Phosphorylation of Fli1 at Threonine 312 by Protein Kinase C δ Promotes Its Interaction with p300/CREB-Binding Protein-Associated Factor and Subsequent Acetylation in Response to Transforming Growth Factor β

Abstract

Previous studies have shown that transforming growth factor β (TGF-β)-induced collagen gene expression involves acetylation-dependent dissociation from the human α2(I) collagen (COL1A2) promoter of the transcriptional repressor Fli1. The goal of this study was to elucidate the regulatory steps preceding the acetylation of Fli1. We first showed that TGF-β induces Fli1 phosphorylation on a threonine residue(s). The major phosphorylation site was localized to threonine 312 located in the DNA binding domain of Fli1. Using several independent approaches, we demonstrated that Fli1 is directly phosphorylated by protein kinase C δ (PKC δ). Additional experiments showed that in response to TGF-β, PKC δ is recruited to the collagen promoter to phosphorylate Fli1 and that this step is a prerequisite for the subsequent interaction of Fli1 with p300/CREB-binding protein-associated factor (PCAF) and an acetylation event. The phosphorylation of endogenous Fli1 preceded its acetylation in response to TGF-β stimulation, and the blockade of PKC δ abrogated both the phosphorylation and acetylation of Fli1 in dermal fibroblasts. Promoter studies showed that a phosphorylation-deficient mutant of Fli1 exhibited an increased inhibitory effect on the COL1A2 gene, which could not be reversed by the forced expression of PCAF or PKC δ. These data strongly suggest that the phosphorylation-acetylation cascade triggered by PKC δ represents the primary mechanism whereby TGF-β regulates the transcriptional activity of Fli1 in the context of the collagen promoter.

Fli1 is a member of the Ets transcription factor family that was initially identified as a proto-oncogene in Friend virus-induced erythroleukemia in mice (6, 7). In humans, Fli1 is frequently involved in the development of Ewing sarcoma and related subtypes of primitive neuroectodermal tumors (12), suggesting that Fli1 plays an important role in the process of cellular transformation. In normal tissues, Fli1 is preferentially expressed in endothelial and hematopoietic cell lineages (16) and participates in the regulation of development and differentiation of these cell types. Extensive in vitro studies as well as the data obtained from various Fli1 null mice support a crucial role of Fli1 in megakaryocytic differentiation and myelomonocytic, erythroid, and NK cell development (25, 36). Fli1 knockout mice die during embryogenesis with a loss of vascular integrity leading to cerebral hemorrhage, suggesting that Fli1 is involved in the regulation of genes critical for vascular remodeling (14, 36). However, unlike hematopoietic cells, in which the role of Fli1 is well established, little is known about Fli1 function in the vasculature.

Although Fli1 is present in a relatively limited amount in dermal fibroblasts, Fli1 plays a pivotal role in the regulation of extracellular matrix (ECM) genes, including type I collagen (10, 21, 22), tenascin-C (19, 34), ECM-degrading enzyme MMP-1 (20), and the multifunctional matricellular factor CCN2 (28). Most importantly, Fli1 has been shown to be a potent inhibitor of type I collagen production in dermal fibroblasts, and the persistent downregulation of this transcription factor has been implicated in the pathogenesis of cutaneous fibrosis in scleroderma (22). Previous studies have demonstrated that Fli1 occupies the human α2(I) collagen (COL1A2) promoter under physiological conditions but dissociates from the promoter in response to TGF-β stimulation (1, 10, 21). This process is critical for the TGF-β-dependent upregulation of the type I collagen gene, as demonstrated by the gene silencing of Fli1, which was sufficient to induce a robust increase of type I collagen production (28). Our recent study has begun to elucidate the mechanism governing Fli1 interaction with the COL1A2 promoter (1). Upon TGF-β stimulation, p300/CREB-binding protein-associated factor (PCAF) forms a complex with Fli1 leading to the acetylation of Fli1 at lysine 380. Acetylated Fli1 dissociates from the COL1A2 promoter and is targeted for degradation, resulting in a decreased steady-state level of Fli1 protein and a derepression of collagen gene transcription. Conversely, the Fli1 mutant resistant to PCAF-dependent acetylation has an increased DNA binding ability, resulting in a potent inhibition of COL1A2 promoter activity, further underscoring the importance of acetylation in the regulation of Fli1 function.

TGF-β is a multifunctional polypeptide growth factor that regulates cell proliferation, functional differentiation, cell motility, and apoptosis (24). TGF-β is a potent inducer of the ECM and is required under physiologic conditions, such as wound repair, to induce fibroblasts to produce and contract the ECM (42). On the other hand, deregulated TGF-β signaling has long been postulated to underlie pathological fibrosis, but the specific mechanisms involved in this process have not been fully delineated (42). TGF-β signaling is initiated by ligand binding to a heteromeric complex of transmembrane serine/threonine kinases (type I and type II), and the subsequent activation of transcriptional coregulators Smad2 and Smad3 (24). Although the Smad pathway constitutes a major mode of TGF-β signaling, recent evidence suggests that alternative non-Smad pathways are activated in parallel with Smad signaling and also mediate TGF-β responses (27). The non-Smad pathways include mitogen-activated protein kinases (13, 15, 33); protein kinase C (PKC) (21, 26); the phosphatidyl-inositol-3 kinase pathway and its downstream target, Akt (43); and the nonreceptor kinases c-Abl and c-Src (11, 26). There is also evidence that some of these pathways, including ERK1/2, Akt, c-Abl, and PKC δ play a role in pathological fibrosis (41), but the specific mechanisms for any of these signaling molecules in the fibrotic process remain to be elucidated. Given that TGF-β regulates Fli1 acetylation and that several prior studies linked phosphorylation and acetylation (9, 32), there arose an interesting possibility that one of the TGF-β-activated kinases phosphorylates Fli1.

In this study, we sought to further elucidate the molecular mechanism that underlies the TGF-β-mediated regulation of Fli1, focusing on the steps that precede the recruitment of PCAF to Fli1 and the subsequent acetylation of Fli1. Our findings have established that in response to TGF-β stimulation, Fli1 is phosphorylated by PKC δ at threonine 312. Furthermore, phosphorylation at this residue promotes the interaction of Fli1 with PCAF and subsequently facilitates Fli1 acetylation. Because of the previously reported role of PKC δ in the transcriptional regulation of the collagen gene under physiological conditions and in scleroderma fibrosis (18, 21), our observations clarify the mechanism whereby PKC δ regulates the expression of the collagen gene in dermal fibroblasts and provide new insight into the mechanism of cutaneous fibrosis in scleroderma.

MATERIALS AND METHODS

Reagents.

