 expression.
Overexpression of USF2 Reverses PKG-mediated Repression of Glucose-induced TS P1 Expression (Promoter Activity and Protein Levels) and TGF-␤ Bioactivity-To examine the effects of USF2 on TSP1 promoter activity and TGF-␤ bioactivity, we transiently co-transfected the expression vector for USF2 with TSP1-luciferase reporter construct (TSP (Ϫ2033)) in RMCs(tr/ cd). 24 h after transfection, conditioned media were collected to analyze TSP1 protein levels and measure TGF-␤ activity, and cells were harvested to determine the promoter activity by assaying the luciferase activity as described under "Experimental Procedures." USF2 overexpression significantly augmented TSP1 promoter activity and TSP1 protein levels under normal glucose conditions with or without tetracycline induction (data not shown), showing that up-regulation of USF2 enhances TSP1 gene and protein expression. Furthermore, overexpression of USF2 protein abolished PKG-mediated re-pression of TSP1 promoter activity (Fig. 9A) and protein levels (Fig. 9B) under high glucose conditions in a concentration-dependent manner in cells treated with tetracycline to induce PKG activity. Similarly, the PKG-mediated decrease in active TGF-␤ was reversed by the overexpression of USF2 protein (Fig. 9C). However, total TGF-␤ production was not altered by the overexpression of USF2 protein in mesangial cells (Fig.  9D). Taken together, these data suggest that decreased protein levels and DNA binding activity of USF2 mediates the inhibitory effect of PKG on glucose-induced TSP1 gene expression and TSP1-dependent TGF-␤ activation in mesangial cells. DISCUSSION Accumulating evidence suggests that TSP1 plays an important role in diabetes and diabetic nephropathy (17, 46 -51). Previously we demonstrated that high glucose concentrations (30 mM) mediate increases in TSP1 expression and TSP1-dependent TGF-␤ bioactivity in glomerular mesangial cells through down-modulation of NO/cGMP-dependent protein kinase signaling (18). Moreover, increased PKG activity repressed glucose-induced TSP1 expression at the level of transcription (19). In this present study, we further investigated the molecular mechanisms by which glucose induces and PKG represses TSP1 gene transcription in glomerular mesangial cells. We identified a glucose-responsive region in TSP1 promoter from Ϫ932 bp to Ϫ915 bp (18bp), which is also involved in PKG-mediated repression of TSP1 transcription under high glucose conditions. Furthermore, we showed that this 18-bp region regulates TSP1 promoter activity by binding USF1 and USF2. Glucose stimulation of USF2 protein expression through PKC, extracellular signal-regulated kinase, or p38 MAPK pathways leads to increased DNA binding activity and contributes to glucose-induced TSP1 transcription and increased TGF-␤ bioactivity in our studies. PKG inhibits the effects of glucose on TSP1 expression and TGF-␤ activation through down-regulation of glucose-induced USF2 protein levels.
USFs belong to the basic helix-loop-helix leucine zipper FIG. 6. Chromatin immunoprecipitation analysis of USF interaction with the TSP1 promoter in vivo. A, schematic representation showing the putative USF binding sites. Two arrows indicate primers used for amplifying the region from Ϫ958 to Ϫ878 spanning the putative USF binding site. Formaldehyde-cross-linked chromatin isolated from primary mesangial cells (B) or a stable cell line of mesangial cells (C) was immunoprecipitated (IP) with antibodies and subjected to PCR as described under "Experimental Procedures." PCR products were electrophoresed on 3.5% agarose gel. Input represents 0.2% of chromatin applied for immunoprecipitation. The experiments were repeated three times, and the representative result is shown. NG, normal glucose; HG, high glucose; tet, tetracycline.
(b-HLH-LZ) family of transcription factors (52). USFs (USF1 and USF2) were originally identified by their ability to bind to the adenovirus major late promoter (37). USF1 and USF2 are also related to the Myc family of transcription factors and have a similar polypeptide structure and a similar DNA binding specificity (53,54). USF1 and USF2 are encoded by two distinct genes in human, rat and mouse, and these proteins have molecular masses of 43 and 44 kDa, respectively (54 -57). Structurally, USF1 and USF2 are related with a highly conserved COOH-terminal domain responsible for their dimerization and DNA binding (52,58,59). The major USF species present in most tissues and cell lines is the heterodimer of USF1 and USF2. USF1 homodimers are less abundant, and USF2 homodimers are usually quite rare (58 -60). Both USF1 and USF2 have been shown to bind to the canonical sequence CANNTG (an E-box motif) as either homo-or heterodimers to regulate gene transcription (55,59,61,62).