Recombinant human TGF-β1 was obtained from Peprotech. The polyclonal rabbit anti-Fli1 antibody was prepared as described previously (22). The polyclonal rabbit antiphosphoserine antibody and antiphosphothreonine antibody were purchased from Zymed. The monoclonal mouse anti-Fli1 antibody and anti-PKC δ antibody were purchased from BD Biosciences. The polyclonal rabbit anti-acetylated lysine antibody, anti-hemagglutinin (HA) tag antibody, anti-phospho-Smad2 antibody, and anti-Smad2 antibody were purchased from Cell Signaling Technology. Antibodies for β-actin and Flag tag were purchased from Sigma. Antibodies for PCAF and lamin A/C were obtained from Santa Cruz Biotechnology. Antibody for calmodulin binding peptide was purchased from Millipore. Rottlerin and anacardic acid were obtained from Calbiochem. PKC δ small interfering RNA (siRNA) (GGCUACAAAUGCAGGCAAU) was synthesized and purified by Dharmacon based on a previously published siRNA sequence (39). Control siRNA was purchased from Santa Cruz.

Cell cultures.

Human dermal fibroblast cultures were established from the foreskins of healthy newborns from the Medical University of South Carolina Hospital in compliance with the Institutional Review Board for Human Studies. All studies used cells from passage number 3 to 6. Human embryonic kidney 293T cells were purchased from ATCC and maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.

Plasmid construction.

pSG5-Fli1 and pCTAP-Fli1 were generated as described previously (1, 34). Various Fli1 constructs with point mutations were generated with the use of the QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The Fli1 deletion fragments were subcloned into the pCTAP vector. A −772 COL1A2/CAT construct was generated as previously described (38). The expression vector for PCAF and PCAF/ΔHAT was a gift from T. Kouzarides (Cambridge University, Cambridge, United Kingdom). The expression vector for the constitutive active TGF-β type I receptor (TβRI T204D) was a gift from K. Miyazono (University of Tokyo, Tokyo, Japan). TβRI T204D with mutations in the L45 loop (TβRI T204D mL45) was generated by PCR-based mutagenesis as described previously (44). Plasmids were purified twice on cesium chloride gradients. At least two different plasmid preparations were used for each experiment. Adenoviral PCAF siRNA was previously described (1).

Cell fractionation.

Cells were fractionated as described previously (23). Briefly, cells were lysed using buffer A (10 mM HEPES [pH 7.9], 1.5 mM MgCl₂, 10 mM KCl, 1 mM dithiothreitol [DTT], 0.5 mM phenylmethylsulfonyl fluoride [PMSF]) to isolate the cytoplasmic fraction. The remaining pellet was then resuspended in buffer C (20 mM HEPES [pH 7.9], 25% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 1 mM DTT, 0.5 mM PMSF) to extract the nuclear contents.

Immunoblotting.

Whole-cell extracts were prepared using lysis buffer with the following contents: 1% Triton X-100, 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 3 mM MgCl₂, 1 mM CaCl₂, proteinase and phosphatase inhibitor cocktails (Calbiochem), 1 mM PMSF, and trichostatin A (100 ng/ml). The protein extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The membranes were incubated overnight with primary antibody, washed, and incubated for 1 h with secondary antibody. After washing, visualization was performed by enhanced chemiluminescence (Pierce).

Immunoprecipitation.

In order to precipitate endogenous Fli1 in dermal fibroblasts, nuclear extracts were preadsorbed with protein G Sepharose beads (GE Healthcare) and incubated with 0.5 μg of monoclonal mouse anti-Fli1 antibody and then with protein G Sepharose beads. Streptavidin-coupled agarose beads (Sigma) were used instead of protein G Sepharose beads for the immunoprecipitation of ectopically expressed and tagged Fli1. The precipitated proteins were subjected to immunoblotting.

Generation of phospho-Fli1 (Thr312)-specific antiserum.

The antibody was custom generated by GenScript Corporation. Briefly, the phosphopeptide CTNGEFKM(pT)DPDEVA was synthesized, coupled to keyhole limpet hemocyanin, and used as an immunogen in rabbits to generate antiserum directed against phosphothreonine 312. Antiserum was precleared by passing over a resin of immobilized, unphosphorylated peptide. For Western analyses, antiserum was preincubated with 10 μM of nonphosphorylated peptide for 1 h and then added to the blots.

In vitro phosphorylation assay.

Tagged-Fli1 proteins were precipitated by streptavidin-coupled beads from the nuclear extracts of 293T cells. The precipitated proteins were washed with kinase buffer (25 mM Tris-HCl [pH 7.5], 10 mM MgCl₂, 0.1 mM NaVO₄, 5 mM β-glycerophosphate, 2 mM DTT) three times. Then, 50 ul of kinase buffer with ATP (200 μM), phosphatidyl-serine (50 μg/ml) and phorbol ester (1 μM) was added to Fli1 protein bound to streptavidin-coupled beads. Various amounts of purified recombinant PKC δ-glutathione S transferase (GST) fusion protein was added to each reaction and incubated at 30°C for 30 min. The reaction mixtures were boiled in 6× sodium dodecyl sulfate sample buffer for 5 min, and the proteins were subjected to immunoblotting with anti-phospho-Fli1 (Thr312) antibody.

Reporter gene assay.

Foreskin fibroblasts were grown to 50% confluence in 100-mm dishes, transfected with the indicated constructs along with pSV-β-galactosidase using FuGENE6 (Roche), and incubated for 48 h. Extracts, normalized for protein content, were incubated with butyl-coenzyme A and [¹⁴C]-chloramphenicol for 90 min at 37°C. Butylated chloramphenicol was extracted using an organic solvent (a 2:1 mixture of tetramethylpentadecane and xylene) and quantitated by scintillation counting.

In vivo acetylation assay.

293T cells were transfected with expression vectors encoding tagged-Fli1 (0.1 μg) and the indicated HAT protein (2 μg), and incubated for 48 h. Fli1 was precipitated from whole-cell lysates by streptavidin-coupled agarose beads and subjected to immunoblotting using polyclonal rabbit anti-acetylated lysine antibody. After the development, the membrane was stripped and reprobed with anti-calmodulin binding peptide antibody to determine the total levels of ectopically expressed Fli1.

DNA affinity precipitation assay.

The oligonucleotides containing biotin on the 5′ nucleotide of the sense strand were used. The sequences of these oligonucleotides are as follows: (i) COL1A2 Ets binding site (EBS) oligonucleotide, 5′-GAAAGGGCGGGGGAGGGCGGGAGGATGCGGAGGGCGGAG-3′, which corresponds to base pairs −307 to −269 of the COL1A2 promoter, containing both an EBS and a GC box; (ii) COL1A1 EBS oligonucleotide, 5′-AATCATGCCTAGGGTTTGGAGGAAGGCATTTGATTCTGTTCTGG-3′, which corresponds to base pairs −732 to −689 of the COL1A1 promoter, containing a putative EBS predicted by MatInspector professional software (Genomatix); and (iii) COL1A1 EBS-M oligonucleotide, 5′-AATCATGCCTAGGGTTTGGAGTATGGCATTTGATTCTGTTCTGG-3′, which has a mutated EBS of COL1A1 EBS oligonucleotide. These oligonucleotides were annealed to their respective complementary oligonucleotides, and double-stranded oligonucleotides were gel purified and used. Whole-cell extracts (500 μg) prepared from 293T cells were incubated for 10 min at 4°C with gel shift binding buffer (10 mM Tris-HCl [pH 8.0], 40 mM KCl, 1 mM DTT, 6% glycerol, 0.05% NP-40), and 20 μg of poly(dI-dC) (Pierce) in a final volume of 1 ml. Preclearing was performed by adding streptavidin-coupled agarose beads and incubating the mixture for 30 min with gentle rocking at 4°C. After centrifugation, the supernatant was incubated with 500 pmol of each double-stranded oligonucleotide overnight at 4°C with gentle rocking. Then 65 μl of streptavidin-coated agarose beads was added, followed by a further 2-h incubation at 4°C. The protein-DNA-streptavidin-agarose complex was washed twice with Tris-EDTA (100 mM NaCl), twice with gel shift binding buffer, and once with phosphate-buffered saline. The precipitates were subjected to immunoblotting using the indicated antibodies. The specific binding of Fli1 to COL1A2 EBS oligonucleotide through EBS was demonstrated previously (1, 21).

ChIP assay.

The chromatin immunoprecipitation (ChIP) assay was carried out essentially as described previously (1, 28). Briefly, cells were treated with 1% formaldehyde for 10 min. The cross-linked chromatin was then prepared and sonicated to an average size of 300 to 500 bp. The DNA fragments were immunoprecipitated overnight with or without polyclonal anti-Fli1 antibody at 4°C. After reversal of the cross-linking, the immunoprecipitated chromatin was amplified by PCR amplification of specific regions of the COL1A1 and COL1A2 genomic locus. The primers were as follows: COL1A1/F-750, 5′-CTCCCCACTCCATCTCTCAA-3′; COL1A1/R-541, 5′-GTCTTCTGGTGTGGCTAGGG-3′; COL1A2/F-404, 5′-CTGGACAGCTCCTGCTTTGAT-3′; and COL1A2/R-237, 5′-CAAGGGGAAACTCTGACTCG-3′. The amplified DNA products were resolved by agarose gel electrophoresis.

Statistical analysis.

Data presented as bar graphs are the means ± standard deviations (SD) of at least three independent experiments. Statistical analysis was performed using the Mann Whitney-U test (a P of <0.05 was considered significant).

RESULTS

TGF-β mediates Fli1 phosphorylation at a threonine residue(s) through a Smad-independent pathway.

In an initial experiment, we investigated whether the activation of TGF-β signaling mediates Fli1 phosphorylation. The pCTAP-Fli1 expression vector, containing a streptavidin binding peptide and a calmodulin binding peptide, was transfected into 293T cells along with a constitutively active TGF-β type I receptor (TβRI T204D) or a control empty vector, and whole-cell extracts were subjected to immunoprecipitation using streptavidin-coupled agarose beads, followed by immunoblotting with antiphosphoserine antibody or antiphosphothreonine antibody. Consistent with previous observations for human T cells (47), Fli1 was highly phosphorylated on a serine residue(s) even in the absence of TβRI T204D (Fig. 1A, left panel). The forced expression of TβRI T204D did not affect the serine phosphorylation levels of Fli1. In contrast, Fli1 was not phosphorylated on a threonine residue(s) in the absence of TβRI T204D, which is also in agreement with a previous report (Fig. 1A, right panel) (47). However, in the presence of TβRI T204D, Fli1 phosphorylation on a threonine residue(s) was markedly increased. Since Smad2/3 are the primary mediators of TGF-β signaling, we asked if Smad signaling is involved in TGF-β-dependent Fli1 phosphorylation. For this purpose, we used a TβRI T204D mL45 construct, which has a constitutively active kinase domain but lacks the ability to phosphorylate Smad2/3 (44). As shown in Fig. 1B, in the presence of TβRI T204D mL45, Fli1 threonine phosphorylation was comparable with that induced by TβRI T204D. Taken together, these results indicate that TGF-β induces Fli1 threonine phosphorylation through a Smad-independent pathway.

FIG. 1.

TGF-β mediates Fli1 phosphorylation at a threonine residue(s) through a Smad-independent pathway. 293T cells were transfected with pCTAP-Fli1 along with TβRI T204D, TβRI T204D mL45, or empty vector and incubated for 48 h. Equal amounts of protein from each whole-cell extract were subjected to immunoprecipitation (IP) using streptavidin-coupled agarose beads (SA beads), followed by immunoblotting using antiphosphoserine antibody (p-Ser) or antiphosphothreonine antibody (p-Thr). In order to visualize the total levels of ectopically expressed Fli1, the same membrane was stripped and reprobed with the anti-calmodulin binding peptide antibody. The levels of HA-tagged proteins in cell lysates were determined by Western blotting. To confirm the activation status of TGF-β signaling, the levels of phospho-Smad2 were evaluated by Western blotting. WT, wild type.

Threonine 312 is a major site of Fli1 phosphorylation by TGF-β.

In order to map the region of the Fli1 protein undergoing phosphorylation in response to TGF-β stimulation, we constructed three sequential deletion mutants of the Fli1 protein (Fig. 2A). As shown in Fig. 2B, Fli1_{1 to 384}, which lacks the C-terminal activation domain (CTA), and Fli1_{1 to 373}, which additionally lacks the region between the DNA binding domain, termed the Ets binding domain (EBD), and the CTA domain, remain responsive to the TβRI T204D-induced phosphorylation on a threonine residue(s). However, TβRI T204D-dependent phosphorylation on a threonine residue(s) was almost completely abrogated in the Fli1_{1 to 275} construct, which has an additional deletion of EBD. These results suggest that a major phosphorylation site(s) mediating TGF-β response resides within the EBD. This region contains four threonine residues, T301, T305, T312, and T349. To further determine which threonine residue is phosphorylated in response to TGF-β, we replaced the threonine residues with alanine either individually or in a set of two threonine residues (Fli1 T301/305A, Fli1 T312A, and Fli1 T349A) and carried out an in vivo phosphorylation assay using mutated constructs. As shown in Fig. 3A, Fli1 T301/305A and Fli1 T349A were phosphorylated by TβRI T204D at levels comparable to those of wild-type Fli1. In contrast, TβRI T204D failed to phosphorylate the Fli1 T312A mutant. These results strongly suggest that the major phosphorylation site of Fli1 is located at threonine 312.

FIG. 2.

The TGF-β-dependent phosphorylation site(s) of Fli1 is located within the EBD. (A) Schematic representations of pCTAP-Fli1 and three distinct sequential deletion constructs. A major acetylation site, lysine 380, is indicated by an arrow. ATA, A-terminal activation domain. (B) Phosphorylation levels of each Fli1 deletion construct were determined as described in the Fig. 1 legend.

FIG. 3.