Although USF1 and USF2 genes are ubiquitously expressed in mammalian cells, the relative abundance of USF1 and USF2 transcripts and protein levels varies among different cell types (59,61). It has also been shown that the function of USFs is modulated in a cell-specific manner (62). In liver, Vaulont and co-workers (63) demonstrate the contribution of USF1 and USF2 to hepatic glucose responses in mice possessing homologous disruption of these two transcription factor genes. They show that USF1 and USF2 bind in vitro glucose/carbohydrate response elements of glycolytic and lipogenic genes, such as L-type pyruvate kinase and spot 14 genes, and mediate glucose-induced transcriptional activation of these genes (63). Furthermore, in the liver of USF1-deficient mice, enhanced USF2 expression and an increase in the levels of USF2 homodimers compensated for lack of USF1 in the transcriptional FIG. 7. Mutation of USF binding site-reduced glucose induced TSP1 promoter activity. A, the competition assay was performed using labeled oligo V as the probe and 100ϫ cold wild type (WT) oligo V or mutant oligo V as the competitor. Nuclear extract (N.E.) for this assay were from RMCs(tr/cd) treated with high glucose media (HG) in the absence or presence of tetracycline (Tet). The results shown are the representative of three separate experiments. B, mutation of TSP (Ϫ2033) (mTSP) was generated by introducing two point mutations as shown in Fig. 6A. RMCs(tr/cd) were cultured and made quiescent in serum-free media in the absence or presence of tetracycline for 2 days. Then cells were transiently transfected with wild type TSP1 (Ϫ2033) or mutant TSP1 (Ϫ2033) as well as internal control pRL-SV40 under normal (NG) or high glucose conditions for 1 day. The promoter activity was quantified by assaying luciferase activity as described under "Experimental Procedures" and normalized to the Renilla luciferase activity of internal control. The experiments were repeated three times, and representative data are shown. Data are the means of three replicates Ϯ S.D. RLU, relative light units. , p Ͻ 0.05 for 30 mM glucose with PD98059 or SB202190 compounds without tetracycline versus 30 mM glucose without PD98059 or SB202190 compounds without tetracycline. RLU, relative light units.
response of L-type pyruvate kinase and spot 14 genes to glucose. However, in USF2 Ϫ/Ϫ mice, an impaired glucose responsiveness was observed (64), suggesting that USF2 is essential for transcriptional responses of liver genes to glucose. Our present work shows for the first time that glucose upregulates USF2 protein accumulation in glomerular mesangial cells and enhances protein-DNA interactions on the TSP1 promoter, resulting in increased TSP1 protein levels and TSP1-de-pendent TGF-␤ activation. Furthermore, when the PKC, extracellular signal-regulated kinase, or p38 MAPK pathways are inhibited, glucose-induced USF2 expression is down-regulated, suggesting that the PKC, extracellular signal-regulated kinase, and p38 MAPK pathways are important for glucose-mediated stimulation of USF2 expression. In contrast to USF2, we failed to observe glucose stimulation of USF1 protein accumulation. This differs from the results of Bidder et al. (65) and Weigert et al. (66). FIG. 9. Overexpression of USF2 protein reverses PKG-mediated repression of glucose-induced TSP1 expression (promoter activity and protein levels) and TGF-␤ activity. RMCs(tr/cd) were cultured and made quiescent in serum-free media for 2 days in the absence or presence of tetracycline (Tet). Then cells were co-transfected with 1 g of TSP (Ϫ2033) luciferase construct and expression vectors for USF2 in normal (NG) or high glucose (HG) media as indicated for 1 day. Empty expression vector (pSG5) was used as control. As described under "Experimental Procedures," conditioned media were collected, TSP1 protein levels were assayed by immunoblotting, and TGF-␤ activity was measured in the PAI/1 promoter luciferase assay; cells were harvested, and the promoter activity was quantified by assaying luciferase activity. The experiments were repeated three times, and a representative result is shown. Data are represented as the mean of three replicates Ϯ S.D. *, p Ͻ 0.05 for 30 mM glucose versus 5 mM glucose; **, p Ͻ 0.05 for high glucose with tetracycline versus NG with tetracycline; #, p Ͻ 0.05 for high glucose with tetracycline versus high glucose without tetracycline; ##, p Ͻ 0.05 for high glucose ϩ USF versus high glucose ϩ empty vector. RLU, relative light units. Bidder et al. (65) show that 30 mM glucose treatment of aortic smooth muscle cells up-regulates USF1 protein levels with no alteration of USF2 protein levels. Weigert et al. (66) show that 30 mM glucose treatment of porcine mesangial cells enhanced both USF1 and USF2 protein levels through the hexosamine biosynthetic pathway. The reasons for these differences are not clear. However, these findings suggest that USF regulation is complex, with possible cell-specific and species-specific regulation of USF proteins by glucose.
In our studies, USF2 protein levels are down-regulated by increased PKG activity, which contributes to decreased USF binding activity, resulting in repression of TSP1 gene transcription under high glucose conditions. Regulation of USF2 at the level of transcription, mRNA stabilization, translation, protein degradation, or some combination thereof probably contributes to the decreased USF2 protein levels mediated by PKG. We found no evidence for the involvement of PKC, extracellular signal-regulated kinase and p38 MAPK pathways in PKG-mediated down-regulation of USF2, and the mechanisms by which PKG down-regulates USF2 remain to be determined.
Our studies show that overexpression of USF2 reverses PKGmediatedrepressionofglucose-inducedTSP1expressionandTSP1dependent TGF-␤ activation. However, total TGF-␤ production is not altered by the overexpression of USF2 protein in mesangial cells (Fig. 9D). This differs from the recent studies of Weigert et al. (66), who showed that overexpression of USF1 and/or USF2 enhanced TGF-␤1 promoter activity in a simian virus 40 (SV40)transformed mouse mesangial cell line. Experimental conditions such as transfection efficiency, SV40 transformation, or speciesspecific regulation of USF-mediated TGF-␤ expression in mesangial cells might contribute to this difference.
In summary, our data provide the first evidence that USF2 mediates glucose stimulation of TSP1 gene transcription and protein expression and TSP1-dependent TGF-␤ bioactivity in mesangial cells, suggesting that USF2 is an important transcriptional regulator of diabetic nephropathy. Moreover, increased PKG activity down-regulates glucose-induced USF2 protein, resulting in a decrease in protein-DNA activity, contributing to PKG-mediated repression of glucose-induced TSP1 gene expression and TSP1-dependent TGF-␤ activation.