Fli1 is phosphorylated at threonine 312 in response to TGF-β stimulation. (A) Phosphorylation levels of Fli1 mutants carrying alanine substitutions for threonines 301, 305, 312, and 349 were determined in 293T cells as described in the legend to Fig. 1. (B) Amino acid sequence of a peptide containing phosphorylated threonine, which was used to generate the antiphosphorylation of threonine 312 Fli1 antibody. (C) Determination of the specificity of the phospho-Fli1 (Thr312) antiserum. The blots were prepared as described in the legend to Fig. 1. Strips of the membrane were incubated with phospho-Fli1 (Thr312) antiserum with no additions (left panel), in the presence of 10 μM nonphosphopeptide (non-P-pep) (middle panel) or in the presence of 10 μM nonphosphopeptide and phosphopeptide (P-pep) (right panel). Blots were reprobed with anti-calmodulin binding peptide antibody.

To further corroborate that TGF-β induces the phosphorylation of threonine 312, we developed a rabbit polyclonal antibody that selectively reacts with Fli1 phosphorylated on this threonine. The reactivity of the antibody was tested in 293T cells cotransfected with pCTAP-Fli1 and TβRI T204D. An immunoblotting analysis using the phosphothreonine 312 antiserum showed moderate reactivity toward Fli1 extracted from 293T cells in the absence of TβRI T204D, and a markedly increased reactivity in cells expressing TβRI T204D (left panels in Fig. 3C). The reactivity toward Fli1 in the absence of TβRI T204D was almost completely eliminated by the addition of 10 μM nonphosphorylated peptide, while the reactivity in the presence of TβRI T204D was not affected by this treatment (middle panels in Fig. 3C). Furthermore, competition with phosphorylated peptide in addition to nonphosphorylated peptide resulted in a total loss of reactivity under both conditions (Fig. 3C, right panel). These results indicate that the antibody specifically recognizes phosphorylated Fli1 at threonine 312 and confirms that threonine 312 located within the EBD constitutes a major TGF-β-induced phosphorylation site of Fli1.

Fli1 is phosphorylated at threonine 312 through a PKC δ-dependent pathway in response to TGF-β stimulation.

We next sought to determine the nature of the kinase involved in the TGF-β-induced phosphorylation of Fli1. To this end, we inspected the amino acid sequence of Fli1 surrounding threonine 312. As shown in Fig. 4A, threonine 312 lies within consensus target sites for several different protein kinases, including PKC, cyclic-AMP-dependent protein kinase, Rho-kinase (KXT motif), and casein kinase II (TD or TXXD motif), all of which can be activated by TGF-β stimulation (8, 45, 46). Based on previous reports suggesting that PKC δ affects the transcriptional activity of Fli1 in the context of the human COL1A2 promoter (21), we first focused on PKC δ. We initially examined the effect of a blockade of PKC δ on the TGF-β-induced threonine phosphorylation of Fli1 using a selective pharmacological inhibitor of PKC δ, rottlerin. As shown in Fig. 4B, Fli1 phosphorylation levels in 293T cells expressing TβRI T204D were dramatically decreased by treatment with rottlerin (1 μM). Since the specificity of this inhibitor has recently been questioned (35), we also evaluated the effect of a dominant negative PKC δ or PKC δ siRNA on the phosphorylation levels of Fli1. As shown in Fig. 4C and D, both treatments almost completely abrogated Fli1 phosphorylation induced by TβRI T204D. Furthermore, the forced expression of PKC δ induced a marked phosphorylation of wild-type Fli1, even in the absence of TβRI T204D, while having no effect on the phosphorylation levels of the Fli1 T312A mutant (Fig. 4E). Taken together, these results indicate that TGF-β elicits Fli1 phosphorylation at threonine 312 through a PKC δ-dependent pathway.

FIG. 4.

Fli1 is phosphorylated at threonine 312 through a PKC δ-dependent pathway in response to TGF-β stimulation. (A) The consensus target sites including threonine 312 for serine/threonine kinases. cAMP, cyclic AMP. (B to D) 293T cells were transfected with wild-type pCTAP-Fli1 along with TβRI T204D or empty vector and incubated for 48 h. In some experiments, the cells were treated with 1 μM of rottlerin (B) or transduced with dominant-negative PKC δ-expressing adenovirus (C) for the last 24 h. In the gene silencing experiments, the cells were treated with 10 nM control or PKC δ siRNA using HiPerFect reagent for 24 h prior to plasmid transfection (D). Phosphothreonine levels of Fli1 in each cell lysate were determined by immunoprecipitation (IP) using streptavidin-coupled agarose beads (SA beads), followed by immunoblotting using antiphosphothreonine antibody (p-Thr). The total levels of ectopically expressed Fli1 were determined on the same membrane using anti-calmodulin binding peptide antibody. (E) 293T cells were transfected with expression vectors encoding wild-type Fli1 or the Fli1 T312A mutant and incubated for 48 h. Some cells were transduced with wild-type PKC δ adenovirus for the last 24 h. The levels of Fli1 phosphorylation were determined as described above.

PKC δ directly phosphorylates Fli1 at threonine 312.

We next investigated whether PKC δ directly phosphorylates Fli1. Since Fli1 is primarily located in the nucleus (5) (Fig. 5A), this suggests that the phosphorylation of Fli1 most likely also occurs in this cellular compartment. Therefore, the subcellular localization of PKC δ was examined in 293T cells in the presence or absence of TβRI T204D. As shown in Fig. 5B, in the absence of TβRI T204D, PKC δ was mainly localized in the cytoplasm, with a small amount present in the nucleus. However, in 293T cells expressing TβRI T204D, the amount of nuclear PKC δ was increased, which is consistent with the previously reported upregulation of PKC δ protein levels by TGF-β (21, 31). We next investigated whether PKC δ interacts with Fli1 in response to TGF-β stimulation. For this experiment, 293T cells were transfected with pCATP-Fli1 along with either TβRI T204D or empty vector, and the interaction of Fli1 with endogenous PKC δ was evaluated by immunoprecipitation using nuclear extracts. As shown in Fig. 5C, ectopically expressed Fli1 interacted with endogenous PKC δ in the presence of TβRI T204D, while the interaction was below the detectable level in the absence of TβRI T204D, indicating that TGF-β promotes the interaction of PKC δ with Fli1 in the nucleus. Since in unstimulated cells Fli1 is bound to the collagen promoter, we next used a DNA affinity precipitation assay to investigate whether PKC δ forms complexes with Fli1 on the human COL1A2 promoter. We utilized previously described COL1A2 oligonucleotide, which contains a well-characterized EBS (1, 2, 21). An expression vector encoding untagged wild-type Fli1 was expressed in 293T cells in the presence or absence of TβRI T204D, and the nuclear extracts were subjected to a DNA affinity precipitation assay. Endogenous PKC δ was precipitated by COL1A2 EBS oligonucleotide in the presence, but not in the absence, of TβRI T204D, suggesting that PKC δ is recruited to the COL1A2 promoter in response to TGF-β stimulation (Fig. 5D). To further confirm the functional role of PKC δ in Fli1 phosphorylation, we performed an in vitro phosphorylation assay. Tagged Fli1 was isolated from 293T cells using streptavidin-coupled beads and incubated with purified recombinant PKC δ-GST fusion protein. The reaction products were subjected to immunoblotting with anti-phospho-Fli1 (Thr312) antibody. Fli1 was phosphorylated at threonine 312 by recombinant PKC δ-GST fusion protein in a dose-dependent manner (Fig. 5E). We concluded from these experiments that Fli1 is directly phosphorylated by PKC δ in response to TGF-β stimulation.

FIG. 5.

PKC δ directly phosphorylates Fli1 at threonine 312. (A) Cultured dermal fibroblasts and human microvascular endothelial cells were stained with anti-Fli1 antibody. Positive signals were developed with diaminobenzidine. (B) 293T cells were transfected with TβRI T204D or empty vector for 48 h and then cytoplasmic extracts (500 μl of buffer A) and nuclear extracts (500 μl of buffer C) were prepared. In order to evaluate the ratio of PKC δ localized in the cytoplasm (C) and nucleus (N) accurately, a 10% volume of each extract was subjected to immunoblotting with anti-PKC δ antibody. To confirm that the cytoplasm and nucleus were properly separated, the levels of β-actin and lamin A/C were determined. The bottom panel shows the relative expression levels of PKC δ in each compartment. (C) 293T cells were transfected with the indicated expression vectors for 48 h. Nuclear extracts and whole-cell lysates were prepared under the same conditions. Tagged-Fli1 was precipitated from nuclear extracts with streptavidin-coupled agarose beads (SA beads), and the precipitates were subjected to immunoblotting using anti-PKC δ antibody and anti-calmodulin binding peptide antibody. The levels of PKC δ in the nuclear extracts and the levels of TβRI T204D in the whole-cell lysates were determined by immunoblotting. WT, wild type. (D) Nuclear extracts (NE) and whole-cell lysates were prepared under the same conditions. Nuclear extracts were subjected to DNA affinity precipitation (DNAP) with COL1A2 EBS oligonucleotide. The levels of PKC δ in the precipitates were determined by immunoblotting. (E) Streptavidin bead-bound tagged Fli1 prepared from 293T cells was incubated with recombinant PKC δ-GST fusion protein. In order to activate PKC δ, phosphatidyl serine and phorbol ester were added to the reaction. The whole reaction was subjected to immunoblotting using anti-phospho-Fli1 (Thr312) antibody. The levels of tagged Fli1 and PKC δ were confirmed by immunoblotting on the same membrane.

The phosphorylation at threonine 312 promotes the interaction of Fli1 with PCAF and the subsequent acetylation event.

So far, we have established that in response to TGF-β stimulation, Fli1 becomes phosphorylated by PKC δ at threonine 312 and acetylated by PCAF at lysine 380 (1). Since phosphorylation was shown to regulate the acetylation of selected transcription factors such as p53 (32) and NF-κB (9), we investigated the effect of Fli1 phosphorylation on threonine 312 on the acetylation status of Fli1. We first determined the acetylation levels of wild-type Fli1 and the Fli1 T312A mutant in 293T cells expressing either PCAF or PCAF/ΔHAT. As previously shown (1), the expression of PCAF resulted in a large increase in the acetylation levels of wild-type Fli1 (Fig. 6A, lane 3), and this effect was further augmented in the presence of TβRI T204D (lane 4). Consistently, the interaction of Fli1 with PCAF was increased in the presence of TβRI T204D (compare lane 1 to 2 and 3 to 4). The interaction of PCAF/ΔHAT with Fli1 was not affected by the absence of the HAT domain; however, PCAF/ΔHAT did not show any effect on Fli1 acetylation (lanes 1 and 2). When the Fli1 T312A mutant was used, a faint residual signal was detected in the presence of PCAF (lane 7). In contrast to that of wild-type Fli1, TβRI T204D did not affect the signal intensity of acetylated Fli1 T312A (lane 8). Importantly, the interaction of the Fli1 T312A mutant with PCAF was below the detectable level. Taken together, these results support the notion that the phosphorylation of threonine 312 is a prerequisite for the interaction of Fli1 with PCAF and the subsequent acetylation of Fli1 by the HAT activity of PCAF.

FIG. 6.

Phosphorylation at threonine 312 increases the interaction of Fli1 with PCAF and subsequently promotes its acetylation. (A) Wild-type Fli1 or Fli1-T312A constructs were transfected into 293T cells along with the indicated expression vectors, including TβRI T204D, PCAF, and PCAF/ΔHAT, and incubated for 48 h. Total cell extracts were subjected to immunoprecipitation (IP) using streptavidin-coupled agarose beads (SA beads), followed by immunoblotting using anti-acetylated lysine antibody (AcK) or anti-Flag antibody. In order to visualize the total levels of ectopically expressed Fli1, the same membrane was stripped and reprobed with the anti-calmodulin binding peptide antibody. The levels of Flag-tagged and HA-tagged proteins in cell lysates were determined by Western blotting. (B) Wild-type Fli1 or Fli1-T312A constructs were transfected into 293T cells along with PCAF or PCAF/ΔHAT, and the cells were incubated for 48 h. The DNA binding ability of each Fli1 construct was evaluated by DNA affinity precipitation (DNAP) assay. The levels of Fli1 and Flag-tagged proteins were determined by immunoblotting. (C) 293T cells were transfected with the −772 human COL1A2 promoter, along with the indicated expression vectors. After 48 h incubation, formaldehyde-cross-linked, moderately sheared chromatin was prepared. The DNA fragments were immunoprecipitated using rabbit polyclonal anti-Fli1 antibody, and the presence of COL1A2 promoter fragments were detected using PCR.

We have previously demonstrated that acetylation at lysine 380 regulates the DNA binding status of Fli1 on the human COL1A2 promoter (1). To investigate whether the phosphorylation of threonine 312 is involved in the regulation of the DNA binding status of Fli1, we performed a DNA affinity precipitation assay. Untagged-Fli1 constructs, those of the wild type and the T312A mutant, were transfected into 293T cells along with PCAF or PCAF/ΔHAT. The binding of each Fli1 construct to the COL1A2 EBS oligonucleotide was examined. As shown in Fig. 6B, there was no difference in DNA binding ability between wild-type Fli1 and the Fli1 T312A mutant in the presence of PCAF/ΔHAT. Consistent with our previous data (1), PCAF substantially decreased the DNA binding ability of wild-type Fli1 (lanes 1 and 2). In contrast, the DNA binding status of the Fli1 T312A mutant was minimally affected by PCAF (lanes 3 and 4). These results support our previous observation that the DNA binding ability of Fli1 is regulated by PCAF-dependent acetylation and indicate that the phosphorylation of threonine 312 is required for the acetylation-dependent regulation of the DNA binding ability of Fli1. In order to further confirm these observations in vivo, we employed ChIP analysis. Each Fli1 construct was cotransfected with PCAF or PCAF/ΔHAT along with the −772 COL1A2 promoter construct into 293T cells. Cross-linked chromatin was immunoprecipitated with polyclonal rabbit anti-Fli1 antibody, which recognizes the CTA of the Fli1 protein, and the purified genomic DNA was amplified with primers specific to the COL1A2 promoter. As shown in Fig. 6C, both of the Fli1 constructs occupied the −404 to −237 region of the human COL1A2 promoter in the presence of PCAF/ΔHAT (lanes 1 to 3 and lanes 7 to 9). Consistent with our previous observation, the binding of wild-type Fli1 was dramatically decreased in the presence of PCAF (Fig. 6C, lanes 4 to 6). In contrast, PCAF only slightly affected the DNA binding status of the Fli1 T312A mutant (Fig. 6C, lanes 10 to 12). These data indicate that the phosphorylation of threonine 312 is required for the acetylation-dependent dissociation of Fli1 from the COL1A2 promoter in vivo.

The phosphorylation of Fli1 at threonine 312 precedes its acetylation in dermal fibroblasts in response to TGF-β.

The data presented so far employed various overexpression systems. Therefore, we next examined the levels of phosphorylation and acetylation of endogenous Fli1 in dermal fibroblasts after TGF-β stimulation. As shown in Fig. 7A, Fli1 phosphorylation was below the detectable level in unstimulated cells under regular exposure time. Upon TGF-β treatment, the phosphorylation levels of Fli1 were dramatically increased at 2 h. Consistent with our previous report (1), the total levels of Fli1 were decreased at 24 h after TGF-β stimulation. The lower graph in Fig. 7A shows the phosphorylation levels of Fli1 normalized by its total levels. After TGF-β stimulation, Fli1 remained phosphorylated up to 24 h. In order to investigate the relationship between phosphorylation and acetylation, we focused on the early time points (up to 3 h) and examined the effect of TGF-β on the levels of phosphorylation and acetylation of Fli1. For an accurate evaluation, phosphorylation and acetylation were determined on the same membrane, which was probed first with anti-acetylated lysine antibody, stripped, and then reprobed with anti-phospho-Fli1 (Thr312) antibody. As shown in Fig. 7B, consistent with our previous observation (1), Fli1 was highly acetylated at 3 h after TGF-β stimulation. In contrast to phosphorylation, however, Fli1 was faintly acetylated at 2 h. These results indicate that the phosphorylation of endogenous Fli1 precedes its acetylation in response to TGF-β. Consistently, PKC δ interacted with Fli1 in the nucleus at 2 h after TGF-β stimulation (Fig. 7C), supporting the notion that PKC δ-mediated phosphorylation is required for the subsequent acetylation of Fli1. Previous studies have demonstrated that the blockade of PKC δ activity by rottlerin markedly decreased the basal expression levels of COL1A1 and COL1A2 in quiescent dermal fibroblasts (18, 21). Furthermore, the study by Jinnin et al. (21) correlated the inhibitory effect of rottlerin with increased Fli1 binding to the COL1A2 promoter. Therefore, we examined the effect of rottlerin on the TGF-β-induced levels of Fli1 phosphorylation and acetylation. Treatment with rottlerin completely abrogated the phosphorylation and acetylation of Fli1 (Fig. 7D). As expected, treatment with the histone acetylase inhibitor anacardic acid inhibited acetylation but had no effect on the phosphorylation of Fli1. Consistent with these data, the siRNA-mediated depletion of PCAF prevented TGF-β-induced Fli1 acetylation (Fig. 7E), whereas the depletion of PKC δ prevented both the acetylation and phosphorylation of Fli1 (Fig. 7F). In addition, the inhibition of PKC δ using either rottlerin or specific siRNA prevented complex formation between Fli1 and PCAF (Fig. 7G and H). As shown in Fig. 7I, basal levels of Fli1 phosphorylation and acetylation were detectable after a longer exposure time and were also significantly decreased by the treatment with rottlerin. Similar results were obtained by depleting endogenous PKC δ using specific siRNA (Fig. 7J). Consistent with a previous observation (21), rottlerin increased the expression levels of Fli1, which was more convincing in straight Western blotting (shown in Fig. 7I and J, bottom panels). We have previously shown that the siRNA-mediated depletion of PCAF interfered with the TGF-β-induced dissociation of Fli1 from the endogenous COL1A2 promoter (1). To examine the effect of endogenous PKC δ on the DNA binding status of Fli1, we performed ChIP analysis of the COL1A2 promoter using human dermal fibroblasts. As shown in Fig. 7K, in cells treated with control siRNA, TGF-β decreased the association of Fli1 with the COL1A2 promoter at 3 h after stimulation. In contrast, in fibroblasts treated with PKC δ siRNA, we observed the increased binding of Fli1 to the COL1A2 promoter in untreated cells, and Fli1 remained associated with the collagen promoter, although at slightly reduced levels, after TGF-β stimulation. The partial effect of TGF-β on Fli1 binding may be due to the incomplete depletion of endogenous PKC δ. Together, these results suggest that the basal activity of PKC δ is required to maintain the steady-state level of the Fli1 protein under physiological conditions through the phosphorylation-acetylation-dependent mechanism.

FIG. 7.

The phosphorylation of Fli1 precedes its acetylation in dermal fibroblasts in response to TGF-β. (A) Subconfluent dermal fibroblasts were serum starved for 24 h and then treated with TGF-β1 (2.5 ng/ml) for the indicated periods of time. Fli1 was precipitated from nuclear extracts using mouse monoclonal anti-Fli1 antibody, and the phosphorylation levels of Fli1 at threonine 312 were determined by immunoblotting with anti-phospho-Fli1 (Thr312) antibody. The membrane was stripped and reprobed with rabbit polyclonal anti-Fli1 antibody. The graph shows the levels of phosphorylation normalized by total Fli1 levels. (B) Subconfluent dermal fibroblasts were serum starved for 24 h and then treated with TGF-β1 (2.5 ng/ml) for the indicated periods of time. The phosphorylation and acetylation levels of Fli1 were visualized on the same membrane by sequential Western blots. (C) Fli1 was precipitated from nuclear extracts using monoclonal mouse anti-Fli1 antibody, and the precipitates were subjected to immunoblotting with anti-PKC δ antibody. The membrane was stripped and reprobed with rabbit polyclonal anti-Fli1 antibody. (D) Subconfluent dermal fibroblasts were serum starved for 48 h. Where indicated, for the last 24 h, cells were treated with rottlerin (3 μM), anacardic acid (10 μM), or a corresponding volume of dimethyl sulfoxide. The levels of Fli1 phosphorylation and acetylation were determined as described above. AcK, anti-acetylated lysine antibody. (E and F) Dermal fibroblasts were transduced with PCAF siRNA adenovirus (with scrambled siRNA adenovirus as a control) or transfected with PKC δ siRNA oligonucleotide (with control siRNA oligonucleotide as a control) using HiPerFect reagent for 72 h. Where indicated, cells were treated with TGF-β for the last 3 h. The levels of Fli1 phosphorylation and acetylation were determined as described above. (G and H) Dermal fibroblasts were incubated for 3 h with TGF-β (2.5 ng/ml) or a vehicle in the presence (+) or absence (−) of rottlerin (3 μM). For the gene silencing experiments, cells were treated with PKC δ siRNA or control siRNA for 72 h. Fli1 was precipitated from nuclear extracts using monoclonal mouse anti-Fli1 antibody, and the precipitates were subjected to immunoblotting with anti-PCAF antibody. The membrane was stripped and reprobed with rabbit polyclonal anti-Fli1 antibody. (I and J) Dermal fibroblasts were treated overnight with rottlerin (3 μM) or dimethyl sulfoxide (DMSO). For the gene silencing experiments, cells were treated with PKC δ siRNA or control siRNA for 72 h. The levels of Fli1 phosphorylation and acetylation were determined as described above. (K) Dermal fibroblasts were treated with PKC δ siRNA or control siRNA for 72 h. In some experiments, cells were stimulated with TGF-β1 (2.5 ng/ml) for the last 3 h. Formaldehyde-cross-linked, moderately sheared chromatin was prepared. The DNA fragments were immunoprecipitated using rabbit anti-Fli1 antibody (Anti-Fli1), and the presence of the human COL1A2 promoter fragments were detected using PCR. No-Ab, beads alone.

Phosphorylation at threonine 312 regulates the transcriptional activity of Fli1 in the context of the human COL1A2 promoter.

To further confirm the role of PKC δ in the regulation of the collagen gene in dermal fibroblasts, we examined the expression of collagen protein and mRNA in cells with an increased or decreased expression of PKC δ. As shown in Fig. 8A, PKC δ overexpression resulted in increased protein and mRNA levels of collagen, whereas the depletion of endogenous PKC δ had an opposite effect on collagen gene expression (Fig. 8B). To assess the impact of T312 phosphorylation on Fli1 transactivation potential, we performed reporter gene analysis of dermal fibroblasts using a −772 COL1A2/CAT construct. As a control, we used 293 cells that lack endogenous Fli1. As expected, COL1A2 promoter activity was unresponsive to TGF-β stimulation in these cells (data not shown). We first examined the effect of PCAF on the transcriptional activity of wild-type Fli1 and the T312A mutant. As shown in Fig. 8C, consistent with our previous results (1), a TGF-β1-dependent increase of COL1A2 promoter activity was significantly suppressed by the expression of Fli1, and this inhibitory effect was almost completely reversed by the coexpression of PCAF. The Fli1 T312A mutant showed a more potent inhibition of COL1A2 promoter activity in the presence or absence of TGF-β1. In contrast to that of wild-type Fli1, PCAF overexpression failed to reverse the inhibition of COL1A2 promoter activity by the Fli1 T312A mutant. A similar effect was previously observed using the Fli1 K380R mutant, which lacks a major acetylation site by PCAF (1). We also assessed the effect of PKC δ on the transcriptional activity of Fli1. As shown in Fig. 8D, the forced expression of PKC δ reversed the inhibitory effect of Fli1 in a dose-dependent manner. In contrast, the inhibitory effect of the Fli1 T312A mutant was only modestly reversed by the ectopic expression of PKC δ. These results support the functional role of the PKC δ-dependent phosphorylation of threonine 312 in the regulation of Fli1 transcriptional activity in the context of the COL1A2 promoter.

FIG. 8.

Phosphorylation at threonine 312 is essential for the regulation of the transcriptional activity of Fli1 in the context of the human COL1A2 promoter. (A and B) Dermal fibroblasts were transduced with PKC δ-expressing adenovirus (with green fluorescent protein-expressing adenovirus as a control) or transfected with PKC δ siRNA oligonucleotide (with control siRNA oligonucleotide as a control) for 72 h. The protein levels of type I collagen, PKC δ, and β-actin were determined by immunoblotting using whole-cell extracts. mRNA levels of COL1A2 gene were determined by quantitative real-time PCR under the same conditions. (C) Fibroblasts were transfected with the −772 COL1A2/CAT construct (2 μg), along with Fli1 constructs (0.1 μg), either the wild-type or the T312A mutant, and either PCAF or empty vector (0.5 μg) for 48 h. For the last 24 h, some cells were stimulated with TGF-β1 (2.5 ng/ml). Values represent CAT activities relative to those of untreated cells (100 arbitrary units [AU]). The mean and SD of the results from three separate experiments are shown. *, P < 0.05 versus control cells stimulated with TGF-β1; **, P < 0.05 versus cells transfected with Fli1 only and stimulated with TGF-β1. (D) Fibroblasts were transfected with 2 μg of the −772 COL1A2/CAT construct, along with empty vector or the indicated Fli1 constructs, and incubated for 48 h. At 6 h after transfection, some cells were transduced with PKC δ-expressing adenovirus (at multiplicities of infection of 5 and 10). In addition, some cells were treated with TGF-β1 (2.5 ng/ml) for the last 24 h. Values represent CAT activities relative to those of untreated cells transfected with empty vector (100 AU). The means and SD of the results from three separate experiments are shown. *, P < 0.05 versus control cells stimulated with TGF-β1; **, P < 0.05 versus cells transfected with wild-type Fli1 in the absence of PKC δ-expressing adenovirus and stimulated with TGF-β1; #, P < 0.05 versus cells transfected with wild-type Fli1 in the presence of PKC δ-expressing adenovirus and stimulated with TGF-β1.

DISCUSSION

The transcription regulation of the collagen type I gene, the most abundant component of the ECM, is tightly controlled to maintain proper tissue homeostasis. Disruption of this homeostasis leads to various pathological conditions associated with enhanced collagen deposition and scar tissue formation. Recent studies provide strong evidence for a critical role of the transcription factor Fli1 in the repression of collagen type I genes and in the inhibition of the TGF-β-induced profibrotic gene program (1, 4, 10, 21, 22, 28). Here we report that TGF-β abrogates the function of this repressor through the sequential posttranslational modifications that involves phosphorylation and acetylation. Our previous study has demonstrated that the PCAF-dependent acetylation of Fli1 at lysine 380 represents a principal mechanism whereby TGF-β abolishes the function of this repressor (1). In this study, we delineated the role of phosphorylation in this process. We showed that TGF-β elicits the phosphorylation of Fli1 at a threonine residue(s) in a Smad-independent manner. A major phosphorylation site was mapped to the threonine 312 located in the EBD region of Fli1. We utilized several independent approaches to demonstrate that this phosphorylation event is directly mediated by PKC δ. Importantly, the phosphorylation of Fli1 at this site is a prerequisite for its interaction with PCAF and the subsequent acetylation event, leading to the dissociation of Fli1 from the promoter of target genes, such as COL1A2. Conversely, the Fli1 T312A mutant, shown to be resistant to phosphorylation, is unable to dissociate from the promoter. As a result, the Fli1 T312A mutant is more potent than the wild-type Fli1 in repressing COL1A2 promoter activity and, unlike wild-type Fli1, is unresponsive to the reversal of this inhibitory effect by the coexpression of PCAF or PKC δ. Experiments using primary dermal fibroblasts showed that in response to TGF-β, endogenous Fli1 is phosphorylated and acetylated in a sequential manner, with phosphorylation peaking at 2 h, followed by acetylation at 3 h poststimulation. We have previously shown that TGF-β treatment leads to the dissociation of endogenous Fli1 from the COL1A2 promoter and the derepression of collagen gene transcription. Together, these results indicate that the transcriptional activity of Fli1 is strictly regulated by a phosphorylation-acetylation cascade in response to TGF-β stimulation. Similar regulatory events were previously described for p53 and NF-κB RelA, which were phosphorylated and acetylated in response to DNA damage and tumor necrosis factor alpha, respectively (9, 32). As a result, the DNA binding ability and transcriptional activity of those factors were enhanced, while in the case of Fli1, these posttranslational modifications lead to DNA dissociation and decreased protein stability. Such distinct functional consequences are consistent with the repressor role of Fli1 in the context of TGF-β signaling. Since Fli1 functions as both an activator and repressor of gene transcription depending on the cellular context, it is likely that additional phosphorylation and/or acetylation sites will be identified in response to different stimuli.

This study has provided additional support for the pathological role of PKC δ in the process of dermal fibrosis. PKC δ is a member of the novel PKC subfamily (ηPKC) of serine-threonine kinases. It is widely expressed in adult tissues and has been shown to play an important role in growth regulation and tissue remodeling (37). Several previous studies have also linked PKC δ to collagen regulation and fibrosis (18, 21, 26). It was shown that TGF-β-induced collagen synthesis was dependent on the activation of a signaling cascade, including PKC δ, Src, and ERK1/2, in human mesangial cells (26). Other studies have shown that Src and another nonreceptor tyrosine kinase, c-Abl, activate PKC δ through the phosphorylation of specific tyrosine residues (17). While we have not yet investigated the upstream signaling events leading to the activation and nuclear translocation of PKC δ in response to TGF-β, the nonreceptor tyrosine kinases are likely candidates to contribute to this process. Increased expression levels of both PKC δ and c-Abl were observed in scleroderma fibroblasts (18, 41). Furthermore, consistent with our current findings, the elevated expression of PKC δ was present in the nucleus of scleroderma cells (18). Jimenez et al. have also demonstrated that in scleroderma fibroblasts, rottlerin strongly suppressed the expression of the COL1A1 and COL3A1 genes at the transcriptional level and mapped the corresponding responsive element to the 129-bp segment encompassing nucleotides −675 to −804 of the COL1A1 gene promoter (18). The analysis of this sequence using MatInspector software (Genomatix) predicted the presence of a putative transcription EBS at nucleotides −709 to −712. We confirmed the occupancy of this site by Fli1 using ChIP analysis in quiescent dermal fibroblasts (Fig. 9A). Furthermore, the inhibition of endogenous PKC δ, further enhanced Fli1 binding to the COL1A1 promoter (Fig. 9A and B). These data indicate that the PKC δ-dependent phosphorylation of Fli1 coordinately regulates COL1A1 and COL1A2 expression in dermal fibroblasts. Significantly, we have found that the phosphorylation levels of Fli1 are increased in lesional scleroderma fibroblast cell lines compared with healthy control fibroblasts (A. Bujor and M. Trojanowska, unpublished data), providing further support for the deregulation of this pathway in scleroderma fibrosis.

FIG. 9.

Rottlerin increases the binding of Fli1 to the COL1A1 promoter. (A) Chromatin was isolated from adult dermal fibroblasts treated with PKC δ siRNA or control siRNA and immunoprecipitated using rabbit anti-Fli1 antibody (5 μg) or beads alone (No-Ab). After the isolation of bound DNA, PCR amplification was carried out using the COL1A1 promoter-specific primers. Input DNA was taken from each sample before the addition of an antibody. (B) Confluent fibroblasts were serum starved for 48 h. Some cells were treated with rottlerin (3 μM) for the last 24 h. Nuclear extracts were incubated with biotin-labeled oligonucleotides. Proteins bound to these nucleotides were isolated with streptavidin-coupled agarose beads, and Fli1 was detected by immunoblotting. The levels of Fli1 in nuclear extracts were determined by Western blotting. WT, wild type; DNAP, DNA affinity precipitation.

The discovery of Smad proteins have provided a basic framework of the molecular events that are required for TGF-β-dependent gene expression; however, due to the complexity of the intracellular signaling mechanisms elicited by this pleiotropic factor, the current understanding of this process is still far from complete. The identification and detailed characterization of the non-Smad pathways, which may function in a context-specific manner, are important to understanding the signaling processes linking TGF-β receptors to cellular responses under physiological and pathological conditions. With respect to scleroderma fibrosis, which is characterized by the persistent activation of TGF-β signaling, existing evidence is consistent with the activation of both Smad and non-Smad signaling pathways contributing to the fibrotic process (41). Our current study has provided new insights into the mechanism explaining how such distinct signaling events are brought together to coordinately regulate a specific gene program. Here we show that in parallel with the well-characterized phosphorylation and nuclear translocation of Smads, TGF-β, through a yet-unknown mechanism, activates PKC δ (21), which translocates into the nucleus and phosphorylates Fli1, which in turn triggers its acetylation by PCAF and the subsequent dissociation from the promoter of collagen genes (Fig. 10). Thus, the TGF-β-dependent synchronized regulation of the DNA binding activity of transcriptional activators, such as Smads, and repressors, such as Fli1, ultimately results in the increased expression of ECM genes. Clinical studies of cultured scleroderma fibroblasts and skin biopsies strongly support the primary role of TGF-β in the development of pathological fibrosis (3, 18, 22, 29, 30, 40). The non-Smad signaling molecules, such as PKC δ, represent attractive novel therapeutic targets for the treatment of this disease.

FIG. 10.

Schematic model for the TGF-β-induced posttranslational modifications of Fli1. In quiescent cells, collagen gene expression is repressed by Fli1 occupancy of the collagen promoter (1). In response to TGF-β stimulation, PKC δ is translocated into the nucleus and recruited to the COL1A2 promoter, leading to the phosphorylation of Fli1 at threonine 312. The phosphorylation of Fli1 is required for the subsequent PCAF-mediated acetylation of Fli1 at lysine 380. Acetylated Fli1 dissociates from the collagen promoter, resulting in the enhancement of the collagen gene transcription.